diff options
Diffstat (limited to 'src')
70 files changed, 9738 insertions, 3497 deletions
diff --git a/src/build-remote/build-remote.cc b/src/build-remote/build-remote.cc index abf3669b5b35..279ae62f69cc 100644 --- a/src/build-remote/build-remote.cc +++ b/src/build-remote/build-remote.cc @@ -38,6 +38,12 @@ static AutoCloseFD openSlotLock(const Machine & m, unsigned long long slot) return openLockFile(fmt("%s/%s-%d", currentLoad, escapeUri(m.storeUri), slot), true); } +static bool allSupportedLocally(const std::set<std::string>& requiredFeatures) { + for (auto & feature : requiredFeatures) + if (!settings.systemFeatures.get().count(feature)) return false; + return true; +} + static int _main(int argc, char * * argv) { { @@ -97,9 +103,10 @@ static int _main(int argc, char * * argv) source >> drvPath; auto requiredFeatures = readStrings<std::set<std::string>>(source); - auto canBuildLocally = amWilling - && ( neededSystem == settings.thisSystem - || settings.extraPlatforms.get().count(neededSystem) > 0); + auto canBuildLocally = amWilling + && ( neededSystem == settings.thisSystem + || settings.extraPlatforms.get().count(neededSystem) > 0) + && allSupportedLocally(requiredFeatures); /* Error ignored here, will be caught later */ mkdir(currentLoad.c_str(), 0777); diff --git a/src/cpptoml/cpptoml.h b/src/cpptoml/cpptoml.h index 07de010b1520..5a00da3b4cda 100644 --- a/src/cpptoml/cpptoml.h +++ b/src/cpptoml/cpptoml.h @@ -4,8 +4,8 @@ * @date May 2013 */ -#ifndef _CPPTOML_H_ -#define _CPPTOML_H_ +#ifndef CPPTOML_H +#define CPPTOML_H #include <algorithm> #include <cassert> @@ -84,11 +84,12 @@ class option return &value_; } - const T& value_or(const T& alternative) const + template <class U> + T value_or(U&& alternative) const { if (!empty_) return value_; - return alternative; + return static_cast<T>(std::forward<U>(alternative)); } private: @@ -295,13 +296,12 @@ struct valid_value_or_string_convertible }; template <class T> -struct value_traits<T, typename std:: - enable_if<valid_value_or_string_convertible<T>:: - value>::type> +struct value_traits<T, typename std::enable_if< + valid_value_or_string_convertible<T>::value>::type> { - using value_type = typename std:: - conditional<valid_value<typename std::decay<T>::type>::value, - typename std::decay<T>::type, std::string>::type; + using value_type = typename std::conditional< + valid_value<typename std::decay<T>::type>::value, + typename std::decay<T>::type, std::string>::type; using type = value<value_type>; @@ -312,12 +312,11 @@ struct value_traits<T, typename std:: }; template <class T> -struct value_traits<T, - typename std:: - enable_if<!valid_value_or_string_convertible<T>::value - && std::is_floating_point< - typename std::decay<T>::type>::value>:: - type> +struct value_traits< + T, + typename std::enable_if< + !valid_value_or_string_convertible<T>::value + && std::is_floating_point<typename std::decay<T>::type>::value>::type> { using value_type = typename std::decay<T>::type; @@ -330,11 +329,11 @@ struct value_traits<T, }; template <class T> -struct value_traits<T, - typename std:: - enable_if<!valid_value_or_string_convertible<T>::value - && std::is_signed<typename std::decay<T>:: - type>::value>::type> +struct value_traits< + T, typename std::enable_if< + !valid_value_or_string_convertible<T>::value + && !std::is_floating_point<typename std::decay<T>::type>::value + && std::is_signed<typename std::decay<T>::type>::value>::type> { using value_type = int64_t; @@ -356,11 +355,10 @@ struct value_traits<T, }; template <class T> -struct value_traits<T, - typename std:: - enable_if<!valid_value_or_string_convertible<T>::value - && std::is_unsigned<typename std::decay<T>:: - type>::value>::type> +struct value_traits< + T, typename std::enable_if< + !valid_value_or_string_convertible<T>::value + && std::is_unsigned<typename std::decay<T>::type>::value>::type> { using value_type = int64_t; @@ -395,10 +393,15 @@ struct array_of_trait<array> template <class T> inline std::shared_ptr<typename value_traits<T>::type> make_value(T&& val); inline std::shared_ptr<array> make_array(); + +namespace detail +{ template <class T> inline std::shared_ptr<T> make_element(); +} + inline std::shared_ptr<table> make_table(); -inline std::shared_ptr<table_array> make_table_array(); +inline std::shared_ptr<table_array> make_table_array(bool is_inline = false); #if defined(CPPTOML_NO_RTTI) /// Base type used to store underlying data type explicitly if RTTI is disabled @@ -576,7 +579,7 @@ class base : public std::enable_shared_from_this<base> #if defined(CPPTOML_NO_RTTI) base_type type() const { - return type_; + return type_; } protected: @@ -698,7 +701,7 @@ inline std::shared_ptr<value<double>> base::as() if (type() == base_type::INT) { auto v = std::static_pointer_cast<value<int64_t>>(shared_from_this()); - return make_value<double>(static_cast<double>(v->get()));; + return make_value<double>(static_cast<double>(v->get())); } #else if (auto v = std::dynamic_pointer_cast<value<double>>(shared_from_this())) @@ -731,7 +734,8 @@ inline std::shared_ptr<const value<double>> base::as() const { #if defined(CPPTOML_NO_RTTI) if (type() == base_type::FLOAT) - return std::static_pointer_cast<const value<double>>(shared_from_this()); + return std::static_pointer_cast<const value<double>>( + shared_from_this()); if (type() == base_type::INT) { @@ -1027,11 +1031,14 @@ inline std::shared_ptr<array> make_array() return std::make_shared<make_shared_enabler>(); } +namespace detail +{ template <> inline std::shared_ptr<array> make_element<array>() { return make_array(); } +} // namespace detail /** * Obtains a option<vector<T>>. The option will be empty if the array @@ -1060,7 +1067,7 @@ class table; class table_array : public base { friend class table; - friend std::shared_ptr<table_array> make_table_array(); + friend std::shared_ptr<table_array> make_table_array(bool); public: std::shared_ptr<base> clone() const override; @@ -1152,14 +1159,25 @@ class table_array : public base array_.reserve(n); } + /** + * Whether or not the table array is declared inline. This mostly + * matters for parsing, where statically defined arrays cannot be + * appended to using the array-of-table syntax. + */ + bool is_inline() const + { + return is_inline_; + } + private: #if defined(CPPTOML_NO_RTTI) - table_array() : base(base_type::TABLE_ARRAY) + table_array(bool is_inline = false) + : base(base_type::TABLE_ARRAY), is_inline_(is_inline) { // nothing } #else - table_array() + table_array(bool is_inline = false) : is_inline_(is_inline) { // nothing } @@ -1169,26 +1187,30 @@ class table_array : public base table_array& operator=(const table_array& rhs) = delete; std::vector<std::shared_ptr<table>> array_; + const bool is_inline_ = false; }; -inline std::shared_ptr<table_array> make_table_array() +inline std::shared_ptr<table_array> make_table_array(bool is_inline) { struct make_shared_enabler : public table_array { - make_shared_enabler() + make_shared_enabler(bool mse_is_inline) : table_array(mse_is_inline) { // nothing } }; - return std::make_shared<make_shared_enabler>(); + return std::make_shared<make_shared_enabler>(is_inline); } +namespace detail +{ template <> inline std::shared_ptr<table_array> make_element<table_array>() { - return make_table_array(); + return make_table_array(true); } +} // namespace detail // The below are overloads for fetching specific value types out of a value // where special casting behavior (like bounds checking) is desired @@ -1679,11 +1701,14 @@ std::shared_ptr<table> make_table() return std::make_shared<make_shared_enabler>(); } +namespace detail +{ template <> inline std::shared_ptr<table> make_element<table>() { return make_table(); } +} // namespace detail template <class T> std::shared_ptr<base> value<T>::clone() const @@ -1702,7 +1727,7 @@ inline std::shared_ptr<base> array::clone() const inline std::shared_ptr<base> table_array::clone() const { - auto result = make_table_array(); + auto result = make_table_array(is_inline()); result->reserve(array_.size()); for (const auto& ptr : array_) result->array_.push_back(ptr->clone()->as_table()); @@ -1738,6 +1763,11 @@ inline bool is_number(char c) return c >= '0' && c <= '9'; } +inline bool is_hex(char c) +{ + return is_number(c) || (c >= 'a' && c <= 'f') || (c >= 'A' && c <= 'F'); +} + /** * Helper object for consuming expected characters. */ @@ -1766,6 +1796,13 @@ class consumer [&](char c) { (*this)(c); }); } + void eat_or(char a, char b) + { + if (it_ == end_ || (*it_ != a && *it_ != b)) + on_error_(); + ++it_; + } + int eat_digits(int len) { int val = 0; @@ -1830,7 +1867,7 @@ inline std::istream& getline(std::istream& input, std::string& line) line.push_back(static_cast<char>(c)); } } -} +} // namespace detail /** * The parser class. @@ -1914,21 +1951,25 @@ class parser std::string full_table_name; bool inserted = false; - while (it != end && *it != ']') - { - auto part = parse_key(it, end, - [](char c) { return c == '.' || c == ']'; }); + auto key_end = [](char c) { return c == ']'; }; + + auto key_part_handler = [&](const std::string& part) { if (part.empty()) throw_parse_exception("Empty component of table name"); if (!full_table_name.empty()) - full_table_name += "."; + full_table_name += '.'; full_table_name += part; if (curr_table->contains(part)) { +#if !defined(__PGI) auto b = curr_table->get(part); +#else + // Workaround for PGI compiler + std::shared_ptr<base> b = curr_table->get(part); +#endif if (b->is_table()) curr_table = static_cast<table*>(b.get()); else if (b->is_table_array()) @@ -1946,16 +1987,23 @@ class parser curr_table->insert(part, make_table()); curr_table = static_cast<table*>(curr_table->get(part).get()); } - consume_whitespace(it, end); - if (it != end && *it == '.') - ++it; - consume_whitespace(it, end); - } + }; + + key_part_handler(parse_key(it, end, key_end, key_part_handler)); if (it == end) throw_parse_exception( "Unterminated table declaration; did you forget a ']'?"); + if (*it != ']') + { + std::string errmsg{"Unexpected character in table definition: "}; + errmsg += '"'; + errmsg += *it; + errmsg += '"'; + throw_parse_exception(errmsg); + } + // table already existed if (!inserted) { @@ -1969,8 +2017,9 @@ class parser // since it has already been defined. If there aren't any // values, then it was implicitly created by something like // [a.b] - if (curr_table->empty() || std::any_of(curr_table->begin(), - curr_table->end(), is_value)) + if (curr_table->empty() + || std::any_of(curr_table->begin(), curr_table->end(), + is_value)) { throw_parse_exception("Redefinition of table " + full_table_name); @@ -1989,36 +2038,45 @@ class parser if (it == end || *it == ']') throw_parse_exception("Table array name cannot be empty"); - std::string full_ta_name; - while (it != end && *it != ']') - { - auto part = parse_key(it, end, - [](char c) { return c == '.' || c == ']'; }); + auto key_end = [](char c) { return c == ']'; }; + std::string full_ta_name; + auto key_part_handler = [&](const std::string& part) { if (part.empty()) throw_parse_exception("Empty component of table array name"); if (!full_ta_name.empty()) - full_ta_name += "."; + full_ta_name += '.'; full_ta_name += part; - consume_whitespace(it, end); - if (it != end && *it == '.') - ++it; - consume_whitespace(it, end); - if (curr_table->contains(part)) { +#if !defined(__PGI) auto b = curr_table->get(part); +#else + // Workaround for PGI compiler + std::shared_ptr<base> b = curr_table->get(part); +#endif // if this is the end of the table array name, add an - // element to the table array that we just looked up + // element to the table array that we just looked up, + // provided it was not declared inline if (it != end && *it == ']') { if (!b->is_table_array()) + { throw_parse_exception("Key " + full_ta_name + " is not a table array"); + } + auto v = b->as_table_array(); + + if (v->is_inline()) + { + throw_parse_exception("Static array " + full_ta_name + + " cannot be appended to"); + } + v->get().push_back(make_table()); curr_table = v->get().back().get(); } @@ -2059,15 +2117,16 @@ class parser = static_cast<table*>(curr_table->get(part).get()); } } - } + }; + + key_part_handler(parse_key(it, end, key_end, key_part_handler)); // consume the last "]]" - if (it == end) - throw_parse_exception("Unterminated table array name"); - ++it; - if (it == end) + auto eat = make_consumer(it, end, [this]() { throw_parse_exception("Unterminated table array name"); - ++it; + }); + eat(']'); + eat(']'); consume_whitespace(it, end); eol_or_comment(it, end); @@ -2076,7 +2135,35 @@ class parser void parse_key_value(std::string::iterator& it, std::string::iterator& end, table* curr_table) { - auto key = parse_key(it, end, [](char c) { return c == '='; }); + auto key_end = [](char c) { return c == '='; }; + + auto key_part_handler = [&](const std::string& part) { + // two cases: this key part exists already, in which case it must + // be a table, or it doesn't exist in which case we must create + // an implicitly defined table + if (curr_table->contains(part)) + { + auto val = curr_table->get(part); + if (val->is_table()) + { + curr_table = static_cast<table*>(val.get()); + } + else + { + throw_parse_exception("Key " + part + + " already exists as a value"); + } + } + else + { + auto newtable = make_table(); + curr_table->insert(part, newtable); + curr_table = newtable.get(); + } + }; + + auto key = parse_key(it, end, key_end, key_part_handler); + if (curr_table->contains(key)) throw_parse_exception("Key " + key + " already present"); if (it == end || *it != '=') @@ -2087,18 +2174,57 @@ class parser consume_whitespace(it, end); } - template <class Function> - std::string parse_key(std::string::iterator& it, - const std::string::iterator& end, Function&& fun) + template <class KeyEndFinder, class KeyPartHandler> + std::string + parse_key(std::string::iterator& it, const std::string::iterator& end, + KeyEndFinder&& key_end, KeyPartHandler&& key_part_handler) + { + // parse the key as a series of one or more simple-keys joined with '.' + while (it != end && !key_end(*it)) + { + auto part = parse_simple_key(it, end); + consume_whitespace(it, end); + + if (it == end || key_end(*it)) + { + return part; + } + + if (*it != '.') + { + std::string errmsg{"Unexpected character in key: "}; + errmsg += '"'; + errmsg += *it; + errmsg += '"'; + throw_parse_exception(errmsg); + } + + key_part_handler(part); + + // consume the dot + ++it; + } + + throw_parse_exception("Unexpected end of key"); + } + + std::string parse_simple_key(std::string::iterator& it, + const std::string::iterator& end) { consume_whitespace(it, end); - if (*it == '"') + + if (it == end) + throw_parse_exception("Unexpected end of key (blank key?)"); + + if (*it == '"' || *it == '\'') { - return parse_quoted_key(it, end); + return string_literal(it, end, *it); } else { - auto bke = std::find_if(it, end, std::forward<Function>(fun)); + auto bke = std::find_if(it, end, [](char c) { + return c == '.' || c == '=' || c == ']'; + }); return parse_bare_key(it, bke); } } @@ -2142,12 +2268,6 @@ class parser return key; } - std::string parse_quoted_key(std::string::iterator& it, - const std::string::iterator& end) - { - return string_literal(it, end, '"'); - } - enum class parse_type { STRING = 1, @@ -2193,7 +2313,7 @@ class parser parse_type determine_value_type(const std::string::iterator& it, const std::string::iterator& end) { - if(it == end) + if (it == end) { throw_parse_exception("Failed to parse value type"); } @@ -2209,7 +2329,11 @@ class parser { return *dtype; } - else if (is_number(*it) || *it == '-' || *it == '+') + else if (is_number(*it) || *it == '-' || *it == '+' + || (*it == 'i' && it + 1 != end && it[1] == 'n' + && it + 2 != end && it[2] == 'f') + || (*it == 'n' && it + 1 != end && it[1] == 'a' + && it + 2 != end && it[2] == 'n')) { return determine_number_type(it, end); } @@ -2235,6 +2359,13 @@ class parser auto check_it = it; if (*check_it == '-' || *check_it == '+') ++check_it; + + if (check_it == end) + throw_parse_exception("Malformed number"); + + if (*check_it == 'i' || *check_it == 'n') + return parse_type::FLOAT; + while (check_it != end && is_number(*check_it)) ++check_it; if (check_it != end && *check_it == '.') @@ -2283,57 +2414,56 @@ class parser bool consuming = false; std::shared_ptr<value<std::string>> ret; - auto handle_line - = [&](std::string::iterator& local_it, - std::string::iterator& local_end) { - if (consuming) - { - local_it = std::find_if_not(local_it, local_end, is_ws); - - // whole line is whitespace - if (local_it == local_end) - return; - } - - consuming = false; - - while (local_it != local_end) - { - // handle escaped characters - if (delim == '"' && *local_it == '\\') - { - auto check = local_it; - // check if this is an actual escape sequence or a - // whitespace escaping backslash - ++check; - consume_whitespace(check, local_end); - if (check == local_end) - { - consuming = true; - break; - } - - ss << parse_escape_code(local_it, local_end); - continue; - } - - // if we can end the string - if (std::distance(local_it, local_end) >= 3) - { - auto check = local_it; - // check for """ - if (*check++ == delim && *check++ == delim - && *check++ == delim) - { - local_it = check; - ret = make_value<std::string>(ss.str()); - break; - } - } - - ss << *local_it++; - } - }; + auto handle_line = [&](std::string::iterator& local_it, + std::string::iterator& local_end) { + if (consuming) + { + local_it = std::find_if_not(local_it, local_end, is_ws); + + // whole line is whitespace + if (local_it == local_end) + return; + } + + consuming = false; + + while (local_it != local_end) + { + // handle escaped characters + if (delim == '"' && *local_it == '\\') + { + auto check = local_it; + // check if this is an actual escape sequence or a + // whitespace escaping backslash + ++check; + consume_whitespace(check, local_end); + if (check == local_end) + { + consuming = true; + break; + } + + ss << parse_escape_code(local_it, local_end); + continue; + } + + // if we can end the string + if (std::distance(local_it, local_end) >= 3) + { + auto check = local_it; + // check for """ + if (*check++ == delim && *check++ == delim + && *check++ == delim) + { + local_it = check; + ret = make_value<std::string>(ss.str()); + break; + } + } + + ss << *local_it++; + } + }; // handle the remainder of the current line handle_line(it, end); @@ -2514,17 +2644,13 @@ class parser return value; } - bool is_hex(char c) - { - return is_number(c) || (c >= 'a' && c <= 'f') || (c >= 'A' && c <= 'F'); - } - uint32_t hex_to_digit(char c) { if (is_number(c)) return static_cast<uint32_t>(c - '0'); - return 10 + static_cast<uint32_t>( - c - ((c >= 'a' && c <= 'f') ? 'a' : 'A')); + return 10 + + static_cast<uint32_t>(c + - ((c >= 'a' && c <= 'f') ? 'a' : 'A')); } std::shared_ptr<base> parse_number(std::string::iterator& it, @@ -2538,17 +2664,23 @@ class parser ++check_it; }; - eat_sign(); + auto check_no_leading_zero = [&]() { + if (check_it != end && *check_it == '0' && check_it + 1 != check_end + && check_it[1] != '.') + { + throw_parse_exception("Numbers may not have leading zeros"); + } + }; - auto eat_numbers = [&]() { + auto eat_digits = [&](bool (*check_char)(char)) { auto beg = check_it; - while (check_it != end && is_number(*check_it)) + while (check_it != end && check_char(*check_it)) { ++check_it; if (check_it != end && *check_it == '_') { ++check_it; - if (check_it == end || !is_number(*check_it)) + if (check_it == end || !check_char(*check_it)) throw_parse_exception("Malformed number"); } } @@ -2557,15 +2689,63 @@ class parser throw_parse_exception("Malformed number"); }; - auto check_no_leading_zero = [&]() { - if (check_it != end && *check_it == '0' && check_it + 1 != check_end - && check_it[1] != '.') + auto eat_hex = [&]() { eat_digits(&is_hex); }; + + auto eat_numbers = [&]() { eat_digits(&is_number); }; + + if (check_it != end && *check_it == '0' && check_it + 1 != check_end + && (check_it[1] == 'x' || check_it[1] == 'o' || check_it[1] == 'b')) + { + ++check_it; + char base = *check_it; + ++check_it; + if (base == 'x') { - throw_parse_exception("Numbers may not have leading zeros"); + eat_hex(); + return parse_int(it, check_it, 16); } - }; + else if (base == 'o') + { + auto start = check_it; + eat_numbers(); + auto val = parse_int(start, check_it, 8, "0"); + it = start; + return val; + } + else // if (base == 'b') + { + auto start = check_it; + eat_numbers(); + auto val = parse_int(start, check_it, 2); + it = start; + return val; + } + } + eat_sign(); check_no_leading_zero(); + + if (check_it != end && check_it + 1 != end && check_it + 2 != end) + { + if (check_it[0] == 'i' && check_it[1] == 'n' && check_it[2] == 'f') + { + auto val = std::numeric_limits<double>::infinity(); + if (*it == '-') + val = -val; + it = check_it + 3; + return make_value(val); + } + else if (check_it[0] == 'n' && check_it[1] == 'a' + && check_it[2] == 'n') + { + auto val = std::numeric_limits<double>::quiet_NaN(); + if (*it == '-') + val = -val; + it = check_it + 3; + return make_value(val); + } + } + eat_numbers(); if (check_it != end @@ -2604,14 +2784,17 @@ class parser } std::shared_ptr<value<int64_t>> parse_int(std::string::iterator& it, - const std::string::iterator& end) + const std::string::iterator& end, + int base = 10, + const char* prefix = "") { std::string v{it, end}; + v = prefix + v; v.erase(std::remove(v.begin(), v.end(), '_'), v.end()); it = end; try { - return make_value<int64_t>(std::stoll(v)); + return make_value<int64_t>(std::stoll(v, nullptr, base)); } catch (const std::invalid_argument& ex) { @@ -2674,18 +2857,33 @@ class parser std::string::iterator find_end_of_number(std::string::iterator it, std::string::iterator end) { - return std::find_if(it, end, [](char c) { + auto ret = std::find_if(it, end, [](char c) { return !is_number(c) && c != '_' && c != '.' && c != 'e' && c != 'E' - && c != '-' && c != '+'; + && c != '-' && c != '+' && c != 'x' && c != 'o' && c != 'b'; }); + if (ret != end && ret + 1 != end && ret + 2 != end) + { + if ((ret[0] == 'i' && ret[1] == 'n' && ret[2] == 'f') + || (ret[0] == 'n' && ret[1] == 'a' && ret[2] == 'n')) + { + ret = ret + 3; + } + } + return ret; } std::string::iterator find_end_of_date(std::string::iterator it, std::string::iterator end) { - return std::find_if(it, end, [](char c) { - return !is_number(c) && c != 'T' && c != 'Z' && c != ':' && c != '-' - && c != '+' && c != '.'; + auto end_of_date = std::find_if(it, end, [](char c) { + return !is_number(c) && c != '-'; + }); + if (end_of_date != end && *end_of_date == ' ' && end_of_date + 1 != end + && is_number(end_of_date[1])) + end_of_date++; + return std::find_if(end_of_date, end, [](char c) { + return !is_number(c) && c != 'T' && c != 'Z' && c != ':' + && c != '-' && c != '+' && c != '.'; }); } @@ -2754,7 +2952,7 @@ class parser if (it == date_end) return make_value(ldate); - eat('T'); + eat.eat_or('T', ' '); local_datetime ldt; static_cast<local_date&>(ldt) = ldate; @@ -2850,9 +3048,9 @@ class parser auto arr = make_array(); while (it != end && *it != ']') { - auto value = parse_value(it, end); - if (auto v = value->as<Value>()) - arr->get().push_back(value); + auto val = parse_value(it, end); + if (auto v = val->as<Value>()) + arr->get().push_back(val); else throw_parse_exception("Arrays must be homogeneous"); skip_whitespace_and_comments(it, end); @@ -2871,7 +3069,7 @@ class parser std::string::iterator& it, std::string::iterator& end) { - auto arr = make_element<Object>(); + auto arr = detail::make_element<Object>(); while (it != end && *it != ']') { @@ -2881,7 +3079,7 @@ class parser arr->get().push_back(((*this).*fun)(it, end)); skip_whitespace_and_comments(it, end); - if (*it != ',') + if (it == end || *it != ',') break; ++it; @@ -2906,8 +3104,11 @@ class parser throw_parse_exception("Unterminated inline table"); consume_whitespace(it, end); - parse_key_value(it, end, tbl.get()); - consume_whitespace(it, end); + if (it != end && *it != '}') + { + parse_key_value(it, end, tbl.get()); + consume_whitespace(it, end); + } } while (*it == ','); if (it == end || *it != '}') @@ -2987,7 +3188,8 @@ class parser if (it[4] != '-' || it[7] != '-') return {}; - if (len >= 19 && it[10] == 'T' && is_time(it + 11, date_end)) + if (len >= 19 && (it[10] == 'T' || it[10] == ' ') + && is_time(it + 11, date_end)) { // datetime type auto time_end = find_end_of_time(it + 11, date_end); @@ -3243,7 +3445,7 @@ class toml_writer { res += "\\\\"; } - else if ((const uint32_t)*it <= 0x001f) + else if (static_cast<uint32_t>(*it) <= UINT32_C(0x001f)) { res += "\\u"; std::stringstream ss; @@ -3274,12 +3476,21 @@ class toml_writer */ void write(const value<double>& v) { - std::ios::fmtflags flags{stream_.flags()}; - - stream_ << std::showpoint; - write(v.get()); - - stream_.flags(flags); + std::stringstream ss; + ss << std::showpoint + << std::setprecision(std::numeric_limits<double>::max_digits10) + << v.get(); + + auto double_str = ss.str(); + auto pos = double_str.find("e0"); + if (pos != std::string::npos) + double_str.replace(pos, 2, "e"); + pos = double_str.find("e-0"); + if (pos != std::string::npos) + double_str.replace(pos, 3, "e-"); + + stream_ << double_str; + has_naked_endline_ = false; } /** @@ -3287,9 +3498,9 @@ class toml_writer * offset_datetime. */ template <class T> - typename std::enable_if<is_one_of<T, int64_t, local_date, local_time, - local_datetime, - offset_datetime>::value>::type + typename std::enable_if< + is_one_of<T, int64_t, local_date, local_time, local_datetime, + offset_datetime>::value>::type write(const value<T>& v) { write(v.get()); @@ -3453,5 +3664,5 @@ inline std::ostream& operator<<(std::ostream& stream, const array& a) a.accept(writer); return stream; } -} -#endif +} // namespace cpptoml +#endif // CPPTOML_H diff --git a/src/libexpr/common-eval-args.cc b/src/libexpr/common-eval-args.cc index 3e0c78f280f7..13950ab8d169 100644 --- a/src/libexpr/common-eval-args.cc +++ b/src/libexpr/common-eval-args.cc @@ -45,9 +45,11 @@ Bindings * MixEvalArgs::getAutoArgs(EvalState & state) Path lookupFileArg(EvalState & state, string s) { - if (isUri(s)) - return getDownloader()->downloadCached(state.store, s, true); - else if (s.size() > 2 && s.at(0) == '<' && s.at(s.size() - 1) == '>') { + if (isUri(s)) { + CachedDownloadRequest request(s); + request.unpack = true; + return getDownloader()->downloadCached(state.store, request).path; + } else if (s.size() > 2 && s.at(0) == '<' && s.at(s.size() - 1) == '>') { Path p = s.substr(1, s.size() - 2); return state.findFile(p); } else diff --git a/src/libexpr/eval.cc b/src/libexpr/eval.cc index ab407e56907c..d8e10d9f20e1 100644 --- a/src/libexpr/eval.cc +++ b/src/libexpr/eval.cc @@ -130,6 +130,16 @@ std::ostream & operator << (std::ostream & str, const Value & v) } +const Value *getPrimOp(const Value &v) { + const Value * primOp = &v; + while (primOp->type == tPrimOpApp) { + primOp = primOp->primOpApp.left; + } + assert(primOp->type == tPrimOp); + return primOp; +} + + string showType(const Value & v) { switch (v.type) { @@ -144,8 +154,10 @@ string showType(const Value & v) case tApp: return "a function application"; case tLambda: return "a function"; case tBlackhole: return "a black hole"; - case tPrimOp: return "a built-in function"; - case tPrimOpApp: return "a partially applied built-in function"; + case tPrimOp: + return fmt("the built-in function '%s'", string(v.primOp->name)); + case tPrimOpApp: + return fmt("the partially applied built-in function '%s'", string(getPrimOp(v)->primOp->name)); case tExternal: return v.external->showType(); case tFloat: return "a float"; } @@ -757,6 +769,7 @@ void EvalState::evalFile(const Path & path_, Value & v) void EvalState::resetFileCache() { fileEvalCache.clear(); + fileParseCache.clear(); } @@ -1798,6 +1811,7 @@ void EvalState::printStats() gc.attr("totalBytes", totalBytes); } #endif + if (countCalls) { { auto obj = topObj.object("primops"); @@ -1833,6 +1847,11 @@ void EvalState::printStats() } } } + + if (getEnv("NIX_SHOW_SYMBOLS", "0") != "0") { + auto list = topObj.list("symbols"); + symbols.dump([&](const std::string & s) { list.elem(s); }); + } } } diff --git a/src/libexpr/eval.hh b/src/libexpr/eval.hh index d0f298e168e9..a314e01e0a71 100644 --- a/src/libexpr/eval.hh +++ b/src/libexpr/eval.hh @@ -81,7 +81,7 @@ public: /* The allowed filesystem paths in restricted or pure evaluation mode. */ - std::experimental::optional<PathSet> allowedPaths; + std::optional<PathSet> allowedPaths; Value vEmptySet; @@ -316,6 +316,9 @@ private: /* Return a string representing the type of the value `v'. */ string showType(const Value & v); +/* Decode a context string โ!<name>!<path>โ into a pair <path, + name>. */ +std::pair<string, string> decodeContext(const string & s); /* If `path' refers to a directory, then append "/default.nix". */ Path resolveExprPath(Path path); diff --git a/src/libexpr/get-drvs.cc b/src/libexpr/get-drvs.cc index d38ed2df3b18..21a4d7917fce 100644 --- a/src/libexpr/get-drvs.cc +++ b/src/libexpr/get-drvs.cc @@ -295,7 +295,7 @@ static bool getDerivation(EvalState & state, Value & v, } -std::experimental::optional<DrvInfo> getDerivation(EvalState & state, Value & v, +std::optional<DrvInfo> getDerivation(EvalState & state, Value & v, bool ignoreAssertionFailures) { Done done; diff --git a/src/libexpr/get-drvs.hh b/src/libexpr/get-drvs.hh index daaa635fe1b1..d7860fc6a4bc 100644 --- a/src/libexpr/get-drvs.hh +++ b/src/libexpr/get-drvs.hh @@ -78,7 +78,7 @@ typedef list<DrvInfo> DrvInfos; /* If value `v' denotes a derivation, return a DrvInfo object describing it. Otherwise return nothing. */ -std::experimental::optional<DrvInfo> getDerivation(EvalState & state, +std::optional<DrvInfo> getDerivation(EvalState & state, Value & v, bool ignoreAssertionFailures); void getDerivations(EvalState & state, Value & v, const string & pathPrefix, diff --git a/src/libexpr/local.mk b/src/libexpr/local.mk index daa3258f0d3c..ccd5293e4e5e 100644 --- a/src/libexpr/local.mk +++ b/src/libexpr/local.mk @@ -6,7 +6,7 @@ libexpr_DIR := $(d) libexpr_SOURCES := $(wildcard $(d)/*.cc) $(wildcard $(d)/primops/*.cc) $(d)/lexer-tab.cc $(d)/parser-tab.cc -libexpr_LIBS = libutil libstore libformat +libexpr_LIBS = libutil libstore libexpr_LDFLAGS = ifneq ($(OS), FreeBSD) diff --git a/src/libexpr/nix-expr.pc.in b/src/libexpr/nix-expr.pc.in index 79f3e2f4506e..80f7a492b1a1 100644 --- a/src/libexpr/nix-expr.pc.in +++ b/src/libexpr/nix-expr.pc.in @@ -7,4 +7,4 @@ Description: Nix Package Manager Version: @PACKAGE_VERSION@ Requires: nix-store bdw-gc Libs: -L${libdir} -lnixexpr -Cflags: -I${includedir}/nix -std=c++14 +Cflags: -I${includedir}/nix -std=c++17 diff --git a/src/libexpr/parser.y b/src/libexpr/parser.y index cbd576d7d126..967c88d9bc80 100644 --- a/src/libexpr/parser.y +++ b/src/libexpr/parser.y @@ -1,7 +1,7 @@ %glr-parser %pure-parser %locations -%error-verbose +%define parse.error verbose %defines /* %no-lines */ %parse-param { void * scanner } @@ -81,6 +81,8 @@ static void addAttr(ExprAttrs * attrs, AttrPath & attrPath, AttrPath::iterator i; // All attrpaths have at least one attr assert(!attrPath.empty()); + // Checking attrPath validity. + // =========================== for (i = attrPath.begin(); i + 1 < attrPath.end(); i++) { if (i->symbol.set()) { ExprAttrs::AttrDefs::iterator j = attrs->attrs.find(i->symbol); @@ -102,11 +104,29 @@ static void addAttr(ExprAttrs * attrs, AttrPath & attrPath, attrs = nested; } } + // Expr insertion. + // ========================== if (i->symbol.set()) { ExprAttrs::AttrDefs::iterator j = attrs->attrs.find(i->symbol); if (j != attrs->attrs.end()) { - dupAttr(attrPath, pos, j->second.pos); + // This attr path is already defined. However, if both + // e and the expr pointed by the attr path are two attribute sets, + // we want to merge them. + // Otherwise, throw an error. + auto ae = dynamic_cast<ExprAttrs *>(e); + auto jAttrs = dynamic_cast<ExprAttrs *>(j->second.e); + if (jAttrs && ae) { + for (auto & ad : ae->attrs) { + auto j2 = jAttrs->attrs.find(ad.first); + if (j2 != jAttrs->attrs.end()) // Attr already defined in iAttrs, error. + dupAttr(ad.first, j2->second.pos, ad.second.pos); + jAttrs->attrs[ad.first] = ad.second; + } + } else { + dupAttr(attrPath, pos, j->second.pos); + } } else { + // This attr path is not defined. Let's create it. attrs->attrs[i->symbol] = ExprAttrs::AttrDef(e, pos); e->setName(i->symbol); } @@ -657,7 +677,9 @@ std::pair<bool, std::string> EvalState::resolveSearchPathElem(const SearchPathEl if (isUri(elem.second)) { try { - res = { true, getDownloader()->downloadCached(store, elem.second, true) }; + CachedDownloadRequest request(elem.second); + request.unpack = true; + res = { true, getDownloader()->downloadCached(store, request).path }; } catch (DownloadError & e) { printError(format("warning: Nix search path entry '%1%' cannot be downloaded, ignoring") % elem.second); res = { false, "" }; diff --git a/src/libexpr/primops.cc b/src/libexpr/primops.cc index 7372134e2c9b..070e72f3a966 100644 --- a/src/libexpr/primops.cc +++ b/src/libexpr/primops.cc @@ -315,6 +315,12 @@ static void prim_isBool(EvalState & state, const Pos & pos, Value * * args, Valu mkBool(v, args[0]->type == tBool); } +/* Determine whether the argument is a path. */ +static void prim_isPath(EvalState & state, const Pos & pos, Value * * args, Value & v) +{ + state.forceValue(*args[0]); + mkBool(v, args[0]->type == tPath); +} struct CompareValues { @@ -555,7 +561,7 @@ static void prim_derivationStrict(EvalState & state, const Pos & pos, Value * * PathSet context; - std::experimental::optional<std::string> outputHash; + std::optional<std::string> outputHash; std::string outputHashAlgo; bool outputHashRecursive = false; @@ -687,21 +693,12 @@ static void prim_derivationStrict(EvalState & state, const Pos & pos, Value * * } } - /* See prim_unsafeDiscardOutputDependency. */ - else if (path.at(0) == '~') - drv.inputSrcs.insert(string(path, 1)); - /* Handle derivation outputs of the form โ!<name>!<path>โ. */ else if (path.at(0) == '!') { std::pair<string, string> ctx = decodeContext(path); drv.inputDrvs[ctx.first].insert(ctx.second); } - /* Handle derivation contexts returned by - โbuiltins.storePathโ. */ - else if (isDerivation(path)) - drv.inputDrvs[path] = state.store->queryDerivationOutputNames(path); - /* Otherwise it's a source file. */ else drv.inputSrcs.insert(path); @@ -724,16 +721,14 @@ static void prim_derivationStrict(EvalState & state, const Pos & pos, Value * * if (outputs.size() != 1 || *(outputs.begin()) != "out") throw Error(format("multiple outputs are not supported in fixed-output derivations, at %1%") % posDrvName); - HashType ht = parseHashType(outputHashAlgo); - if (ht == htUnknown) - throw EvalError(format("unknown hash algorithm '%1%', at %2%") % outputHashAlgo % posDrvName); + HashType ht = outputHashAlgo.empty() ? htUnknown : parseHashType(outputHashAlgo); Hash h(*outputHash, ht); - outputHash = h.to_string(Base16, false); - if (outputHashRecursive) outputHashAlgo = "r:" + outputHashAlgo; Path outPath = state.store->makeFixedOutputPath(outputHashRecursive, h, drvName); if (!jsonObject) drv.env["out"] = outPath; - drv.outputs["out"] = DerivationOutput(outPath, outputHashAlgo, *outputHash); + drv.outputs["out"] = DerivationOutput(outPath, + (outputHashRecursive ? "r:" : "") + printHashType(h.type), + h.to_string(Base16, false)); } else { @@ -928,6 +923,20 @@ static void prim_findFile(EvalState & state, const Pos & pos, Value * * args, Va mkPath(v, state.checkSourcePath(state.findFile(searchPath, path, pos)).c_str()); } +/* Return the cryptographic hash of a file in base-16. */ +static void prim_hashFile(EvalState & state, const Pos & pos, Value * * args, Value & v) +{ + string type = state.forceStringNoCtx(*args[0], pos); + HashType ht = parseHashType(type); + if (ht == htUnknown) + throw Error(format("unknown hash type '%1%', at %2%") % type % pos); + + PathSet context; // discarded + Path p = state.coerceToPath(pos, *args[1], context); + + mkString(v, hashFile(ht, state.checkSourcePath(p)).to_string(Base16, false), context); +} + /* Read a directory (without . or ..) */ static void prim_readDir(EvalState & state, const Pos & pos, Value * * args, Value & v) { @@ -1006,13 +1015,8 @@ static void prim_toFile(EvalState & state, const Pos & pos, Value * * args, Valu PathSet refs; for (auto path : context) { - if (path.at(0) == '=') path = string(path, 1); - if (isDerivation(path)) { - /* See prim_unsafeDiscardOutputDependency. */ - if (path.at(0) != '~') - throw EvalError(format("in 'toFile': the file '%1%' cannot refer to derivation outputs, at %2%") % name % pos); - path = string(path, 1); - } + if (path.at(0) != '/') + throw EvalError(format("in 'toFile': the file '%1%' cannot refer to derivation outputs, at %2%") % name % pos); refs.insert(path); } @@ -1796,41 +1800,6 @@ static void prim_stringLength(EvalState & state, const Pos & pos, Value * * args } -static void prim_unsafeDiscardStringContext(EvalState & state, const Pos & pos, Value * * args, Value & v) -{ - PathSet context; - string s = state.coerceToString(pos, *args[0], context); - mkString(v, s, PathSet()); -} - - -static void prim_hasContext(EvalState & state, const Pos & pos, Value * * args, Value & v) -{ - PathSet context; - state.forceString(*args[0], context, pos); - mkBool(v, !context.empty()); -} - - -/* Sometimes we want to pass a derivation path (i.e. pkg.drvPath) to a - builder without causing the derivation to be built (for instance, - in the derivation that builds NARs in nix-push, when doing - source-only deployment). This primop marks the string context so - that builtins.derivation adds the path to drv.inputSrcs rather than - drv.inputDrvs. */ -static void prim_unsafeDiscardOutputDependency(EvalState & state, const Pos & pos, Value * * args, Value & v) -{ - PathSet context; - string s = state.coerceToString(pos, *args[0], context); - - PathSet context2; - for (auto & p : context) - context2.insert(p.at(0) == '=' ? "~" + string(p, 1) : p); - - mkString(v, s, context2); -} - - /* Return the cryptographic hash of a string in base-16. */ static void prim_hashString(EvalState & state, const Pos & pos, Value * * args, Value & v) { @@ -2081,9 +2050,9 @@ static void prim_splitVersion(EvalState & state, const Pos & pos, Value * * args void fetch(EvalState & state, const Pos & pos, Value * * args, Value & v, const string & who, bool unpack, const std::string & defaultName) { - string url; - Hash expectedHash; - string name = defaultName; + CachedDownloadRequest request(""); + request.unpack = unpack; + request.name = defaultName; state.forceValue(*args[0]); @@ -2094,27 +2063,27 @@ void fetch(EvalState & state, const Pos & pos, Value * * args, Value & v, for (auto & attr : *args[0]->attrs) { string n(attr.name); if (n == "url") - url = state.forceStringNoCtx(*attr.value, *attr.pos); + request.uri = state.forceStringNoCtx(*attr.value, *attr.pos); else if (n == "sha256") - expectedHash = Hash(state.forceStringNoCtx(*attr.value, *attr.pos), htSHA256); + request.expectedHash = Hash(state.forceStringNoCtx(*attr.value, *attr.pos), htSHA256); else if (n == "name") - name = state.forceStringNoCtx(*attr.value, *attr.pos); + request.name = state.forceStringNoCtx(*attr.value, *attr.pos); else throw EvalError(format("unsupported argument '%1%' to '%2%', at %3%") % attr.name % who % attr.pos); } - if (url.empty()) + if (request.uri.empty()) throw EvalError(format("'url' argument required, at %1%") % pos); } else - url = state.forceStringNoCtx(*args[0], pos); + request.uri = state.forceStringNoCtx(*args[0], pos); - state.checkURI(url); + state.checkURI(request.uri); - if (evalSettings.pureEval && !expectedHash) + if (evalSettings.pureEval && !request.expectedHash) throw Error("in pure evaluation mode, '%s' requires a 'sha256' argument", who); - Path res = getDownloader()->downloadCached(state.store, url, unpack, name, expectedHash); + Path res = getDownloader()->downloadCached(state.store, request).path; if (state.allowedPaths) state.allowedPaths->insert(res); @@ -2220,6 +2189,7 @@ void EvalState::createBaseEnv() addPrimOp("__isInt", 1, prim_isInt); addPrimOp("__isFloat", 1, prim_isFloat); addPrimOp("__isBool", 1, prim_isBool); + addPrimOp("__isPath", 1, prim_isPath); addPrimOp("__genericClosure", 1, prim_genericClosure); addPrimOp("abort", 1, prim_abort); addPrimOp("__addErrorContext", 2, prim_addErrorContext); @@ -2246,6 +2216,7 @@ void EvalState::createBaseEnv() addPrimOp("__readFile", 1, prim_readFile); addPrimOp("__readDir", 1, prim_readDir); addPrimOp("__findFile", 2, prim_findFile); + addPrimOp("__hashFile", 2, prim_hashFile); // Creating files addPrimOp("__toXML", 1, prim_toXML); @@ -2301,9 +2272,6 @@ void EvalState::createBaseEnv() addPrimOp("toString", 1, prim_toString); addPrimOp("__substring", 3, prim_substring); addPrimOp("__stringLength", 1, prim_stringLength); - addPrimOp("__hasContext", 1, prim_hasContext); - addPrimOp("__unsafeDiscardStringContext", 1, prim_unsafeDiscardStringContext); - addPrimOp("__unsafeDiscardOutputDependency", 1, prim_unsafeDiscardOutputDependency); addPrimOp("__hashString", 2, prim_hashString); addPrimOp("__match", 2, prim_match); addPrimOp("__split", 2, prim_split); diff --git a/src/libexpr/primops/context.cc b/src/libexpr/primops/context.cc new file mode 100644 index 000000000000..2d79739ea047 --- /dev/null +++ b/src/libexpr/primops/context.cc @@ -0,0 +1,187 @@ +#include "primops.hh" +#include "eval-inline.hh" +#include "derivations.hh" + +namespace nix { + +static void prim_unsafeDiscardStringContext(EvalState & state, const Pos & pos, Value * * args, Value & v) +{ + PathSet context; + string s = state.coerceToString(pos, *args[0], context); + mkString(v, s, PathSet()); +} + +static RegisterPrimOp r1("__unsafeDiscardStringContext", 1, prim_unsafeDiscardStringContext); + + +static void prim_hasContext(EvalState & state, const Pos & pos, Value * * args, Value & v) +{ + PathSet context; + state.forceString(*args[0], context, pos); + mkBool(v, !context.empty()); +} + +static RegisterPrimOp r2("__hasContext", 1, prim_hasContext); + + +/* Sometimes we want to pass a derivation path (i.e. pkg.drvPath) to a + builder without causing the derivation to be built (for instance, + in the derivation that builds NARs in nix-push, when doing + source-only deployment). This primop marks the string context so + that builtins.derivation adds the path to drv.inputSrcs rather than + drv.inputDrvs. */ +static void prim_unsafeDiscardOutputDependency(EvalState & state, const Pos & pos, Value * * args, Value & v) +{ + PathSet context; + string s = state.coerceToString(pos, *args[0], context); + + PathSet context2; + for (auto & p : context) + context2.insert(p.at(0) == '=' ? string(p, 1) : p); + + mkString(v, s, context2); +} + +static RegisterPrimOp r3("__unsafeDiscardOutputDependency", 1, prim_unsafeDiscardOutputDependency); + + +/* Extract the context of a string as a structured Nix value. + + The context is represented as an attribute set whose keys are the + paths in the context set and whose values are attribute sets with + the following keys: + path: True if the relevant path is in the context as a plain store + path (i.e. the kind of context you get when interpolating + a Nix path (e.g. ./.) into a string). False if missing. + allOutputs: True if the relevant path is a derivation and it is + in the context as a drv file with all of its outputs + (i.e. the kind of context you get when referencing + .drvPath of some derivation). False if missing. + outputs: If a non-empty list, the relevant path is a derivation + and the provided outputs are referenced in the context + (i.e. the kind of context you get when referencing + .outPath of some derivation). Empty list if missing. + Note that for a given path any combination of the above attributes + may be present. +*/ +static void prim_getContext(EvalState & state, const Pos & pos, Value * * args, Value & v) +{ + struct ContextInfo { + bool path = false; + bool allOutputs = false; + Strings outputs; + }; + PathSet context; + state.forceString(*args[0], context, pos); + auto contextInfos = std::map<Path, ContextInfo>(); + for (const auto & p : context) { + Path drv; + string output; + const Path * path = &p; + if (p.at(0) == '=') { + drv = string(p, 1); + path = &drv; + } else if (p.at(0) == '!') { + std::pair<string, string> ctx = decodeContext(p); + drv = ctx.first; + output = ctx.second; + path = &drv; + } + auto isPath = drv.empty(); + auto isAllOutputs = (!drv.empty()) && output.empty(); + + auto iter = contextInfos.find(*path); + if (iter == contextInfos.end()) { + contextInfos.emplace(*path, ContextInfo{isPath, isAllOutputs, output.empty() ? Strings{} : Strings{std::move(output)}}); + } else { + if (isPath) + iter->second.path = true; + else if (isAllOutputs) + iter->second.allOutputs = true; + else + iter->second.outputs.emplace_back(std::move(output)); + } + } + + state.mkAttrs(v, contextInfos.size()); + + auto sPath = state.symbols.create("path"); + auto sAllOutputs = state.symbols.create("allOutputs"); + for (const auto & info : contextInfos) { + auto & infoVal = *state.allocAttr(v, state.symbols.create(info.first)); + state.mkAttrs(infoVal, 3); + if (info.second.path) + mkBool(*state.allocAttr(infoVal, sPath), true); + if (info.second.allOutputs) + mkBool(*state.allocAttr(infoVal, sAllOutputs), true); + if (!info.second.outputs.empty()) { + auto & outputsVal = *state.allocAttr(infoVal, state.sOutputs); + state.mkList(outputsVal, info.second.outputs.size()); + size_t i = 0; + for (const auto & output : info.second.outputs) { + mkString(*(outputsVal.listElems()[i++] = state.allocValue()), output); + } + } + infoVal.attrs->sort(); + } + v.attrs->sort(); +} + +static RegisterPrimOp r4("__getContext", 1, prim_getContext); + + +/* Append the given context to a given string. + + See the commentary above unsafeGetContext for details of the + context representation. +*/ +static void prim_appendContext(EvalState & state, const Pos & pos, Value * * args, Value & v) +{ + PathSet context; + auto orig = state.forceString(*args[0], context, pos); + + state.forceAttrs(*args[1], pos); + + auto sPath = state.symbols.create("path"); + auto sAllOutputs = state.symbols.create("allOutputs"); + for (auto & i : *args[1]->attrs) { + if (!state.store->isStorePath(i.name)) + throw EvalError("Context key '%s' is not a store path, at %s", i.name, i.pos); + if (!settings.readOnlyMode) + state.store->ensurePath(i.name); + state.forceAttrs(*i.value, *i.pos); + auto iter = i.value->attrs->find(sPath); + if (iter != i.value->attrs->end()) { + if (state.forceBool(*iter->value, *iter->pos)) + context.insert(i.name); + } + + iter = i.value->attrs->find(sAllOutputs); + if (iter != i.value->attrs->end()) { + if (state.forceBool(*iter->value, *iter->pos)) { + if (!isDerivation(i.name)) { + throw EvalError("Tried to add all-outputs context of %s, which is not a derivation, to a string, at %s", i.name, i.pos); + } + context.insert("=" + string(i.name)); + } + } + + iter = i.value->attrs->find(state.sOutputs); + if (iter != i.value->attrs->end()) { + state.forceList(*iter->value, *iter->pos); + if (iter->value->listSize() && !isDerivation(i.name)) { + throw EvalError("Tried to add derivation output context of %s, which is not a derivation, to a string, at %s", i.name, i.pos); + } + for (unsigned int n = 0; n < iter->value->listSize(); ++n) { + auto name = state.forceStringNoCtx(*iter->value->listElems()[n], *iter->pos); + context.insert("!" + name + "!" + string(i.name)); + } + } + } + + mkString(v, orig, context); +} + +static RegisterPrimOp r5("__appendContext", 2, prim_appendContext); + +} diff --git a/src/libexpr/primops/fetchGit.cc b/src/libexpr/primops/fetchGit.cc index 588b0fa4d53b..6229fef8d02e 100644 --- a/src/libexpr/primops/fetchGit.cc +++ b/src/libexpr/primops/fetchGit.cc @@ -26,7 +26,7 @@ struct GitInfo std::regex revRegex("^[0-9a-fA-F]{40}$"); GitInfo exportGit(ref<Store> store, const std::string & uri, - std::experimental::optional<std::string> ref, std::string rev, + std::optional<std::string> ref, std::string rev, const std::string & name) { if (evalSettings.pureEval && rev == "") @@ -120,7 +120,7 @@ GitInfo exportGit(ref<Store> store, const std::string & uri, git fetch to update the local ref to the remote ref. */ struct stat st; doFetch = stat(localRefFile.c_str(), &st) != 0 || - st.st_mtime + settings.tarballTtl <= now; + (uint64_t) st.st_mtime + settings.tarballTtl <= (uint64_t) now; } if (doFetch) { @@ -194,7 +194,7 @@ GitInfo exportGit(ref<Store> store, const std::string & uri, static void prim_fetchGit(EvalState & state, const Pos & pos, Value * * args, Value & v) { std::string url; - std::experimental::optional<std::string> ref; + std::optional<std::string> ref; std::string rev; std::string name = "source"; PathSet context; @@ -239,7 +239,7 @@ static void prim_fetchGit(EvalState & state, const Pos & pos, Value * * args, Va v.attrs->sort(); if (state.allowedPaths) - state.allowedPaths->insert(gitInfo.storePath); + state.allowedPaths->insert(state.store->toRealPath(gitInfo.storePath)); } static RegisterPrimOp r("fetchGit", 1, prim_fetchGit); diff --git a/src/libexpr/primops/fetchMercurial.cc b/src/libexpr/primops/fetchMercurial.cc index 66f49f374321..a907d0e1cd82 100644 --- a/src/libexpr/primops/fetchMercurial.cc +++ b/src/libexpr/primops/fetchMercurial.cc @@ -80,7 +80,7 @@ HgInfo exportMercurial(ref<Store> store, const std::string & uri, time_t now = time(0); struct stat st; if (stat(stampFile.c_str(), &st) != 0 || - st.st_mtime + settings.tarballTtl <= now) + (uint64_t) st.st_mtime + settings.tarballTtl <= (uint64_t) now) { /* Except that if this is a commit hash that we already have, we don't have to pull again. */ @@ -96,17 +96,14 @@ HgInfo exportMercurial(ref<Store> store, const std::string & uri, try { runProgram("hg", true, { "pull", "-R", cacheDir, "--", uri }); } - catch (ExecError & e){ + catch (ExecError & e) { string transJournal = cacheDir + "/.hg/store/journal"; /* hg throws "abandoned transaction" error only if this file exists */ - if (pathExists(transJournal)) - { + if (pathExists(transJournal)) { runProgram("hg", true, { "recover", "-R", cacheDir }); runProgram("hg", true, { "pull", "-R", cacheDir, "--", uri }); - } - else - { - throw ExecError(e.status, fmt("program hg '%1%' ", statusToString(e.status))); + } else { + throw ExecError(e.status, fmt("'hg pull' %s", statusToString(e.status))); } } } else { @@ -214,7 +211,7 @@ static void prim_fetchMercurial(EvalState & state, const Pos & pos, Value * * ar v.attrs->sort(); if (state.allowedPaths) - state.allowedPaths->insert(hgInfo.storePath); + state.allowedPaths->insert(state.store->toRealPath(hgInfo.storePath)); } static RegisterPrimOp r("fetchMercurial", 1, prim_fetchMercurial); diff --git a/src/libexpr/primops/fromTOML.cc b/src/libexpr/primops/fromTOML.cc index 4128de05d0cf..a84e569e944d 100644 --- a/src/libexpr/primops/fromTOML.cc +++ b/src/libexpr/primops/fromTOML.cc @@ -49,6 +49,19 @@ static void prim_fromTOML(EvalState & state, const Pos & pos, Value * * args, Va visit(*(v.listElems()[i] = state.allocValue()), t2->get()[i]); } + // Handle cases like 'a = [[{ a = true }]]', which IMHO should be + // parsed as a array containing an array containing a table, + // but instead are parsed as an array containing a table array + // containing a table. + else if (auto t2 = t->as_table_array()) { + size_t size = t2->get().size(); + + state.mkList(v, size); + + for (size_t j = 0; j < size; ++j) + visit(*(v.listElems()[j] = state.allocValue()), t2->get()[j]); + } + else if (t->is_value()) { if (auto val = t->as<int64_t>()) mkInt(v, val->get()); diff --git a/src/libexpr/symbol-table.hh b/src/libexpr/symbol-table.hh index 44929f7eea06..91faea122ce1 100644 --- a/src/libexpr/symbol-table.hh +++ b/src/libexpr/symbol-table.hh @@ -75,6 +75,13 @@ public: } size_t totalSize() const; + + template<typename T> + void dump(T callback) + { + for (auto & s : symbols) + callback(s); + } }; } diff --git a/src/libmain/common-args.cc b/src/libmain/common-args.cc index 4c35a4199590..9e1d7cee60e6 100644 --- a/src/libmain/common-args.cc +++ b/src/libmain/common-args.cc @@ -35,6 +35,15 @@ MixCommonArgs::MixCommonArgs(const string & programName) } }); + mkFlag() + .longName("max-jobs") + .shortName('j') + .label("jobs") + .description("maximum number of parallel builds") + .handler([=](std::string s) { + settings.set("max-jobs", s); + }); + std::string cat = "config"; globalConfig.convertToArgs(*this, cat); diff --git a/src/libmain/local.mk b/src/libmain/local.mk index f1fd3eb72424..0c80f5a0a037 100644 --- a/src/libmain/local.mk +++ b/src/libmain/local.mk @@ -8,7 +8,7 @@ libmain_SOURCES := $(wildcard $(d)/*.cc) libmain_LDFLAGS = $(OPENSSL_LIBS) -libmain_LIBS = libstore libutil libformat +libmain_LIBS = libstore libutil libmain_ALLOW_UNDEFINED = 1 diff --git a/src/libmain/nix-main.pc.in b/src/libmain/nix-main.pc.in index 38bc85c484eb..37b03dcd42c0 100644 --- a/src/libmain/nix-main.pc.in +++ b/src/libmain/nix-main.pc.in @@ -6,4 +6,4 @@ Name: Nix Description: Nix Package Manager Version: @PACKAGE_VERSION@ Libs: -L${libdir} -lnixmain -Cflags: -I${includedir}/nix -std=c++14 +Cflags: -I${includedir}/nix -std=c++17 diff --git a/src/libmain/shared.cc b/src/libmain/shared.cc index 4ed34e54dc55..cd752f4678a0 100644 --- a/src/libmain/shared.cc +++ b/src/libmain/shared.cc @@ -175,10 +175,6 @@ LegacyArgs::LegacyArgs(const std::string & programName, .description("build from source if substitution fails") .set(&(bool&) settings.tryFallback, true); - mkFlag1('j', "max-jobs", "jobs", "maximum number of parallel builds", [=](std::string s) { - settings.set("max-jobs", s); - }); - auto intSettingAlias = [&](char shortName, const std::string & longName, const std::string & description, const std::string & dest) { mkFlag<unsigned int>(shortName, longName, description, [=](unsigned int n) { diff --git a/src/libstore/binary-cache-store.cc b/src/libstore/binary-cache-store.cc index 4527ee6ba660..8b736056e01d 100644 --- a/src/libstore/binary-cache-store.cc +++ b/src/libstore/binary-cache-store.cc @@ -10,6 +10,8 @@ #include "nar-info-disk-cache.hh" #include "nar-accessor.hh" #include "json.hh" +#include "retry.hh" +#include "download.hh" #include <chrono> @@ -79,13 +81,15 @@ void BinaryCacheStore::getFile(const std::string & path, Sink & sink) std::shared_ptr<std::string> BinaryCacheStore::getFile(const std::string & path) { - StringSink sink; - try { - getFile(path, sink); - } catch (NoSuchBinaryCacheFile &) { - return nullptr; - } - return sink.s; + return retry<std::shared_ptr<std::string>>(downloadSettings.tries, [&]() -> std::shared_ptr<std::string> { + StringSink sink; + try { + getFile(path, sink); + } catch (NoSuchBinaryCacheFile &) { + return nullptr; + } + return sink.s; + }); } Path BinaryCacheStore::narInfoFileFor(const Path & storePath) diff --git a/src/libstore/binary-cache-store.hh b/src/libstore/binary-cache-store.hh index 6bc83fc50ca1..953f3b90af46 100644 --- a/src/libstore/binary-cache-store.hh +++ b/src/libstore/binary-cache-store.hh @@ -72,24 +72,11 @@ public: bool isValidPathUncached(const Path & path) override; - PathSet queryAllValidPaths() override - { unsupported(); } - void queryPathInfoUncached(const Path & path, Callback<std::shared_ptr<ValidPathInfo>> callback) override; - void queryReferrers(const Path & path, - PathSet & referrers) override - { unsupported(); } - - PathSet queryDerivationOutputs(const Path & path) override - { unsupported(); } - - StringSet queryDerivationOutputNames(const Path & path) override - { unsupported(); } - Path queryPathFromHashPart(const string & hashPart) override - { unsupported(); } + { unsupported("queryPathFromHashPart"); } bool wantMassQuery() override { return wantMassQuery_; } @@ -108,22 +95,10 @@ public: BuildResult buildDerivation(const Path & drvPath, const BasicDerivation & drv, BuildMode buildMode) override - { unsupported(); } + { unsupported("buildDerivation"); } void ensurePath(const Path & path) override - { unsupported(); } - - void addTempRoot(const Path & path) override - { unsupported(); } - - void addIndirectRoot(const Path & path) override - { unsupported(); } - - Roots findRoots() override - { unsupported(); } - - void collectGarbage(const GCOptions & options, GCResults & results) override - { unsupported(); } + { unsupported("ensurePath"); } ref<FSAccessor> getFSAccessor() override; diff --git a/src/libstore/build.cc b/src/libstore/build.cc index 59abae9b90db..5b38bcf3c5ec 100644 --- a/src/libstore/build.cc +++ b/src/libstore/build.cc @@ -12,6 +12,7 @@ #include "json.hh" #include "nar-info.hh" #include "parsed-derivations.hh" +#include "machines.hh" #include <algorithm> #include <iostream> @@ -37,6 +38,7 @@ #include <unistd.h> #include <errno.h> #include <cstring> +#include <termios.h> #include <pwd.h> #include <grp.h> @@ -460,6 +462,28 @@ static void commonChildInit(Pipe & logPipe) close(fdDevNull); } +void handleDiffHook(uid_t uid, uid_t gid, Path tryA, Path tryB, Path drvPath, Path tmpDir) +{ + auto diffHook = settings.diffHook; + if (diffHook != "" && settings.runDiffHook) { + try { + RunOptions diffHookOptions(diffHook,{tryA, tryB, drvPath, tmpDir}); + diffHookOptions.searchPath = true; + diffHookOptions.uid = uid; + diffHookOptions.gid = gid; + diffHookOptions.chdir = "/"; + + auto diffRes = runProgram(diffHookOptions); + if (!statusOk(diffRes.first)) + throw ExecError(diffRes.first, fmt("diff-hook program '%1%' %2%", diffHook, statusToString(diffRes.first))); + + if (diffRes.second != "") + printError(chomp(diffRes.second)); + } catch (Error & error) { + printError("diff hook execution failed: %s", error.what()); + } + } +} ////////////////////////////////////////////////////////////////////// @@ -881,6 +905,9 @@ public: Worker & worker, BuildMode buildMode = bmNormal); ~DerivationGoal(); + /* Whether we need to perform hash rewriting if there are valid output paths. */ + bool needsHashRewrite(); + void timedOut() override; string key() override @@ -1033,6 +1060,17 @@ DerivationGoal::~DerivationGoal() } +inline bool DerivationGoal::needsHashRewrite() +{ +#if __linux__ + return !useChroot; +#else + /* Darwin requires hash rewriting even when sandboxing is enabled. */ + return true; +#endif +} + + void DerivationGoal::killChild() { if (pid != -1) { @@ -1521,8 +1559,8 @@ void DerivationGoal::buildDone() if (hook) { hook->builderOut.readSide = -1; hook->fromHook.readSide = -1; - } - else builderOut.readSide = -1; + } else + builderOut.readSide = -1; /* Close the log file. */ closeLogFile(); @@ -2072,7 +2110,7 @@ void DerivationGoal::startBuilder() #endif } - else { + if (needsHashRewrite()) { if (pathExists(homeDir)) throw Error(format("directory '%1%' exists; please remove it") % homeDir); @@ -2144,7 +2182,48 @@ void DerivationGoal::startBuilder() Path logFile = openLogFile(); /* Create a pipe to get the output of the builder. */ - builderOut.create(); + //builderOut.create(); + + builderOut.readSide = posix_openpt(O_RDWR | O_NOCTTY); + if (!builderOut.readSide) + throw SysError("opening pseudoterminal master"); + + std::string slaveName(ptsname(builderOut.readSide.get())); + + if (buildUser) { + if (chmod(slaveName.c_str(), 0600)) + throw SysError("changing mode of pseudoterminal slave"); + + if (chown(slaveName.c_str(), buildUser->getUID(), 0)) + throw SysError("changing owner of pseudoterminal slave"); + } else { + if (grantpt(builderOut.readSide.get())) + throw SysError("granting access to pseudoterminal slave"); + } + + #if 0 + // Mount the pt in the sandbox so that the "tty" command works. + // FIXME: this doesn't work with the new devpts in the sandbox. + if (useChroot) + dirsInChroot[slaveName] = {slaveName, false}; + #endif + + if (unlockpt(builderOut.readSide.get())) + throw SysError("unlocking pseudoterminal"); + + builderOut.writeSide = open(slaveName.c_str(), O_RDWR | O_NOCTTY); + if (!builderOut.writeSide) + throw SysError("opening pseudoterminal slave"); + + // Put the pt into raw mode to prevent \n -> \r\n translation. + struct termios term; + if (tcgetattr(builderOut.writeSide.get(), &term)) + throw SysError("getting pseudoterminal attributes"); + + cfmakeraw(&term); + + if (tcsetattr(builderOut.writeSide.get(), TCSANOW, &term)) + throw SysError("putting pseudoterminal into raw mode"); result.startTime = time(0); @@ -2193,7 +2272,6 @@ void DerivationGoal::startBuilder() userNamespaceSync.create(); options.allowVfork = false; - options.restoreMountNamespace = false; Pid helper = startProcess([&]() { @@ -2260,7 +2338,6 @@ void DerivationGoal::startBuilder() #endif { options.allowVfork = !buildUser && !drv->isBuiltin(); - options.restoreMountNamespace = false; pid = startProcess([&]() { runChild(); }, options); @@ -2371,6 +2448,9 @@ void DerivationGoal::initEnv() may change that in the future. So tell the builder which file descriptor to use for that. */ env["NIX_LOG_FD"] = "2"; + + /* Trigger colored output in various tools. */ + env["TERM"] = "xterm-256color"; } @@ -2415,7 +2495,7 @@ void DerivationGoal::writeStructuredAttrs() objects consisting entirely of those values. (So nested arrays or objects are not supported.) */ - auto handleSimpleType = [](const nlohmann::json & value) -> std::experimental::optional<std::string> { + auto handleSimpleType = [](const nlohmann::json & value) -> std::optional<std::string> { if (value.is_string()) return shellEscape(value); @@ -2501,17 +2581,17 @@ void setupSeccomp() seccomp_release(ctx); }); - if (settings.thisSystem == "x86_64-linux" && + if (nativeSystem == "x86_64-linux" && seccomp_arch_add(ctx, SCMP_ARCH_X86) != 0) throw SysError("unable to add 32-bit seccomp architecture"); - if (settings.thisSystem == "x86_64-linux" && + if (nativeSystem == "x86_64-linux" && seccomp_arch_add(ctx, SCMP_ARCH_X32) != 0) throw SysError("unable to add X32 seccomp architecture"); - if (settings.thisSystem == "aarch64-linux" && + if (nativeSystem == "aarch64-linux" && seccomp_arch_add(ctx, SCMP_ARCH_ARM) != 0) - printError("unsable to add ARM seccomp architecture; this may result in spurious build failures if running 32-bit ARM processes."); + printError("unable to add ARM seccomp architecture; this may result in spurious build failures if running 32-bit ARM processes"); /* Prevent builders from creating setuid/setgid binaries. */ for (int perm : { S_ISUID, S_ISGID }) { @@ -2874,6 +2954,10 @@ void DerivationGoal::runChild() for (auto & i : missingPaths) { sandboxProfile += (format("\t(subpath \"%1%\")\n") % i.c_str()).str(); } + /* Also add redirected outputs to the chroot */ + for (auto & i : redirectedOutputs) { + sandboxProfile += (format("\t(subpath \"%1%\")\n") % i.second.c_str()).str(); + } sandboxProfile += ")\n"; /* Our inputs (transitive dependencies and any impurities computed above) @@ -3026,8 +3110,7 @@ void DerivationGoal::registerOutputs() InodesSeen inodesSeen; Path checkSuffix = ".check"; - bool runDiffHook = settings.runDiffHook; - bool keepPreviousRound = settings.keepFailed || runDiffHook; + bool keepPreviousRound = settings.keepFailed || settings.runDiffHook; std::exception_ptr delayedException; @@ -3052,7 +3135,9 @@ void DerivationGoal::registerOutputs() throw SysError(format("moving build output '%1%' from the sandbox to the Nix store") % path); } if (buildMode != bmCheck) actualPath = worker.store.toRealPath(path); - } else { + } + + if (needsHashRewrite()) { Path redirected = redirectedOutputs[path]; if (buildMode == bmRepair && redirectedBadOutputs.find(path) != redirectedBadOutputs.end() @@ -3170,11 +3255,17 @@ void DerivationGoal::registerOutputs() if (!worker.store.isValidPath(path)) continue; auto info = *worker.store.queryPathInfo(path); if (hash.first != info.narHash) { - if (settings.keepFailed) { + if (settings.runDiffHook || settings.keepFailed) { Path dst = worker.store.toRealPath(path + checkSuffix); deletePath(dst); if (rename(actualPath.c_str(), dst.c_str())) throw SysError(format("renaming '%1%' to '%2%'") % actualPath % dst); + + handleDiffHook( + buildUser ? buildUser->getUID() : getuid(), + buildUser ? buildUser->getGID() : getgid(), + path, dst, drvPath, tmpDir); + throw Error(format("derivation '%1%' may not be deterministic: output '%2%' differs from '%3%'") % drvPath % path % dst); } else @@ -3239,16 +3330,10 @@ void DerivationGoal::registerOutputs() ? fmt("output '%1%' of '%2%' differs from '%3%' from previous round", i->second.path, drvPath, prev) : fmt("output '%1%' of '%2%' differs from previous round", i->second.path, drvPath); - auto diffHook = settings.diffHook; - if (prevExists && diffHook != "" && runDiffHook) { - try { - auto diff = runProgram(diffHook, true, {prev, i->second.path}); - if (diff != "") - printError(chomp(diff)); - } catch (Error & error) { - printError("diff hook execution failed: %s", error.what()); - } - } + handleDiffHook( + buildUser ? buildUser->getUID() : getuid(), + buildUser ? buildUser->getGID() : getgid(), + prev, i->second.path, drvPath, tmpDir); if (settings.enforceDeterminism) throw NotDeterministic(msg); @@ -3313,8 +3398,8 @@ void DerivationGoal::checkOutputs(const std::map<Path, ValidPathInfo> & outputs) struct Checks { bool ignoreSelfRefs = false; - std::experimental::optional<uint64_t> maxSize, maxClosureSize; - std::experimental::optional<Strings> allowedReferences, allowedRequisites, disallowedReferences, disallowedRequisites; + std::optional<uint64_t> maxSize, maxClosureSize; + std::optional<Strings> allowedReferences, allowedRequisites, disallowedReferences, disallowedRequisites; }; /* Compute the closure and closure size of some output. This @@ -3361,7 +3446,7 @@ void DerivationGoal::checkOutputs(const std::map<Path, ValidPathInfo> & outputs) info.path, closureSize, *checks.maxClosureSize); } - auto checkRefs = [&](const std::experimental::optional<Strings> & value, bool allowed, bool recursive) + auto checkRefs = [&](const std::optional<Strings> & value, bool allowed, bool recursive) { if (!value) return; @@ -3415,7 +3500,7 @@ void DerivationGoal::checkOutputs(const std::map<Path, ValidPathInfo> & outputs) if (maxClosureSize != output->end()) checks.maxClosureSize = maxClosureSize->get<uint64_t>(); - auto get = [&](const std::string & name) -> std::experimental::optional<Strings> { + auto get = [&](const std::string & name) -> std::optional<Strings> { auto i = output->find(name); if (i != output->end()) { Strings res; @@ -4321,14 +4406,15 @@ void Worker::waitForInput() for (auto & k : fds2) { if (FD_ISSET(k, &fds)) { ssize_t rd = read(k, buffer.data(), buffer.size()); - if (rd == -1) { - if (errno != EINTR) - throw SysError(format("reading from %1%") - % goal->getName()); - } else if (rd == 0) { + // FIXME: is there a cleaner way to handle pt close + // than EIO? Is this even standard? + if (rd == 0 || (rd == -1 && errno == EIO)) { debug(format("%1%: got EOF") % goal->getName()); goal->handleEOF(k); j->fds.erase(k); + } else if (rd == -1) { + if (errno != EINTR) + throw SysError("%s: read failed", goal->getName()); } else { printMsg(lvlVomit, format("%1%: read %2% bytes") % goal->getName() % rd); @@ -4413,6 +4499,11 @@ static void primeCache(Store & store, const PathSet & paths) PathSet willBuild, willSubstitute, unknown; unsigned long long downloadSize, narSize; store.queryMissing(paths, willBuild, willSubstitute, unknown, downloadSize, narSize); + + if (!willBuild.empty() && 0 == settings.maxBuildJobs && getMachines().empty()) + throw Error( + "%d derivations need to be built, but neither local builds ('--max-jobs') " + "nor remote builds ('--builders') are enabled", willBuild.size()); } diff --git a/src/libstore/builtins/fetchurl.cc b/src/libstore/builtins/fetchurl.cc index 92aec63a0379..b1af3b4fc316 100644 --- a/src/libstore/builtins/fetchurl.cc +++ b/src/libstore/builtins/fetchurl.cc @@ -64,7 +64,8 @@ void builtinFetchurl(const BasicDerivation & drv, const std::string & netrcData) try { if (!hasSuffix(hashedMirror, "/")) hashedMirror += '/'; auto ht = parseHashType(getAttr("outputHashAlgo")); - fetch(hashedMirror + printHashType(ht) + "/" + Hash(getAttr("outputHash"), ht).to_string(Base16, false)); + auto h = Hash(getAttr("outputHash"), ht); + fetch(hashedMirror + printHashType(h.type) + "/" + h.to_string(Base16, false)); return; } catch (Error & e) { debug(e.what()); diff --git a/src/libstore/download.cc b/src/libstore/download.cc index 467f570bbf05..0c5a73ea3c51 100644 --- a/src/libstore/download.cc +++ b/src/libstore/download.cc @@ -8,6 +8,7 @@ #include "compression.hh" #include "pathlocks.hh" #include "finally.hh" +#include "retry.hh" #ifdef ENABLE_S3 #include <aws/core/client/ClientConfiguration.h> @@ -19,34 +20,16 @@ #include <curl/curl.h> #include <algorithm> -#include <cmath> #include <cstring> #include <iostream> #include <queue> -#include <random> #include <thread> using namespace std::string_literals; namespace nix { -struct DownloadSettings : Config -{ - Setting<bool> enableHttp2{this, true, "http2", - "Whether to enable HTTP/2 support."}; - - Setting<std::string> userAgentSuffix{this, "", "user-agent-suffix", - "String appended to the user agent in HTTP requests."}; - - Setting<size_t> httpConnections{this, 25, "http-connections", - "Number of parallel HTTP connections.", - {"binary-caches-parallel-connections"}}; - - Setting<unsigned long> connectTimeout{this, 0, "connect-timeout", - "Timeout for connecting to servers during downloads. 0 means use curl's builtin default."}; -}; - -static DownloadSettings downloadSettings; +DownloadSettings downloadSettings; static GlobalConfig::Register r1(&downloadSettings); @@ -62,9 +45,6 @@ struct CurlDownloader : public Downloader { CURLM * curlm = 0; - std::random_device rd; - std::mt19937 mt19937; - struct DownloadItem : public std::enable_shared_from_this<DownloadItem> { CurlDownloader & downloader; @@ -77,12 +57,6 @@ struct CurlDownloader : public Downloader bool active = false; // whether the handle has been added to the multi object std::string status; - unsigned int attempt = 0; - - /* Don't start this download until the specified time point - has been reached. */ - std::chrono::steady_clock::time_point embargo; - struct curl_slist * requestHeaders = 0; std::string encoding; @@ -270,6 +244,8 @@ struct CurlDownloader : public Downloader #if LIBCURL_VERSION_NUM >= 0x072f00 if (downloadSettings.enableHttp2) curl_easy_setopt(req, CURLOPT_HTTP_VERSION, CURL_HTTP_VERSION_2TLS); + else + curl_easy_setopt(req, CURLOPT_HTTP_VERSION, CURL_HTTP_VERSION_1_1); #endif curl_easy_setopt(req, CURLOPT_WRITEFUNCTION, DownloadItem::writeCallbackWrapper); curl_easy_setopt(req, CURLOPT_WRITEDATA, this); @@ -319,16 +295,21 @@ struct CurlDownloader : public Downloader long httpStatus = 0; curl_easy_getinfo(req, CURLINFO_RESPONSE_CODE, &httpStatus); - char * effectiveUrlCStr; - curl_easy_getinfo(req, CURLINFO_EFFECTIVE_URL, &effectiveUrlCStr); - if (effectiveUrlCStr) - result.effectiveUrl = effectiveUrlCStr; + char * effectiveUriCStr; + curl_easy_getinfo(req, CURLINFO_EFFECTIVE_URL, &effectiveUriCStr); + if (effectiveUriCStr) + result.effectiveUri = effectiveUriCStr; debug("finished %s of '%s'; curl status = %d, HTTP status = %d, body = %d bytes", request.verb(), request.uri, code, httpStatus, result.bodySize); - if (decompressionSink) - decompressionSink->finish(); + if (decompressionSink) { + try { + decompressionSink->finish(); + } catch (...) { + writeException = std::current_exception(); + } + } if (code == CURLE_WRITE_ERROR && result.etag == request.expectedETag) { code = CURLE_OK; @@ -396,9 +377,7 @@ struct CurlDownloader : public Downloader } } - attempt++; - - auto exc = + fail( code == CURLE_ABORTED_BY_CALLBACK && _isInterrupted ? DownloadError(Interrupted, fmt("%s of '%s' was interrupted", request.verb(), request.uri)) : httpStatus != 0 @@ -409,31 +388,15 @@ struct CurlDownloader : public Downloader ) : DownloadError(err, fmt("unable to %s '%s': %s (%d)", - request.verb(), request.uri, curl_easy_strerror(code), code)); - - /* If this is a transient error, then maybe retry the - download after a while. */ - if (err == Transient && attempt < request.tries) { - int ms = request.baseRetryTimeMs * std::pow(2.0f, attempt - 1 + std::uniform_real_distribution<>(0.0, 0.5)(downloader.mt19937)); - printError(format("warning: %s; retrying in %d ms") % exc.what() % ms); - embargo = std::chrono::steady_clock::now() + std::chrono::milliseconds(ms); - downloader.enqueueItem(shared_from_this()); - } - else - fail(exc); + request.verb(), request.uri, curl_easy_strerror(code), code))); } } }; struct State { - struct EmbargoComparator { - bool operator() (const std::shared_ptr<DownloadItem> & i1, const std::shared_ptr<DownloadItem> & i2) { - return i1->embargo > i2->embargo; - } - }; bool quit = false; - std::priority_queue<std::shared_ptr<DownloadItem>, std::vector<std::shared_ptr<DownloadItem>>, EmbargoComparator> incoming; + std::vector<std::shared_ptr<DownloadItem>> incoming; }; Sync<State> state_; @@ -446,7 +409,6 @@ struct CurlDownloader : public Downloader std::thread workerThread; CurlDownloader() - : mt19937(rd()) { static std::once_flag globalInit; std::call_once(globalInit, curl_global_init, CURL_GLOBAL_ALL); @@ -540,9 +502,7 @@ struct CurlDownloader : public Downloader nextWakeup = std::chrono::steady_clock::time_point(); - /* Add new curl requests from the incoming requests queue, - except for requests that are embargoed (waiting for a - retry timeout to expire). */ + /* Add new curl requests from the incoming requests queue. */ if (extraFDs[0].revents & CURL_WAIT_POLLIN) { char buf[1024]; auto res = read(extraFDs[0].fd, buf, sizeof(buf)); @@ -551,22 +511,9 @@ struct CurlDownloader : public Downloader } std::vector<std::shared_ptr<DownloadItem>> incoming; - auto now = std::chrono::steady_clock::now(); - { auto state(state_.lock()); - while (!state->incoming.empty()) { - auto item = state->incoming.top(); - if (item->embargo <= now) { - incoming.push_back(item); - state->incoming.pop(); - } else { - if (nextWakeup == std::chrono::steady_clock::time_point() - || item->embargo < nextWakeup) - nextWakeup = item->embargo; - break; - } - } + std::swap(state->incoming, incoming); quit = state->quit; } @@ -593,7 +540,7 @@ struct CurlDownloader : public Downloader { auto state(state_.lock()); - while (!state->incoming.empty()) state->incoming.pop(); + state->incoming.clear(); state->quit = true; } } @@ -609,11 +556,27 @@ struct CurlDownloader : public Downloader auto state(state_.lock()); if (state->quit) throw nix::Error("cannot enqueue download request because the download thread is shutting down"); - state->incoming.push(item); + state->incoming.push_back(item); } writeFull(wakeupPipe.writeSide.get(), " "); } +#ifdef ENABLE_S3 + std::tuple<std::string, std::string, Store::Params> parseS3Uri(std::string uri) + { + auto [path, params] = splitUriAndParams(uri); + + auto slash = path.find('/', 5); // 5 is the length of "s3://" prefix + if (slash == std::string::npos) + throw nix::Error("bad S3 URI '%s'", path); + + std::string bucketName(path, 5, slash - 5); + std::string key(path, slash + 1); + + return {bucketName, key, params}; + } +#endif + void enqueueDownload(const DownloadRequest & request, Callback<DownloadResult> callback) override { @@ -622,12 +585,15 @@ struct CurlDownloader : public Downloader // FIXME: do this on a worker thread try { #ifdef ENABLE_S3 - S3Helper s3Helper("", Aws::Region::US_EAST_1, "", ""); // FIXME: make configurable - auto slash = request.uri.find('/', 5); - if (slash == std::string::npos) - throw nix::Error("bad S3 URI '%s'", request.uri); - std::string bucketName(request.uri, 5, slash - 5); - std::string key(request.uri, slash + 1); + auto [bucketName, key, params] = parseS3Uri(request.uri); + + std::string profile = get(params, "profile", ""); + std::string region = get(params, "region", Aws::Region::US_EAST_1); + std::string scheme = get(params, "scheme", ""); + std::string endpoint = get(params, "endpoint", ""); + + S3Helper s3Helper(profile, region, scheme, endpoint); + // FIXME: implement ETag auto s3Res = s3Helper.getObject(bucketName, key); DownloadResult res; @@ -673,7 +639,9 @@ std::future<DownloadResult> Downloader::enqueueDownload(const DownloadRequest & DownloadResult Downloader::download(const DownloadRequest & request) { - return enqueueDownload(request).get(); + return retry<DownloadResult>(request.tries, [&]() { + return enqueueDownload(request).get(); + }); } void Downloader::download(DownloadRequest && request, Sink & sink) @@ -771,20 +739,26 @@ void Downloader::download(DownloadRequest && request, Sink & sink) } } -Path Downloader::downloadCached(ref<Store> store, const string & url_, bool unpack, string name, const Hash & expectedHash, string * effectiveUrl, int ttl) +CachedDownloadResult Downloader::downloadCached( + ref<Store> store, const CachedDownloadRequest & request) { - auto url = resolveUri(url_); + auto url = resolveUri(request.uri); + auto name = request.name; if (name == "") { auto p = url.rfind('/'); if (p != string::npos) name = string(url, p + 1); } Path expectedStorePath; - if (expectedHash) { - expectedStorePath = store->makeFixedOutputPath(unpack, expectedHash, name); - if (store->isValidPath(expectedStorePath)) - return store->toRealPath(expectedStorePath); + if (request.expectedHash) { + expectedStorePath = store->makeFixedOutputPath(request.unpack, request.expectedHash, name); + if (store->isValidPath(expectedStorePath)) { + CachedDownloadResult result; + result.storePath = expectedStorePath; + result.path = store->toRealPath(expectedStorePath); + return result; + } } Path cacheDir = getCacheDir() + "/nix/tarballs"; @@ -803,6 +777,8 @@ Path Downloader::downloadCached(ref<Store> store, const string & url_, bool unpa bool skip = false; + CachedDownloadResult result; + if (pathExists(fileLink) && pathExists(dataFile)) { storePath = readLink(fileLink); store->addTempRoot(storePath); @@ -810,10 +786,10 @@ Path Downloader::downloadCached(ref<Store> store, const string & url_, bool unpa auto ss = tokenizeString<vector<string>>(readFile(dataFile), "\n"); if (ss.size() >= 3 && ss[0] == url) { time_t lastChecked; - if (string2Int(ss[2], lastChecked) && lastChecked + ttl >= time(0)) { + if (string2Int(ss[2], lastChecked) && (uint64_t) lastChecked + request.ttl >= (uint64_t) time(0)) { skip = true; - if (effectiveUrl) - *effectiveUrl = url_; + result.effectiveUri = request.uri; + result.etag = ss[1]; } else if (!ss[1].empty()) { debug(format("verifying previous ETag '%1%'") % ss[1]); expectedETag = ss[1]; @@ -826,17 +802,17 @@ Path Downloader::downloadCached(ref<Store> store, const string & url_, bool unpa if (!skip) { try { - DownloadRequest request(url); - request.expectedETag = expectedETag; - auto res = download(request); - if (effectiveUrl) - *effectiveUrl = res.effectiveUrl; + DownloadRequest request2(url); + request2.expectedETag = expectedETag; + auto res = download(request2); + result.effectiveUri = res.effectiveUri; + result.etag = res.etag; if (!res.cached) { ValidPathInfo info; StringSink sink; dumpString(*res.data, sink); - Hash hash = hashString(expectedHash ? expectedHash.type : htSHA256, *res.data); + Hash hash = hashString(request.expectedHash ? request.expectedHash.type : htSHA256, *res.data); info.path = store->makeFixedOutputPath(false, hash, name); info.narHash = hashString(htSHA256, *sink.s); info.narSize = sink.s->size(); @@ -851,11 +827,12 @@ Path Downloader::downloadCached(ref<Store> store, const string & url_, bool unpa writeFile(dataFile, url + "\n" + res.etag + "\n" + std::to_string(time(0)) + "\n"); } catch (DownloadError & e) { if (storePath.empty()) throw; - printError(format("warning: %1%; using cached result") % e.msg()); + warn("%s; using cached result", e.msg()); + result.etag = expectedETag; } } - if (unpack) { + if (request.unpack) { Path unpackedLink = cacheDir + "/" + baseNameOf(storePath) + "-unpacked"; PathLocks lock2({unpackedLink}, fmt("waiting for lock on '%1%'...", unpackedLink)); Path unpackedStorePath; @@ -878,14 +855,16 @@ Path Downloader::downloadCached(ref<Store> store, const string & url_, bool unpa } if (expectedStorePath != "" && storePath != expectedStorePath) { - Hash gotHash = unpack - ? hashPath(expectedHash.type, store->toRealPath(storePath)).first - : hashFile(expectedHash.type, store->toRealPath(storePath)); + Hash gotHash = request.unpack + ? hashPath(request.expectedHash.type, store->toRealPath(storePath)).first + : hashFile(request.expectedHash.type, store->toRealPath(storePath)); throw nix::Error("hash mismatch in file downloaded from '%s':\n wanted: %s\n got: %s", - url, expectedHash.to_string(), gotHash.to_string()); + url, request.expectedHash.to_string(), gotHash.to_string()); } - return store->toRealPath(storePath); + result.storePath = storePath; + result.path = store->toRealPath(storePath); + return result; } @@ -898,5 +877,4 @@ bool isUri(const string & s) return scheme == "http" || scheme == "https" || scheme == "file" || scheme == "channel" || scheme == "git" || scheme == "s3" || scheme == "ssh"; } - } diff --git a/src/libstore/download.hh b/src/libstore/download.hh index f0228f7d053a..9e965b506d0a 100644 --- a/src/libstore/download.hh +++ b/src/libstore/download.hh @@ -9,13 +9,34 @@ namespace nix { +struct DownloadSettings : Config +{ + Setting<bool> enableHttp2{this, true, "http2", + "Whether to enable HTTP/2 support."}; + + Setting<std::string> userAgentSuffix{this, "", "user-agent-suffix", + "String appended to the user agent in HTTP requests."}; + + Setting<size_t> httpConnections{this, 25, "http-connections", + "Number of parallel HTTP connections.", + {"binary-caches-parallel-connections"}}; + + Setting<unsigned long> connectTimeout{this, 0, "connect-timeout", + "Timeout for connecting to servers during downloads. 0 means use curl's builtin default."}; + + Setting<unsigned int> tries{this, 5, "download-attempts", + "How often Nix will attempt to download a file before giving up."}; +}; + +extern DownloadSettings downloadSettings; + struct DownloadRequest { std::string uri; std::string expectedETag; bool verifyTLS = true; bool head = false; - size_t tries = 5; + size_t tries = downloadSettings.tries; unsigned int baseRetryTimeMs = 250; ActivityId parentAct; bool decompress = true; @@ -36,11 +57,33 @@ struct DownloadResult { bool cached = false; std::string etag; - std::string effectiveUrl; + std::string effectiveUri; std::shared_ptr<std::string> data; uint64_t bodySize = 0; }; +struct CachedDownloadRequest +{ + std::string uri; + bool unpack = false; + std::string name; + Hash expectedHash; + unsigned int ttl = settings.tarballTtl; + + CachedDownloadRequest(const std::string & uri) + : uri(uri) { } +}; + +struct CachedDownloadResult +{ + // Note: 'storePath' may be different from 'path' when using a + // chroot store. + Path storePath; + Path path; + std::optional<std::string> etag; + std::string effectiveUri; +}; + class Store; struct Downloader @@ -53,19 +96,20 @@ struct Downloader std::future<DownloadResult> enqueueDownload(const DownloadRequest & request); - /* Synchronously download a file. */ + /* Synchronously download a file. The request will be retried in + case of transient failures. */ DownloadResult download(const DownloadRequest & request); /* Download a file, writing its data to a sink. The sink will be - invoked on the thread of the caller. */ + invoked on the thread of the caller. The request will not be + retried in case of transient failures. */ void download(DownloadRequest && request, Sink & sink); /* Check if the specified file is already in ~/.cache/nix/tarballs and is more recent than โtarball-ttlโ seconds. Otherwise, use the recorded ETag to verify if the server has a more recent version, and if so, download it to the Nix store. */ - Path downloadCached(ref<Store> store, const string & uri, bool unpack, string name = "", - const Hash & expectedHash = Hash(), string * effectiveUri = nullptr, int ttl = settings.tarballTtl); + CachedDownloadResult downloadCached(ref<Store> store, const CachedDownloadRequest & request); enum Error { NotFound, Forbidden, Misc, Transient, Interrupted }; }; @@ -84,6 +128,11 @@ public: DownloadError(Downloader::Error error, const FormatOrString & fs) : Error(fs), error(error) { } + + bool isTransient() override + { + return error == Downloader::Error::Transient; + } }; bool isUri(const string & s); diff --git a/src/libstore/gc.cc b/src/libstore/gc.cc index b415d5421476..26e2b0dca7ca 100644 --- a/src/libstore/gc.cc +++ b/src/libstore/gc.cc @@ -129,8 +129,8 @@ Path LocalFSStore::addPermRoot(const Path & _storePath, check if the root is in a directory in or linked from the gcroots directory. */ if (settings.checkRootReachability) { - Roots roots = findRoots(); - if (roots.find(gcRoot) == roots.end()) + Roots roots = findRoots(false); + if (roots[storePath].count(gcRoot) == 0) printError( format( "warning: '%1%' is not in a directory where the garbage collector looks for roots; " @@ -197,10 +197,11 @@ void LocalStore::addTempRoot(const Path & path) } -std::set<std::pair<pid_t, Path>> LocalStore::readTempRoots(FDs & fds) -{ - std::set<std::pair<pid_t, Path>> tempRoots; +static std::string censored = "{censored}"; + +void LocalStore::findTempRoots(FDs & fds, Roots & tempRoots, bool censor) +{ /* Read the `temproots' directory for per-process temporary root files. */ for (auto & i : readDirectory(tempRootsDir)) { @@ -250,14 +251,12 @@ std::set<std::pair<pid_t, Path>> LocalStore::readTempRoots(FDs & fds) Path root(contents, pos, end - pos); debug("got temporary root '%s'", root); assertStorePath(root); - tempRoots.emplace(pid, root); + tempRoots[root].emplace(censor ? censored : fmt("{temp:%d}", pid)); pos = end + 1; } fds.push_back(fd); /* keep open */ } - - return tempRoots; } @@ -266,7 +265,7 @@ void LocalStore::findRoots(const Path & path, unsigned char type, Roots & roots) auto foundRoot = [&](const Path & path, const Path & target) { Path storePath = toStorePath(target); if (isStorePath(storePath) && isValidPath(storePath)) - roots[path] = storePath; + roots[storePath].emplace(path); else printInfo(format("skipping invalid root from '%1%' to '%2%'") % path % storePath); }; @@ -306,7 +305,7 @@ void LocalStore::findRoots(const Path & path, unsigned char type, Roots & roots) else if (type == DT_REG) { Path storePath = storeDir + "/" + baseNameOf(path); if (isStorePath(storePath) && isValidPath(storePath)) - roots[path] = storePath; + roots[storePath].emplace(path); } } @@ -321,44 +320,31 @@ void LocalStore::findRoots(const Path & path, unsigned char type, Roots & roots) } -Roots LocalStore::findRootsNoTemp() +void LocalStore::findRootsNoTemp(Roots & roots, bool censor) { - Roots roots; - /* Process direct roots in {gcroots,profiles}. */ findRoots(stateDir + "/" + gcRootsDir, DT_UNKNOWN, roots); findRoots(stateDir + "/profiles", DT_UNKNOWN, roots); - /* Add additional roots returned by the program specified by the - NIX_ROOT_FINDER environment variable. This is typically used - to add running programs to the set of roots (to prevent them - from being garbage collected). */ - size_t n = 0; - for (auto & root : findRuntimeRoots()) - roots[fmt("{memory:%d}", n++)] = root; - - return roots; + /* Add additional roots returned by different platforms-specific + heuristics. This is typically used to add running programs to + the set of roots (to prevent them from being garbage collected). */ + findRuntimeRoots(roots, censor); } -Roots LocalStore::findRoots() +Roots LocalStore::findRoots(bool censor) { - Roots roots = findRootsNoTemp(); + Roots roots; + findRootsNoTemp(roots, censor); FDs fds; - pid_t prev = -1; - size_t n = 0; - for (auto & root : readTempRoots(fds)) { - if (prev != root.first) n = 0; - prev = root.first; - roots[fmt("{temp:%d:%d}", root.first, n++)] = root.second; - } + findTempRoots(fds, roots, censor); return roots; } - -static void readProcLink(const string & file, StringSet & paths) +static void readProcLink(const string & file, Roots & roots) { /* 64 is the starting buffer size gnu readlink uses... */ auto bufsiz = ssize_t{64}; @@ -377,8 +363,8 @@ try_again: goto try_again; } if (res > 0 && buf[0] == '/') - paths.emplace(static_cast<char *>(buf), res); - return; + roots[std::string(static_cast<char *>(buf), res)] + .emplace(file); } static string quoteRegexChars(const string & raw) @@ -387,20 +373,20 @@ static string quoteRegexChars(const string & raw) return std::regex_replace(raw, specialRegex, R"(\$&)"); } -static void readFileRoots(const char * path, StringSet & paths) +static void readFileRoots(const char * path, Roots & roots) { try { - paths.emplace(readFile(path)); + roots[readFile(path)].emplace(path); } catch (SysError & e) { if (e.errNo != ENOENT && e.errNo != EACCES) throw; } } -PathSet LocalStore::findRuntimeRoots() +void LocalStore::findRuntimeRoots(Roots & roots, bool censor) { - PathSet roots; - StringSet paths; + Roots unchecked; + auto procDir = AutoCloseDir{opendir("/proc")}; if (procDir) { struct dirent * ent; @@ -410,10 +396,10 @@ PathSet LocalStore::findRuntimeRoots() while (errno = 0, ent = readdir(procDir.get())) { checkInterrupt(); if (std::regex_match(ent->d_name, digitsRegex)) { - readProcLink((format("/proc/%1%/exe") % ent->d_name).str(), paths); - readProcLink((format("/proc/%1%/cwd") % ent->d_name).str(), paths); + readProcLink(fmt("/proc/%s/exe" ,ent->d_name), unchecked); + readProcLink(fmt("/proc/%s/cwd", ent->d_name), unchecked); - auto fdStr = (format("/proc/%1%/fd") % ent->d_name).str(); + auto fdStr = fmt("/proc/%s/fd", ent->d_name); auto fdDir = AutoCloseDir(opendir(fdStr.c_str())); if (!fdDir) { if (errno == ENOENT || errno == EACCES) @@ -422,9 +408,8 @@ PathSet LocalStore::findRuntimeRoots() } struct dirent * fd_ent; while (errno = 0, fd_ent = readdir(fdDir.get())) { - if (fd_ent->d_name[0] != '.') { - readProcLink((format("%1%/%2%") % fdStr % fd_ent->d_name).str(), paths); - } + if (fd_ent->d_name[0] != '.') + readProcLink(fmt("%s/%s", fdStr, fd_ent->d_name), unchecked); } if (errno) { if (errno == ESRCH) @@ -434,18 +419,19 @@ PathSet LocalStore::findRuntimeRoots() fdDir.reset(); try { - auto mapLines = - tokenizeString<std::vector<string>>(readFile((format("/proc/%1%/maps") % ent->d_name).str(), true), "\n"); - for (const auto& line : mapLines) { + auto mapFile = fmt("/proc/%s/maps", ent->d_name); + auto mapLines = tokenizeString<std::vector<string>>(readFile(mapFile, true), "\n"); + for (const auto & line : mapLines) { auto match = std::smatch{}; if (std::regex_match(line, match, mapRegex)) - paths.emplace(match[1]); + unchecked[match[1]].emplace(mapFile); } - auto envString = readFile((format("/proc/%1%/environ") % ent->d_name).str(), true); + auto envFile = fmt("/proc/%s/environ", ent->d_name); + auto envString = readFile(envFile, true); auto env_end = std::sregex_iterator{}; for (auto i = std::sregex_iterator{envString.begin(), envString.end(), storePathRegex}; i != env_end; ++i) - paths.emplace(i->str()); + unchecked[i->str()].emplace(envFile); } catch (SysError & e) { if (errno == ENOENT || errno == EACCES || errno == ESRCH) continue; @@ -465,7 +451,7 @@ PathSet LocalStore::findRuntimeRoots() for (const auto & line : lsofLines) { std::smatch match; if (std::regex_match(line, match, lsofRegex)) - paths.emplace(match[1]); + unchecked[match[1]].emplace("{lsof}"); } } catch (ExecError & e) { /* lsof not installed, lsof failed */ @@ -473,21 +459,23 @@ PathSet LocalStore::findRuntimeRoots() #endif #if defined(__linux__) - readFileRoots("/proc/sys/kernel/modprobe", paths); - readFileRoots("/proc/sys/kernel/fbsplash", paths); - readFileRoots("/proc/sys/kernel/poweroff_cmd", paths); + readFileRoots("/proc/sys/kernel/modprobe", unchecked); + readFileRoots("/proc/sys/kernel/fbsplash", unchecked); + readFileRoots("/proc/sys/kernel/poweroff_cmd", unchecked); #endif - for (auto & i : paths) - if (isInStore(i)) { - Path path = toStorePath(i); - if (roots.find(path) == roots.end() && isStorePath(path) && isValidPath(path)) { + for (auto & [target, links] : unchecked) { + if (isInStore(target)) { + Path path = toStorePath(target); + if (isStorePath(path) && isValidPath(path)) { debug(format("got additional root '%1%'") % path); - roots.insert(path); + if (censor) + roots[path].insert(censored); + else + roots[path].insert(links.begin(), links.end()); } } - - return roots; + } } @@ -754,16 +742,20 @@ void LocalStore::collectGarbage(const GCOptions & options, GCResults & results) /* Find the roots. Since we've grabbed the GC lock, the set of permanent roots cannot increase now. */ printError(format("finding garbage collector roots...")); - Roots rootMap = options.ignoreLiveness ? Roots() : findRootsNoTemp(); + Roots rootMap; + if (!options.ignoreLiveness) + findRootsNoTemp(rootMap, true); - for (auto & i : rootMap) state.roots.insert(i.second); + for (auto & i : rootMap) state.roots.insert(i.first); /* Read the temporary roots. This acquires read locks on all per-process temporary root files. So after this point no paths can be added to the set of temporary roots. */ FDs fds; - for (auto & root : readTempRoots(fds)) - state.tempRoots.insert(root.second); + Roots tempRoots; + findTempRoots(fds, tempRoots, true); + for (auto & root : tempRoots) + state.tempRoots.insert(root.first); state.roots.insert(state.tempRoots.begin(), state.tempRoots.end()); /* After this point the set of roots or temporary roots cannot diff --git a/src/libstore/globals.hh b/src/libstore/globals.hh index 53efc6a90fb6..0af8215d1fd8 100644 --- a/src/libstore/globals.hh +++ b/src/libstore/globals.hh @@ -255,7 +255,7 @@ public: "Secret keys with which to sign local builds."}; Setting<unsigned int> tarballTtl{this, 60 * 60, "tarball-ttl", - "How soon to expire files fetched by builtins.fetchTarball and builtins.fetchurl."}; + "How long downloaded files are considered up-to-date."}; Setting<bool> requireSigs{this, true, "require-sigs", "Whether to check that any non-content-addressed path added to the " diff --git a/src/libstore/http-binary-cache-store.cc b/src/libstore/http-binary-cache-store.cc index 8da0e2f9d82a..5633b4355d25 100644 --- a/src/libstore/http-binary-cache-store.cc +++ b/src/libstore/http-binary-cache-store.cc @@ -2,6 +2,7 @@ #include "download.hh" #include "globals.hh" #include "nar-info-disk-cache.hh" +#include "retry.hh" namespace nix { @@ -84,7 +85,6 @@ protected: try { DownloadRequest request(cacheUri + "/" + path); request.head = true; - request.tries = 5; getDownloader()->download(request); return true; } catch (DownloadError & e) { @@ -114,7 +114,6 @@ protected: DownloadRequest makeRequest(const std::string & path) { DownloadRequest request(cacheUri + "/" + path); - request.tries = 8; return request; } @@ -137,21 +136,46 @@ protected: { checkEnabled(); - auto request(makeRequest(path)); - - getDownloader()->enqueueDownload(request, - {[callback, this](std::future<DownloadResult> result) { - try { - callback(result.get().data); - } catch (DownloadError & e) { - if (e.error == Downloader::NotFound || e.error == Downloader::Forbidden) - return callback(std::shared_ptr<std::string>()); - maybeDisable(); - callback.rethrow(); - } catch (...) { - callback.rethrow(); - } - }}); + struct State + { + DownloadRequest request; + std::function<void()> tryDownload; + unsigned int attempt = 0; + State(DownloadRequest && request) : request(request) {} + }; + + auto state = std::make_shared<State>(makeRequest(path)); + + state->tryDownload = [callback, state, this]() { + getDownloader()->enqueueDownload(state->request, + {[callback, state, this](std::future<DownloadResult> result) { + try { + callback(result.get().data); + } catch (DownloadError & e) { + if (e.error == Downloader::NotFound || e.error == Downloader::Forbidden) + return callback(std::shared_ptr<std::string>()); + ++state->attempt; + if (state->attempt < state->request.tries && e.isTransient()) { + auto ms = retrySleepTime(state->attempt); + warn("%s; retrying in %d ms", e.what(), ms); + /* We can't sleep here because that would + block the download thread. So use a + separate thread for sleeping. */ + std::thread([state, ms]() { + std::this_thread::sleep_for(std::chrono::milliseconds(ms)); + state->tryDownload(); + }).detach(); + } else { + maybeDisable(); + callback.rethrow(); + } + } catch (...) { + callback.rethrow(); + } + }}); + }; + + state->tryDownload(); } }; diff --git a/src/libstore/legacy-ssh-store.cc b/src/libstore/legacy-ssh-store.cc index 26e1851981db..7c9bc2b68ba8 100644 --- a/src/libstore/legacy-ssh-store.cc +++ b/src/libstore/legacy-ssh-store.cc @@ -187,28 +187,17 @@ struct LegacySSHStore : public Store copyNAR(conn->from, sink); } - PathSet queryAllValidPaths() override { unsupported(); } - - void queryReferrers(const Path & path, PathSet & referrers) override - { unsupported(); } - - PathSet queryDerivationOutputs(const Path & path) override - { unsupported(); } - - StringSet queryDerivationOutputNames(const Path & path) override - { unsupported(); } - Path queryPathFromHashPart(const string & hashPart) override - { unsupported(); } + { unsupported("queryPathFromHashPart"); } Path addToStore(const string & name, const Path & srcPath, bool recursive, HashType hashAlgo, PathFilter & filter, RepairFlag repair) override - { unsupported(); } + { unsupported("addToStore"); } Path addTextToStore(const string & name, const string & s, const PathSet & references, RepairFlag repair) override - { unsupported(); } + { unsupported("addTextToStore"); } BuildResult buildDerivation(const Path & drvPath, const BasicDerivation & drv, BuildMode buildMode) override @@ -242,25 +231,7 @@ struct LegacySSHStore : public Store } void ensurePath(const Path & path) override - { unsupported(); } - - void addTempRoot(const Path & path) override - { unsupported(); } - - void addIndirectRoot(const Path & path) override - { unsupported(); } - - Roots findRoots() override - { unsupported(); } - - void collectGarbage(const GCOptions & options, GCResults & results) override - { unsupported(); } - - ref<FSAccessor> getFSAccessor() override - { unsupported(); } - - void addSignatures(const Path & storePath, const StringSet & sigs) override - { unsupported(); } + { unsupported("ensurePath"); } void computeFSClosure(const PathSet & paths, PathSet & out, bool flipDirection = false, diff --git a/src/libstore/local-store.cc b/src/libstore/local-store.cc index 5b4e7ca4ca99..485fdd691932 100644 --- a/src/libstore/local-store.cc +++ b/src/libstore/local-store.cc @@ -366,8 +366,6 @@ void LocalStore::makeStoreWritable() throw SysError("getting info about the Nix store mount point"); if (stat.f_flag & ST_RDONLY) { - saveMountNamespace(); - if (unshare(CLONE_NEWNS) == -1) throw SysError("setting up a private mount namespace"); diff --git a/src/libstore/local-store.hh b/src/libstore/local-store.hh index fce963433a5e..6b655647b031 100644 --- a/src/libstore/local-store.hh +++ b/src/libstore/local-store.hh @@ -180,11 +180,11 @@ private: typedef std::shared_ptr<AutoCloseFD> FDPtr; typedef list<FDPtr> FDs; - std::set<std::pair<pid_t, Path>> readTempRoots(FDs & fds); + void findTempRoots(FDs & fds, Roots & roots, bool censor); public: - Roots findRoots() override; + Roots findRoots(bool censor) override; void collectGarbage(const GCOptions & options, GCResults & results) override; @@ -267,9 +267,9 @@ private: void findRoots(const Path & path, unsigned char type, Roots & roots); - Roots findRootsNoTemp(); + void findRootsNoTemp(Roots & roots, bool censor); - PathSet findRuntimeRoots(); + void findRuntimeRoots(Roots & roots, bool censor); void removeUnusedLinks(const GCState & state); diff --git a/src/libstore/local.mk b/src/libstore/local.mk index 3799257f83ff..89fc918c30fd 100644 --- a/src/libstore/local.mk +++ b/src/libstore/local.mk @@ -6,7 +6,7 @@ libstore_DIR := $(d) libstore_SOURCES := $(wildcard $(d)/*.cc $(d)/builtins/*.cc) -libstore_LIBS = libutil libformat +libstore_LIBS = libutil libstore_LDFLAGS = $(SQLITE3_LIBS) -lbz2 $(LIBCURL_LIBS) $(SODIUM_LIBS) -pthread ifneq ($(OS), FreeBSD) diff --git a/src/libstore/machines.cc b/src/libstore/machines.cc index edd03d147832..f848582dafd4 100644 --- a/src/libstore/machines.cc +++ b/src/libstore/machines.cc @@ -89,10 +89,11 @@ void parseMachines(const std::string & s, Machines & machines) Machines getMachines() { - Machines machines; - - parseMachines(settings.builders, machines); - + static auto machines = [&]() { + Machines machines; + parseMachines(settings.builders, machines); + return machines; + }(); return machines; } diff --git a/src/libstore/nix-store.pc.in b/src/libstore/nix-store.pc.in index 5cf22faadcbe..6d67b1e03808 100644 --- a/src/libstore/nix-store.pc.in +++ b/src/libstore/nix-store.pc.in @@ -6,4 +6,4 @@ Name: Nix Description: Nix Package Manager Version: @PACKAGE_VERSION@ Libs: -L${libdir} -lnixstore -lnixutil -Cflags: -I${includedir}/nix -std=c++14 +Cflags: -I${includedir}/nix -std=c++17 diff --git a/src/libstore/parsed-derivations.cc b/src/libstore/parsed-derivations.cc index dc3286482736..17fde00a0167 100644 --- a/src/libstore/parsed-derivations.cc +++ b/src/libstore/parsed-derivations.cc @@ -16,7 +16,7 @@ ParsedDerivation::ParsedDerivation(const Path & drvPath, BasicDerivation & drv) } } -std::experimental::optional<std::string> ParsedDerivation::getStringAttr(const std::string & name) const +std::optional<std::string> ParsedDerivation::getStringAttr(const std::string & name) const { if (structuredAttrs) { auto i = structuredAttrs->find(name); @@ -56,7 +56,7 @@ bool ParsedDerivation::getBoolAttr(const std::string & name, bool def) const } } -std::experimental::optional<Strings> ParsedDerivation::getStringsAttr(const std::string & name) const +std::optional<Strings> ParsedDerivation::getStringsAttr(const std::string & name) const { if (structuredAttrs) { auto i = structuredAttrs->find(name); diff --git a/src/libstore/parsed-derivations.hh b/src/libstore/parsed-derivations.hh index 0a82c146172b..ed07dc652e8d 100644 --- a/src/libstore/parsed-derivations.hh +++ b/src/libstore/parsed-derivations.hh @@ -8,22 +8,22 @@ class ParsedDerivation { Path drvPath; BasicDerivation & drv; - std::experimental::optional<nlohmann::json> structuredAttrs; + std::optional<nlohmann::json> structuredAttrs; public: ParsedDerivation(const Path & drvPath, BasicDerivation & drv); - const std::experimental::optional<nlohmann::json> & getStructuredAttrs() const + const std::optional<nlohmann::json> & getStructuredAttrs() const { return structuredAttrs; } - std::experimental::optional<std::string> getStringAttr(const std::string & name) const; + std::optional<std::string> getStringAttr(const std::string & name) const; bool getBoolAttr(const std::string & name, bool def = false) const; - std::experimental::optional<Strings> getStringsAttr(const std::string & name) const; + std::optional<Strings> getStringsAttr(const std::string & name) const; StringSet getRequiredSystemFeatures() const; diff --git a/src/libstore/remote-store.cc b/src/libstore/remote-store.cc index def140cfbe18..15faf78a526d 100644 --- a/src/libstore/remote-store.cc +++ b/src/libstore/remote-store.cc @@ -596,7 +596,7 @@ void RemoteStore::syncWithGC() } -Roots RemoteStore::findRoots() +Roots RemoteStore::findRoots(bool censor) { auto conn(getConnection()); conn->to << wopFindRoots; @@ -606,7 +606,7 @@ Roots RemoteStore::findRoots() while (count--) { Path link = readString(conn->from); Path target = readStorePath(*this, conn->from); - result[link] = target; + result[target].emplace(link); } return result; } diff --git a/src/libstore/remote-store.hh b/src/libstore/remote-store.hh index 4f554b5980e8..80f18ab715d9 100644 --- a/src/libstore/remote-store.hh +++ b/src/libstore/remote-store.hh @@ -82,7 +82,7 @@ public: void syncWithGC() override; - Roots findRoots() override; + Roots findRoots(bool censor) override; void collectGarbage(const GCOptions & options, GCResults & results) override; @@ -149,7 +149,7 @@ public: private: ref<RemoteStore::Connection> openConnection() override; - std::experimental::optional<std::string> path; + std::optional<std::string> path; }; diff --git a/src/libstore/s3-binary-cache-store.cc b/src/libstore/s3-binary-cache-store.cc index 51de89e0d92f..cd547a964850 100644 --- a/src/libstore/s3-binary-cache-store.cc +++ b/src/libstore/s3-binary-cache-store.cc @@ -126,6 +126,7 @@ ref<Aws::Client::ClientConfiguration> S3Helper::makeConfig(const string & region res->endpointOverride = endpoint; } res->requestTimeoutMs = 600 * 1000; + res->connectTimeoutMs = 5 * 1000; res->retryStrategy = std::make_shared<RetryStrategy>(); res->caFile = settings.caFile; return res; diff --git a/src/libstore/ssh.cc b/src/libstore/ssh.cc index cf133b57cb20..5e0e44935cca 100644 --- a/src/libstore/ssh.cc +++ b/src/libstore/ssh.cc @@ -1,5 +1,4 @@ #include "ssh.hh" -#include "affinity.hh" namespace nix { @@ -35,9 +34,7 @@ std::unique_ptr<SSHMaster::Connection> SSHMaster::startCommand(const std::string auto conn = std::make_unique<Connection>(); conn->sshPid = startProcess([&]() { - restoreAffinity(); restoreSignals(); - restoreMountNamespace(); close(in.writeSide.get()); close(out.readSide.get()); diff --git a/src/libstore/store-api.cc b/src/libstore/store-api.cc index dc54c735fdb1..28ad7c019a94 100644 --- a/src/libstore/store-api.cc +++ b/src/libstore/store-api.cc @@ -6,10 +6,11 @@ #include "thread-pool.hh" #include "json.hh" #include "derivations.hh" +#include "retry.hh" +#include "download.hh" #include <future> - namespace nix { @@ -562,64 +563,67 @@ void Store::buildPaths(const PathSet & paths, BuildMode buildMode) { for (auto & path : paths) if (isDerivation(path)) - unsupported(); + unsupported("buildPaths"); if (queryValidPaths(paths).size() != paths.size()) - unsupported(); + unsupported("buildPaths"); } void copyStorePath(ref<Store> srcStore, ref<Store> dstStore, const Path & storePath, RepairFlag repair, CheckSigsFlag checkSigs) { - auto srcUri = srcStore->getUri(); - auto dstUri = dstStore->getUri(); - - Activity act(*logger, lvlInfo, actCopyPath, - srcUri == "local" || srcUri == "daemon" - ? fmt("copying path '%s' to '%s'", storePath, dstUri) - : dstUri == "local" || dstUri == "daemon" - ? fmt("copying path '%s' from '%s'", storePath, srcUri) - : fmt("copying path '%s' from '%s' to '%s'", storePath, srcUri, dstUri), - {storePath, srcUri, dstUri}); - PushActivity pact(act.id); - - auto info = srcStore->queryPathInfo(storePath); - - uint64_t total = 0; - - if (!info->narHash) { - StringSink sink; - srcStore->narFromPath({storePath}, sink); - auto info2 = make_ref<ValidPathInfo>(*info); - info2->narHash = hashString(htSHA256, *sink.s); - if (!info->narSize) info2->narSize = sink.s->size(); - if (info->ultimate) info2->ultimate = false; - info = info2; - - StringSource source(*sink.s); - dstStore->addToStore(*info, source, repair, checkSigs); - return; - } + retry<void>(downloadSettings.tries, [&]() { - if (info->ultimate) { - auto info2 = make_ref<ValidPathInfo>(*info); - info2->ultimate = false; - info = info2; - } + auto srcUri = srcStore->getUri(); + auto dstUri = dstStore->getUri(); + + Activity act(*logger, lvlInfo, actCopyPath, + srcUri == "local" || srcUri == "daemon" + ? fmt("copying path '%s' to '%s'", storePath, dstUri) + : dstUri == "local" || dstUri == "daemon" + ? fmt("copying path '%s' from '%s'", storePath, srcUri) + : fmt("copying path '%s' from '%s' to '%s'", storePath, srcUri, dstUri), + {storePath, srcUri, dstUri}); + PushActivity pact(act.id); + + auto info = srcStore->queryPathInfo(storePath); + + uint64_t total = 0; + + if (!info->narHash) { + StringSink sink; + srcStore->narFromPath({storePath}, sink); + auto info2 = make_ref<ValidPathInfo>(*info); + info2->narHash = hashString(htSHA256, *sink.s); + if (!info->narSize) info2->narSize = sink.s->size(); + if (info->ultimate) info2->ultimate = false; + info = info2; + + StringSource source(*sink.s); + dstStore->addToStore(*info, source, repair, checkSigs); + return; + } + + if (info->ultimate) { + auto info2 = make_ref<ValidPathInfo>(*info); + info2->ultimate = false; + info = info2; + } - auto source = sinkToSource([&](Sink & sink) { - LambdaSink wrapperSink([&](const unsigned char * data, size_t len) { - sink(data, len); - total += len; - act.progress(total, info->narSize); + auto source = sinkToSource([&](Sink & sink) { + LambdaSink wrapperSink([&](const unsigned char * data, size_t len) { + sink(data, len); + total += len; + act.progress(total, info->narSize); + }); + srcStore->narFromPath({storePath}, wrapperSink); + }, [&]() { + throw EndOfFile("NAR for '%s' fetched from '%s' is incomplete", storePath, srcStore->getUri()); }); - srcStore->narFromPath({storePath}, wrapperSink); - }, [&]() { - throw EndOfFile("NAR for '%s' fetched from '%s' is incomplete", storePath, srcStore->getUri()); - }); - dstStore->addToStore(*info, *source, repair, checkSigs); + dstStore->addToStore(*info, *source, repair, checkSigs); + }); } @@ -842,12 +846,11 @@ namespace nix { RegisterStoreImplementation::Implementations * RegisterStoreImplementation::implementations = 0; - -ref<Store> openStore(const std::string & uri_, - const Store::Params & extraParams) +/* Split URI into protocol+hierarchy part and its parameter set. */ +std::pair<std::string, Store::Params> splitUriAndParams(const std::string & uri_) { auto uri(uri_); - Store::Params params(extraParams); + Store::Params params; auto q = uri.find('?'); if (q != std::string::npos) { for (auto s : tokenizeString<Strings>(uri.substr(q + 1), "&")) { @@ -873,6 +876,15 @@ ref<Store> openStore(const std::string & uri_, } uri = uri_.substr(0, q); } + return {uri, params}; +} + +ref<Store> openStore(const std::string & uri_, + const Store::Params & extraParams) +{ + auto [uri, uriParams] = splitUriAndParams(uri_); + auto params = extraParams; + params.insert(uriParams.begin(), uriParams.end()); for (auto fun : *RegisterStoreImplementation::implementations) { auto store = fun(uri, params); diff --git a/src/libstore/store-api.hh b/src/libstore/store-api.hh index 106b2be5e6b2..59967737670d 100644 --- a/src/libstore/store-api.hh +++ b/src/libstore/store-api.hh @@ -11,6 +11,8 @@ #include <atomic> #include <limits> #include <map> +#include <unordered_map> +#include <unordered_set> #include <memory> #include <string> @@ -47,7 +49,7 @@ const size_t storePathHashLen = 32; // i.e. 160 bits const uint32_t exportMagic = 0x4558494e; -typedef std::map<Path, Path> Roots; +typedef std::unordered_map<Path, std::unordered_set<std::string>> Roots; struct GCOptions @@ -349,7 +351,8 @@ public: (i.e. you'll get /nix/store/<hash> rather than /nix/store/<hash>-<name>). Use queryPathInfo() to obtain the full store path. */ - virtual PathSet queryAllValidPaths() = 0; + virtual PathSet queryAllValidPaths() + { unsupported("queryAllValidPaths"); } /* Query information about a valid path. It is permitted to omit the name part of the store path. */ @@ -368,8 +371,8 @@ public: /* Queries the set of incoming FS references for a store path. The result is not cleared. */ - virtual void queryReferrers(const Path & path, - PathSet & referrers) = 0; + virtual void queryReferrers(const Path & path, PathSet & referrers) + { unsupported("queryReferrers"); } /* Return all currently valid derivations that have `path' as an output. (Note that the result of `queryDeriver()' is the @@ -378,10 +381,12 @@ public: virtual PathSet queryValidDerivers(const Path & path) { return {}; }; /* Query the outputs of the derivation denoted by `path'. */ - virtual PathSet queryDerivationOutputs(const Path & path) = 0; + virtual PathSet queryDerivationOutputs(const Path & path) + { unsupported("queryDerivationOutputs"); } /* Query the output names of the derivation denoted by `path'. */ - virtual StringSet queryDerivationOutputNames(const Path & path) = 0; + virtual StringSet queryDerivationOutputNames(const Path & path) + { unsupported("queryDerivationOutputNames"); } /* Query the full store path given the hash part of a valid store path, or "" if the path doesn't exist. */ @@ -447,14 +452,16 @@ public: /* Add a store path as a temporary root of the garbage collector. The root disappears as soon as we exit. */ - virtual void addTempRoot(const Path & path) = 0; + virtual void addTempRoot(const Path & path) + { unsupported("addTempRoot"); } /* Add an indirect root, which is merely a symlink to `path' from /nix/var/nix/gcroots/auto/<hash of `path'>. `path' is supposed to be a symlink to a store path. The garbage collector will automatically remove the indirect root when it finds that `path' has disappeared. */ - virtual void addIndirectRoot(const Path & path) = 0; + virtual void addIndirectRoot(const Path & path) + { unsupported("addIndirectRoot"); } /* Acquire the global GC lock, then immediately release it. This function must be called after registering a new permanent root, @@ -478,11 +485,15 @@ public: /* Find the roots of the garbage collector. Each root is a pair (link, storepath) where `link' is the path of the symlink - outside of the Nix store that point to `storePath'. */ - virtual Roots findRoots() = 0; + outside of the Nix store that point to `storePath'. If + 'censor' is true, privacy-sensitive information about roots + found in /proc is censored. */ + virtual Roots findRoots(bool censor) + { unsupported("findRoots"); } /* Perform a garbage collection. */ - virtual void collectGarbage(const GCOptions & options, GCResults & results) = 0; + virtual void collectGarbage(const GCOptions & options, GCResults & results) + { unsupported("collectGarbage"); } /* Return a string representing information about the path that can be loaded into the database using `nix-store --load-db' or @@ -513,11 +524,13 @@ public: virtual bool verifyStore(bool checkContents, RepairFlag repair = NoRepair) { return false; }; /* Return an object to access files in the Nix store. */ - virtual ref<FSAccessor> getFSAccessor() = 0; + virtual ref<FSAccessor> getFSAccessor() + { unsupported("getFSAccessor"); } /* Add signatures to the specified store path. The signatures are not verified. */ - virtual void addSignatures(const Path & storePath, const StringSet & sigs) = 0; + virtual void addSignatures(const Path & storePath, const StringSet & sigs) + { unsupported("addSignatures"); } /* Utility functions. */ @@ -620,9 +633,9 @@ protected: Stats stats; /* Unsupported methods. */ - [[noreturn]] void unsupported() + [[noreturn]] void unsupported(const std::string & op) { - throw Unsupported("requested operation is not supported by store '%s'", getUri()); + throw Unsupported("operation '%s' is not supported by store '%s'", op, getUri()); } }; @@ -753,8 +766,7 @@ StoreType getStoreType(const std::string & uri = settings.storeUri.get(), const std::string & stateDir = settings.nixStateDir); /* Return the default substituter stores, defined by the - โsubstitutersโ option and various legacy options like - โbinary-cachesโ. */ + โsubstitutersโ option and various legacy options. */ std::list<ref<Store>> getDefaultSubstituters(); @@ -789,4 +801,8 @@ ValidPathInfo decodeValidPathInfo(std::istream & str, for paths created by makeFixedOutputPath() / addToStore(). */ std::string makeFixedOutputCA(bool recursive, const Hash & hash); + +/* Split URI into protocol+hierarchy part and its parameter set. */ +std::pair<std::string, Store::Params> splitUriAndParams(const std::string & uri); + } diff --git a/src/libutil/archive.cc b/src/libutil/archive.cc index bb68e82886d0..3aa120270970 100644 --- a/src/libutil/archive.cc +++ b/src/libutil/archive.cc @@ -331,7 +331,7 @@ struct RestoreSink : ParseSink filesystem doesn't support preallocation (e.g. on OpenSolaris). Since preallocation is just an optimisation, ignore it. */ - if (errno && errno != EINVAL) + if (errno && errno != EINVAL && errno != EOPNOTSUPP && errno != ENOSYS) throw SysError(format("preallocating file of %1% bytes") % len); } #endif diff --git a/src/libutil/hash.cc b/src/libutil/hash.cc index 9d82f13a5e38..1c14ebb187cc 100644 --- a/src/libutil/hash.cc +++ b/src/libutil/hash.cc @@ -105,9 +105,9 @@ string printHash16or32(const Hash & hash) std::string Hash::to_string(Base base, bool includeType) const { std::string s; - if (includeType) { + if (base == SRI || includeType) { s += printHashType(type); - s += ':'; + s += base == SRI ? '-' : ':'; } switch (base) { case Base16: @@ -117,6 +117,7 @@ std::string Hash::to_string(Base base, bool includeType) const s += printHash32(*this); break; case Base64: + case SRI: s += base64Encode(std::string((const char *) hash, hashSize)); break; } @@ -127,28 +128,33 @@ std::string Hash::to_string(Base base, bool includeType) const Hash::Hash(const std::string & s, HashType type) : type(type) { - auto colon = s.find(':'); - size_t pos = 0; - - if (colon == string::npos) { - if (type == htUnknown) + bool isSRI = false; + + auto sep = s.find(':'); + if (sep == string::npos) { + sep = s.find('-'); + if (sep != string::npos) { + isSRI = true; + } else if (type == htUnknown) throw BadHash("hash '%s' does not include a type", s); - } else { - string hts = string(s, 0, colon); + } + + if (sep != string::npos) { + string hts = string(s, 0, sep); this->type = parseHashType(hts); if (this->type == htUnknown) throw BadHash("unknown hash type '%s'", hts); if (type != htUnknown && type != this->type) throw BadHash("hash '%s' should have type '%s'", s, printHashType(type)); - pos = colon + 1; + pos = sep + 1; } init(); size_t size = s.size() - pos; - if (size == base16Len()) { + if (!isSRI && size == base16Len()) { auto parseHexDigit = [&](char c) { if (c >= '0' && c <= '9') return c - '0'; @@ -164,7 +170,7 @@ Hash::Hash(const std::string & s, HashType type) } } - else if (size == base32Len()) { + else if (!isSRI && size == base32Len()) { for (unsigned int n = 0; n < size; ++n) { char c = s[pos + size - n - 1]; @@ -187,10 +193,10 @@ Hash::Hash(const std::string & s, HashType type) } } - else if (size == base64Len()) { + else if (isSRI || size == base64Len()) { auto d = base64Decode(std::string(s, pos)); if (d.size() != hashSize) - throw BadHash("invalid base-64 hash '%s'", s); + throw BadHash("invalid %s hash '%s'", isSRI ? "SRI" : "base-64", s); assert(hashSize); memcpy(hash, d.data(), hashSize); } diff --git a/src/libutil/hash.hh b/src/libutil/hash.hh index fd7a61df8e46..2dbc3b630814 100644 --- a/src/libutil/hash.hh +++ b/src/libutil/hash.hh @@ -20,7 +20,7 @@ const int sha512HashSize = 64; extern const string base32Chars; -enum Base : int { Base64, Base32, Base16 }; +enum Base : int { Base64, Base32, Base16, SRI }; struct Hash @@ -38,8 +38,9 @@ struct Hash Hash(HashType type) : type(type) { init(); }; /* Initialize the hash from a string representation, in the format - "[<type>:]<base16|base32|base64>". If the 'type' argument is - htUnknown, then the hash type must be specified in the + "[<type>:]<base16|base32|base64>" or "<type>-<base64>" (a + Subresource Integrity hash expression). If the 'type' argument + is htUnknown, then the hash type must be specified in the string. */ Hash(const std::string & s, HashType type = htUnknown); diff --git a/src/libutil/local.mk b/src/libutil/local.mk index 824f48fbfc9f..3ccc23fd5c1b 100644 --- a/src/libutil/local.mk +++ b/src/libutil/local.mk @@ -7,7 +7,3 @@ libutil_DIR := $(d) libutil_SOURCES := $(wildcard $(d)/*.cc) libutil_LDFLAGS = $(LIBLZMA_LIBS) -lbz2 -pthread $(OPENSSL_LIBS) $(LIBBROTLI_LIBS) -lboost_context - -libutil_LIBS = libformat - -libutil_CXXFLAGS = -DBROTLI=\"$(brotli)\" diff --git a/src/libutil/logging.cc b/src/libutil/logging.cc index 799c6e1ae441..b379306f6ec0 100644 --- a/src/libutil/logging.cc +++ b/src/libutil/logging.cc @@ -21,7 +21,7 @@ Logger * logger = makeDefaultLogger(); void Logger::warn(const std::string & msg) { - log(lvlInfo, ANSI_RED "warning:" ANSI_NORMAL " " + msg); + log(lvlWarn, ANSI_RED "warning:" ANSI_NORMAL " " + msg); } class SimpleLogger : public Logger @@ -46,6 +46,7 @@ public: char c; switch (lvl) { case lvlError: c = '3'; break; + case lvlWarn: c = '4'; break; case lvlInfo: c = '5'; break; case lvlTalkative: case lvlChatty: c = '6'; break; default: c = '7'; diff --git a/src/libutil/logging.hh b/src/libutil/logging.hh index 678703102e9b..5f221944594b 100644 --- a/src/libutil/logging.hh +++ b/src/libutil/logging.hh @@ -6,6 +6,7 @@ namespace nix { typedef enum { lvlError = 0, + lvlWarn, lvlInfo, lvlTalkative, lvlChatty, diff --git a/src/libutil/lru-cache.hh b/src/libutil/lru-cache.hh index 9b8290e634c9..8b83f842c324 100644 --- a/src/libutil/lru-cache.hh +++ b/src/libutil/lru-cache.hh @@ -2,7 +2,7 @@ #include <map> #include <list> -#include <experimental/optional> +#include <optional> namespace nix { @@ -64,7 +64,7 @@ public: /* Look up an item in the cache. If it exists, it becomes the most recently used item. */ - std::experimental::optional<Value> get(const Key & key) + std::optional<Value> get(const Key & key) { auto i = data.find(key); if (i == data.end()) return {}; diff --git a/src/libutil/retry.hh b/src/libutil/retry.hh new file mode 100644 index 000000000000..b45cb37f736b --- /dev/null +++ b/src/libutil/retry.hh @@ -0,0 +1,38 @@ +#pragma once + +#include "logging.hh" + +#include <functional> +#include <cmath> +#include <random> +#include <thread> + +namespace nix { + +inline unsigned int retrySleepTime(unsigned int attempt) +{ + std::random_device rd; + std::mt19937 mt19937; + return 250.0 * std::pow(2.0f, + attempt - 1 + std::uniform_real_distribution<>(0.0, 0.5)(mt19937)); +} + +template<typename C> +C retry(unsigned int attempts, std::function<C()> && f) +{ + unsigned int attempt = 0; + while (true) { + try { + return f(); + } catch (BaseError & e) { + ++attempt; + if (attempt >= attempts || !e.isTransient()) + throw; + auto ms = retrySleepTime(attempt); + warn("%s; retrying in %d ms", e.what(), ms); + std::this_thread::sleep_for(std::chrono::milliseconds(ms)); + } + } +} + +} diff --git a/src/libutil/serialise.cc b/src/libutil/serialise.cc index 0e75eeec2bfe..8201549fd7d0 100644 --- a/src/libutil/serialise.cc +++ b/src/libutil/serialise.cc @@ -171,7 +171,7 @@ std::unique_ptr<Source> sinkToSource( std::function<void(Sink &)> fun; std::function<void()> eof; - std::experimental::optional<coro_t::pull_type> coro; + std::optional<coro_t::pull_type> coro; bool started = false; SinkToSource(std::function<void(Sink &)> fun, std::function<void()> eof) diff --git a/src/libutil/types.hh b/src/libutil/types.hh index 92bf469b5c6f..88e3243f47a5 100644 --- a/src/libutil/types.hh +++ b/src/libutil/types.hh @@ -109,6 +109,8 @@ public: const string & msg() const { return err; } const string & prefix() const { return prefix_; } BaseError & addPrefix(const FormatOrString & fs); + + virtual bool isTransient() { return false; } }; #define MakeError(newClass, superClass) \ diff --git a/src/libutil/util.cc b/src/libutil/util.cc index ce50334e1e62..17aee2d5c3d0 100644 --- a/src/libutil/util.cc +++ b/src/libutil/util.cc @@ -16,6 +16,7 @@ #include <future> #include <fcntl.h> +#include <grp.h> #include <limits.h> #include <pwd.h> #include <sys/ioctl.h> @@ -38,6 +39,9 @@ extern char * * environ; namespace nix { +const std::string nativeSystem = SYSTEM; + + BaseError & BaseError::addPrefix(const FormatOrString & fs) { prefix_ = fs.s + prefix_; @@ -936,8 +940,6 @@ pid_t startProcess(std::function<void()> fun, const ProcessOptions & options) throw SysError("setting death signal"); #endif restoreAffinity(); - if (options.restoreMountNamespace) - restoreMountNamespace(); fun(); } catch (std::exception & e) { try { @@ -967,7 +969,7 @@ std::vector<char *> stringsToCharPtrs(const Strings & ss) string runProgram(Path program, bool searchPath, const Strings & args, - const std::experimental::optional<std::string> & input) + const std::optional<std::string> & input) { RunOptions opts(program, args); opts.searchPath = searchPath; @@ -1024,6 +1026,16 @@ void runProgram2(const RunOptions & options) if (source && dup2(in.readSide.get(), STDIN_FILENO) == -1) throw SysError("dupping stdin"); + if (options.chdir && chdir((*options.chdir).c_str()) == -1) + throw SysError("chdir failed"); + if (options.gid && setgid(*options.gid) == -1) + throw SysError("setgid failed"); + /* Drop all other groups if we're setgid. */ + if (options.gid && setgroups(0, 0) == -1) + throw SysError("setgroups failed"); + if (options.uid && setuid(*options.uid) == -1) + throw SysError("setuid failed"); + Strings args_(options.args); args_.push_front(options.program); @@ -1506,26 +1518,4 @@ std::unique_ptr<InterruptCallback> createInterruptCallback(std::function<void()> return std::unique_ptr<InterruptCallback>(res.release()); } -static AutoCloseFD fdSavedMountNamespace; - -void saveMountNamespace() -{ -#if __linux__ - std::once_flag done; - std::call_once(done, []() { - fdSavedMountNamespace = open("/proc/self/ns/mnt", O_RDONLY); - if (!fdSavedMountNamespace) - throw SysError("saving parent mount namespace"); - }); -#endif -} - -void restoreMountNamespace() -{ -#if __linux__ - if (fdSavedMountNamespace && setns(fdSavedMountNamespace.get(), CLONE_NEWNS) == -1) - throw SysError("restoring parent mount namespace"); -#endif -} - } diff --git a/src/libutil/util.hh b/src/libutil/util.hh index d67bddc138c8..fce3cab8def5 100644 --- a/src/libutil/util.hh +++ b/src/libutil/util.hh @@ -14,7 +14,7 @@ #include <cstdio> #include <map> #include <sstream> -#include <experimental/optional> +#include <optional> #include <future> #ifndef HAVE_STRUCT_DIRENT_D_TYPE @@ -30,6 +30,10 @@ struct Sink; struct Source; +/* The system for which Nix is compiled. */ +extern const std::string nativeSystem; + + /* Return an environment variable. */ string getEnv(const string & key, const string & def = ""); @@ -250,7 +254,6 @@ struct ProcessOptions bool dieWithParent = true; bool runExitHandlers = false; bool allowVfork = true; - bool restoreMountNamespace = true; }; pid_t startProcess(std::function<void()> fun, const ProcessOptions & options = ProcessOptions()); @@ -260,14 +263,17 @@ pid_t startProcess(std::function<void()> fun, const ProcessOptions & options = P shell backtick operator). */ string runProgram(Path program, bool searchPath = false, const Strings & args = Strings(), - const std::experimental::optional<std::string> & input = {}); + const std::optional<std::string> & input = {}); struct RunOptions { + std::optional<uid_t> uid; + std::optional<uid_t> gid; + std::optional<Path> chdir; Path program; bool searchPath = true; Strings args; - std::experimental::optional<std::string> input; + std::optional<std::string> input; Source * standardIn = nullptr; Sink * standardOut = nullptr; bool _killStderr = false; @@ -402,6 +408,7 @@ void ignoreException(); /* Some ANSI escape sequences. */ #define ANSI_NORMAL "\e[0m" #define ANSI_BOLD "\e[1m" +#define ANSI_FAINT "\e[2m" #define ANSI_RED "\e[31;1m" #define ANSI_GREEN "\e[32;1m" #define ANSI_BLUE "\e[34;1m" @@ -515,13 +522,4 @@ typedef std::function<bool(const Path & path)> PathFilter; extern PathFilter defaultPathFilter; -/* Save the current mount namespace. Ignored if called more than - once. */ -void saveMountNamespace(); - -/* Restore the mount namespace saved by saveMountNamespace(). Ignored - if saveMountNamespace() was never called. */ -void restoreMountNamespace(); - - } diff --git a/src/nix-build/nix-build.cc b/src/nix-build/nix-build.cc index 11ea3b1f7ae1..c6a4d416648f 100755 --- a/src/nix-build/nix-build.cc +++ b/src/nix-build/nix-build.cc @@ -274,19 +274,21 @@ static void _main(int argc, char * * argv) exprs = {state->parseStdin()}; else for (auto i : left) { - auto absolute = i; - try { - absolute = canonPath(absPath(i), true); - } catch (Error e) {}; if (fromArgs) exprs.push_back(state->parseExprFromString(i, absPath("."))); - else if (store->isStorePath(absolute) && std::regex_match(absolute, std::regex(".*\\.drv(!.*)?"))) + else { + auto absolute = i; + try { + absolute = canonPath(absPath(i), true); + } catch (Error e) {}; + if (store->isStorePath(absolute) && std::regex_match(absolute, std::regex(".*\\.drv(!.*)?"))) drvs.push_back(DrvInfo(*state, store, absolute)); else /* If we're in a #! script, interpret filenames relative to the script. */ exprs.push_back(state->parseExprFromFile(resolveExprPath(state->checkSourcePath(lookupFileArg(*state, inShebang && !packages ? absPath(i, absPath(dirOf(script))) : i))))); + } } /* Evaluate them into derivations. */ @@ -401,6 +403,8 @@ static void _main(int argc, char * * argv) } else env[var.first] = var.second; + restoreAffinity(); + /* Run a shell using the derivation's environment. For convenience, source $stdenv/setup to setup additional environment variables and shell functions. Also don't @@ -444,9 +448,7 @@ static void _main(int argc, char * * argv) auto argPtrs = stringsToCharPtrs(args); - restoreAffinity(); restoreSignals(); - restoreMountNamespace(); execvp(shell.c_str(), argPtrs.data()); diff --git a/src/nix-channel/nix-channel.cc b/src/nix-channel/nix-channel.cc index 8b66cc7e314e..06eb3d23ba83 100755 --- a/src/nix-channel/nix-channel.cc +++ b/src/nix-channel/nix-channel.cc @@ -86,10 +86,12 @@ static void update(const StringSet & channelNames) // We want to download the url to a file to see if it's a tarball while also checking if we // got redirected in the process, so that we can grab the various parts of a nix channel // definition from a consistent location if the redirect changes mid-download. - std::string effectiveUrl; + CachedDownloadRequest request(url); + request.ttl = 0; auto dl = getDownloader(); - auto filename = dl->downloadCached(store, url, false, "", Hash(), &effectiveUrl, 0); - url = chomp(std::move(effectiveUrl)); + auto result = dl->downloadCached(store, request); + auto filename = result.path; + url = chomp(result.effectiveUri); // If the URL contains a version number, append it to the name // attribute (so that "nix-env -q" on the channels profile @@ -111,22 +113,11 @@ static void update(const StringSet & channelNames) } if (!unpacked) { - // The URL doesn't unpack directly, so let's try treating it like a full channel folder with files in it - // Check if the channel advertises a binary cache. - DownloadRequest request(url + "/binary-cache-url"); - try { - auto dlRes = dl->download(request); - extraAttrs = "binaryCacheURL = \"" + *dlRes.data + "\";"; - } catch (DownloadError & e) { - } - // Download the channel tarball. - auto fullURL = url + "/nixexprs.tar.xz"; try { - filename = dl->downloadCached(store, fullURL, false); + filename = dl->downloadCached(store, CachedDownloadRequest(url + "/nixexprs.tar.xz")).path; } catch (DownloadError & e) { - fullURL = url + "/nixexprs.tar.bz2"; - filename = dl->downloadCached(store, fullURL, false); + filename = dl->downloadCached(store, CachedDownloadRequest(url + "/nixexprs.tar.bz2")).path; } chomp(filename); } diff --git a/src/nix-daemon/nix-daemon.cc b/src/nix-daemon/nix-daemon.cc index 8368c3266142..e88aaf636444 100644 --- a/src/nix-daemon/nix-daemon.cc +++ b/src/nix-daemon/nix-daemon.cc @@ -475,11 +475,19 @@ static void performOp(TunnelLogger * logger, ref<Store> store, case wopFindRoots: { logger->startWork(); - Roots roots = store->findRoots(); + Roots roots = store->findRoots(!trusted); logger->stopWork(); - to << roots.size(); + + size_t size = 0; for (auto & i : roots) - to << i.first << i.second; + size += i.second.size(); + + to << size; + + for (auto & [target, links] : roots) + for (auto & link : links) + to << link << target; + break; } @@ -566,7 +574,7 @@ static void performOp(TunnelLogger * logger, ref<Store> store, else if (setSubstituters(settings.extraSubstituters)) ; else - debug("ignoring untrusted setting '%s'", name); + warn("ignoring the user-specified setting '%s', because it is a restricted setting and you are not a trusted user", name); } catch (UsageError & e) { warn(e.what()); } diff --git a/src/nix-store/nix-store.cc b/src/nix-store/nix-store.cc index a9ad14762e62..f324056bb3a1 100644 --- a/src/nix-store/nix-store.cc +++ b/src/nix-store/nix-store.cc @@ -427,10 +427,11 @@ static void opQuery(Strings opFlags, Strings opArgs) maybeUseOutputs(store->followLinksToStorePath(i), useOutput, forceRealise), referrers, true, settings.gcKeepOutputs, settings.gcKeepDerivations); } - Roots roots = store->findRoots(); - for (auto & i : roots) - if (referrers.find(i.second) != referrers.end()) - cout << format("%1%\n") % i.first; + Roots roots = store->findRoots(false); + for (auto & [target, links] : roots) + if (referrers.find(target) != referrers.end()) + for (auto & link : links) + cout << format("%1% -> %2%\n") % link % target; break; } @@ -485,11 +486,16 @@ static void opReadLog(Strings opFlags, Strings opArgs) static void opDumpDB(Strings opFlags, Strings opArgs) { if (!opFlags.empty()) throw UsageError("unknown flag"); - if (!opArgs.empty()) - throw UsageError("no arguments expected"); - PathSet validPaths = store->queryAllValidPaths(); - for (auto & i : validPaths) - cout << store->makeValidityRegistration({i}, true, true); + if (!opArgs.empty()) { + for (auto & i : opArgs) + i = store->followLinksToStorePath(i); + for (auto & i : opArgs) + cout << store->makeValidityRegistration({i}, true, true); + } else { + PathSet validPaths = store->queryAllValidPaths(); + for (auto & i : validPaths) + cout << store->makeValidityRegistration({i}, true, true); + } } @@ -585,9 +591,14 @@ static void opGC(Strings opFlags, Strings opArgs) if (!opArgs.empty()) throw UsageError("no arguments expected"); if (printRoots) { - Roots roots = store->findRoots(); - for (auto & i : roots) - cout << i.first << " -> " << i.second << std::endl; + Roots roots = store->findRoots(false); + std::set<std::pair<Path, Path>> roots2; + // Transpose and sort the roots. + for (auto & [target, links] : roots) + for (auto & link : links) + roots2.emplace(link, target); + for (auto & [link, target] : roots2) + std::cout << link << " -> " << target << "\n"; } else { diff --git a/src/nix/add-to-store.cc b/src/nix/add-to-store.cc index d0003790d3b9..e86b96e3f3f2 100644 --- a/src/nix/add-to-store.cc +++ b/src/nix/add-to-store.cc @@ -8,7 +8,7 @@ using namespace nix; struct CmdAddToStore : MixDryRun, StoreCommand { Path path; - std::experimental::optional<std::string> namePart; + std::optional<std::string> namePart; CmdAddToStore() { diff --git a/src/nix/edit.cc b/src/nix/edit.cc index d8d5895bd867..c9671f76d0fa 100644 --- a/src/nix/edit.cc +++ b/src/nix/edit.cc @@ -3,7 +3,6 @@ #include "eval.hh" #include "attr-path.hh" #include "progress-bar.hh" -#include "affinity.hh" #include <unistd.h> @@ -73,10 +72,6 @@ struct CmdEdit : InstallableCommand stopProgressBar(); - restoreAffinity(); - restoreSignals(); - restoreMountNamespace(); - execvp(args.front().c_str(), stringsToCharPtrs(args).data()); throw SysError("cannot run editor '%s'", editor); diff --git a/src/nix/hash.cc b/src/nix/hash.cc index 64062fb97955..af4105e28904 100644 --- a/src/nix/hash.cc +++ b/src/nix/hash.cc @@ -9,13 +9,14 @@ struct CmdHash : Command { enum Mode { mFile, mPath }; Mode mode; - Base base = Base16; + Base base = SRI; bool truncate = false; - HashType ht = htSHA512; + HashType ht = htSHA256; std::vector<std::string> paths; CmdHash(Mode mode) : mode(mode) { + mkFlag(0, "sri", "print hash in SRI format", &base, SRI); mkFlag(0, "base64", "print hash in base-64", &base, Base64); mkFlag(0, "base32", "print hash in base-32 (Nix-specific)", &base, Base32); mkFlag(0, "base16", "print hash in base-16", &base, Base16); @@ -43,7 +44,7 @@ struct CmdHash : Command Hash h = mode == mFile ? hashFile(ht, path) : hashPath(ht, path).first; if (truncate && h.hashSize > 20) h = compressHash(h, 20); std::cout << format("%1%\n") % - h.to_string(base, false); + h.to_string(base, base == SRI); } } }; @@ -54,7 +55,7 @@ static RegisterCommand r2(make_ref<CmdHash>(CmdHash::mPath)); struct CmdToBase : Command { Base base; - HashType ht = htSHA512; + HashType ht = htUnknown; std::vector<std::string> args; CmdToBase(Base base) : base(base) @@ -70,26 +71,30 @@ struct CmdToBase : Command return base == Base16 ? "to-base16" : base == Base32 ? "to-base32" : - "to-base64"; + base == Base64 ? "to-base64" : + "to-sri"; } std::string description() override { - return fmt("convert a hash to base-%d representation", - base == Base16 ? 16 : - base == Base32 ? 32 : 64); + return fmt("convert a hash to %s representation", + base == Base16 ? "base-16" : + base == Base32 ? "base-32" : + base == Base64 ? "base-64" : + "SRI"); } void run() override { for (auto s : args) - std::cout << fmt("%s\n", Hash(s, ht).to_string(base, false)); + std::cout << fmt("%s\n", Hash(s, ht).to_string(base, base == SRI)); } }; static RegisterCommand r3(make_ref<CmdToBase>(Base16)); static RegisterCommand r4(make_ref<CmdToBase>(Base32)); static RegisterCommand r5(make_ref<CmdToBase>(Base64)); +static RegisterCommand r6(make_ref<CmdToBase>(SRI)); /* Legacy nix-hash command. */ static int compatNixHash(int argc, char * * argv) diff --git a/src/nix/local.mk b/src/nix/local.mk index 40a0e8d6bde1..ca4604d566c3 100644 --- a/src/nix/local.mk +++ b/src/nix/local.mk @@ -15,7 +15,7 @@ nix_SOURCES := \ $(wildcard src/nix-prefetch-url/*.cc) \ $(wildcard src/nix-store/*.cc) \ -nix_LIBS = libexpr libmain libstore libutil libformat +nix_LIBS = libexpr libmain libstore libutil nix_LDFLAGS = -pthread $(SODIUM_LIBS) $(EDITLINE_LIBS) diff --git a/src/nix/main.cc b/src/nix/main.cc index 64c1dc35787c..a80fd0ea62fc 100644 --- a/src/nix/main.cc +++ b/src/nix/main.cc @@ -8,18 +8,53 @@ #include "shared.hh" #include "store-api.hh" #include "progress-bar.hh" +#include "download.hh" #include "finally.hh" +#include <sys/types.h> +#include <sys/socket.h> +#include <ifaddrs.h> +#include <netdb.h> + extern std::string chrootHelperName; void chrootHelper(int argc, char * * argv); namespace nix { +/* Check if we have a non-loopback/link-local network interface. */ +static bool haveInternet() +{ + struct ifaddrs * addrs; + + if (getifaddrs(&addrs)) + return true; + + Finally free([&]() { freeifaddrs(addrs); }); + + for (auto i = addrs; i; i = i->ifa_next) { + if (!i->ifa_addr) continue; + if (i->ifa_addr->sa_family == AF_INET) { + if (ntohl(((sockaddr_in *) i->ifa_addr)->sin_addr.s_addr) != INADDR_LOOPBACK) { + return true; + } + } else if (i->ifa_addr->sa_family == AF_INET6) { + if (!IN6_IS_ADDR_LOOPBACK(&((sockaddr_in6 *) i->ifa_addr)->sin6_addr) && + !IN6_IS_ADDR_LINKLOCAL(&((sockaddr_in6 *) i->ifa_addr)->sin6_addr)) + return true; + } + } + + return false; +} + std::string programPath; struct NixArgs : virtual MultiCommand, virtual MixCommonArgs { + bool printBuildLogs = false; + bool useNet = true; + NixArgs() : MultiCommand(*RegisterCommand::commands), MixCommonArgs("nix") { mkFlag() @@ -42,9 +77,20 @@ struct NixArgs : virtual MultiCommand, virtual MixCommonArgs }); mkFlag() + .longName("print-build-logs") + .shortName('L') + .description("print full build logs on stderr") + .set(&printBuildLogs, true); + + mkFlag() .longName("version") .description("show version information") .handler([&]() { printVersion(programName); }); + + mkFlag() + .longName("no-net") + .description("disable substituters and consider all previously downloaded files up-to-date") + .handler([&]() { useNet = false; }); } void printFlags(std::ostream & out) override @@ -85,7 +131,7 @@ void mainWrapped(int argc, char * * argv) if (legacy) return legacy(argc, argv); } - verbosity = lvlError; + verbosity = lvlWarn; settings.verboseBuild = false; NixArgs args; @@ -98,8 +144,24 @@ void mainWrapped(int argc, char * * argv) Finally f([]() { stopProgressBar(); }); - if (isatty(STDERR_FILENO)) - startProgressBar(); + startProgressBar(args.printBuildLogs); + + if (args.useNet && !haveInternet()) { + warn("you don't have Internet access; disabling some network-dependent features"); + args.useNet = false; + } + + if (!args.useNet) { + // FIXME: should check for command line overrides only. + if (!settings.useSubstitutes.overriden) + settings.useSubstitutes = false; + if (!settings.tarballTtl.overriden) + settings.tarballTtl = std::numeric_limits<unsigned int>::max(); + if (!downloadSettings.tries.overriden) + downloadSettings.tries = 0; + if (!downloadSettings.connectTimeout.overriden) + downloadSettings.connectTimeout = 1; + } args.command->prepare(); args.command->run(); diff --git a/src/nix/progress-bar.cc b/src/nix/progress-bar.cc index 40b905ba3243..b1c1d87de1a2 100644 --- a/src/nix/progress-bar.cc +++ b/src/nix/progress-bar.cc @@ -2,6 +2,7 @@ #include "util.hh" #include "sync.hh" #include "store-api.hh" +#include "names.hh" #include <atomic> #include <map> @@ -38,6 +39,7 @@ private: std::map<ActivityType, uint64_t> expectedByType; bool visible = true; ActivityId parent; + std::optional<std::string> name; }; struct ActivitiesByType @@ -60,6 +62,7 @@ private: uint64_t corruptedPaths = 0, untrustedPaths = 0; bool active = true; + bool haveUpdate = true; }; Sync<State> state_; @@ -68,14 +71,21 @@ private: std::condition_variable quitCV, updateCV; + bool printBuildLogs; + bool isTTY; + public: - ProgressBar() + ProgressBar(bool printBuildLogs, bool isTTY) + : printBuildLogs(printBuildLogs) + , isTTY(isTTY) { + state_.lock()->active = isTTY; updateThread = std::thread([&]() { auto state(state_.lock()); while (state->active) { - state.wait(updateCV); + if (!state->haveUpdate) + state.wait(updateCV); draw(*state); state.wait_for(quitCV, std::chrono::milliseconds(50)); } @@ -109,8 +119,14 @@ public: void log(State & state, Verbosity lvl, const std::string & s) { - writeToStderr("\r\e[K" + s + ANSI_NORMAL "\n"); - draw(state); + if (state.active) { + writeToStderr("\r\e[K" + s + ANSI_NORMAL "\n"); + draw(state); + } else { + auto s2 = s + ANSI_NORMAL "\n"; + if (!isTTY) s2 = filterANSIEscapes(s2, true); + writeToStderr(s2); + } } void startActivity(ActivityId act, Verbosity lvl, ActivityType type, @@ -141,6 +157,7 @@ public: auto nrRounds = getI(fields, 3); if (nrRounds != 1) i->s += fmt(" (round %d/%d)", curRound, nrRounds); + i->name = DrvName(name).name; } if (type == actSubstitute) { @@ -163,7 +180,7 @@ public: || (type == actCopyPath && hasAncestor(*state, actSubstitute, parent))) i->visible = false; - update(); + update(*state); } /* Check whether an activity has an ancestore with the specified @@ -198,7 +215,7 @@ public: state->its.erase(i); } - update(); + update(*state); } void result(ActivityId act, ResultType type, const std::vector<Field> & fields) override @@ -208,7 +225,7 @@ public: if (type == resFileLinked) { state->filesLinked++; state->bytesLinked += getI(fields, 0); - update(); + update(*state); } else if (type == resBuildLogLine) { @@ -217,29 +234,33 @@ public: auto i = state->its.find(act); assert(i != state->its.end()); ActInfo info = *i->second; - state->activities.erase(i->second); - info.lastLine = lastLine; - state->activities.emplace_back(info); - i->second = std::prev(state->activities.end()); - update(); + if (printBuildLogs) { + log(*state, lvlInfo, ANSI_FAINT + info.name.value_or("unnamed") + "> " + ANSI_NORMAL + lastLine); + } else { + state->activities.erase(i->second); + info.lastLine = lastLine; + state->activities.emplace_back(info); + i->second = std::prev(state->activities.end()); + update(*state); + } } } else if (type == resUntrustedPath) { state->untrustedPaths++; - update(); + update(*state); } else if (type == resCorruptedPath) { state->corruptedPaths++; - update(); + update(*state); } else if (type == resSetPhase) { auto i = state->its.find(act); assert(i != state->its.end()); i->second->phase = getS(fields, 0); - update(); + update(*state); } else if (type == resProgress) { @@ -250,7 +271,7 @@ public: actInfo.expected = getI(fields, 1); actInfo.running = getI(fields, 2); actInfo.failed = getI(fields, 3); - update(); + update(*state); } else if (type == resSetExpected) { @@ -262,17 +283,19 @@ public: state->activitiesByType[type].expected -= j; j = getI(fields, 1); state->activitiesByType[type].expected += j; - update(); + update(*state); } } - void update() + void update(State & state) { + state.haveUpdate = true; updateCV.notify_one(); } void draw(State & state) { + state.haveUpdate = false; if (!state.active) return; std::string line; @@ -333,11 +356,18 @@ public: if (running || done || expected || failed) { if (running) - s = fmt(ANSI_BLUE + numberFmt + ANSI_NORMAL "/" ANSI_GREEN + numberFmt + ANSI_NORMAL "/" + numberFmt, - running / unit, done / unit, expected / unit); + if (expected != 0) + s = fmt(ANSI_BLUE + numberFmt + ANSI_NORMAL "/" ANSI_GREEN + numberFmt + ANSI_NORMAL "/" + numberFmt, + running / unit, done / unit, expected / unit); + else + s = fmt(ANSI_BLUE + numberFmt + ANSI_NORMAL "/" ANSI_GREEN + numberFmt + ANSI_NORMAL, + running / unit, done / unit); else if (expected != done) - s = fmt(ANSI_GREEN + numberFmt + ANSI_NORMAL "/" + numberFmt, - done / unit, expected / unit); + if (expected != 0) + s = fmt(ANSI_GREEN + numberFmt + ANSI_NORMAL "/" + numberFmt, + done / unit, expected / unit); + else + s = fmt(ANSI_GREEN + numberFmt + ANSI_NORMAL, done / unit); else s = fmt(done ? ANSI_GREEN + numberFmt + ANSI_NORMAL : numberFmt, done / unit); s = fmt(itemFmt, s); @@ -395,9 +425,9 @@ public: } }; -void startProgressBar() +void startProgressBar(bool printBuildLogs) { - logger = new ProgressBar(); + logger = new ProgressBar(printBuildLogs, isatty(STDERR_FILENO)); } void stopProgressBar() diff --git a/src/nix/progress-bar.hh b/src/nix/progress-bar.hh index af8eda5a84fd..4d61175c24e4 100644 --- a/src/nix/progress-bar.hh +++ b/src/nix/progress-bar.hh @@ -4,7 +4,7 @@ namespace nix { -void startProgressBar(); +void startProgressBar(bool printBuildLogs = false); void stopProgressBar(); diff --git a/src/nix/repl.cc b/src/nix/repl.cc index d4806d74adb8..d8f812149069 100644 --- a/src/nix/repl.cc +++ b/src/nix/repl.cc @@ -192,6 +192,14 @@ static int listPossibleCallback(char *s, char ***avp) { return ac; } +namespace { + // Used to communicate to NixRepl::getLine whether a signal occurred in ::readline. + volatile sig_atomic_t g_signal_received = 0; + + void sigintHandler(int signo) { + g_signal_received = signo; + } +} void NixRepl::mainLoop(const std::vector<std::string> & files) { @@ -251,8 +259,40 @@ void NixRepl::mainLoop(const std::vector<std::string> & files) bool NixRepl::getLine(string & input, const std::string &prompt) { + struct sigaction act, old; + sigset_t savedSignalMask, set; + + auto setupSignals = [&]() { + act.sa_handler = sigintHandler; + sigfillset(&act.sa_mask); + act.sa_flags = 0; + if (sigaction(SIGINT, &act, &old)) + throw SysError("installing handler for SIGINT"); + + sigemptyset(&set); + sigaddset(&set, SIGINT); + if (sigprocmask(SIG_UNBLOCK, &set, &savedSignalMask)) + throw SysError("unblocking SIGINT"); + }; + auto restoreSignals = [&]() { + if (sigprocmask(SIG_SETMASK, &savedSignalMask, nullptr)) + throw SysError("restoring signals"); + + if (sigaction(SIGINT, &old, 0)) + throw SysError("restoring handler for SIGINT"); + }; + + setupSignals(); char * s = readline(prompt.c_str()); Finally doFree([&]() { free(s); }); + restoreSignals(); + + if (g_signal_received) { + g_signal_received = 0; + input.clear(); + return true; + } + if (!s) return false; input += s; @@ -337,8 +377,6 @@ static int runProgram(const string & program, const Strings & args) if (pid == -1) throw SysError("forking"); if (pid == 0) { restoreAffinity(); - restoreSignals(); - restoreMountNamespace(); execvp(program.c_str(), stringsToCharPtrs(args2).data()); _exit(1); } @@ -643,30 +681,13 @@ std::ostream & NixRepl::printValue(std::ostream & str, Value & v, unsigned int m for (auto & i : *v.attrs) sorted[i.name] = i.value; - /* If this is a derivation, then don't show the - self-references ("all", "out", etc.). */ - StringSet hidden; - if (isDrv) { - hidden.insert("all"); - Bindings::iterator i = v.attrs->find(state.sOutputs); - if (i == v.attrs->end()) - hidden.insert("out"); - else { - state.forceList(*i->value); - for (unsigned int j = 0; j < i->value->listSize(); ++j) - hidden.insert(state.forceStringNoCtx(*i->value->listElems()[j])); - } - } - for (auto & i : sorted) { if (isVarName(i.first)) str << i.first; else printStringValue(str, i.first.c_str()); str << " = "; - if (hidden.find(i.first) != hidden.end()) - str << "ยซ...ยป"; - else if (seen.find(i.second) != seen.end()) + if (seen.find(i.second) != seen.end()) str << "ยซrepeatedยป"; else try { diff --git a/src/nix/run.cc b/src/nix/run.cc index 1297072989b9..35b763345872 100644 --- a/src/nix/run.cc +++ b/src/nix/run.cc @@ -153,9 +153,9 @@ struct CmdRun : InstallablesCommand stopProgressBar(); - restoreAffinity(); restoreSignals(); - restoreMountNamespace(); + + restoreAffinity(); /* If this is a diverted store (i.e. its "logical" location (typically /nix/store) differs from its "physical" location diff --git a/src/nlohmann/json.hpp b/src/nlohmann/json.hpp index 5b0b0ea5b301..c9af0bed36d6 100644 --- a/src/nlohmann/json.hpp +++ b/src/nlohmann/json.hpp @@ -1,11 +1,12 @@ /* __ _____ _____ _____ __| | __| | | | JSON for Modern C++ -| | |__ | | | | | | version 3.0.1 +| | |__ | | | | | | version 3.5.0 |_____|_____|_____|_|___| https://github.com/nlohmann/json Licensed under the MIT License <http://opensource.org/licenses/MIT>. -Copyright (c) 2013-2017 Niels Lohmann <http://nlohmann.me>. +SPDX-License-Identifier: MIT +Copyright (c) 2013-2018 Niels Lohmann <http://nlohmann.me>. Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal @@ -29,42 +30,104 @@ SOFTWARE. #ifndef NLOHMANN_JSON_HPP #define NLOHMANN_JSON_HPP -#include <algorithm> // all_of, copy, fill, find, for_each, generate_n, none_of, remove, reverse, transform -#include <array> // array +#define NLOHMANN_JSON_VERSION_MAJOR 3 +#define NLOHMANN_JSON_VERSION_MINOR 5 +#define NLOHMANN_JSON_VERSION_PATCH 0 + +#include <algorithm> // all_of, find, for_each #include <cassert> // assert #include <ciso646> // and, not, or -#include <clocale> // lconv, localeconv -#include <cmath> // isfinite, labs, ldexp, signbit #include <cstddef> // nullptr_t, ptrdiff_t, size_t -#include <cstdint> // int64_t, uint64_t -#include <cstdlib> // abort, strtod, strtof, strtold, strtoul, strtoll, strtoull -#include <cstring> // memcpy, strlen -#include <forward_list> // forward_list -#include <functional> // function, hash, less +#include <functional> // hash, less #include <initializer_list> // initializer_list -#include <iomanip> // hex -#include <iosfwd> // istream, ostream -#include <iterator> // advance, begin, back_inserter, bidirectional_iterator_tag, distance, end, inserter, iterator, iterator_traits, next, random_access_iterator_tag, reverse_iterator -#include <limits> // numeric_limits -#include <locale> // locale -#include <map> // map -#include <memory> // addressof, allocator, allocator_traits, unique_ptr +#include <iosfwd> // istream, ostream +#include <iterator> // random_access_iterator_tag #include <numeric> // accumulate -#include <sstream> // stringstream -#include <string> // getline, stoi, string, to_string -#include <type_traits> // add_pointer, conditional, decay, enable_if, false_type, integral_constant, is_arithmetic, is_base_of, is_const, is_constructible, is_convertible, is_default_constructible, is_enum, is_floating_point, is_integral, is_nothrow_move_assignable, is_nothrow_move_constructible, is_pointer, is_reference, is_same, is_scalar, is_signed, remove_const, remove_cv, remove_pointer, remove_reference, true_type, underlying_type -#include <utility> // declval, forward, make_pair, move, pair, swap -#include <valarray> // valarray +#include <string> // string, stoi, to_string +#include <utility> // declval, forward, move, pair, swap + +// #include <nlohmann/json_fwd.hpp> +#ifndef NLOHMANN_JSON_FWD_HPP +#define NLOHMANN_JSON_FWD_HPP + +#include <cstdint> // int64_t, uint64_t +#include <map> // map +#include <memory> // allocator +#include <string> // string #include <vector> // vector +/*! +@brief namespace for Niels Lohmann +@see https://github.com/nlohmann +@since version 1.0.0 +*/ +namespace nlohmann +{ +/*! +@brief default JSONSerializer template argument + +This serializer ignores the template arguments and uses ADL +([argument-dependent lookup](https://en.cppreference.com/w/cpp/language/adl)) +for serialization. +*/ +template<typename T = void, typename SFINAE = void> +struct adl_serializer; + +template<template<typename U, typename V, typename... Args> class ObjectType = + std::map, + template<typename U, typename... Args> class ArrayType = std::vector, + class StringType = std::string, class BooleanType = bool, + class NumberIntegerType = std::int64_t, + class NumberUnsignedType = std::uint64_t, + class NumberFloatType = double, + template<typename U> class AllocatorType = std::allocator, + template<typename T, typename SFINAE = void> class JSONSerializer = + adl_serializer> +class basic_json; + +/*! +@brief JSON Pointer + +A JSON pointer defines a string syntax for identifying a specific value +within a JSON document. It can be used with functions `at` and +`operator[]`. Furthermore, JSON pointers are the base for JSON patches. + +@sa [RFC 6901](https://tools.ietf.org/html/rfc6901) + +@since version 2.0.0 +*/ +template<typename BasicJsonType> +class json_pointer; + +/*! +@brief default JSON class + +This type is the default specialization of the @ref basic_json class which +uses the standard template types. + +@since version 1.0.0 +*/ +using json = basic_json<>; +} // namespace nlohmann + +#endif + +// #include <nlohmann/detail/macro_scope.hpp> + + +// This file contains all internal macro definitions +// You MUST include macro_unscope.hpp at the end of json.hpp to undef all of them + // exclude unsupported compilers -#if defined(__clang__) - #if (__clang_major__ * 10000 + __clang_minor__ * 100 + __clang_patchlevel__) < 30400 - #error "unsupported Clang version - see https://github.com/nlohmann/json#supported-compilers" - #endif -#elif defined(__GNUC__) && !(defined(__ICC) || defined(__INTEL_COMPILER)) - #if (__GNUC__ * 10000 + __GNUC_MINOR__ * 100 + __GNUC_PATCHLEVEL__) < 40900 - #error "unsupported GCC version - see https://github.com/nlohmann/json#supported-compilers" +#if !defined(JSON_SKIP_UNSUPPORTED_COMPILER_CHECK) + #if defined(__clang__) + #if (__clang_major__ * 10000 + __clang_minor__ * 100 + __clang_patchlevel__) < 30400 + #error "unsupported Clang version - see https://github.com/nlohmann/json#supported-compilers" + #endif + #elif defined(__GNUC__) && !(defined(__ICC) || defined(__INTEL_COMPILER)) + #if (__GNUC__ * 10000 + __GNUC_MINOR__ * 100 + __GNUC_PATCHLEVEL__) < 40800 + #error "unsupported GCC version - see https://github.com/nlohmann/json#supported-compilers" + #endif #endif #endif @@ -90,14 +153,36 @@ SOFTWARE. #endif // allow to disable exceptions -#if (defined(__cpp_exceptions) || defined(__EXCEPTIONS) || defined(_CPPUNWIND)) && not defined(JSON_NOEXCEPTION) +#if (defined(__cpp_exceptions) || defined(__EXCEPTIONS) || defined(_CPPUNWIND)) && !defined(JSON_NOEXCEPTION) #define JSON_THROW(exception) throw exception #define JSON_TRY try #define JSON_CATCH(exception) catch(exception) + #define JSON_INTERNAL_CATCH(exception) catch(exception) #else #define JSON_THROW(exception) std::abort() #define JSON_TRY if(true) #define JSON_CATCH(exception) if(false) + #define JSON_INTERNAL_CATCH(exception) if(false) +#endif + +// override exception macros +#if defined(JSON_THROW_USER) + #undef JSON_THROW + #define JSON_THROW JSON_THROW_USER +#endif +#if defined(JSON_TRY_USER) + #undef JSON_TRY + #define JSON_TRY JSON_TRY_USER +#endif +#if defined(JSON_CATCH_USER) + #undef JSON_CATCH + #define JSON_CATCH JSON_CATCH_USER + #undef JSON_INTERNAL_CATCH + #define JSON_INTERNAL_CATCH JSON_CATCH_USER +#endif +#if defined(JSON_INTERNAL_CATCH_USER) + #undef JSON_INTERNAL_CATCH + #define JSON_INTERNAL_CATCH JSON_INTERNAL_CATCH_USER #endif // manual branch prediction @@ -118,25 +203,35 @@ SOFTWARE. #endif /*! -@brief namespace for Niels Lohmann -@see https://github.com/nlohmann -@since version 1.0.0 +@brief macro to briefly define a mapping between an enum and JSON +@def NLOHMANN_JSON_SERIALIZE_ENUM +@since version 3.4.0 */ -namespace nlohmann -{ -template<typename = void, typename = void> -struct adl_serializer; - -// forward declaration of basic_json (required to split the class) -template<template<typename, typename, typename...> class ObjectType = std::map, - template<typename, typename...> class ArrayType = std::vector, - class StringType = std::string, class BooleanType = bool, - class NumberIntegerType = std::int64_t, - class NumberUnsignedType = std::uint64_t, - class NumberFloatType = double, - template<typename> class AllocatorType = std::allocator, - template<typename, typename = void> class JSONSerializer = adl_serializer> -class basic_json; +#define NLOHMANN_JSON_SERIALIZE_ENUM(ENUM_TYPE, ...) \ + template<typename BasicJsonType> \ + inline void to_json(BasicJsonType& j, const ENUM_TYPE& e) \ + { \ + static_assert(std::is_enum<ENUM_TYPE>::value, #ENUM_TYPE " must be an enum!"); \ + static const std::pair<ENUM_TYPE, BasicJsonType> m[] = __VA_ARGS__; \ + auto it = std::find_if(std::begin(m), std::end(m), \ + [e](const std::pair<ENUM_TYPE, BasicJsonType>& ej_pair) -> bool \ + { \ + return ej_pair.first == e; \ + }); \ + j = ((it != std::end(m)) ? it : std::begin(m))->second; \ + } \ + template<typename BasicJsonType> \ + inline void from_json(const BasicJsonType& j, ENUM_TYPE& e) \ + { \ + static_assert(std::is_enum<ENUM_TYPE>::value, #ENUM_TYPE " must be an enum!"); \ + static const std::pair<ENUM_TYPE, BasicJsonType> m[] = __VA_ARGS__; \ + auto it = std::find_if(std::begin(m), std::end(m), \ + [j](const std::pair<ENUM_TYPE, BasicJsonType>& ej_pair) -> bool \ + { \ + return ej_pair.second == j; \ + }); \ + e = ((it != std::end(m)) ? it : std::begin(m))->first; \ + } // Ugly macros to avoid uglier copy-paste when specializing basic_json. They // may be removed in the future once the class is split. @@ -154,17 +249,590 @@ class basic_json; NumberIntegerType, NumberUnsignedType, NumberFloatType, \ AllocatorType, JSONSerializer> +// #include <nlohmann/detail/meta/cpp_future.hpp> + + +#include <ciso646> // not +#include <cstddef> // size_t +#include <type_traits> // conditional, enable_if, false_type, integral_constant, is_constructible, is_integral, is_same, remove_cv, remove_reference, true_type + +namespace nlohmann +{ +namespace detail +{ +// alias templates to reduce boilerplate +template<bool B, typename T = void> +using enable_if_t = typename std::enable_if<B, T>::type; + +template<typename T> +using uncvref_t = typename std::remove_cv<typename std::remove_reference<T>::type>::type; + +// implementation of C++14 index_sequence and affiliates +// source: https://stackoverflow.com/a/32223343 +template<std::size_t... Ints> +struct index_sequence +{ + using type = index_sequence; + using value_type = std::size_t; + static constexpr std::size_t size() noexcept + { + return sizeof...(Ints); + } +}; + +template<class Sequence1, class Sequence2> +struct merge_and_renumber; + +template<std::size_t... I1, std::size_t... I2> +struct merge_and_renumber<index_sequence<I1...>, index_sequence<I2...>> + : index_sequence < I1..., (sizeof...(I1) + I2)... > {}; + +template<std::size_t N> +struct make_index_sequence + : merge_and_renumber < typename make_index_sequence < N / 2 >::type, + typename make_index_sequence < N - N / 2 >::type > {}; + +template<> struct make_index_sequence<0> : index_sequence<> {}; +template<> struct make_index_sequence<1> : index_sequence<0> {}; + +template<typename... Ts> +using index_sequence_for = make_index_sequence<sizeof...(Ts)>; + +// dispatch utility (taken from ranges-v3) +template<unsigned N> struct priority_tag : priority_tag < N - 1 > {}; +template<> struct priority_tag<0> {}; + +// taken from ranges-v3 +template<typename T> +struct static_const +{ + static constexpr T value{}; +}; + +template<typename T> +constexpr T static_const<T>::value; +} // namespace detail +} // namespace nlohmann + +// #include <nlohmann/detail/meta/type_traits.hpp> + + +#include <ciso646> // not +#include <limits> // numeric_limits +#include <type_traits> // false_type, is_constructible, is_integral, is_same, true_type +#include <utility> // declval + +// #include <nlohmann/json_fwd.hpp> + +// #include <nlohmann/detail/iterators/iterator_traits.hpp> + + +#include <iterator> // random_access_iterator_tag + +// #include <nlohmann/detail/meta/void_t.hpp> + + +namespace nlohmann +{ +namespace detail +{ +template <typename ...Ts> struct make_void +{ + using type = void; +}; +template <typename ...Ts> using void_t = typename make_void<Ts...>::type; +} // namespace detail +} // namespace nlohmann + +// #include <nlohmann/detail/meta/cpp_future.hpp> + +namespace nlohmann +{ +namespace detail +{ +template <typename It, typename = void> +struct iterator_types {}; + +template <typename It> +struct iterator_types < + It, + void_t<typename It::difference_type, typename It::value_type, typename It::pointer, + typename It::reference, typename It::iterator_category >> +{ + using difference_type = typename It::difference_type; + using value_type = typename It::value_type; + using pointer = typename It::pointer; + using reference = typename It::reference; + using iterator_category = typename It::iterator_category; +}; + +// This is required as some compilers implement std::iterator_traits in a way that +// doesn't work with SFINAE. See https://github.com/nlohmann/json/issues/1341. +template <typename T, typename = void> +struct iterator_traits +{ +}; + +template <typename T> +struct iterator_traits < T, enable_if_t < !std::is_pointer<T>::value >> + : iterator_types<T> +{ +}; + +template <typename T> +struct iterator_traits<T*, enable_if_t<std::is_object<T>::value>> +{ + using iterator_category = std::random_access_iterator_tag; + using value_type = T; + using difference_type = ptrdiff_t; + using pointer = T*; + using reference = T&; +}; +} +} + +// #include <nlohmann/detail/meta/cpp_future.hpp> + +// #include <nlohmann/detail/meta/detected.hpp> + + +#include <type_traits> + +// #include <nlohmann/detail/meta/void_t.hpp> + + +// http://en.cppreference.com/w/cpp/experimental/is_detected +namespace nlohmann +{ +namespace detail +{ +struct nonesuch +{ + nonesuch() = delete; + ~nonesuch() = delete; + nonesuch(nonesuch const&) = delete; + void operator=(nonesuch const&) = delete; +}; + +template <class Default, + class AlwaysVoid, + template <class...> class Op, + class... Args> +struct detector +{ + using value_t = std::false_type; + using type = Default; +}; + +template <class Default, template <class...> class Op, class... Args> +struct detector<Default, void_t<Op<Args...>>, Op, Args...> +{ + using value_t = std::true_type; + using type = Op<Args...>; +}; + +template <template <class...> class Op, class... Args> +using is_detected = typename detector<nonesuch, void, Op, Args...>::value_t; + +template <template <class...> class Op, class... Args> +using detected_t = typename detector<nonesuch, void, Op, Args...>::type; + +template <class Default, template <class...> class Op, class... Args> +using detected_or = detector<Default, void, Op, Args...>; + +template <class Default, template <class...> class Op, class... Args> +using detected_or_t = typename detected_or<Default, Op, Args...>::type; + +template <class Expected, template <class...> class Op, class... Args> +using is_detected_exact = std::is_same<Expected, detected_t<Op, Args...>>; + +template <class To, template <class...> class Op, class... Args> +using is_detected_convertible = + std::is_convertible<detected_t<Op, Args...>, To>; +} // namespace detail +} // namespace nlohmann + +// #include <nlohmann/detail/macro_scope.hpp> + + +namespace nlohmann +{ /*! -@brief unnamed namespace with internal helper functions +@brief detail namespace with internal helper functions -This namespace collects some functions that could not be defined inside the -@ref basic_json class. +This namespace collects functions that should not be exposed, +implementations of some @ref basic_json methods, and meta-programming helpers. @since version 2.1.0 */ namespace detail { +///////////// +// helpers // +///////////// + +// Note to maintainers: +// +// Every trait in this file expects a non CV-qualified type. +// The only exceptions are in the 'aliases for detected' section +// (i.e. those of the form: decltype(T::member_function(std::declval<T>()))) +// +// In this case, T has to be properly CV-qualified to constraint the function arguments +// (e.g. to_json(BasicJsonType&, const T&)) + +template<typename> struct is_basic_json : std::false_type {}; + +NLOHMANN_BASIC_JSON_TPL_DECLARATION +struct is_basic_json<NLOHMANN_BASIC_JSON_TPL> : std::true_type {}; + +////////////////////////// +// aliases for detected // +////////////////////////// + +template <typename T> +using mapped_type_t = typename T::mapped_type; + +template <typename T> +using key_type_t = typename T::key_type; + +template <typename T> +using value_type_t = typename T::value_type; + +template <typename T> +using difference_type_t = typename T::difference_type; + +template <typename T> +using pointer_t = typename T::pointer; + +template <typename T> +using reference_t = typename T::reference; + +template <typename T> +using iterator_category_t = typename T::iterator_category; + +template <typename T> +using iterator_t = typename T::iterator; + +template <typename T, typename... Args> +using to_json_function = decltype(T::to_json(std::declval<Args>()...)); + +template <typename T, typename... Args> +using from_json_function = decltype(T::from_json(std::declval<Args>()...)); + +template <typename T, typename U> +using get_template_function = decltype(std::declval<T>().template get<U>()); + +// trait checking if JSONSerializer<T>::from_json(json const&, udt&) exists +template <typename BasicJsonType, typename T, typename = void> +struct has_from_json : std::false_type {}; + +template <typename BasicJsonType, typename T> +struct has_from_json<BasicJsonType, T, + enable_if_t<not is_basic_json<T>::value>> +{ + using serializer = typename BasicJsonType::template json_serializer<T, void>; + + static constexpr bool value = + is_detected_exact<void, from_json_function, serializer, + const BasicJsonType&, T&>::value; +}; + +// This trait checks if JSONSerializer<T>::from_json(json const&) exists +// this overload is used for non-default-constructible user-defined-types +template <typename BasicJsonType, typename T, typename = void> +struct has_non_default_from_json : std::false_type {}; + +template<typename BasicJsonType, typename T> +struct has_non_default_from_json<BasicJsonType, T, enable_if_t<not is_basic_json<T>::value>> +{ + using serializer = typename BasicJsonType::template json_serializer<T, void>; + + static constexpr bool value = + is_detected_exact<T, from_json_function, serializer, + const BasicJsonType&>::value; +}; + +// This trait checks if BasicJsonType::json_serializer<T>::to_json exists +// Do not evaluate the trait when T is a basic_json type, to avoid template instantiation infinite recursion. +template <typename BasicJsonType, typename T, typename = void> +struct has_to_json : std::false_type {}; + +template <typename BasicJsonType, typename T> +struct has_to_json<BasicJsonType, T, enable_if_t<not is_basic_json<T>::value>> +{ + using serializer = typename BasicJsonType::template json_serializer<T, void>; + + static constexpr bool value = + is_detected_exact<void, to_json_function, serializer, BasicJsonType&, + T>::value; +}; + + +/////////////////// +// is_ functions // +/////////////////// + +template <typename T, typename = void> +struct is_iterator_traits : std::false_type {}; + +template <typename T> +struct is_iterator_traits<iterator_traits<T>> +{ + private: + using traits = iterator_traits<T>; + + public: + static constexpr auto value = + is_detected<value_type_t, traits>::value && + is_detected<difference_type_t, traits>::value && + is_detected<pointer_t, traits>::value && + is_detected<iterator_category_t, traits>::value && + is_detected<reference_t, traits>::value; +}; + +// source: https://stackoverflow.com/a/37193089/4116453 + +template <typename T, typename = void> +struct is_complete_type : std::false_type {}; + +template <typename T> +struct is_complete_type<T, decltype(void(sizeof(T)))> : std::true_type {}; + +template <typename BasicJsonType, typename CompatibleObjectType, + typename = void> +struct is_compatible_object_type_impl : std::false_type {}; + +template <typename BasicJsonType, typename CompatibleObjectType> +struct is_compatible_object_type_impl < + BasicJsonType, CompatibleObjectType, + enable_if_t<is_detected<mapped_type_t, CompatibleObjectType>::value and + is_detected<key_type_t, CompatibleObjectType>::value >> +{ + + using object_t = typename BasicJsonType::object_t; + + // macOS's is_constructible does not play well with nonesuch... + static constexpr bool value = + std::is_constructible<typename object_t::key_type, + typename CompatibleObjectType::key_type>::value and + std::is_constructible<typename object_t::mapped_type, + typename CompatibleObjectType::mapped_type>::value; +}; + +template <typename BasicJsonType, typename CompatibleObjectType> +struct is_compatible_object_type + : is_compatible_object_type_impl<BasicJsonType, CompatibleObjectType> {}; + +template <typename BasicJsonType, typename ConstructibleObjectType, + typename = void> +struct is_constructible_object_type_impl : std::false_type {}; + +template <typename BasicJsonType, typename ConstructibleObjectType> +struct is_constructible_object_type_impl < + BasicJsonType, ConstructibleObjectType, + enable_if_t<is_detected<mapped_type_t, ConstructibleObjectType>::value and + is_detected<key_type_t, ConstructibleObjectType>::value >> +{ + using object_t = typename BasicJsonType::object_t; + + static constexpr bool value = + (std::is_constructible<typename ConstructibleObjectType::key_type, typename object_t::key_type>::value and + std::is_same<typename object_t::mapped_type, typename ConstructibleObjectType::mapped_type>::value) or + (has_from_json<BasicJsonType, typename ConstructibleObjectType::mapped_type>::value or + has_non_default_from_json<BasicJsonType, typename ConstructibleObjectType::mapped_type >::value); +}; + +template <typename BasicJsonType, typename ConstructibleObjectType> +struct is_constructible_object_type + : is_constructible_object_type_impl<BasicJsonType, + ConstructibleObjectType> {}; + +template <typename BasicJsonType, typename CompatibleStringType, + typename = void> +struct is_compatible_string_type_impl : std::false_type {}; + +template <typename BasicJsonType, typename CompatibleStringType> +struct is_compatible_string_type_impl < + BasicJsonType, CompatibleStringType, + enable_if_t<is_detected_exact<typename BasicJsonType::string_t::value_type, + value_type_t, CompatibleStringType>::value >> +{ + static constexpr auto value = + std::is_constructible<typename BasicJsonType::string_t, CompatibleStringType>::value; +}; + +template <typename BasicJsonType, typename ConstructibleStringType> +struct is_compatible_string_type + : is_compatible_string_type_impl<BasicJsonType, ConstructibleStringType> {}; + +template <typename BasicJsonType, typename ConstructibleStringType, + typename = void> +struct is_constructible_string_type_impl : std::false_type {}; + +template <typename BasicJsonType, typename ConstructibleStringType> +struct is_constructible_string_type_impl < + BasicJsonType, ConstructibleStringType, + enable_if_t<is_detected_exact<typename BasicJsonType::string_t::value_type, + value_type_t, ConstructibleStringType>::value >> +{ + static constexpr auto value = + std::is_constructible<ConstructibleStringType, + typename BasicJsonType::string_t>::value; +}; + +template <typename BasicJsonType, typename ConstructibleStringType> +struct is_constructible_string_type + : is_constructible_string_type_impl<BasicJsonType, ConstructibleStringType> {}; + +template <typename BasicJsonType, typename CompatibleArrayType, typename = void> +struct is_compatible_array_type_impl : std::false_type {}; + +template <typename BasicJsonType, typename CompatibleArrayType> +struct is_compatible_array_type_impl < + BasicJsonType, CompatibleArrayType, + enable_if_t<is_detected<value_type_t, CompatibleArrayType>::value and + is_detected<iterator_t, CompatibleArrayType>::value and +// This is needed because json_reverse_iterator has a ::iterator type... +// Therefore it is detected as a CompatibleArrayType. +// The real fix would be to have an Iterable concept. + not is_iterator_traits< + iterator_traits<CompatibleArrayType>>::value >> +{ + static constexpr bool value = + std::is_constructible<BasicJsonType, + typename CompatibleArrayType::value_type>::value; +}; + +template <typename BasicJsonType, typename CompatibleArrayType> +struct is_compatible_array_type + : is_compatible_array_type_impl<BasicJsonType, CompatibleArrayType> {}; + +template <typename BasicJsonType, typename ConstructibleArrayType, typename = void> +struct is_constructible_array_type_impl : std::false_type {}; + +template <typename BasicJsonType, typename ConstructibleArrayType> +struct is_constructible_array_type_impl < + BasicJsonType, ConstructibleArrayType, + enable_if_t<std::is_same<ConstructibleArrayType, + typename BasicJsonType::value_type>::value >> + : std::true_type {}; + +template <typename BasicJsonType, typename ConstructibleArrayType> +struct is_constructible_array_type_impl < + BasicJsonType, ConstructibleArrayType, + enable_if_t<not std::is_same<ConstructibleArrayType, + typename BasicJsonType::value_type>::value and + is_detected<value_type_t, ConstructibleArrayType>::value and + is_detected<iterator_t, ConstructibleArrayType>::value and + is_complete_type< + detected_t<value_type_t, ConstructibleArrayType>>::value >> +{ + static constexpr bool value = + // This is needed because json_reverse_iterator has a ::iterator type, + // furthermore, std::back_insert_iterator (and other iterators) have a base class `iterator`... + // Therefore it is detected as a ConstructibleArrayType. + // The real fix would be to have an Iterable concept. + not is_iterator_traits < + iterator_traits<ConstructibleArrayType >>::value and + + (std::is_same<typename ConstructibleArrayType::value_type, typename BasicJsonType::array_t::value_type>::value or + has_from_json<BasicJsonType, + typename ConstructibleArrayType::value_type>::value or + has_non_default_from_json < + BasicJsonType, typename ConstructibleArrayType::value_type >::value); +}; + +template <typename BasicJsonType, typename ConstructibleArrayType> +struct is_constructible_array_type + : is_constructible_array_type_impl<BasicJsonType, ConstructibleArrayType> {}; + +template <typename RealIntegerType, typename CompatibleNumberIntegerType, + typename = void> +struct is_compatible_integer_type_impl : std::false_type {}; + +template <typename RealIntegerType, typename CompatibleNumberIntegerType> +struct is_compatible_integer_type_impl < + RealIntegerType, CompatibleNumberIntegerType, + enable_if_t<std::is_integral<RealIntegerType>::value and + std::is_integral<CompatibleNumberIntegerType>::value and + not std::is_same<bool, CompatibleNumberIntegerType>::value >> +{ + // is there an assert somewhere on overflows? + using RealLimits = std::numeric_limits<RealIntegerType>; + using CompatibleLimits = std::numeric_limits<CompatibleNumberIntegerType>; + + static constexpr auto value = + std::is_constructible<RealIntegerType, + CompatibleNumberIntegerType>::value and + CompatibleLimits::is_integer and + RealLimits::is_signed == CompatibleLimits::is_signed; +}; + +template <typename RealIntegerType, typename CompatibleNumberIntegerType> +struct is_compatible_integer_type + : is_compatible_integer_type_impl<RealIntegerType, + CompatibleNumberIntegerType> {}; + +template <typename BasicJsonType, typename CompatibleType, typename = void> +struct is_compatible_type_impl: std::false_type {}; + +template <typename BasicJsonType, typename CompatibleType> +struct is_compatible_type_impl < + BasicJsonType, CompatibleType, + enable_if_t<is_complete_type<CompatibleType>::value >> +{ + static constexpr bool value = + has_to_json<BasicJsonType, CompatibleType>::value; +}; + +template <typename BasicJsonType, typename CompatibleType> +struct is_compatible_type + : is_compatible_type_impl<BasicJsonType, CompatibleType> {}; +} // namespace detail +} // namespace nlohmann + +// #include <nlohmann/detail/exceptions.hpp> + + +#include <exception> // exception +#include <stdexcept> // runtime_error +#include <string> // to_string + +// #include <nlohmann/detail/input/position_t.hpp> + + +#include <cstddef> // size_t + +namespace nlohmann +{ +namespace detail +{ +/// struct to capture the start position of the current token +struct position_t +{ + /// the total number of characters read + std::size_t chars_read_total = 0; + /// the number of characters read in the current line + std::size_t chars_read_current_line = 0; + /// the number of lines read + std::size_t lines_read = 0; + + /// conversion to size_t to preserve SAX interface + constexpr operator size_t() const + { + return chars_read_total; + } +}; + +} +} + + +namespace nlohmann +{ +namespace detail +{ //////////////// // exceptions // //////////////// @@ -248,6 +916,7 @@ json.exception.parse_error.109 | parse error: array index 'one' is not a number json.exception.parse_error.110 | parse error at 1: cannot read 2 bytes from vector | When parsing CBOR or MessagePack, the byte vector ends before the complete value has been read. json.exception.parse_error.112 | parse error at 1: error reading CBOR; last byte: 0xF8 | Not all types of CBOR or MessagePack are supported. This exception occurs if an unsupported byte was read. json.exception.parse_error.113 | parse error at 2: expected a CBOR string; last byte: 0x98 | While parsing a map key, a value that is not a string has been read. +json.exception.parse_error.114 | parse error: Unsupported BSON record type 0x0F | The parsing of the corresponding BSON record type is not implemented (yet). @note For an input with n bytes, 1 is the index of the first character and n+1 is the index of the terminating null byte or the end of file. This also @@ -271,15 +940,23 @@ class parse_error : public exception /*! @brief create a parse error exception @param[in] id_ the id of the exception - @param[in] byte_ the byte index where the error occurred (or 0 if the - position cannot be determined) + @param[in] position the position where the error occurred (or with + chars_read_total=0 if the position cannot be + determined) @param[in] what_arg the explanatory string @return parse_error object */ + static parse_error create(int id_, const position_t& pos, const std::string& what_arg) + { + std::string w = exception::name("parse_error", id_) + "parse error" + + position_string(pos) + ": " + what_arg; + return parse_error(id_, pos.chars_read_total, w.c_str()); + } + static parse_error create(int id_, std::size_t byte_, const std::string& what_arg) { std::string w = exception::name("parse_error", id_) + "parse error" + - (byte_ != 0 ? (" at " + std::to_string(byte_)) : "") + + (byte_ != 0 ? (" at byte " + std::to_string(byte_)) : "") + ": " + what_arg; return parse_error(id_, byte_, w.c_str()); } @@ -298,6 +975,12 @@ class parse_error : public exception private: parse_error(int id_, std::size_t byte_, const char* what_arg) : exception(id_, what_arg), byte(byte_) {} + + static std::string position_string(const position_t& pos) + { + return " at line " + std::to_string(pos.lines_read + 1) + + ", column " + std::to_string(pos.chars_read_current_line); + } }; /*! @@ -377,6 +1060,7 @@ json.exception.type_error.313 | invalid value to unflatten | The @ref unflatten json.exception.type_error.314 | only objects can be unflattened | The @ref unflatten function only works for an object whose keys are JSON Pointers. json.exception.type_error.315 | values in object must be primitive | The @ref unflatten function only works for an object whose keys are JSON Pointers and whose values are primitive. json.exception.type_error.316 | invalid UTF-8 byte at index 10: 0x7E | The @ref dump function only works with UTF-8 encoded strings; that is, if you assign a `std::string` to a JSON value, make sure it is UTF-8 encoded. | +json.exception.type_error.317 | JSON value cannot be serialized to requested format | The dynamic type of the object cannot be represented in the requested serialization format (e.g. a raw `true` or `null` JSON object cannot be serialized to BSON) | @liveexample{The following code shows how a `type_error` exception can be caught.,type_error} @@ -419,6 +1103,9 @@ json.exception.out_of_range.403 | key 'foo' not found | The provided key was not json.exception.out_of_range.404 | unresolved reference token 'foo' | A reference token in a JSON Pointer could not be resolved. json.exception.out_of_range.405 | JSON pointer has no parent | The JSON Patch operations 'remove' and 'add' can not be applied to the root element of the JSON value. json.exception.out_of_range.406 | number overflow parsing '10E1000' | A parsed number could not be stored as without changing it to NaN or INF. +json.exception.out_of_range.407 | number overflow serializing '9223372036854775808' | UBJSON and BSON only support integer numbers up to 9223372036854775807. | +json.exception.out_of_range.408 | excessive array size: 8658170730974374167 | The size (following `#`) of an UBJSON array or object exceeds the maximal capacity. | +json.exception.out_of_range.409 | BSON key cannot contain code point U+0000 (at byte 2) | Key identifiers to be serialized to BSON cannot contain code point U+0000, since the key is stored as zero-terminated c-string | @liveexample{The following code shows how an `out_of_range` exception can be caught.,out_of_range} @@ -481,9 +1168,21 @@ class other_error : public exception private: other_error(int id_, const char* what_arg) : exception(id_, what_arg) {} }; +} // namespace detail +} // namespace nlohmann + +// #include <nlohmann/detail/value_t.hpp> +#include <array> // array +#include <ciso646> // and +#include <cstddef> // size_t +#include <cstdint> // uint8_t +namespace nlohmann +{ +namespace detail +{ /////////////////////////// // JSON type enumeration // /////////////////////////// @@ -512,7 +1211,7 @@ value with the default value for a given type @since version 1.0.0 */ -enum class value_t : uint8_t +enum class value_t : std::uint8_t { null, ///< null value object, ///< object (unordered set of name/value pairs) @@ -537,7 +1236,7 @@ Returns an ordering that is similar to Python: */ inline bool operator<(const value_t lhs, const value_t rhs) noexcept { - static constexpr std::array<uint8_t, 8> order = {{ + static constexpr std::array<std::uint8_t, 8> order = {{ 0 /* null */, 3 /* object */, 4 /* array */, 5 /* string */, 1 /* boolean */, 2 /* integer */, 2 /* unsigned */, 2 /* float */ } @@ -547,80 +1246,573 @@ inline bool operator<(const value_t lhs, const value_t rhs) noexcept const auto r_index = static_cast<std::size_t>(rhs); return l_index < order.size() and r_index < order.size() and order[l_index] < order[r_index]; } +} // namespace detail +} // namespace nlohmann +// #include <nlohmann/detail/conversions/from_json.hpp> -///////////// -// helpers // -///////////// -template<typename> struct is_basic_json : std::false_type {}; +#include <algorithm> // transform +#include <array> // array +#include <ciso646> // and, not +#include <forward_list> // forward_list +#include <iterator> // inserter, front_inserter, end +#include <map> // map +#include <string> // string +#include <tuple> // tuple, make_tuple +#include <type_traits> // is_arithmetic, is_same, is_enum, underlying_type, is_convertible +#include <unordered_map> // unordered_map +#include <utility> // pair, declval +#include <valarray> // valarray -NLOHMANN_BASIC_JSON_TPL_DECLARATION -struct is_basic_json<NLOHMANN_BASIC_JSON_TPL> : std::true_type {}; +// #include <nlohmann/detail/exceptions.hpp> -// alias templates to reduce boilerplate -template<bool B, typename T = void> -using enable_if_t = typename std::enable_if<B, T>::type; +// #include <nlohmann/detail/macro_scope.hpp> -template<typename T> -using uncvref_t = typename std::remove_cv<typename std::remove_reference<T>::type>::type; +// #include <nlohmann/detail/meta/cpp_future.hpp> -// implementation of C++14 index_sequence and affiliates -// source: https://stackoverflow.com/a/32223343 -template<std::size_t... Ints> -struct index_sequence +// #include <nlohmann/detail/meta/type_traits.hpp> + +// #include <nlohmann/detail/value_t.hpp> + + +namespace nlohmann { - using type = index_sequence; - using value_type = std::size_t; - static constexpr std::size_t size() noexcept +namespace detail +{ +template<typename BasicJsonType> +void from_json(const BasicJsonType& j, typename std::nullptr_t& n) +{ + if (JSON_UNLIKELY(not j.is_null())) { - return sizeof...(Ints); + JSON_THROW(type_error::create(302, "type must be null, but is " + std::string(j.type_name()))); + } + n = nullptr; +} + +// overloads for basic_json template parameters +template<typename BasicJsonType, typename ArithmeticType, + enable_if_t<std::is_arithmetic<ArithmeticType>::value and + not std::is_same<ArithmeticType, typename BasicJsonType::boolean_t>::value, + int> = 0> +void get_arithmetic_value(const BasicJsonType& j, ArithmeticType& val) +{ + switch (static_cast<value_t>(j)) + { + case value_t::number_unsigned: + { + val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::number_unsigned_t*>()); + break; + } + case value_t::number_integer: + { + val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::number_integer_t*>()); + break; + } + case value_t::number_float: + { + val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::number_float_t*>()); + break; + } + + default: + JSON_THROW(type_error::create(302, "type must be number, but is " + std::string(j.type_name()))); + } +} + +template<typename BasicJsonType> +void from_json(const BasicJsonType& j, typename BasicJsonType::boolean_t& b) +{ + if (JSON_UNLIKELY(not j.is_boolean())) + { + JSON_THROW(type_error::create(302, "type must be boolean, but is " + std::string(j.type_name()))); + } + b = *j.template get_ptr<const typename BasicJsonType::boolean_t*>(); +} + +template<typename BasicJsonType> +void from_json(const BasicJsonType& j, typename BasicJsonType::string_t& s) +{ + if (JSON_UNLIKELY(not j.is_string())) + { + JSON_THROW(type_error::create(302, "type must be string, but is " + std::string(j.type_name()))); + } + s = *j.template get_ptr<const typename BasicJsonType::string_t*>(); +} + +template < + typename BasicJsonType, typename ConstructibleStringType, + enable_if_t < + is_constructible_string_type<BasicJsonType, ConstructibleStringType>::value and + not std::is_same<typename BasicJsonType::string_t, + ConstructibleStringType>::value, + int > = 0 > +void from_json(const BasicJsonType& j, ConstructibleStringType& s) +{ + if (JSON_UNLIKELY(not j.is_string())) + { + JSON_THROW(type_error::create(302, "type must be string, but is " + std::string(j.type_name()))); + } + + s = *j.template get_ptr<const typename BasicJsonType::string_t*>(); +} + +template<typename BasicJsonType> +void from_json(const BasicJsonType& j, typename BasicJsonType::number_float_t& val) +{ + get_arithmetic_value(j, val); +} + +template<typename BasicJsonType> +void from_json(const BasicJsonType& j, typename BasicJsonType::number_unsigned_t& val) +{ + get_arithmetic_value(j, val); +} + +template<typename BasicJsonType> +void from_json(const BasicJsonType& j, typename BasicJsonType::number_integer_t& val) +{ + get_arithmetic_value(j, val); +} + +template<typename BasicJsonType, typename EnumType, + enable_if_t<std::is_enum<EnumType>::value, int> = 0> +void from_json(const BasicJsonType& j, EnumType& e) +{ + typename std::underlying_type<EnumType>::type val; + get_arithmetic_value(j, val); + e = static_cast<EnumType>(val); +} + +// forward_list doesn't have an insert method +template<typename BasicJsonType, typename T, typename Allocator, + enable_if_t<std::is_convertible<BasicJsonType, T>::value, int> = 0> +void from_json(const BasicJsonType& j, std::forward_list<T, Allocator>& l) +{ + if (JSON_UNLIKELY(not j.is_array())) + { + JSON_THROW(type_error::create(302, "type must be array, but is " + std::string(j.type_name()))); + } + std::transform(j.rbegin(), j.rend(), + std::front_inserter(l), [](const BasicJsonType & i) + { + return i.template get<T>(); + }); +} + +// valarray doesn't have an insert method +template<typename BasicJsonType, typename T, + enable_if_t<std::is_convertible<BasicJsonType, T>::value, int> = 0> +void from_json(const BasicJsonType& j, std::valarray<T>& l) +{ + if (JSON_UNLIKELY(not j.is_array())) + { + JSON_THROW(type_error::create(302, "type must be array, but is " + std::string(j.type_name()))); + } + l.resize(j.size()); + std::copy(j.m_value.array->begin(), j.m_value.array->end(), std::begin(l)); +} + +template<typename BasicJsonType> +void from_json_array_impl(const BasicJsonType& j, typename BasicJsonType::array_t& arr, priority_tag<3> /*unused*/) +{ + arr = *j.template get_ptr<const typename BasicJsonType::array_t*>(); +} + +template <typename BasicJsonType, typename T, std::size_t N> +auto from_json_array_impl(const BasicJsonType& j, std::array<T, N>& arr, + priority_tag<2> /*unused*/) +-> decltype(j.template get<T>(), void()) +{ + for (std::size_t i = 0; i < N; ++i) + { + arr[i] = j.at(i).template get<T>(); + } +} + +template<typename BasicJsonType, typename ConstructibleArrayType> +auto from_json_array_impl(const BasicJsonType& j, ConstructibleArrayType& arr, priority_tag<1> /*unused*/) +-> decltype( + arr.reserve(std::declval<typename ConstructibleArrayType::size_type>()), + j.template get<typename ConstructibleArrayType::value_type>(), + void()) +{ + using std::end; + + arr.reserve(j.size()); + std::transform(j.begin(), j.end(), + std::inserter(arr, end(arr)), [](const BasicJsonType & i) + { + // get<BasicJsonType>() returns *this, this won't call a from_json + // method when value_type is BasicJsonType + return i.template get<typename ConstructibleArrayType::value_type>(); + }); +} + +template <typename BasicJsonType, typename ConstructibleArrayType> +void from_json_array_impl(const BasicJsonType& j, ConstructibleArrayType& arr, + priority_tag<0> /*unused*/) +{ + using std::end; + + std::transform( + j.begin(), j.end(), std::inserter(arr, end(arr)), + [](const BasicJsonType & i) + { + // get<BasicJsonType>() returns *this, this won't call a from_json + // method when value_type is BasicJsonType + return i.template get<typename ConstructibleArrayType::value_type>(); + }); +} + +template <typename BasicJsonType, typename ConstructibleArrayType, + enable_if_t < + is_constructible_array_type<BasicJsonType, ConstructibleArrayType>::value and + not is_constructible_object_type<BasicJsonType, ConstructibleArrayType>::value and + not is_constructible_string_type<BasicJsonType, ConstructibleArrayType>::value and + not is_basic_json<ConstructibleArrayType>::value, + int > = 0 > + +auto from_json(const BasicJsonType& j, ConstructibleArrayType& arr) +-> decltype(from_json_array_impl(j, arr, priority_tag<3> {}), +j.template get<typename ConstructibleArrayType::value_type>(), +void()) +{ + if (JSON_UNLIKELY(not j.is_array())) + { + JSON_THROW(type_error::create(302, "type must be array, but is " + + std::string(j.type_name()))); + } + + from_json_array_impl(j, arr, priority_tag<3> {}); +} + +template<typename BasicJsonType, typename ConstructibleObjectType, + enable_if_t<is_constructible_object_type<BasicJsonType, ConstructibleObjectType>::value, int> = 0> +void from_json(const BasicJsonType& j, ConstructibleObjectType& obj) +{ + if (JSON_UNLIKELY(not j.is_object())) + { + JSON_THROW(type_error::create(302, "type must be object, but is " + std::string(j.type_name()))); + } + + auto inner_object = j.template get_ptr<const typename BasicJsonType::object_t*>(); + using value_type = typename ConstructibleObjectType::value_type; + std::transform( + inner_object->begin(), inner_object->end(), + std::inserter(obj, obj.begin()), + [](typename BasicJsonType::object_t::value_type const & p) + { + return value_type(p.first, p.second.template get<typename ConstructibleObjectType::mapped_type>()); + }); +} + +// overload for arithmetic types, not chosen for basic_json template arguments +// (BooleanType, etc..); note: Is it really necessary to provide explicit +// overloads for boolean_t etc. in case of a custom BooleanType which is not +// an arithmetic type? +template<typename BasicJsonType, typename ArithmeticType, + enable_if_t < + std::is_arithmetic<ArithmeticType>::value and + not std::is_same<ArithmeticType, typename BasicJsonType::number_unsigned_t>::value and + not std::is_same<ArithmeticType, typename BasicJsonType::number_integer_t>::value and + not std::is_same<ArithmeticType, typename BasicJsonType::number_float_t>::value and + not std::is_same<ArithmeticType, typename BasicJsonType::boolean_t>::value, + int> = 0> +void from_json(const BasicJsonType& j, ArithmeticType& val) +{ + switch (static_cast<value_t>(j)) + { + case value_t::number_unsigned: + { + val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::number_unsigned_t*>()); + break; + } + case value_t::number_integer: + { + val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::number_integer_t*>()); + break; + } + case value_t::number_float: + { + val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::number_float_t*>()); + break; + } + case value_t::boolean: + { + val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::boolean_t*>()); + break; + } + + default: + JSON_THROW(type_error::create(302, "type must be number, but is " + std::string(j.type_name()))); + } +} + +template<typename BasicJsonType, typename A1, typename A2> +void from_json(const BasicJsonType& j, std::pair<A1, A2>& p) +{ + p = {j.at(0).template get<A1>(), j.at(1).template get<A2>()}; +} + +template<typename BasicJsonType, typename Tuple, std::size_t... Idx> +void from_json_tuple_impl(const BasicJsonType& j, Tuple& t, index_sequence<Idx...> /*unused*/) +{ + t = std::make_tuple(j.at(Idx).template get<typename std::tuple_element<Idx, Tuple>::type>()...); +} + +template<typename BasicJsonType, typename... Args> +void from_json(const BasicJsonType& j, std::tuple<Args...>& t) +{ + from_json_tuple_impl(j, t, index_sequence_for<Args...> {}); +} + +template <typename BasicJsonType, typename Key, typename Value, typename Compare, typename Allocator, + typename = enable_if_t<not std::is_constructible< + typename BasicJsonType::string_t, Key>::value>> +void from_json(const BasicJsonType& j, std::map<Key, Value, Compare, Allocator>& m) +{ + if (JSON_UNLIKELY(not j.is_array())) + { + JSON_THROW(type_error::create(302, "type must be array, but is " + std::string(j.type_name()))); + } + for (const auto& p : j) + { + if (JSON_UNLIKELY(not p.is_array())) + { + JSON_THROW(type_error::create(302, "type must be array, but is " + std::string(p.type_name()))); + } + m.emplace(p.at(0).template get<Key>(), p.at(1).template get<Value>()); + } +} + +template <typename BasicJsonType, typename Key, typename Value, typename Hash, typename KeyEqual, typename Allocator, + typename = enable_if_t<not std::is_constructible< + typename BasicJsonType::string_t, Key>::value>> +void from_json(const BasicJsonType& j, std::unordered_map<Key, Value, Hash, KeyEqual, Allocator>& m) +{ + if (JSON_UNLIKELY(not j.is_array())) + { + JSON_THROW(type_error::create(302, "type must be array, but is " + std::string(j.type_name()))); + } + for (const auto& p : j) + { + if (JSON_UNLIKELY(not p.is_array())) + { + JSON_THROW(type_error::create(302, "type must be array, but is " + std::string(p.type_name()))); + } + m.emplace(p.at(0).template get<Key>(), p.at(1).template get<Value>()); + } +} + +struct from_json_fn +{ + template<typename BasicJsonType, typename T> + auto operator()(const BasicJsonType& j, T& val) const + noexcept(noexcept(from_json(j, val))) + -> decltype(from_json(j, val), void()) + { + return from_json(j, val); } }; +} // namespace detail -template<class Sequence1, class Sequence2> -struct merge_and_renumber; +/// namespace to hold default `from_json` function +/// to see why this is required: +/// http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2015/n4381.html +namespace +{ +constexpr const auto& from_json = detail::static_const<detail::from_json_fn>::value; +} // namespace +} // namespace nlohmann -template<std::size_t... I1, std::size_t... I2> -struct merge_and_renumber<index_sequence<I1...>, index_sequence<I2...>> - : index_sequence < I1..., (sizeof...(I1) + I2)... > {}; +// #include <nlohmann/detail/conversions/to_json.hpp> -template<std::size_t N> -struct make_index_sequence - : merge_and_renumber < typename make_index_sequence < N / 2 >::type, - typename make_index_sequence < N - N / 2 >::type > {}; -template<> struct make_index_sequence<0> : index_sequence<> {}; -template<> struct make_index_sequence<1> : index_sequence<0> {}; +#include <ciso646> // or, and, not +#include <iterator> // begin, end +#include <tuple> // tuple, get +#include <type_traits> // is_same, is_constructible, is_floating_point, is_enum, underlying_type +#include <utility> // move, forward, declval, pair +#include <valarray> // valarray +#include <vector> // vector -template<typename... Ts> -using index_sequence_for = make_index_sequence<sizeof...(Ts)>; +// #include <nlohmann/detail/meta/cpp_future.hpp> -/* -Implementation of two C++17 constructs: conjunction, negation. This is needed -to avoid evaluating all the traits in a condition +// #include <nlohmann/detail/meta/type_traits.hpp> -For example: not std::is_same<void, T>::value and has_value_type<T>::value -will not compile when T = void (on MSVC at least). Whereas -conjunction<negation<std::is_same<void, T>>, has_value_type<T>>::value will -stop evaluating if negation<...>::value == false +// #include <nlohmann/detail/value_t.hpp> -Please note that those constructs must be used with caution, since symbols can -become very long quickly (which can slow down compilation and cause MSVC -internal compiler errors). Only use it when you have to (see example ahead). -*/ -template<class...> struct conjunction : std::true_type {}; -template<class B1> struct conjunction<B1> : B1 {}; -template<class B1, class... Bn> -struct conjunction<B1, Bn...> : std::conditional<bool(B1::value), conjunction<Bn...>, B1>::type {}; +// #include <nlohmann/detail/iterators/iteration_proxy.hpp> -template<class B> struct negation : std::integral_constant<bool, not B::value> {}; -// dispatch utility (taken from ranges-v3) -template<unsigned N> struct priority_tag : priority_tag < N - 1 > {}; -template<> struct priority_tag<0> {}; +#include <cstddef> // size_t +#include <string> // string, to_string +#include <iterator> // input_iterator_tag +#include <tuple> // tuple_size, get, tuple_element + +// #include <nlohmann/detail/value_t.hpp> + +// #include <nlohmann/detail/meta/type_traits.hpp> + + +namespace nlohmann +{ +namespace detail +{ +template <typename IteratorType> class iteration_proxy_value +{ + public: + using difference_type = std::ptrdiff_t; + using value_type = iteration_proxy_value; + using pointer = value_type * ; + using reference = value_type & ; + using iterator_category = std::input_iterator_tag; + + private: + /// the iterator + IteratorType anchor; + /// an index for arrays (used to create key names) + std::size_t array_index = 0; + /// last stringified array index + mutable std::size_t array_index_last = 0; + /// a string representation of the array index + mutable std::string array_index_str = "0"; + /// an empty string (to return a reference for primitive values) + const std::string empty_str = ""; + + public: + explicit iteration_proxy_value(IteratorType it) noexcept : anchor(it) {} + + /// dereference operator (needed for range-based for) + iteration_proxy_value& operator*() + { + return *this; + } + + /// increment operator (needed for range-based for) + iteration_proxy_value& operator++() + { + ++anchor; + ++array_index; + + return *this; + } + + /// equality operator (needed for InputIterator) + bool operator==(const iteration_proxy_value& o) const noexcept + { + return anchor == o.anchor; + } + + /// inequality operator (needed for range-based for) + bool operator!=(const iteration_proxy_value& o) const noexcept + { + return anchor != o.anchor; + } + + /// return key of the iterator + const std::string& key() const + { + assert(anchor.m_object != nullptr); + + switch (anchor.m_object->type()) + { + // use integer array index as key + case value_t::array: + { + if (array_index != array_index_last) + { + array_index_str = std::to_string(array_index); + array_index_last = array_index; + } + return array_index_str; + } + + // use key from the object + case value_t::object: + return anchor.key(); + + // use an empty key for all primitive types + default: + return empty_str; + } + } + + /// return value of the iterator + typename IteratorType::reference value() const + { + return anchor.value(); + } +}; + +/// proxy class for the items() function +template<typename IteratorType> class iteration_proxy +{ + private: + /// the container to iterate + typename IteratorType::reference container; + + public: + /// construct iteration proxy from a container + explicit iteration_proxy(typename IteratorType::reference cont) noexcept + : container(cont) {} + + /// return iterator begin (needed for range-based for) + iteration_proxy_value<IteratorType> begin() noexcept + { + return iteration_proxy_value<IteratorType>(container.begin()); + } + + /// return iterator end (needed for range-based for) + iteration_proxy_value<IteratorType> end() noexcept + { + return iteration_proxy_value<IteratorType>(container.end()); + } +}; +// Structured Bindings Support +// For further reference see https://blog.tartanllama.xyz/structured-bindings/ +// And see https://github.com/nlohmann/json/pull/1391 +template <std::size_t N, typename IteratorType, enable_if_t<N == 0, int> = 0> +auto get(const nlohmann::detail::iteration_proxy_value<IteratorType>& i) -> decltype(i.key()) +{ + return i.key(); +} +// Structured Bindings Support +// For further reference see https://blog.tartanllama.xyz/structured-bindings/ +// And see https://github.com/nlohmann/json/pull/1391 +template <std::size_t N, typename IteratorType, enable_if_t<N == 1, int> = 0> +auto get(const nlohmann::detail::iteration_proxy_value<IteratorType>& i) -> decltype(i.value()) +{ + return i.value(); +} +} // namespace detail +} // namespace nlohmann +// The Addition to the STD Namespace is required to add +// Structured Bindings Support to the iteration_proxy_value class +// For further reference see https://blog.tartanllama.xyz/structured-bindings/ +// And see https://github.com/nlohmann/json/pull/1391 +namespace std +{ +template <typename IteratorType> +class tuple_size<::nlohmann::detail::iteration_proxy_value<IteratorType>> + : public std::integral_constant<std::size_t, 2> {}; +template <std::size_t N, typename IteratorType> +class tuple_element<N, ::nlohmann::detail::iteration_proxy_value<IteratorType >> +{ + public: + using type = decltype( + get<N>(std::declval < + ::nlohmann::detail::iteration_proxy_value<IteratorType >> ())); +}; +} + +namespace nlohmann +{ +namespace detail +{ ////////////////// // constructors // ////////////////// @@ -657,6 +1849,16 @@ struct external_constructor<value_t::string> j.m_value = std::move(s); j.assert_invariant(); } + + template<typename BasicJsonType, typename CompatibleStringType, + enable_if_t<not std::is_same<CompatibleStringType, typename BasicJsonType::string_t>::value, + int> = 0> + static void construct(BasicJsonType& j, const CompatibleStringType& str) + { + j.m_type = value_t::string; + j.m_value.string = j.template create<typename BasicJsonType::string_t>(str); + j.assert_invariant(); + } }; template<> @@ -783,159 +1985,6 @@ struct external_constructor<value_t::object> } }; - -//////////////////////// -// has_/is_ functions // -//////////////////////// - -/*! -@brief Helper to determine whether there's a key_type for T. - -This helper is used to tell associative containers apart from other containers -such as sequence containers. For instance, `std::map` passes the test as it -contains a `mapped_type`, whereas `std::vector` fails the test. - -@sa http://stackoverflow.com/a/7728728/266378 -@since version 1.0.0, overworked in version 2.0.6 -*/ -#define NLOHMANN_JSON_HAS_HELPER(type) \ - template<typename T> struct has_##type { \ - private: \ - template<typename U, typename = typename U::type> \ - static int detect(U &&); \ - static void detect(...); \ - public: \ - static constexpr bool value = \ - std::is_integral<decltype(detect(std::declval<T>()))>::value; \ - } - -NLOHMANN_JSON_HAS_HELPER(mapped_type); -NLOHMANN_JSON_HAS_HELPER(key_type); -NLOHMANN_JSON_HAS_HELPER(value_type); -NLOHMANN_JSON_HAS_HELPER(iterator); - -#undef NLOHMANN_JSON_HAS_HELPER - - -template<bool B, class RealType, class CompatibleObjectType> -struct is_compatible_object_type_impl : std::false_type {}; - -template<class RealType, class CompatibleObjectType> -struct is_compatible_object_type_impl<true, RealType, CompatibleObjectType> -{ - static constexpr auto value = - std::is_constructible<typename RealType::key_type, typename CompatibleObjectType::key_type>::value and - std::is_constructible<typename RealType::mapped_type, typename CompatibleObjectType::mapped_type>::value; -}; - -template<class BasicJsonType, class CompatibleObjectType> -struct is_compatible_object_type -{ - static auto constexpr value = is_compatible_object_type_impl < - conjunction<negation<std::is_same<void, CompatibleObjectType>>, - has_mapped_type<CompatibleObjectType>, - has_key_type<CompatibleObjectType>>::value, - typename BasicJsonType::object_t, CompatibleObjectType >::value; -}; - -template<typename BasicJsonType, typename T> -struct is_basic_json_nested_type -{ - static auto constexpr value = std::is_same<T, typename BasicJsonType::iterator>::value or - std::is_same<T, typename BasicJsonType::const_iterator>::value or - std::is_same<T, typename BasicJsonType::reverse_iterator>::value or - std::is_same<T, typename BasicJsonType::const_reverse_iterator>::value; -}; - -template<class BasicJsonType, class CompatibleArrayType> -struct is_compatible_array_type -{ - static auto constexpr value = - conjunction<negation<std::is_same<void, CompatibleArrayType>>, - negation<is_compatible_object_type< - BasicJsonType, CompatibleArrayType>>, - negation<std::is_constructible<typename BasicJsonType::string_t, - CompatibleArrayType>>, - negation<is_basic_json_nested_type<BasicJsonType, CompatibleArrayType>>, - has_value_type<CompatibleArrayType>, - has_iterator<CompatibleArrayType>>::value; -}; - -template<bool, typename, typename> -struct is_compatible_integer_type_impl : std::false_type {}; - -template<typename RealIntegerType, typename CompatibleNumberIntegerType> -struct is_compatible_integer_type_impl<true, RealIntegerType, CompatibleNumberIntegerType> -{ - // is there an assert somewhere on overflows? - using RealLimits = std::numeric_limits<RealIntegerType>; - using CompatibleLimits = std::numeric_limits<CompatibleNumberIntegerType>; - - static constexpr auto value = - std::is_constructible<RealIntegerType, CompatibleNumberIntegerType>::value and - CompatibleLimits::is_integer and - RealLimits::is_signed == CompatibleLimits::is_signed; -}; - -template<typename RealIntegerType, typename CompatibleNumberIntegerType> -struct is_compatible_integer_type -{ - static constexpr auto value = - is_compatible_integer_type_impl < - std::is_integral<CompatibleNumberIntegerType>::value and - not std::is_same<bool, CompatibleNumberIntegerType>::value, - RealIntegerType, CompatibleNumberIntegerType >::value; -}; - - -// trait checking if JSONSerializer<T>::from_json(json const&, udt&) exists -template<typename BasicJsonType, typename T> -struct has_from_json -{ - private: - // also check the return type of from_json - template<typename U, typename = enable_if_t<std::is_same<void, decltype(uncvref_t<U>::from_json( - std::declval<BasicJsonType>(), std::declval<T&>()))>::value>> - static int detect(U&&); - static void detect(...); - - public: - static constexpr bool value = std::is_integral<decltype( - detect(std::declval<typename BasicJsonType::template json_serializer<T, void>>()))>::value; -}; - -// This trait checks if JSONSerializer<T>::from_json(json const&) exists -// this overload is used for non-default-constructible user-defined-types -template<typename BasicJsonType, typename T> -struct has_non_default_from_json -{ - private: - template<typename U, typename = - enable_if_t<std::is_same<T, decltype(uncvref_t<U>::from_json(std::declval<BasicJsonType>()))>::value>> - static int detect(U&&); - static void detect(...); - - public: - static constexpr bool value = std::is_integral<decltype(detect( - std::declval<typename BasicJsonType::template json_serializer<T, void>>()))>::value; -}; - -// This trait checks if BasicJsonType::json_serializer<T>::to_json exists -template<typename BasicJsonType, typename T> -struct has_to_json -{ - private: - template<typename U, typename = decltype(uncvref_t<U>::to_json( - std::declval<BasicJsonType&>(), std::declval<T>()))> - static int detect(U&&); - static void detect(...); - - public: - static constexpr bool value = std::is_integral<decltype(detect( - std::declval<typename BasicJsonType::template json_serializer<T, void>>()))>::value; -}; - - ///////////// // to_json // ///////////// @@ -995,10 +2044,14 @@ void to_json(BasicJsonType& j, const std::vector<bool>& e) external_constructor<value_t::array>::construct(j, e); } -template<typename BasicJsonType, typename CompatibleArrayType, - enable_if_t<is_compatible_array_type<BasicJsonType, CompatibleArrayType>::value or - std::is_same<typename BasicJsonType::array_t, CompatibleArrayType>::value, - int> = 0> +template <typename BasicJsonType, typename CompatibleArrayType, + enable_if_t<is_compatible_array_type<BasicJsonType, + CompatibleArrayType>::value and + not is_compatible_object_type< + BasicJsonType, CompatibleArrayType>::value and + not is_compatible_string_type<BasicJsonType, CompatibleArrayType>::value and + not is_basic_json<CompatibleArrayType>::value, + int> = 0> void to_json(BasicJsonType& j, const CompatibleArrayType& arr) { external_constructor<value_t::array>::construct(j, arr); @@ -1006,7 +2059,7 @@ void to_json(BasicJsonType& j, const CompatibleArrayType& arr) template<typename BasicJsonType, typename T, enable_if_t<std::is_convertible<T, BasicJsonType>::value, int> = 0> -void to_json(BasicJsonType& j, std::valarray<T> arr) +void to_json(BasicJsonType& j, const std::valarray<T>& arr) { external_constructor<value_t::array>::construct(j, std::move(arr)); } @@ -1018,7 +2071,7 @@ void to_json(BasicJsonType& j, typename BasicJsonType::array_t&& arr) } template<typename BasicJsonType, typename CompatibleObjectType, - enable_if_t<is_compatible_object_type<BasicJsonType, CompatibleObjectType>::value, int> = 0> + enable_if_t<is_compatible_object_type<BasicJsonType, CompatibleObjectType>::value and not is_basic_json<CompatibleObjectType>::value, int> = 0> void to_json(BasicJsonType& j, const CompatibleObjectType& obj) { external_constructor<value_t::object>::construct(j, obj); @@ -1030,9 +2083,12 @@ void to_json(BasicJsonType& j, typename BasicJsonType::object_t&& obj) external_constructor<value_t::object>::construct(j, std::move(obj)); } -template<typename BasicJsonType, typename T, std::size_t N, - enable_if_t<not std::is_constructible<typename BasicJsonType::string_t, T (&)[N]>::value, int> = 0> -void to_json(BasicJsonType& j, T (&arr)[N]) +template < + typename BasicJsonType, typename T, std::size_t N, + enable_if_t<not std::is_constructible<typename BasicJsonType::string_t, + const T(&)[N]>::value, + int> = 0 > +void to_json(BasicJsonType& j, const T(&arr)[N]) { external_constructor<value_t::array>::construct(j, arr); } @@ -1040,351 +2096,71 @@ void to_json(BasicJsonType& j, T (&arr)[N]) template<typename BasicJsonType, typename... Args> void to_json(BasicJsonType& j, const std::pair<Args...>& p) { - j = {p.first, p.second}; -} - -template<typename BasicJsonType, typename Tuple, std::size_t... Idx> -void to_json_tuple_impl(BasicJsonType& j, const Tuple& t, index_sequence<Idx...>) -{ - j = {std::get<Idx>(t)...}; -} - -template<typename BasicJsonType, typename... Args> -void to_json(BasicJsonType& j, const std::tuple<Args...>& t) -{ - to_json_tuple_impl(j, t, index_sequence_for<Args...> {}); -} - -/////////////// -// from_json // -/////////////// - -// overloads for basic_json template parameters -template<typename BasicJsonType, typename ArithmeticType, - enable_if_t<std::is_arithmetic<ArithmeticType>::value and - not std::is_same<ArithmeticType, typename BasicJsonType::boolean_t>::value, - int> = 0> -void get_arithmetic_value(const BasicJsonType& j, ArithmeticType& val) -{ - switch (static_cast<value_t>(j)) - { - case value_t::number_unsigned: - { - val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::number_unsigned_t*>()); - break; - } - case value_t::number_integer: - { - val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::number_integer_t*>()); - break; - } - case value_t::number_float: - { - val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::number_float_t*>()); - break; - } - - default: - JSON_THROW(type_error::create(302, "type must be number, but is " + std::string(j.type_name()))); - } -} - -template<typename BasicJsonType> -void from_json(const BasicJsonType& j, typename BasicJsonType::boolean_t& b) -{ - if (JSON_UNLIKELY(not j.is_boolean())) - { - JSON_THROW(type_error::create(302, "type must be boolean, but is " + std::string(j.type_name()))); - } - b = *j.template get_ptr<const typename BasicJsonType::boolean_t*>(); -} - -template<typename BasicJsonType> -void from_json(const BasicJsonType& j, typename BasicJsonType::string_t& s) -{ - if (JSON_UNLIKELY(not j.is_string())) - { - JSON_THROW(type_error::create(302, "type must be string, but is " + std::string(j.type_name()))); - } - s = *j.template get_ptr<const typename BasicJsonType::string_t*>(); -} - -template<typename BasicJsonType> -void from_json(const BasicJsonType& j, typename BasicJsonType::number_float_t& val) -{ - get_arithmetic_value(j, val); -} - -template<typename BasicJsonType> -void from_json(const BasicJsonType& j, typename BasicJsonType::number_unsigned_t& val) -{ - get_arithmetic_value(j, val); -} - -template<typename BasicJsonType> -void from_json(const BasicJsonType& j, typename BasicJsonType::number_integer_t& val) -{ - get_arithmetic_value(j, val); -} - -template<typename BasicJsonType, typename EnumType, - enable_if_t<std::is_enum<EnumType>::value, int> = 0> -void from_json(const BasicJsonType& j, EnumType& e) -{ - typename std::underlying_type<EnumType>::type val; - get_arithmetic_value(j, val); - e = static_cast<EnumType>(val); -} - -template<typename BasicJsonType> -void from_json(const BasicJsonType& j, typename BasicJsonType::array_t& arr) -{ - if (JSON_UNLIKELY(not j.is_array())) - { - JSON_THROW(type_error::create(302, "type must be array, but is " + std::string(j.type_name()))); - } - arr = *j.template get_ptr<const typename BasicJsonType::array_t*>(); -} - -// forward_list doesn't have an insert method -template<typename BasicJsonType, typename T, typename Allocator, - enable_if_t<std::is_convertible<BasicJsonType, T>::value, int> = 0> -void from_json(const BasicJsonType& j, std::forward_list<T, Allocator>& l) -{ - if (JSON_UNLIKELY(not j.is_array())) - { - JSON_THROW(type_error::create(302, "type must be array, but is " + std::string(j.type_name()))); - } - std::transform(j.rbegin(), j.rend(), - std::front_inserter(l), [](const BasicJsonType & i) - { - return i.template get<T>(); - }); -} - -// valarray doesn't have an insert method -template<typename BasicJsonType, typename T, - enable_if_t<std::is_convertible<BasicJsonType, T>::value, int> = 0> -void from_json(const BasicJsonType& j, std::valarray<T>& l) -{ - if (JSON_UNLIKELY(not j.is_array())) - { - JSON_THROW(type_error::create(302, "type must be array, but is " + std::string(j.type_name()))); - } - l.resize(j.size()); - std::copy(j.m_value.array->begin(), j.m_value.array->end(), std::begin(l)); -} - -template<typename BasicJsonType, typename CompatibleArrayType> -void from_json_array_impl(const BasicJsonType& j, CompatibleArrayType& arr, priority_tag<0> /*unused*/) -{ - using std::end; - - std::transform(j.begin(), j.end(), - std::inserter(arr, end(arr)), [](const BasicJsonType & i) - { - // get<BasicJsonType>() returns *this, this won't call a from_json - // method when value_type is BasicJsonType - return i.template get<typename CompatibleArrayType::value_type>(); - }); -} - -template<typename BasicJsonType, typename CompatibleArrayType> -auto from_json_array_impl(const BasicJsonType& j, CompatibleArrayType& arr, priority_tag<1> /*unused*/) --> decltype( - arr.reserve(std::declval<typename CompatibleArrayType::size_type>()), - void()) -{ - using std::end; - - arr.reserve(j.size()); - std::transform(j.begin(), j.end(), - std::inserter(arr, end(arr)), [](const BasicJsonType & i) - { - // get<BasicJsonType>() returns *this, this won't call a from_json - // method when value_type is BasicJsonType - return i.template get<typename CompatibleArrayType::value_type>(); - }); -} - -template<typename BasicJsonType, typename T, std::size_t N> -void from_json_array_impl(const BasicJsonType& j, std::array<T, N>& arr, priority_tag<2> /*unused*/) -{ - for (std::size_t i = 0; i < N; ++i) - { - arr[i] = j.at(i).template get<T>(); - } -} - -template<typename BasicJsonType, typename CompatibleArrayType, - enable_if_t<is_compatible_array_type<BasicJsonType, CompatibleArrayType>::value and - std::is_convertible<BasicJsonType, typename CompatibleArrayType::value_type>::value and - not std::is_same<typename BasicJsonType::array_t, CompatibleArrayType>::value, int> = 0> -void from_json(const BasicJsonType& j, CompatibleArrayType& arr) -{ - if (JSON_UNLIKELY(not j.is_array())) - { - JSON_THROW(type_error::create(302, "type must be array, but is " + std::string(j.type_name()))); - } - - from_json_array_impl(j, arr, priority_tag<2> {}); -} - -template<typename BasicJsonType, typename CompatibleObjectType, - enable_if_t<is_compatible_object_type<BasicJsonType, CompatibleObjectType>::value, int> = 0> -void from_json(const BasicJsonType& j, CompatibleObjectType& obj) -{ - if (JSON_UNLIKELY(not j.is_object())) - { - JSON_THROW(type_error::create(302, "type must be object, but is " + std::string(j.type_name()))); - } - - auto inner_object = j.template get_ptr<const typename BasicJsonType::object_t*>(); - using value_type = typename CompatibleObjectType::value_type; - std::transform( - inner_object->begin(), inner_object->end(), - std::inserter(obj, obj.begin()), - [](typename BasicJsonType::object_t::value_type const & p) - { - return value_type(p.first, p.second.template get<typename CompatibleObjectType::mapped_type>()); - }); -} - -// overload for arithmetic types, not chosen for basic_json template arguments -// (BooleanType, etc..); note: Is it really necessary to provide explicit -// overloads for boolean_t etc. in case of a custom BooleanType which is not -// an arithmetic type? -template<typename BasicJsonType, typename ArithmeticType, - enable_if_t < - std::is_arithmetic<ArithmeticType>::value and - not std::is_same<ArithmeticType, typename BasicJsonType::number_unsigned_t>::value and - not std::is_same<ArithmeticType, typename BasicJsonType::number_integer_t>::value and - not std::is_same<ArithmeticType, typename BasicJsonType::number_float_t>::value and - not std::is_same<ArithmeticType, typename BasicJsonType::boolean_t>::value, - int> = 0> -void from_json(const BasicJsonType& j, ArithmeticType& val) -{ - switch (static_cast<value_t>(j)) - { - case value_t::number_unsigned: - { - val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::number_unsigned_t*>()); - break; - } - case value_t::number_integer: - { - val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::number_integer_t*>()); - break; - } - case value_t::number_float: - { - val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::number_float_t*>()); - break; - } - case value_t::boolean: - { - val = static_cast<ArithmeticType>(*j.template get_ptr<const typename BasicJsonType::boolean_t*>()); - break; - } - - default: - JSON_THROW(type_error::create(302, "type must be number, but is " + std::string(j.type_name()))); - } + j = { p.first, p.second }; } -template<typename BasicJsonType, typename A1, typename A2> -void from_json(const BasicJsonType& j, std::pair<A1, A2>& p) +// for https://github.com/nlohmann/json/pull/1134 +template < typename BasicJsonType, typename T, + enable_if_t<std::is_same<T, iteration_proxy_value<typename BasicJsonType::iterator>>::value, int> = 0> +void to_json(BasicJsonType& j, const T& b) { - p = {j.at(0).template get<A1>(), j.at(1).template get<A2>()}; + j = { {b.key(), b.value()} }; } template<typename BasicJsonType, typename Tuple, std::size_t... Idx> -void from_json_tuple_impl(const BasicJsonType& j, Tuple& t, index_sequence<Idx...>) +void to_json_tuple_impl(BasicJsonType& j, const Tuple& t, index_sequence<Idx...> /*unused*/) { - t = std::make_tuple(j.at(Idx).template get<typename std::tuple_element<Idx, Tuple>::type>()...); + j = { std::get<Idx>(t)... }; } template<typename BasicJsonType, typename... Args> -void from_json(const BasicJsonType& j, std::tuple<Args...>& t) +void to_json(BasicJsonType& j, const std::tuple<Args...>& t) { - from_json_tuple_impl(j, t, index_sequence_for<Args...> {}); + to_json_tuple_impl(j, t, index_sequence_for<Args...> {}); } struct to_json_fn { - private: template<typename BasicJsonType, typename T> - auto call(BasicJsonType& j, T&& val, priority_tag<1> /*unused*/) const noexcept(noexcept(to_json(j, std::forward<T>(val)))) + auto operator()(BasicJsonType& j, T&& val) const noexcept(noexcept(to_json(j, std::forward<T>(val)))) -> decltype(to_json(j, std::forward<T>(val)), void()) { return to_json(j, std::forward<T>(val)); } +}; +} // namespace detail - template<typename BasicJsonType, typename T> - void call(BasicJsonType& /*unused*/, T&& /*unused*/, priority_tag<0> /*unused*/) const noexcept - { - static_assert(sizeof(BasicJsonType) == 0, - "could not find to_json() method in T's namespace"); +/// namespace to hold default `to_json` function +namespace +{ +constexpr const auto& to_json = detail::static_const<detail::to_json_fn>::value; +} // namespace +} // namespace nlohmann -#ifdef _MSC_VER - // MSVC does not show a stacktrace for the above assert - using decayed = uncvref_t<T>; - static_assert(sizeof(typename decayed::force_msvc_stacktrace) == 0, - "forcing MSVC stacktrace to show which T we're talking about."); -#endif - } +// #include <nlohmann/detail/input/input_adapters.hpp> - public: - template<typename BasicJsonType, typename T> - void operator()(BasicJsonType& j, T&& val) const - noexcept(noexcept(std::declval<to_json_fn>().call(j, std::forward<T>(val), priority_tag<1> {}))) - { - return call(j, std::forward<T>(val), priority_tag<1> {}); - } -}; -struct from_json_fn -{ - private: - template<typename BasicJsonType, typename T> - auto call(const BasicJsonType& j, T& val, priority_tag<1> /*unused*/) const - noexcept(noexcept(from_json(j, val))) - -> decltype(from_json(j, val), void()) - { - return from_json(j, val); - } +#include <cassert> // assert +#include <cstddef> // size_t +#include <cstring> // strlen +#include <istream> // istream +#include <iterator> // begin, end, iterator_traits, random_access_iterator_tag, distance, next +#include <memory> // shared_ptr, make_shared, addressof +#include <numeric> // accumulate +#include <string> // string, char_traits +#include <type_traits> // enable_if, is_base_of, is_pointer, is_integral, remove_pointer +#include <utility> // pair, declval +#include <cstdio> //FILE * - template<typename BasicJsonType, typename T> - void call(const BasicJsonType& /*unused*/, T& /*unused*/, priority_tag<0> /*unused*/) const noexcept - { - static_assert(sizeof(BasicJsonType) == 0, - "could not find from_json() method in T's namespace"); -#ifdef _MSC_VER - // MSVC does not show a stacktrace for the above assert - using decayed = uncvref_t<T>; - static_assert(sizeof(typename decayed::force_msvc_stacktrace) == 0, - "forcing MSVC stacktrace to show which T we're talking about."); -#endif - } +// #include <nlohmann/detail/macro_scope.hpp> - public: - template<typename BasicJsonType, typename T> - void operator()(const BasicJsonType& j, T& val) const - noexcept(noexcept(std::declval<from_json_fn>().call(j, val, priority_tag<1> {}))) - { - return call(j, val, priority_tag<1> {}); - } -}; -// taken from ranges-v3 -template<typename T> -struct static_const +namespace nlohmann { - static constexpr T value{}; -}; - -template<typename T> -constexpr T static_const<T>::value; +namespace detail +{ +/// the supported input formats +enum class input_format_t { json, cbor, msgpack, ubjson, bson }; //////////////////// // input adapters // @@ -1394,19 +2170,17 @@ constexpr T static_const<T>::value; @brief abstract input adapter interface Produces a stream of std::char_traits<char>::int_type characters from a -std::istream, a buffer, or some other input type. Accepts the return of exactly -one non-EOF character for future input. The int_type characters returned -consist of all valid char values as positive values (typically unsigned char), -plus an EOF value outside that range, specified by the value of the function -std::char_traits<char>::eof(). This value is typically -1, but could be any -arbitrary value which is not a valid char value. +std::istream, a buffer, or some other input type. Accepts the return of +exactly one non-EOF character for future input. The int_type characters +returned consist of all valid char values as positive values (typically +unsigned char), plus an EOF value outside that range, specified by the value +of the function std::char_traits<char>::eof(). This value is typically -1, but +could be any arbitrary value which is not a valid char value. */ struct input_adapter_protocol { /// get a character [0,255] or std::char_traits<char>::eof(). virtual std::char_traits<char>::int_type get_character() = 0; - /// restore the last non-eof() character to input - virtual void unget_character() = 0; virtual ~input_adapter_protocol() = default; }; @@ -1414,6 +2188,27 @@ struct input_adapter_protocol using input_adapter_t = std::shared_ptr<input_adapter_protocol>; /*! +Input adapter for stdio file access. This adapter read only 1 byte and do not use any + buffer. This adapter is a very low level adapter. +*/ +class file_input_adapter : public input_adapter_protocol +{ + public: + explicit file_input_adapter(std::FILE* f) noexcept + : m_file(f) + {} + + std::char_traits<char>::int_type get_character() noexcept override + { + return std::fgetc(m_file); + } + private: + /// the file pointer to read from + std::FILE* m_file; +}; + + +/*! Input adapter for a (caching) istream. Ignores a UFT Byte Order Mark at beginning of input. Does not support changing the underlying std::streambuf in mid-input. Maintains underlying std::istream and std::streambuf to support @@ -1428,56 +2223,32 @@ class input_stream_adapter : public input_adapter_protocol ~input_stream_adapter() override { // clear stream flags; we use underlying streambuf I/O, do not - // maintain ifstream flags - is.clear(); + // maintain ifstream flags, except eof + is.clear(is.rdstate() & std::ios::eofbit); } explicit input_stream_adapter(std::istream& i) : is(i), sb(*i.rdbuf()) - { - // skip byte order mark - std::char_traits<char>::int_type c; - if ((c = get_character()) == 0xEF) - { - if ((c = get_character()) == 0xBB) - { - if ((c = get_character()) == 0xBF) - { - return; // Ignore BOM - } - else if (c != std::char_traits<char>::eof()) - { - is.unget(); - } - is.putback('\xBB'); - } - else if (c != std::char_traits<char>::eof()) - { - is.unget(); - } - is.putback('\xEF'); - } - else if (c != std::char_traits<char>::eof()) - { - is.unget(); // no byte order mark; process as usual - } - } + {} // delete because of pointer members input_stream_adapter(const input_stream_adapter&) = delete; input_stream_adapter& operator=(input_stream_adapter&) = delete; + input_stream_adapter(input_stream_adapter&&) = delete; + input_stream_adapter& operator=(input_stream_adapter&&) = delete; // std::istream/std::streambuf use std::char_traits<char>::to_int_type, to // ensure that std::char_traits<char>::eof() and the character 0xFF do not // end up as the same value, eg. 0xFFFFFFFF. std::char_traits<char>::int_type get_character() override { - return sb.sbumpc(); - } - - void unget_character() override - { - sb.sungetc(); // is.unget() avoided for performance + auto res = sb.sbumpc(); + // set eof manually, as we don't use the istream interface. + if (res == EOF) + { + is.clear(is.rdstate() | std::ios::eofbit); + } + return res; } private: @@ -1490,19 +2261,16 @@ class input_stream_adapter : public input_adapter_protocol class input_buffer_adapter : public input_adapter_protocol { public: - input_buffer_adapter(const char* b, const std::size_t l) - : cursor(b), limit(b + l), start(b) - { - // skip byte order mark - if (l >= 3 and b[0] == '\xEF' and b[1] == '\xBB' and b[2] == '\xBF') - { - cursor += 3; - } - } + input_buffer_adapter(const char* b, const std::size_t l) noexcept + : cursor(b), limit(b + l) + {} // delete because of pointer members input_buffer_adapter(const input_buffer_adapter&) = delete; input_buffer_adapter& operator=(input_buffer_adapter&) = delete; + input_buffer_adapter(input_buffer_adapter&&) = delete; + input_buffer_adapter& operator=(input_buffer_adapter&&) = delete; + ~input_buffer_adapter() override = default; std::char_traits<char>::int_type get_character() noexcept override { @@ -1514,28 +2282,182 @@ class input_buffer_adapter : public input_adapter_protocol return std::char_traits<char>::eof(); } - void unget_character() noexcept override + private: + /// pointer to the current character + const char* cursor; + /// pointer past the last character + const char* const limit; +}; + +template<typename WideStringType, size_t T> +struct wide_string_input_helper +{ + // UTF-32 + static void fill_buffer(const WideStringType& str, size_t& current_wchar, std::array<std::char_traits<char>::int_type, 4>& utf8_bytes, size_t& utf8_bytes_index, size_t& utf8_bytes_filled) { - if (JSON_LIKELY(cursor > start)) + utf8_bytes_index = 0; + + if (current_wchar == str.size()) { - --cursor; + utf8_bytes[0] = std::char_traits<char>::eof(); + utf8_bytes_filled = 1; + } + else + { + // get the current character + const auto wc = static_cast<int>(str[current_wchar++]); + + // UTF-32 to UTF-8 encoding + if (wc < 0x80) + { + utf8_bytes[0] = wc; + utf8_bytes_filled = 1; + } + else if (wc <= 0x7FF) + { + utf8_bytes[0] = 0xC0 | ((wc >> 6) & 0x1F); + utf8_bytes[1] = 0x80 | (wc & 0x3F); + utf8_bytes_filled = 2; + } + else if (wc <= 0xFFFF) + { + utf8_bytes[0] = 0xE0 | ((wc >> 12) & 0x0F); + utf8_bytes[1] = 0x80 | ((wc >> 6) & 0x3F); + utf8_bytes[2] = 0x80 | (wc & 0x3F); + utf8_bytes_filled = 3; + } + else if (wc <= 0x10FFFF) + { + utf8_bytes[0] = 0xF0 | ((wc >> 18) & 0x07); + utf8_bytes[1] = 0x80 | ((wc >> 12) & 0x3F); + utf8_bytes[2] = 0x80 | ((wc >> 6) & 0x3F); + utf8_bytes[3] = 0x80 | (wc & 0x3F); + utf8_bytes_filled = 4; + } + else + { + // unknown character + utf8_bytes[0] = wc; + utf8_bytes_filled = 1; + } } } +}; + +template<typename WideStringType> +struct wide_string_input_helper<WideStringType, 2> +{ + // UTF-16 + static void fill_buffer(const WideStringType& str, size_t& current_wchar, std::array<std::char_traits<char>::int_type, 4>& utf8_bytes, size_t& utf8_bytes_index, size_t& utf8_bytes_filled) + { + utf8_bytes_index = 0; + + if (current_wchar == str.size()) + { + utf8_bytes[0] = std::char_traits<char>::eof(); + utf8_bytes_filled = 1; + } + else + { + // get the current character + const auto wc = static_cast<int>(str[current_wchar++]); + + // UTF-16 to UTF-8 encoding + if (wc < 0x80) + { + utf8_bytes[0] = wc; + utf8_bytes_filled = 1; + } + else if (wc <= 0x7FF) + { + utf8_bytes[0] = 0xC0 | ((wc >> 6)); + utf8_bytes[1] = 0x80 | (wc & 0x3F); + utf8_bytes_filled = 2; + } + else if (0xD800 > wc or wc >= 0xE000) + { + utf8_bytes[0] = 0xE0 | ((wc >> 12)); + utf8_bytes[1] = 0x80 | ((wc >> 6) & 0x3F); + utf8_bytes[2] = 0x80 | (wc & 0x3F); + utf8_bytes_filled = 3; + } + else + { + if (current_wchar < str.size()) + { + const auto wc2 = static_cast<int>(str[current_wchar++]); + const int charcode = 0x10000 + (((wc & 0x3FF) << 10) | (wc2 & 0x3FF)); + utf8_bytes[0] = 0xf0 | (charcode >> 18); + utf8_bytes[1] = 0x80 | ((charcode >> 12) & 0x3F); + utf8_bytes[2] = 0x80 | ((charcode >> 6) & 0x3F); + utf8_bytes[3] = 0x80 | (charcode & 0x3F); + utf8_bytes_filled = 4; + } + else + { + // unknown character + ++current_wchar; + utf8_bytes[0] = wc; + utf8_bytes_filled = 1; + } + } + } + } +}; + +template<typename WideStringType> +class wide_string_input_adapter : public input_adapter_protocol +{ + public: + explicit wide_string_input_adapter(const WideStringType& w) noexcept + : str(w) + {} + + std::char_traits<char>::int_type get_character() noexcept override + { + // check if buffer needs to be filled + if (utf8_bytes_index == utf8_bytes_filled) + { + fill_buffer<sizeof(typename WideStringType::value_type)>(); + + assert(utf8_bytes_filled > 0); + assert(utf8_bytes_index == 0); + } + + // use buffer + assert(utf8_bytes_filled > 0); + assert(utf8_bytes_index < utf8_bytes_filled); + return utf8_bytes[utf8_bytes_index++]; + } private: - /// pointer to the current character - const char* cursor; - /// pointer past the last character - const char* limit; - /// pointer to the first character - const char* start; + template<size_t T> + void fill_buffer() + { + wide_string_input_helper<WideStringType, T>::fill_buffer(str, current_wchar, utf8_bytes, utf8_bytes_index, utf8_bytes_filled); + } + + /// the wstring to process + const WideStringType& str; + + /// index of the current wchar in str + std::size_t current_wchar = 0; + + /// a buffer for UTF-8 bytes + std::array<std::char_traits<char>::int_type, 4> utf8_bytes = {{0, 0, 0, 0}}; + + /// index to the utf8_codes array for the next valid byte + std::size_t utf8_bytes_index = 0; + /// number of valid bytes in the utf8_codes array + std::size_t utf8_bytes_filled = 0; }; class input_adapter { public: // native support - + input_adapter(std::FILE* file) + : ia(std::make_shared<file_input_adapter>(file)) {} /// input adapter for input stream input_adapter(std::istream& i) : ia(std::make_shared<input_stream_adapter>(i)) {} @@ -1544,6 +2466,15 @@ class input_adapter input_adapter(std::istream&& i) : ia(std::make_shared<input_stream_adapter>(i)) {} + input_adapter(const std::wstring& ws) + : ia(std::make_shared<wide_string_input_adapter<std::wstring>>(ws)) {} + + input_adapter(const std::u16string& ws) + : ia(std::make_shared<wide_string_input_adapter<std::u16string>>(ws)) {} + + input_adapter(const std::u32string& ws) + : ia(std::make_shared<wide_string_input_adapter<std::u32string>>(ws)) {} + /// input adapter for buffer template<typename CharT, typename std::enable_if< @@ -1570,23 +2501,26 @@ class input_adapter /// input adapter for iterator range with contiguous storage template<class IteratorType, typename std::enable_if< - std::is_same<typename std::iterator_traits<IteratorType>::iterator_category, std::random_access_iterator_tag>::value, + std::is_same<typename iterator_traits<IteratorType>::iterator_category, std::random_access_iterator_tag>::value, int>::type = 0> input_adapter(IteratorType first, IteratorType last) { +#ifndef NDEBUG // assertion to check that the iterator range is indeed contiguous, // see http://stackoverflow.com/a/35008842/266378 for more discussion - assert(std::accumulate( - first, last, std::pair<bool, int>(true, 0), - [&first](std::pair<bool, int> res, decltype(*first) val) + const auto is_contiguous = std::accumulate( + first, last, std::pair<bool, int>(true, 0), + [&first](std::pair<bool, int> res, decltype(*first) val) { res.first &= (val == *(std::next(std::addressof(*first), res.second++))); return res; - }).first); + }).first; + assert(is_contiguous); +#endif // assertion to check that each element is 1 byte long static_assert( - sizeof(typename std::iterator_traits<IteratorType>::value_type) == 1, + sizeof(typename iterator_traits<IteratorType>::value_type) == 1, "each element in the iterator range must have the size of 1 byte"); const auto len = static_cast<size_t>(std::distance(first, last)); @@ -1610,7 +2544,7 @@ class input_adapter /// input adapter for contiguous container template<class ContiguousContainer, typename std::enable_if<not std::is_pointer<ContiguousContainer>::value and - std::is_base_of<std::random_access_iterator_tag, typename std::iterator_traits<decltype(std::begin(std::declval<ContiguousContainer const>()))>::iterator_category>::value, + std::is_base_of<std::random_access_iterator_tag, typename iterator_traits<decltype(std::begin(std::declval<ContiguousContainer const>()))>::iterator_category>::value, int>::type = 0> input_adapter(const ContiguousContainer& c) : input_adapter(std::begin(c), std::end(c)) {} @@ -1624,10 +2558,34 @@ class input_adapter /// the actual adapter input_adapter_t ia = nullptr; }; +} // namespace detail +} // namespace nlohmann -////////////////////// -// lexer and parser // -////////////////////// +// #include <nlohmann/detail/input/lexer.hpp> + + +#include <clocale> // localeconv +#include <cstddef> // size_t +#include <cstdlib> // strtof, strtod, strtold, strtoll, strtoull +#include <cstdio> // snprintf +#include <initializer_list> // initializer_list +#include <string> // char_traits, string +#include <vector> // vector + +// #include <nlohmann/detail/macro_scope.hpp> + +// #include <nlohmann/detail/input/input_adapters.hpp> + +// #include <nlohmann/detail/input/position_t.hpp> + + +namespace nlohmann +{ +namespace detail +{ +/////////// +// lexer // +/////////// /*! @brief lexical analysis @@ -1640,6 +2598,7 @@ class lexer using number_integer_t = typename BasicJsonType::number_integer_t; using number_unsigned_t = typename BasicJsonType::number_unsigned_t; using number_float_t = typename BasicJsonType::number_float_t; + using string_t = typename BasicJsonType::string_t; public: /// token types for the parser @@ -1701,17 +2660,22 @@ class lexer return "end of input"; case token_type::literal_or_value: return "'[', '{', or a literal"; + // LCOV_EXCL_START default: // catch non-enum values - return "unknown token"; // LCOV_EXCL_LINE + return "unknown token"; + // LCOV_EXCL_STOP } } - explicit lexer(detail::input_adapter_t adapter) + explicit lexer(detail::input_adapter_t&& adapter) : ia(std::move(adapter)), decimal_point_char(get_decimal_point()) {} // delete because of pointer members lexer(const lexer&) = delete; + lexer(lexer&&) = delete; lexer& operator=(lexer&) = delete; + lexer& operator=(lexer&&) = delete; + ~lexer() = default; private: ///////////////////// @@ -1819,9 +2783,10 @@ class lexer @brief scan a string literal This function scans a string according to Sect. 7 of RFC 7159. While - scanning, bytes are escaped and copied into buffer yytext. Then the function - returns successfully, yytext is *not* null-terminated (as it may contain \0 - bytes), and yytext.size() is the number of bytes in the string. + scanning, bytes are escaped and copied into buffer token_buffer. Then the + function returns successfully, token_buffer is *not* null-terminated (as it + may contain \0 bytes), and token_buffer.size() is the number of bytes in the + string. @return token_type::value_string if string could be successfully scanned, token_type::parse_error otherwise @@ -1831,7 +2796,7 @@ class lexer */ token_type scan_string() { - // reset yytext (ignore opening quote) + // reset token_buffer (ignore opening quote) reset(); // we entered the function by reading an open quote @@ -1999,39 +2964,194 @@ class lexer // invalid control characters case 0x00: + { + error_message = "invalid string: control character U+0000 (NUL) must be escaped to \\u0000"; + return token_type::parse_error; + } + case 0x01: + { + error_message = "invalid string: control character U+0001 (SOH) must be escaped to \\u0001"; + return token_type::parse_error; + } + case 0x02: + { + error_message = "invalid string: control character U+0002 (STX) must be escaped to \\u0002"; + return token_type::parse_error; + } + case 0x03: + { + error_message = "invalid string: control character U+0003 (ETX) must be escaped to \\u0003"; + return token_type::parse_error; + } + case 0x04: + { + error_message = "invalid string: control character U+0004 (EOT) must be escaped to \\u0004"; + return token_type::parse_error; + } + case 0x05: + { + error_message = "invalid string: control character U+0005 (ENQ) must be escaped to \\u0005"; + return token_type::parse_error; + } + case 0x06: + { + error_message = "invalid string: control character U+0006 (ACK) must be escaped to \\u0006"; + return token_type::parse_error; + } + case 0x07: + { + error_message = "invalid string: control character U+0007 (BEL) must be escaped to \\u0007"; + return token_type::parse_error; + } + case 0x08: + { + error_message = "invalid string: control character U+0008 (BS) must be escaped to \\u0008 or \\b"; + return token_type::parse_error; + } + case 0x09: + { + error_message = "invalid string: control character U+0009 (HT) must be escaped to \\u0009 or \\t"; + return token_type::parse_error; + } + case 0x0A: + { + error_message = "invalid string: control character U+000A (LF) must be escaped to \\u000A or \\n"; + return token_type::parse_error; + } + case 0x0B: + { + error_message = "invalid string: control character U+000B (VT) must be escaped to \\u000B"; + return token_type::parse_error; + } + case 0x0C: + { + error_message = "invalid string: control character U+000C (FF) must be escaped to \\u000C or \\f"; + return token_type::parse_error; + } + case 0x0D: + { + error_message = "invalid string: control character U+000D (CR) must be escaped to \\u000D or \\r"; + return token_type::parse_error; + } + case 0x0E: + { + error_message = "invalid string: control character U+000E (SO) must be escaped to \\u000E"; + return token_type::parse_error; + } + case 0x0F: + { + error_message = "invalid string: control character U+000F (SI) must be escaped to \\u000F"; + return token_type::parse_error; + } + case 0x10: + { + error_message = "invalid string: control character U+0010 (DLE) must be escaped to \\u0010"; + return token_type::parse_error; + } + case 0x11: + { + error_message = "invalid string: control character U+0011 (DC1) must be escaped to \\u0011"; + return token_type::parse_error; + } + case 0x12: + { + error_message = "invalid string: control character U+0012 (DC2) must be escaped to \\u0012"; + return token_type::parse_error; + } + case 0x13: + { + error_message = "invalid string: control character U+0013 (DC3) must be escaped to \\u0013"; + return token_type::parse_error; + } + case 0x14: + { + error_message = "invalid string: control character U+0014 (DC4) must be escaped to \\u0014"; + return token_type::parse_error; + } + case 0x15: + { + error_message = "invalid string: control character U+0015 (NAK) must be escaped to \\u0015"; + return token_type::parse_error; + } + case 0x16: + { + error_message = "invalid string: control character U+0016 (SYN) must be escaped to \\u0016"; + return token_type::parse_error; + } + case 0x17: + { + error_message = "invalid string: control character U+0017 (ETB) must be escaped to \\u0017"; + return token_type::parse_error; + } + case 0x18: + { + error_message = "invalid string: control character U+0018 (CAN) must be escaped to \\u0018"; + return token_type::parse_error; + } + case 0x19: + { + error_message = "invalid string: control character U+0019 (EM) must be escaped to \\u0019"; + return token_type::parse_error; + } + case 0x1A: + { + error_message = "invalid string: control character U+001A (SUB) must be escaped to \\u001A"; + return token_type::parse_error; + } + case 0x1B: + { + error_message = "invalid string: control character U+001B (ESC) must be escaped to \\u001B"; + return token_type::parse_error; + } + case 0x1C: + { + error_message = "invalid string: control character U+001C (FS) must be escaped to \\u001C"; + return token_type::parse_error; + } + case 0x1D: + { + error_message = "invalid string: control character U+001D (GS) must be escaped to \\u001D"; + return token_type::parse_error; + } + case 0x1E: + { + error_message = "invalid string: control character U+001E (RS) must be escaped to \\u001E"; + return token_type::parse_error; + } + case 0x1F: { - error_message = "invalid string: control character must be escaped"; + error_message = "invalid string: control character U+001F (US) must be escaped to \\u001F"; return token_type::parse_error; } @@ -2303,7 +3423,7 @@ class lexer contains cycles, but any cycle can be left when EOF is read. Therefore, the function is guaranteed to terminate. - During scanning, the read bytes are stored in yytext. This string is + During scanning, the read bytes are stored in token_buffer. This string is then converted to a signed integer, an unsigned integer, or a floating-point number. @@ -2315,9 +3435,9 @@ class lexer locale's decimal point is used instead of `.` to work with the locale-dependent converters. */ - token_type scan_number() + token_type scan_number() // lgtm [cpp/use-of-goto] { - // reset yytext to store the number's bytes + // reset token_buffer to store the number's bytes reset(); // the type of the parsed number; initially set to unsigned; will be @@ -2353,11 +3473,13 @@ class lexer goto scan_number_any1; } + // LCOV_EXCL_START default: { // all other characters are rejected outside scan_number() - assert(false); // LCOV_EXCL_LINE + assert(false); } + // LCOV_EXCL_STOP } scan_number_minus: @@ -2601,10 +3723,10 @@ scan_number_done: // try to parse integers first and fall back to floats if (number_type == token_type::value_unsigned) { - const auto x = std::strtoull(yytext.data(), &endptr, 10); + const auto x = std::strtoull(token_buffer.data(), &endptr, 10); // we checked the number format before - assert(endptr == yytext.data() + yytext.size()); + assert(endptr == token_buffer.data() + token_buffer.size()); if (errno == 0) { @@ -2617,10 +3739,10 @@ scan_number_done: } else if (number_type == token_type::value_integer) { - const auto x = std::strtoll(yytext.data(), &endptr, 10); + const auto x = std::strtoll(token_buffer.data(), &endptr, 10); // we checked the number format before - assert(endptr == yytext.data() + yytext.size()); + assert(endptr == token_buffer.data() + token_buffer.size()); if (errno == 0) { @@ -2634,10 +3756,10 @@ scan_number_done: // this code is reached if we parse a floating-point number or if an // integer conversion above failed - strtof(value_float, yytext.data(), &endptr); + strtof(value_float, token_buffer.data(), &endptr); // we checked the number format before - assert(endptr == yytext.data() + yytext.size()); + assert(endptr == token_buffer.data() + token_buffer.size()); return token_type::value_float; } @@ -2666,10 +3788,10 @@ scan_number_done: // input management ///////////////////// - /// reset yytext; current character is beginning of token + /// reset token_buffer; current character is beginning of token void reset() noexcept { - yytext.clear(); + token_buffer.clear(); token_string.clear(); token_string.push_back(std::char_traits<char>::to_char_type(current)); } @@ -2686,31 +3808,71 @@ scan_number_done: */ std::char_traits<char>::int_type get() { - ++chars_read; - current = ia->get_character(); + ++position.chars_read_total; + ++position.chars_read_current_line; + + if (next_unget) + { + // just reset the next_unget variable and work with current + next_unget = false; + } + else + { + current = ia->get_character(); + } + if (JSON_LIKELY(current != std::char_traits<char>::eof())) { token_string.push_back(std::char_traits<char>::to_char_type(current)); } + + if (current == '\n') + { + ++position.lines_read; + ++position.chars_read_current_line = 0; + } + return current; } - /// unget current character (return it again on next get) + /*! + @brief unget current character (read it again on next get) + + We implement unget by setting variable next_unget to true. The input is not + changed - we just simulate ungetting by modifying chars_read_total, + chars_read_current_line, and token_string. The next call to get() will + behave as if the unget character is read again. + */ void unget() { - --chars_read; + next_unget = true; + + --position.chars_read_total; + + // in case we "unget" a newline, we have to also decrement the lines_read + if (position.chars_read_current_line == 0) + { + if (position.lines_read > 0) + { + --position.lines_read; + } + } + else + { + --position.chars_read_current_line; + } + if (JSON_LIKELY(current != std::char_traits<char>::eof())) { - ia->unget_character(); assert(token_string.size() != 0); token_string.pop_back(); } } - /// add a character to yytext + /// add a character to token_buffer void add(int c) { - yytext.push_back(std::char_traits<char>::to_char_type(c)); + token_buffer.push_back(std::char_traits<char>::to_char_type(c)); } public: @@ -2737,9 +3899,9 @@ scan_number_done: } /// return current string value (implicitly resets the token; useful only once) - std::string move_string() + string_t& get_string() { - return std::move(yytext); + return token_buffer; } ///////////////////// @@ -2747,9 +3909,9 @@ scan_number_done: ///////////////////// /// return position of last read token - constexpr std::size_t get_position() const noexcept + constexpr position_t get_position() const noexcept { - return chars_read; + return position; } /// return the last read token (for errors only). Will never contain EOF @@ -2764,10 +3926,9 @@ scan_number_done: if ('\x00' <= c and c <= '\x1F') { // escape control characters - std::stringstream ss; - ss << "<U+" << std::setw(4) << std::uppercase << std::setfill('0') - << std::hex << static_cast<int>(c) << ">"; - result += ss.str(); + char cs[9]; + (std::snprintf)(cs, 9, "<U+%.4X>", static_cast<unsigned char>(c)); + result += cs; } else { @@ -2789,8 +3950,33 @@ scan_number_done: // actual scanner ///////////////////// + /*! + @brief skip the UTF-8 byte order mark + @return true iff there is no BOM or the correct BOM has been skipped + */ + bool skip_bom() + { + if (get() == 0xEF) + { + // check if we completely parse the BOM + return get() == 0xBB and get() == 0xBF; + } + + // the first character is not the beginning of the BOM; unget it to + // process is later + unget(); + return true; + } + token_type scan() { + // initially, skip the BOM + if (position.chars_read_total == 0 and not skip_bom()) + { + error_message = "invalid BOM; must be 0xEF 0xBB 0xBF if given"; + return token_type::parse_error; + } + // read next character and ignore whitespace do { @@ -2860,14 +4046,17 @@ scan_number_done: /// the current character std::char_traits<char>::int_type current = std::char_traits<char>::eof(); - /// the number of characters read - std::size_t chars_read = 0; + /// whether the next get() call should just return current + bool next_unget = false; + + /// the start position of the current token + position_t position; /// raw input token string (for error messages) std::vector<char> token_string {}; /// buffer for variable-length tokens (numbers, strings) - std::string yytext {}; + string_t token_buffer {}; /// a description of occurred lexer errors const char* error_message = ""; @@ -2880,6 +4069,887 @@ scan_number_done: /// the decimal point const char decimal_point_char = '.'; }; +} // namespace detail +} // namespace nlohmann + +// #include <nlohmann/detail/input/parser.hpp> + + +#include <cassert> // assert +#include <cmath> // isfinite +#include <cstdint> // uint8_t +#include <functional> // function +#include <string> // string +#include <utility> // move + +// #include <nlohmann/detail/exceptions.hpp> + +// #include <nlohmann/detail/macro_scope.hpp> + +// #include <nlohmann/detail/meta/is_sax.hpp> + + +#include <cstdint> // size_t +#include <utility> // declval + +// #include <nlohmann/detail/meta/detected.hpp> + +// #include <nlohmann/detail/meta/type_traits.hpp> + + +namespace nlohmann +{ +namespace detail +{ +template <typename T> +using null_function_t = decltype(std::declval<T&>().null()); + +template <typename T> +using boolean_function_t = + decltype(std::declval<T&>().boolean(std::declval<bool>())); + +template <typename T, typename Integer> +using number_integer_function_t = + decltype(std::declval<T&>().number_integer(std::declval<Integer>())); + +template <typename T, typename Unsigned> +using number_unsigned_function_t = + decltype(std::declval<T&>().number_unsigned(std::declval<Unsigned>())); + +template <typename T, typename Float, typename String> +using number_float_function_t = decltype(std::declval<T&>().number_float( + std::declval<Float>(), std::declval<const String&>())); + +template <typename T, typename String> +using string_function_t = + decltype(std::declval<T&>().string(std::declval<String&>())); + +template <typename T> +using start_object_function_t = + decltype(std::declval<T&>().start_object(std::declval<std::size_t>())); + +template <typename T, typename String> +using key_function_t = + decltype(std::declval<T&>().key(std::declval<String&>())); + +template <typename T> +using end_object_function_t = decltype(std::declval<T&>().end_object()); + +template <typename T> +using start_array_function_t = + decltype(std::declval<T&>().start_array(std::declval<std::size_t>())); + +template <typename T> +using end_array_function_t = decltype(std::declval<T&>().end_array()); + +template <typename T, typename Exception> +using parse_error_function_t = decltype(std::declval<T&>().parse_error( + std::declval<std::size_t>(), std::declval<const std::string&>(), + std::declval<const Exception&>())); + +template <typename SAX, typename BasicJsonType> +struct is_sax +{ + private: + static_assert(is_basic_json<BasicJsonType>::value, + "BasicJsonType must be of type basic_json<...>"); + + using number_integer_t = typename BasicJsonType::number_integer_t; + using number_unsigned_t = typename BasicJsonType::number_unsigned_t; + using number_float_t = typename BasicJsonType::number_float_t; + using string_t = typename BasicJsonType::string_t; + using exception_t = typename BasicJsonType::exception; + + public: + static constexpr bool value = + is_detected_exact<bool, null_function_t, SAX>::value && + is_detected_exact<bool, boolean_function_t, SAX>::value && + is_detected_exact<bool, number_integer_function_t, SAX, + number_integer_t>::value && + is_detected_exact<bool, number_unsigned_function_t, SAX, + number_unsigned_t>::value && + is_detected_exact<bool, number_float_function_t, SAX, number_float_t, + string_t>::value && + is_detected_exact<bool, string_function_t, SAX, string_t>::value && + is_detected_exact<bool, start_object_function_t, SAX>::value && + is_detected_exact<bool, key_function_t, SAX, string_t>::value && + is_detected_exact<bool, end_object_function_t, SAX>::value && + is_detected_exact<bool, start_array_function_t, SAX>::value && + is_detected_exact<bool, end_array_function_t, SAX>::value && + is_detected_exact<bool, parse_error_function_t, SAX, exception_t>::value; +}; + +template <typename SAX, typename BasicJsonType> +struct is_sax_static_asserts +{ + private: + static_assert(is_basic_json<BasicJsonType>::value, + "BasicJsonType must be of type basic_json<...>"); + + using number_integer_t = typename BasicJsonType::number_integer_t; + using number_unsigned_t = typename BasicJsonType::number_unsigned_t; + using number_float_t = typename BasicJsonType::number_float_t; + using string_t = typename BasicJsonType::string_t; + using exception_t = typename BasicJsonType::exception; + + public: + static_assert(is_detected_exact<bool, null_function_t, SAX>::value, + "Missing/invalid function: bool null()"); + static_assert(is_detected_exact<bool, boolean_function_t, SAX>::value, + "Missing/invalid function: bool boolean(bool)"); + static_assert(is_detected_exact<bool, boolean_function_t, SAX>::value, + "Missing/invalid function: bool boolean(bool)"); + static_assert( + is_detected_exact<bool, number_integer_function_t, SAX, + number_integer_t>::value, + "Missing/invalid function: bool number_integer(number_integer_t)"); + static_assert( + is_detected_exact<bool, number_unsigned_function_t, SAX, + number_unsigned_t>::value, + "Missing/invalid function: bool number_unsigned(number_unsigned_t)"); + static_assert(is_detected_exact<bool, number_float_function_t, SAX, + number_float_t, string_t>::value, + "Missing/invalid function: bool number_float(number_float_t, const string_t&)"); + static_assert( + is_detected_exact<bool, string_function_t, SAX, string_t>::value, + "Missing/invalid function: bool string(string_t&)"); + static_assert(is_detected_exact<bool, start_object_function_t, SAX>::value, + "Missing/invalid function: bool start_object(std::size_t)"); + static_assert(is_detected_exact<bool, key_function_t, SAX, string_t>::value, + "Missing/invalid function: bool key(string_t&)"); + static_assert(is_detected_exact<bool, end_object_function_t, SAX>::value, + "Missing/invalid function: bool end_object()"); + static_assert(is_detected_exact<bool, start_array_function_t, SAX>::value, + "Missing/invalid function: bool start_array(std::size_t)"); + static_assert(is_detected_exact<bool, end_array_function_t, SAX>::value, + "Missing/invalid function: bool end_array()"); + static_assert( + is_detected_exact<bool, parse_error_function_t, SAX, exception_t>::value, + "Missing/invalid function: bool parse_error(std::size_t, const " + "std::string&, const exception&)"); +}; +} // namespace detail +} // namespace nlohmann + +// #include <nlohmann/detail/input/input_adapters.hpp> + +// #include <nlohmann/detail/input/json_sax.hpp> + + +#include <cstddef> +#include <string> +#include <vector> + +// #include <nlohmann/detail/input/parser.hpp> + +// #include <nlohmann/detail/exceptions.hpp> + + +namespace nlohmann +{ + +/*! +@brief SAX interface + +This class describes the SAX interface used by @ref nlohmann::json::sax_parse. +Each function is called in different situations while the input is parsed. The +boolean return value informs the parser whether to continue processing the +input. +*/ +template<typename BasicJsonType> +struct json_sax +{ + /// type for (signed) integers + using number_integer_t = typename BasicJsonType::number_integer_t; + /// type for unsigned integers + using number_unsigned_t = typename BasicJsonType::number_unsigned_t; + /// type for floating-point numbers + using number_float_t = typename BasicJsonType::number_float_t; + /// type for strings + using string_t = typename BasicJsonType::string_t; + + /*! + @brief a null value was read + @return whether parsing should proceed + */ + virtual bool null() = 0; + + /*! + @brief a boolean value was read + @param[in] val boolean value + @return whether parsing should proceed + */ + virtual bool boolean(bool val) = 0; + + /*! + @brief an integer number was read + @param[in] val integer value + @return whether parsing should proceed + */ + virtual bool number_integer(number_integer_t val) = 0; + + /*! + @brief an unsigned integer number was read + @param[in] val unsigned integer value + @return whether parsing should proceed + */ + virtual bool number_unsigned(number_unsigned_t val) = 0; + + /*! + @brief an floating-point number was read + @param[in] val floating-point value + @param[in] s raw token value + @return whether parsing should proceed + */ + virtual bool number_float(number_float_t val, const string_t& s) = 0; + + /*! + @brief a string was read + @param[in] val string value + @return whether parsing should proceed + @note It is safe to move the passed string. + */ + virtual bool string(string_t& val) = 0; + + /*! + @brief the beginning of an object was read + @param[in] elements number of object elements or -1 if unknown + @return whether parsing should proceed + @note binary formats may report the number of elements + */ + virtual bool start_object(std::size_t elements) = 0; + + /*! + @brief an object key was read + @param[in] val object key + @return whether parsing should proceed + @note It is safe to move the passed string. + */ + virtual bool key(string_t& val) = 0; + + /*! + @brief the end of an object was read + @return whether parsing should proceed + */ + virtual bool end_object() = 0; + + /*! + @brief the beginning of an array was read + @param[in] elements number of array elements or -1 if unknown + @return whether parsing should proceed + @note binary formats may report the number of elements + */ + virtual bool start_array(std::size_t elements) = 0; + + /*! + @brief the end of an array was read + @return whether parsing should proceed + */ + virtual bool end_array() = 0; + + /*! + @brief a parse error occurred + @param[in] position the position in the input where the error occurs + @param[in] last_token the last read token + @param[in] ex an exception object describing the error + @return whether parsing should proceed (must return false) + */ + virtual bool parse_error(std::size_t position, + const std::string& last_token, + const detail::exception& ex) = 0; + + virtual ~json_sax() = default; +}; + + +namespace detail +{ +/*! +@brief SAX implementation to create a JSON value from SAX events + +This class implements the @ref json_sax interface and processes the SAX events +to create a JSON value which makes it basically a DOM parser. The structure or +hierarchy of the JSON value is managed by the stack `ref_stack` which contains +a pointer to the respective array or object for each recursion depth. + +After successful parsing, the value that is passed by reference to the +constructor contains the parsed value. + +@tparam BasicJsonType the JSON type +*/ +template<typename BasicJsonType> +class json_sax_dom_parser +{ + public: + using number_integer_t = typename BasicJsonType::number_integer_t; + using number_unsigned_t = typename BasicJsonType::number_unsigned_t; + using number_float_t = typename BasicJsonType::number_float_t; + using string_t = typename BasicJsonType::string_t; + + /*! + @param[in, out] r reference to a JSON value that is manipulated while + parsing + @param[in] allow_exceptions_ whether parse errors yield exceptions + */ + explicit json_sax_dom_parser(BasicJsonType& r, const bool allow_exceptions_ = true) + : root(r), allow_exceptions(allow_exceptions_) + {} + + bool null() + { + handle_value(nullptr); + return true; + } + + bool boolean(bool val) + { + handle_value(val); + return true; + } + + bool number_integer(number_integer_t val) + { + handle_value(val); + return true; + } + + bool number_unsigned(number_unsigned_t val) + { + handle_value(val); + return true; + } + + bool number_float(number_float_t val, const string_t& /*unused*/) + { + handle_value(val); + return true; + } + + bool string(string_t& val) + { + handle_value(val); + return true; + } + + bool start_object(std::size_t len) + { + ref_stack.push_back(handle_value(BasicJsonType::value_t::object)); + + if (JSON_UNLIKELY(len != std::size_t(-1) and len > ref_stack.back()->max_size())) + { + JSON_THROW(out_of_range::create(408, + "excessive object size: " + std::to_string(len))); + } + + return true; + } + + bool key(string_t& val) + { + // add null at given key and store the reference for later + object_element = &(ref_stack.back()->m_value.object->operator[](val)); + return true; + } + + bool end_object() + { + ref_stack.pop_back(); + return true; + } + + bool start_array(std::size_t len) + { + ref_stack.push_back(handle_value(BasicJsonType::value_t::array)); + + if (JSON_UNLIKELY(len != std::size_t(-1) and len > ref_stack.back()->max_size())) + { + JSON_THROW(out_of_range::create(408, + "excessive array size: " + std::to_string(len))); + } + + return true; + } + + bool end_array() + { + ref_stack.pop_back(); + return true; + } + + bool parse_error(std::size_t /*unused*/, const std::string& /*unused*/, + const detail::exception& ex) + { + errored = true; + if (allow_exceptions) + { + // determine the proper exception type from the id + switch ((ex.id / 100) % 100) + { + case 1: + JSON_THROW(*reinterpret_cast<const detail::parse_error*>(&ex)); + case 4: + JSON_THROW(*reinterpret_cast<const detail::out_of_range*>(&ex)); + // LCOV_EXCL_START + case 2: + JSON_THROW(*reinterpret_cast<const detail::invalid_iterator*>(&ex)); + case 3: + JSON_THROW(*reinterpret_cast<const detail::type_error*>(&ex)); + case 5: + JSON_THROW(*reinterpret_cast<const detail::other_error*>(&ex)); + default: + assert(false); + // LCOV_EXCL_STOP + } + } + return false; + } + + constexpr bool is_errored() const + { + return errored; + } + + private: + /*! + @invariant If the ref stack is empty, then the passed value will be the new + root. + @invariant If the ref stack contains a value, then it is an array or an + object to which we can add elements + */ + template<typename Value> + BasicJsonType* handle_value(Value&& v) + { + if (ref_stack.empty()) + { + root = BasicJsonType(std::forward<Value>(v)); + return &root; + } + + assert(ref_stack.back()->is_array() or ref_stack.back()->is_object()); + + if (ref_stack.back()->is_array()) + { + ref_stack.back()->m_value.array->emplace_back(std::forward<Value>(v)); + return &(ref_stack.back()->m_value.array->back()); + } + else + { + assert(object_element); + *object_element = BasicJsonType(std::forward<Value>(v)); + return object_element; + } + } + + /// the parsed JSON value + BasicJsonType& root; + /// stack to model hierarchy of values + std::vector<BasicJsonType*> ref_stack; + /// helper to hold the reference for the next object element + BasicJsonType* object_element = nullptr; + /// whether a syntax error occurred + bool errored = false; + /// whether to throw exceptions in case of errors + const bool allow_exceptions = true; +}; + +template<typename BasicJsonType> +class json_sax_dom_callback_parser +{ + public: + using number_integer_t = typename BasicJsonType::number_integer_t; + using number_unsigned_t = typename BasicJsonType::number_unsigned_t; + using number_float_t = typename BasicJsonType::number_float_t; + using string_t = typename BasicJsonType::string_t; + using parser_callback_t = typename BasicJsonType::parser_callback_t; + using parse_event_t = typename BasicJsonType::parse_event_t; + + json_sax_dom_callback_parser(BasicJsonType& r, + const parser_callback_t cb, + const bool allow_exceptions_ = true) + : root(r), callback(cb), allow_exceptions(allow_exceptions_) + { + keep_stack.push_back(true); + } + + bool null() + { + handle_value(nullptr); + return true; + } + + bool boolean(bool val) + { + handle_value(val); + return true; + } + + bool number_integer(number_integer_t val) + { + handle_value(val); + return true; + } + + bool number_unsigned(number_unsigned_t val) + { + handle_value(val); + return true; + } + + bool number_float(number_float_t val, const string_t& /*unused*/) + { + handle_value(val); + return true; + } + + bool string(string_t& val) + { + handle_value(val); + return true; + } + + bool start_object(std::size_t len) + { + // check callback for object start + const bool keep = callback(static_cast<int>(ref_stack.size()), parse_event_t::object_start, discarded); + keep_stack.push_back(keep); + + auto val = handle_value(BasicJsonType::value_t::object, true); + ref_stack.push_back(val.second); + + // check object limit + if (ref_stack.back()) + { + if (JSON_UNLIKELY(len != std::size_t(-1) and len > ref_stack.back()->max_size())) + { + JSON_THROW(out_of_range::create(408, + "excessive object size: " + std::to_string(len))); + } + } + + return true; + } + + bool key(string_t& val) + { + BasicJsonType k = BasicJsonType(val); + + // check callback for key + const bool keep = callback(static_cast<int>(ref_stack.size()), parse_event_t::key, k); + key_keep_stack.push_back(keep); + + // add discarded value at given key and store the reference for later + if (keep and ref_stack.back()) + { + object_element = &(ref_stack.back()->m_value.object->operator[](val) = discarded); + } + + return true; + } + + bool end_object() + { + if (ref_stack.back()) + { + if (not callback(static_cast<int>(ref_stack.size()) - 1, parse_event_t::object_end, *ref_stack.back())) + { + // discard object + *ref_stack.back() = discarded; + } + } + + assert(not ref_stack.empty()); + assert(not keep_stack.empty()); + ref_stack.pop_back(); + keep_stack.pop_back(); + + if (not ref_stack.empty() and ref_stack.back()) + { + // remove discarded value + if (ref_stack.back()->is_object()) + { + for (auto it = ref_stack.back()->begin(); it != ref_stack.back()->end(); ++it) + { + if (it->is_discarded()) + { + ref_stack.back()->erase(it); + break; + } + } + } + } + + return true; + } + + bool start_array(std::size_t len) + { + const bool keep = callback(static_cast<int>(ref_stack.size()), parse_event_t::array_start, discarded); + keep_stack.push_back(keep); + + auto val = handle_value(BasicJsonType::value_t::array, true); + ref_stack.push_back(val.second); + + // check array limit + if (ref_stack.back()) + { + if (JSON_UNLIKELY(len != std::size_t(-1) and len > ref_stack.back()->max_size())) + { + JSON_THROW(out_of_range::create(408, + "excessive array size: " + std::to_string(len))); + } + } + + return true; + } + + bool end_array() + { + bool keep = true; + + if (ref_stack.back()) + { + keep = callback(static_cast<int>(ref_stack.size()) - 1, parse_event_t::array_end, *ref_stack.back()); + if (not keep) + { + // discard array + *ref_stack.back() = discarded; + } + } + + assert(not ref_stack.empty()); + assert(not keep_stack.empty()); + ref_stack.pop_back(); + keep_stack.pop_back(); + + // remove discarded value + if (not keep and not ref_stack.empty()) + { + if (ref_stack.back()->is_array()) + { + ref_stack.back()->m_value.array->pop_back(); + } + } + + return true; + } + + bool parse_error(std::size_t /*unused*/, const std::string& /*unused*/, + const detail::exception& ex) + { + errored = true; + if (allow_exceptions) + { + // determine the proper exception type from the id + switch ((ex.id / 100) % 100) + { + case 1: + JSON_THROW(*reinterpret_cast<const detail::parse_error*>(&ex)); + case 4: + JSON_THROW(*reinterpret_cast<const detail::out_of_range*>(&ex)); + // LCOV_EXCL_START + case 2: + JSON_THROW(*reinterpret_cast<const detail::invalid_iterator*>(&ex)); + case 3: + JSON_THROW(*reinterpret_cast<const detail::type_error*>(&ex)); + case 5: + JSON_THROW(*reinterpret_cast<const detail::other_error*>(&ex)); + default: + assert(false); + // LCOV_EXCL_STOP + } + } + return false; + } + + constexpr bool is_errored() const + { + return errored; + } + + private: + /*! + @param[in] v value to add to the JSON value we build during parsing + @param[in] skip_callback whether we should skip calling the callback + function; this is required after start_array() and + start_object() SAX events, because otherwise we would call the + callback function with an empty array or object, respectively. + + @invariant If the ref stack is empty, then the passed value will be the new + root. + @invariant If the ref stack contains a value, then it is an array or an + object to which we can add elements + + @return pair of boolean (whether value should be kept) and pointer (to the + passed value in the ref_stack hierarchy; nullptr if not kept) + */ + template<typename Value> + std::pair<bool, BasicJsonType*> handle_value(Value&& v, const bool skip_callback = false) + { + assert(not keep_stack.empty()); + + // do not handle this value if we know it would be added to a discarded + // container + if (not keep_stack.back()) + { + return {false, nullptr}; + } + + // create value + auto value = BasicJsonType(std::forward<Value>(v)); + + // check callback + const bool keep = skip_callback or callback(static_cast<int>(ref_stack.size()), parse_event_t::value, value); + + // do not handle this value if we just learnt it shall be discarded + if (not keep) + { + return {false, nullptr}; + } + + if (ref_stack.empty()) + { + root = std::move(value); + return {true, &root}; + } + + // skip this value if we already decided to skip the parent + // (https://github.com/nlohmann/json/issues/971#issuecomment-413678360) + if (not ref_stack.back()) + { + return {false, nullptr}; + } + + // we now only expect arrays and objects + assert(ref_stack.back()->is_array() or ref_stack.back()->is_object()); + + if (ref_stack.back()->is_array()) + { + ref_stack.back()->m_value.array->push_back(std::move(value)); + return {true, &(ref_stack.back()->m_value.array->back())}; + } + else + { + // check if we should store an element for the current key + assert(not key_keep_stack.empty()); + const bool store_element = key_keep_stack.back(); + key_keep_stack.pop_back(); + + if (not store_element) + { + return {false, nullptr}; + } + + assert(object_element); + *object_element = std::move(value); + return {true, object_element}; + } + } + + /// the parsed JSON value + BasicJsonType& root; + /// stack to model hierarchy of values + std::vector<BasicJsonType*> ref_stack; + /// stack to manage which values to keep + std::vector<bool> keep_stack; + /// stack to manage which object keys to keep + std::vector<bool> key_keep_stack; + /// helper to hold the reference for the next object element + BasicJsonType* object_element = nullptr; + /// whether a syntax error occurred + bool errored = false; + /// callback function + const parser_callback_t callback = nullptr; + /// whether to throw exceptions in case of errors + const bool allow_exceptions = true; + /// a discarded value for the callback + BasicJsonType discarded = BasicJsonType::value_t::discarded; +}; + +template<typename BasicJsonType> +class json_sax_acceptor +{ + public: + using number_integer_t = typename BasicJsonType::number_integer_t; + using number_unsigned_t = typename BasicJsonType::number_unsigned_t; + using number_float_t = typename BasicJsonType::number_float_t; + using string_t = typename BasicJsonType::string_t; + + bool null() + { + return true; + } + + bool boolean(bool /*unused*/) + { + return true; + } + + bool number_integer(number_integer_t /*unused*/) + { + return true; + } + + bool number_unsigned(number_unsigned_t /*unused*/) + { + return true; + } + + bool number_float(number_float_t /*unused*/, const string_t& /*unused*/) + { + return true; + } + + bool string(string_t& /*unused*/) + { + return true; + } + + bool start_object(std::size_t /*unused*/ = std::size_t(-1)) + { + return true; + } + + bool key(string_t& /*unused*/) + { + return true; + } + + bool end_object() + { + return true; + } + + bool start_array(std::size_t /*unused*/ = std::size_t(-1)) + { + return true; + } + + bool end_array() + { + return true; + } + + bool parse_error(std::size_t /*unused*/, const std::string& /*unused*/, const detail::exception& /*unused*/) + { + return false; + } +}; +} // namespace detail + +} // namespace nlohmann + +// #include <nlohmann/detail/input/lexer.hpp> + +// #include <nlohmann/detail/value_t.hpp> + + +namespace nlohmann +{ +namespace detail +{ +//////////// +// parser // +//////////// /*! @brief syntax analysis @@ -2892,6 +4962,7 @@ class parser using number_integer_t = typename BasicJsonType::number_integer_t; using number_unsigned_t = typename BasicJsonType::number_unsigned_t; using number_float_t = typename BasicJsonType::number_float_t; + using string_t = typename BasicJsonType::string_t; using lexer_t = lexer<BasicJsonType>; using token_type = typename lexer_t::token_type; @@ -2916,11 +4987,14 @@ class parser std::function<bool(int depth, parse_event_t event, BasicJsonType& parsed)>; /// a parser reading from an input adapter - explicit parser(detail::input_adapter_t adapter, + explicit parser(detail::input_adapter_t&& adapter, const parser_callback_t cb = nullptr, const bool allow_exceptions_ = true) - : callback(cb), m_lexer(adapter), allow_exceptions(allow_exceptions_) - {} + : callback(cb), m_lexer(std::move(adapter)), allow_exceptions(allow_exceptions_) + { + // read first token + get_token(); + } /*! @brief public parser interface @@ -2934,31 +5008,56 @@ class parser */ void parse(const bool strict, BasicJsonType& result) { - // read first token - get_token(); + if (callback) + { + json_sax_dom_callback_parser<BasicJsonType> sdp(result, callback, allow_exceptions); + sax_parse_internal(&sdp); + result.assert_invariant(); - parse_internal(true, result); - result.assert_invariant(); + // in strict mode, input must be completely read + if (strict and (get_token() != token_type::end_of_input)) + { + sdp.parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), + exception_message(token_type::end_of_input, "value"))); + } - // in strict mode, input must be completely read - if (strict) - { - get_token(); - expect(token_type::end_of_input); - } + // in case of an error, return discarded value + if (sdp.is_errored()) + { + result = value_t::discarded; + return; + } - // in case of an error, return discarded value - if (errored) - { - result = value_t::discarded; - return; + // set top-level value to null if it was discarded by the callback + // function + if (result.is_discarded()) + { + result = nullptr; + } } - - // set top-level value to null if it was discarded by the callback - // function - if (result.is_discarded()) + else { - result = nullptr; + json_sax_dom_parser<BasicJsonType> sdp(result, allow_exceptions); + sax_parse_internal(&sdp); + result.assert_invariant(); + + // in strict mode, input must be completely read + if (strict and (get_token() != token_type::end_of_input)) + { + sdp.parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), + exception_message(token_type::end_of_input, "value"))); + } + + // in case of an error, return discarded value + if (sdp.is_errored()) + { + result = value_t::discarded; + return; + } } } @@ -2970,413 +5069,317 @@ class parser */ bool accept(const bool strict = true) { - // read first token - get_token(); + json_sax_acceptor<BasicJsonType> sax_acceptor; + return sax_parse(&sax_acceptor, strict); + } - if (not accept_internal()) + template <typename SAX> + bool sax_parse(SAX* sax, const bool strict = true) + { + (void)detail::is_sax_static_asserts<SAX, BasicJsonType> {}; + const bool result = sax_parse_internal(sax); + + // strict mode: next byte must be EOF + if (result and strict and (get_token() != token_type::end_of_input)) { - return false; + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), + exception_message(token_type::end_of_input, "value"))); } - // strict => last token must be EOF - return not strict or (get_token() == token_type::end_of_input); + return result; } private: - /*! - @brief the actual parser - @throw parse_error.101 in case of an unexpected token - @throw parse_error.102 if to_unicode fails or surrogate error - @throw parse_error.103 if to_unicode fails - */ - void parse_internal(bool keep, BasicJsonType& result) + template <typename SAX> + bool sax_parse_internal(SAX* sax) { - // never parse after a parse error was detected - assert(not errored); + // stack to remember the hierarchy of structured values we are parsing + // true = array; false = object + std::vector<bool> states; + // value to avoid a goto (see comment where set to true) + bool skip_to_state_evaluation = false; - // start with a discarded value - if (not result.is_discarded()) - { - result.m_value.destroy(result.m_type); - result.m_type = value_t::discarded; - } - - switch (last_token) + while (true) { - case token_type::begin_object: + if (not skip_to_state_evaluation) { - if (keep) + // invariant: get_token() was called before each iteration + switch (last_token) { - if (callback) + case token_type::begin_object: { - keep = callback(depth++, parse_event_t::object_start, result); - } + if (JSON_UNLIKELY(not sax->start_object(std::size_t(-1)))) + { + return false; + } - if (not callback or keep) - { - // explicitly set result to object to cope with {} - result.m_type = value_t::object; - result.m_value = value_t::object; - } - } + // closing } -> we are done + if (get_token() == token_type::end_object) + { + if (JSON_UNLIKELY(not sax->end_object())) + { + return false; + } + break; + } - // read next token - get_token(); + // parse key + if (JSON_UNLIKELY(last_token != token_type::value_string)) + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), + exception_message(token_type::value_string, "object key"))); + } + if (JSON_UNLIKELY(not sax->key(m_lexer.get_string()))) + { + return false; + } - // closing } -> we are done - if (last_token == token_type::end_object) - { - if (keep and callback and not callback(--depth, parse_event_t::object_end, result)) - { - result.m_value.destroy(result.m_type); - result.m_type = value_t::discarded; - } - break; - } + // parse separator (:) + if (JSON_UNLIKELY(get_token() != token_type::name_separator)) + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), + exception_message(token_type::name_separator, "object separator"))); + } - // parse values - std::string key; - BasicJsonType value; - while (true) - { - // store key - if (not expect(token_type::value_string)) - { - return; + // remember we are now inside an object + states.push_back(false); + + // parse values + get_token(); + continue; } - key = m_lexer.move_string(); - bool keep_tag = false; - if (keep) + case token_type::begin_array: { - if (callback) + if (JSON_UNLIKELY(not sax->start_array(std::size_t(-1)))) { - BasicJsonType k(key); - keep_tag = callback(depth, parse_event_t::key, k); + return false; } - else + + // closing ] -> we are done + if (get_token() == token_type::end_array) { - keep_tag = true; + if (JSON_UNLIKELY(not sax->end_array())) + { + return false; + } + break; } - } - - // parse separator (:) - get_token(); - if (not expect(token_type::name_separator)) - { - return; - } - // parse and add value - get_token(); - value.m_value.destroy(value.m_type); - value.m_type = value_t::discarded; - parse_internal(keep, value); + // remember we are now inside an array + states.push_back(true); - if (JSON_UNLIKELY(errored)) - { - return; + // parse values (no need to call get_token) + continue; } - if (keep and keep_tag and not value.is_discarded()) + case token_type::value_float: { - result.m_value.object->emplace(std::move(key), std::move(value)); - } + const auto res = m_lexer.get_number_float(); - // comma -> next value - get_token(); - if (last_token == token_type::value_separator) - { - get_token(); - continue; + if (JSON_UNLIKELY(not std::isfinite(res))) + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + out_of_range::create(406, "number overflow parsing '" + m_lexer.get_token_string() + "'")); + } + else + { + if (JSON_UNLIKELY(not sax->number_float(res, m_lexer.get_string()))) + { + return false; + } + break; + } } - // closing } - if (not expect(token_type::end_object)) + case token_type::literal_false: { - return; + if (JSON_UNLIKELY(not sax->boolean(false))) + { + return false; + } + break; } - break; - } - if (keep and callback and not callback(--depth, parse_event_t::object_end, result)) - { - result.m_value.destroy(result.m_type); - result.m_type = value_t::discarded; - } - break; - } - - case token_type::begin_array: - { - if (keep) - { - if (callback) + case token_type::literal_null: { - keep = callback(depth++, parse_event_t::array_start, result); + if (JSON_UNLIKELY(not sax->null())) + { + return false; + } + break; } - if (not callback or keep) + case token_type::literal_true: { - // explicitly set result to array to cope with [] - result.m_type = value_t::array; - result.m_value = value_t::array; + if (JSON_UNLIKELY(not sax->boolean(true))) + { + return false; + } + break; } - } - - // read next token - get_token(); - // closing ] -> we are done - if (last_token == token_type::end_array) - { - if (callback and not callback(--depth, parse_event_t::array_end, result)) + case token_type::value_integer: { - result.m_value.destroy(result.m_type); - result.m_type = value_t::discarded; + if (JSON_UNLIKELY(not sax->number_integer(m_lexer.get_number_integer()))) + { + return false; + } + break; } - break; - } - // parse values - BasicJsonType value; - while (true) - { - // parse value - value.m_value.destroy(value.m_type); - value.m_type = value_t::discarded; - parse_internal(keep, value); - - if (JSON_UNLIKELY(errored)) + case token_type::value_string: { - return; + if (JSON_UNLIKELY(not sax->string(m_lexer.get_string()))) + { + return false; + } + break; } - if (keep and not value.is_discarded()) + case token_type::value_unsigned: { - result.m_value.array->push_back(std::move(value)); + if (JSON_UNLIKELY(not sax->number_unsigned(m_lexer.get_number_unsigned()))) + { + return false; + } + break; } - // comma -> next value - get_token(); - if (last_token == token_type::value_separator) + case token_type::parse_error: { - get_token(); - continue; + // using "uninitialized" to avoid "expected" message + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), + exception_message(token_type::uninitialized, "value"))); } - // closing ] - if (not expect(token_type::end_array)) + default: // the last token was unexpected { - return; + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), + exception_message(token_type::literal_or_value, "value"))); } - break; - } - - if (keep and callback and not callback(--depth, parse_event_t::array_end, result)) - { - result.m_value.destroy(result.m_type); - result.m_type = value_t::discarded; } - break; - } - - case token_type::literal_null: - { - result.m_type = value_t::null; - break; - } - - case token_type::value_string: - { - result.m_type = value_t::string; - result.m_value = m_lexer.move_string(); - break; - } - - case token_type::literal_true: - { - result.m_type = value_t::boolean; - result.m_value = true; - break; - } - - case token_type::literal_false: - { - result.m_type = value_t::boolean; - result.m_value = false; - break; } - - case token_type::value_unsigned: + else { - result.m_type = value_t::number_unsigned; - result.m_value = m_lexer.get_number_unsigned(); - break; + skip_to_state_evaluation = false; } - case token_type::value_integer: + // we reached this line after we successfully parsed a value + if (states.empty()) { - result.m_type = value_t::number_integer; - result.m_value = m_lexer.get_number_integer(); - break; + // empty stack: we reached the end of the hierarchy: done + return true; } - - case token_type::value_float: + else { - result.m_type = value_t::number_float; - result.m_value = m_lexer.get_number_float(); - - // throw in case of infinity or NAN - if (JSON_UNLIKELY(not std::isfinite(result.m_value.number_float))) + if (states.back()) // array { - if (allow_exceptions) + // comma -> next value + if (get_token() == token_type::value_separator) { - JSON_THROW(out_of_range::create(406, "number overflow parsing '" + - m_lexer.get_token_string() + "'")); + // parse a new value + get_token(); + continue; } - expect(token_type::uninitialized); - } - break; - } - - case token_type::parse_error: - { - // using "uninitialized" to avoid "expected" message - if (not expect(token_type::uninitialized)) - { - return; - } - break; // LCOV_EXCL_LINE - } - - default: - { - // the last token was unexpected; we expected a value - if (not expect(token_type::literal_or_value)) - { - return; - } - break; // LCOV_EXCL_LINE - } - } - - if (keep and callback and not callback(depth, parse_event_t::value, result)) - { - result.m_type = value_t::discarded; - } - } - - /*! - @brief the actual acceptor - - @invariant 1. The last token is not yet processed. Therefore, the caller - of this function must make sure a token has been read. - 2. When this function returns, the last token is processed. - That is, the last read character was already considered. - - This invariant makes sure that no token needs to be "unput". - */ - bool accept_internal() - { - switch (last_token) - { - case token_type::begin_object: - { - // read next token - get_token(); - // closing } -> we are done - if (last_token == token_type::end_object) - { - return true; - } - - // parse values - while (true) - { - // parse key - if (last_token != token_type::value_string) + // closing ] + if (JSON_LIKELY(last_token == token_type::end_array)) { - return false; - } + if (JSON_UNLIKELY(not sax->end_array())) + { + return false; + } - // parse separator (:) - get_token(); - if (last_token != token_type::name_separator) - { - return false; + // We are done with this array. Before we can parse a + // new value, we need to evaluate the new state first. + // By setting skip_to_state_evaluation to false, we + // are effectively jumping to the beginning of this if. + assert(not states.empty()); + states.pop_back(); + skip_to_state_evaluation = true; + continue; } - - // parse value - get_token(); - if (not accept_internal()) + else { - return false; + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), + exception_message(token_type::end_array, "array"))); } - + } + else // object + { // comma -> next value - get_token(); - if (last_token == token_type::value_separator) + if (get_token() == token_type::value_separator) { + // parse key + if (JSON_UNLIKELY(get_token() != token_type::value_string)) + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), + exception_message(token_type::value_string, "object key"))); + } + else + { + if (JSON_UNLIKELY(not sax->key(m_lexer.get_string()))) + { + return false; + } + } + + // parse separator (:) + if (JSON_UNLIKELY(get_token() != token_type::name_separator)) + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), + exception_message(token_type::name_separator, "object separator"))); + } + + // parse values get_token(); continue; } // closing } - return (last_token == token_type::end_object); - } - } - - case token_type::begin_array: - { - // read next token - get_token(); - - // closing ] -> we are done - if (last_token == token_type::end_array) - { - return true; - } - - // parse values - while (true) - { - // parse value - if (not accept_internal()) + if (JSON_LIKELY(last_token == token_type::end_object)) { - return false; - } + if (JSON_UNLIKELY(not sax->end_object())) + { + return false; + } - // comma -> next value - get_token(); - if (last_token == token_type::value_separator) - { - get_token(); + // We are done with this object. Before we can parse a + // new value, we need to evaluate the new state first. + // By setting skip_to_state_evaluation to false, we + // are effectively jumping to the beginning of this if. + assert(not states.empty()); + states.pop_back(); + skip_to_state_evaluation = true; continue; } - - // closing ] - return (last_token == token_type::end_array); + else + { + return sax->parse_error(m_lexer.get_position(), + m_lexer.get_token_string(), + parse_error::create(101, m_lexer.get_position(), + exception_message(token_type::end_object, "object"))); + } } } - - case token_type::value_float: - { - // reject infinity or NAN - return std::isfinite(m_lexer.get_number_float()); - } - - case token_type::literal_false: - case token_type::literal_null: - case token_type::literal_true: - case token_type::value_integer: - case token_type::value_string: - case token_type::value_unsigned: - return true; - - default: // the last token was unexpected - return false; } } @@ -3386,31 +5389,17 @@ class parser return (last_token = m_lexer.scan()); } - /*! - @throw parse_error.101 if expected token did not occur - */ - bool expect(token_type t) + std::string exception_message(const token_type expected, const std::string& context) { - if (JSON_UNLIKELY(t != last_token)) + std::string error_msg = "syntax error "; + + if (not context.empty()) { - errored = true; - expected = t; - if (allow_exceptions) - { - throw_exception(); - } - else - { - return false; - } + error_msg += "while parsing " + context + " "; } - return true; - } + error_msg += "- "; - [[noreturn]] void throw_exception() const - { - std::string error_msg = "syntax error - "; if (last_token == token_type::parse_error) { error_msg += std::string(m_lexer.get_error_message()) + "; last read: '" + @@ -3426,31 +5415,33 @@ class parser error_msg += "; expected " + std::string(lexer_t::token_type_name(expected)); } - JSON_THROW(parse_error::create(101, m_lexer.get_position(), error_msg)); + return error_msg; } private: - /// current level of recursion - int depth = 0; /// callback function const parser_callback_t callback = nullptr; /// the type of the last read token token_type last_token = token_type::uninitialized; /// the lexer lexer_t m_lexer; - /// whether a syntax error occurred - bool errored = false; - /// possible reason for the syntax error - token_type expected = token_type::uninitialized; /// whether to throw exceptions in case of errors const bool allow_exceptions = true; }; +} // namespace detail +} // namespace nlohmann -/////////////// -// iterators // -/////////////// +// #include <nlohmann/detail/iterators/primitive_iterator.hpp> -/*! + +#include <cstddef> // ptrdiff_t +#include <limits> // numeric_limits + +namespace nlohmann +{ +namespace detail +{ +/* @brief an iterator for primitive JSON types This class models an iterator for primitive JSON types (boolean, number, @@ -3461,9 +5452,15 @@ end_value (`1`) models past the end. */ class primitive_iterator_t { - public: + private: using difference_type = std::ptrdiff_t; + static constexpr difference_type begin_value = 0; + static constexpr difference_type end_value = begin_value + 1; + /// iterator as signed integer type + difference_type m_it = (std::numeric_limits<std::ptrdiff_t>::min)(); + + public: constexpr difference_type get_value() const noexcept { return m_it; @@ -3503,10 +5500,10 @@ class primitive_iterator_t return lhs.m_it < rhs.m_it; } - primitive_iterator_t operator+(difference_type i) + primitive_iterator_t operator+(difference_type n) noexcept { auto result = *this; - result += i; + result += n; return result; } @@ -3515,57 +5512,57 @@ class primitive_iterator_t return lhs.m_it - rhs.m_it; } - friend std::ostream& operator<<(std::ostream& os, primitive_iterator_t it) - { - return os << it.m_it; - } - - primitive_iterator_t& operator++() + primitive_iterator_t& operator++() noexcept { ++m_it; return *this; } - primitive_iterator_t const operator++(int) + primitive_iterator_t const operator++(int) noexcept { auto result = *this; - m_it++; + ++m_it; return result; } - primitive_iterator_t& operator--() + primitive_iterator_t& operator--() noexcept { --m_it; return *this; } - primitive_iterator_t const operator--(int) + primitive_iterator_t const operator--(int) noexcept { auto result = *this; - m_it--; + --m_it; return result; } - primitive_iterator_t& operator+=(difference_type n) + primitive_iterator_t& operator+=(difference_type n) noexcept { m_it += n; return *this; } - primitive_iterator_t& operator-=(difference_type n) + primitive_iterator_t& operator-=(difference_type n) noexcept { m_it -= n; return *this; } +}; +} // namespace detail +} // namespace nlohmann - private: - static constexpr difference_type begin_value = 0; - static constexpr difference_type end_value = begin_value + 1; +// #include <nlohmann/detail/iterators/internal_iterator.hpp> + + +// #include <nlohmann/detail/iterators/primitive_iterator.hpp> - /// iterator as signed integer type - difference_type m_it = (std::numeric_limits<std::ptrdiff_t>::min)(); -}; +namespace nlohmann +{ +namespace detail +{ /*! @brief an iterator value @@ -3581,26 +5578,50 @@ template<typename BasicJsonType> struct internal_iterator /// generic iterator for all other types primitive_iterator_t primitive_iterator {}; }; +} // namespace detail +} // namespace nlohmann + +// #include <nlohmann/detail/iterators/iter_impl.hpp> + + +#include <ciso646> // not +#include <iterator> // iterator, random_access_iterator_tag, bidirectional_iterator_tag, advance, next +#include <type_traits> // conditional, is_const, remove_const + +// #include <nlohmann/detail/exceptions.hpp> + +// #include <nlohmann/detail/iterators/internal_iterator.hpp> +// #include <nlohmann/detail/iterators/primitive_iterator.hpp> + +// #include <nlohmann/detail/macro_scope.hpp> + +// #include <nlohmann/detail/meta/cpp_future.hpp> + +// #include <nlohmann/detail/value_t.hpp> + + +namespace nlohmann +{ +namespace detail +{ +// forward declare, to be able to friend it later on template<typename IteratorType> class iteration_proxy; +template<typename IteratorType> class iteration_proxy_value; /*! @brief a template for a bidirectional iterator for the @ref basic_json class - This class implements a both iterators (iterator and const_iterator) for the @ref basic_json class. - @note An iterator is called *initialized* when a pointer to a JSON value has been set (e.g., by a constructor or a copy assignment). If the iterator is default-constructed, it is *uninitialized* and most methods are undefined. **The library uses assertions to detect calls on uninitialized iterators.** - @requirement The class satisfies the following concept requirements: - -[BidirectionalIterator](http://en.cppreference.com/w/cpp/concept/BidirectionalIterator): +[BidirectionalIterator](https://en.cppreference.com/w/cpp/named_req/BidirectionalIterator): The iterator that can be moved can be moved in both directions (i.e. incremented and decremented). - @since version 1.0.0, simplified in version 2.0.9, change to bidirectional iterators in version 3.0.0 (see https://github.com/nlohmann/json/issues/593) */ @@ -3611,6 +5632,7 @@ class iter_impl friend iter_impl<typename std::conditional<std::is_const<BasicJsonType>::value, typename std::remove_const<BasicJsonType>::type, const BasicJsonType>::type>; friend BasicJsonType; friend iteration_proxy<iter_impl>; + friend iteration_proxy_value<iter_impl>; using object_t = typename BasicJsonType::object_t; using array_t = typename BasicJsonType::array_t; @@ -4149,7 +6171,7 @@ class iter_impl @brief return the key of an object iterator @pre The iterator is initialized; i.e. `m_object != nullptr`. */ - typename object_t::key_type key() const + const typename object_t::key_type& key() const { assert(m_object != nullptr); @@ -4174,94 +6196,26 @@ class iter_impl /// associated JSON instance pointer m_object = nullptr; /// the actual iterator of the associated instance - internal_iterator<typename std::remove_const<BasicJsonType>::type> m_it = {}; + internal_iterator<typename std::remove_const<BasicJsonType>::type> m_it; }; +} // namespace detail +} // namespace nlohmann +// #include <nlohmann/detail/iterators/iteration_proxy.hpp> -/// proxy class for the iterator_wrapper functions -template<typename IteratorType> class iteration_proxy -{ - private: - /// helper class for iteration - class iteration_proxy_internal - { - private: - /// the iterator - IteratorType anchor; - /// an index for arrays (used to create key names) - std::size_t array_index = 0; - - public: - explicit iteration_proxy_internal(IteratorType it) noexcept : anchor(it) {} - - /// dereference operator (needed for range-based for) - iteration_proxy_internal& operator*() - { - return *this; - } - - /// increment operator (needed for range-based for) - iteration_proxy_internal& operator++() - { - ++anchor; - ++array_index; - - return *this; - } - - /// inequality operator (needed for range-based for) - bool operator!=(const iteration_proxy_internal& o) const noexcept - { - return anchor != o.anchor; - } - - /// return key of the iterator - std::string key() const - { - assert(anchor.m_object != nullptr); - - switch (anchor.m_object->type()) - { - // use integer array index as key - case value_t::array: - return std::to_string(array_index); - - // use key from the object - case value_t::object: - return anchor.key(); - - // use an empty key for all primitive types - default: - return ""; - } - } - - /// return value of the iterator - typename IteratorType::reference value() const - { - return anchor.value(); - } - }; - - /// the container to iterate - typename IteratorType::reference container; +// #include <nlohmann/detail/iterators/json_reverse_iterator.hpp> - public: - /// construct iteration proxy from a container - explicit iteration_proxy(typename IteratorType::reference cont) - : container(cont) {} - /// return iterator begin (needed for range-based for) - iteration_proxy_internal begin() noexcept - { - return iteration_proxy_internal(container.begin()); - } +#include <cstddef> // ptrdiff_t +#include <iterator> // reverse_iterator +#include <utility> // declval - /// return iterator end (needed for range-based for) - iteration_proxy_internal end() noexcept - { - return iteration_proxy_internal(container.end()); - } -}; +namespace nlohmann +{ +namespace detail +{ +////////////////////// +// reverse_iterator // +////////////////////// /*! @brief a template for a reverse iterator class @@ -4272,10 +6226,10 @@ create @ref const_reverse_iterator). @requirement The class satisfies the following concept requirements: - -[BidirectionalIterator](http://en.cppreference.com/w/cpp/concept/BidirectionalIterator): +[BidirectionalIterator](https://en.cppreference.com/w/cpp/named_req/BidirectionalIterator): The iterator that can be moved can be moved in both directions (i.e. incremented and decremented). -- [OutputIterator](http://en.cppreference.com/w/cpp/concept/OutputIterator): +- [OutputIterator](https://en.cppreference.com/w/cpp/named_req/OutputIterator): It is possible to write to the pointed-to element (only if @a Base is @ref iterator). @@ -4292,11 +6246,11 @@ class json_reverse_iterator : public std::reverse_iterator<Base> using reference = typename Base::reference; /// create reverse iterator from iterator - json_reverse_iterator(const typename base_iterator::iterator_type& it) noexcept + explicit json_reverse_iterator(const typename base_iterator::iterator_type& it) noexcept : base_iterator(it) {} /// create reverse iterator from base class - json_reverse_iterator(const base_iterator& it) noexcept : base_iterator(it) {} + explicit json_reverse_iterator(const base_iterator& it) noexcept : base_iterator(it) {} /// post-increment (it++) json_reverse_iterator const operator++(int) @@ -4366,11 +6320,25 @@ class json_reverse_iterator : public std::reverse_iterator<Base> return it.operator * (); } }; +} // namespace detail +} // namespace nlohmann + +// #include <nlohmann/detail/output/output_adapters.hpp> -///////////////////// -// output adapters // -///////////////////// +#include <algorithm> // copy +#include <cstddef> // size_t +#include <ios> // streamsize +#include <iterator> // back_inserter +#include <memory> // shared_ptr, make_shared +#include <ostream> // basic_ostream +#include <string> // basic_string +#include <vector> // vector + +namespace nlohmann +{ +namespace detail +{ /// abstract output adapter interface template<typename CharType> struct output_adapter_protocol { @@ -4388,7 +6356,9 @@ template<typename CharType> class output_vector_adapter : public output_adapter_protocol<CharType> { public: - explicit output_vector_adapter(std::vector<CharType>& vec) : v(vec) {} + explicit output_vector_adapter(std::vector<CharType>& vec) noexcept + : v(vec) + {} void write_character(CharType c) override { @@ -4409,7 +6379,9 @@ template<typename CharType> class output_stream_adapter : public output_adapter_protocol<CharType> { public: - explicit output_stream_adapter(std::basic_ostream<CharType>& s) : stream(s) {} + explicit output_stream_adapter(std::basic_ostream<CharType>& s) noexcept + : stream(s) + {} void write_character(CharType c) override { @@ -4426,11 +6398,13 @@ class output_stream_adapter : public output_adapter_protocol<CharType> }; /// output adapter for basic_string -template<typename CharType> +template<typename CharType, typename StringType = std::basic_string<CharType>> class output_string_adapter : public output_adapter_protocol<CharType> { public: - explicit output_string_adapter(std::basic_string<CharType>& s) : str(s) {} + explicit output_string_adapter(StringType& s) noexcept + : str(s) + {} void write_character(CharType c) override { @@ -4443,10 +6417,10 @@ class output_string_adapter : public output_adapter_protocol<CharType> } private: - std::basic_string<CharType>& str; + StringType& str; }; -template<typename CharType> +template<typename CharType, typename StringType = std::basic_string<CharType>> class output_adapter { public: @@ -4456,8 +6430,8 @@ class output_adapter output_adapter(std::basic_ostream<CharType>& s) : oa(std::make_shared<output_stream_adapter<CharType>>(s)) {} - output_adapter(std::basic_string<CharType>& s) - : oa(std::make_shared<output_string_adapter<CharType>>(s)) {} + output_adapter(StringType& s) + : oa(std::make_shared<output_string_adapter<CharType, StringType>>(s)) {} operator output_adapter_t<CharType>() { @@ -4467,19 +6441,57 @@ class output_adapter private: output_adapter_t<CharType> oa = nullptr; }; +} // namespace detail +} // namespace nlohmann + +// #include <nlohmann/detail/input/binary_reader.hpp> -////////////////////////////// -// binary reader and writer // -////////////////////////////// + +#include <algorithm> // generate_n +#include <array> // array +#include <cassert> // assert +#include <cmath> // ldexp +#include <cstddef> // size_t +#include <cstdint> // uint8_t, uint16_t, uint32_t, uint64_t +#include <cstdio> // snprintf +#include <cstring> // memcpy +#include <iterator> // back_inserter +#include <limits> // numeric_limits +#include <string> // char_traits, string +#include <utility> // make_pair, move + +// #include <nlohmann/detail/input/input_adapters.hpp> + +// #include <nlohmann/detail/input/json_sax.hpp> + +// #include <nlohmann/detail/exceptions.hpp> + +// #include <nlohmann/detail/macro_scope.hpp> + +// #include <nlohmann/detail/meta/is_sax.hpp> + +// #include <nlohmann/detail/value_t.hpp> + + +namespace nlohmann +{ +namespace detail +{ +/////////////////// +// binary reader // +/////////////////// /*! -@brief deserialization of CBOR and MessagePack values +@brief deserialization of CBOR, MessagePack, and UBJSON values */ -template<typename BasicJsonType> +template<typename BasicJsonType, typename SAX = json_sax_dom_parser<BasicJsonType>> class binary_reader { using number_integer_t = typename BasicJsonType::number_integer_t; using number_unsigned_t = typename BasicJsonType::number_unsigned_t; + using number_float_t = typename BasicJsonType::number_float_t; + using string_t = typename BasicJsonType::string_t; + using json_sax_t = SAX; public: /*! @@ -4489,49 +6501,68 @@ class binary_reader */ explicit binary_reader(input_adapter_t adapter) : ia(std::move(adapter)) { + (void)detail::is_sax_static_asserts<SAX, BasicJsonType> {}; assert(ia); } /*! - @brief create a JSON value from CBOR input - + @param[in] format the binary format to parse + @param[in] sax_ a SAX event processor @param[in] strict whether to expect the input to be consumed completed - @return JSON value created from CBOR input - @throw parse_error.110 if input ended unexpectedly or the end of file was - not reached when @a strict was set to true - @throw parse_error.112 if unsupported byte was read + @return */ - BasicJsonType parse_cbor(const bool strict) + bool sax_parse(const input_format_t format, + json_sax_t* sax_, + const bool strict = true) { - const auto res = parse_cbor_internal(); - if (strict) + sax = sax_; + bool result = false; + + switch (format) { - get(); - check_eof(true); - } - return res; - } + case input_format_t::bson: + result = parse_bson_internal(); + break; - /*! - @brief create a JSON value from MessagePack input + case input_format_t::cbor: + result = parse_cbor_internal(); + break; - @param[in] strict whether to expect the input to be consumed completed - @return JSON value created from MessagePack input + case input_format_t::msgpack: + result = parse_msgpack_internal(); + break; - @throw parse_error.110 if input ended unexpectedly or the end of file was - not reached when @a strict was set to true - @throw parse_error.112 if unsupported byte was read - */ - BasicJsonType parse_msgpack(const bool strict) - { - const auto res = parse_msgpack_internal(); - if (strict) + case input_format_t::ubjson: + result = parse_ubjson_internal(); + break; + + // LCOV_EXCL_START + default: + assert(false); + // LCOV_EXCL_STOP + } + + // strict mode: next byte must be EOF + if (result and strict) { - get(); - check_eof(true); + if (format == input_format_t::ubjson) + { + get_ignore_noop(); + } + else + { + get(); + } + + if (JSON_UNLIKELY(current != std::char_traits<char>::eof())) + { + return sax->parse_error(chars_read, get_token_string(), + parse_error::create(110, chars_read, exception_message(format, "expected end of input; last byte: 0x" + get_token_string(), "value"))); + } } - return res; + + return result; } /*! @@ -4547,18 +6578,239 @@ class binary_reader } private: + ////////// + // BSON // + ////////// + + /*! + @brief Reads in a BSON-object and passes it to the SAX-parser. + @return whether a valid BSON-value was passed to the SAX parser + */ + bool parse_bson_internal() + { + std::int32_t document_size; + get_number<std::int32_t, true>(input_format_t::bson, document_size); + + if (JSON_UNLIKELY(not sax->start_object(std::size_t(-1)))) + { + return false; + } + + if (JSON_UNLIKELY(not parse_bson_element_list(/*is_array*/false))) + { + return false; + } + + return sax->end_object(); + } + + /*! + @brief Parses a C-style string from the BSON input. + @param[in, out] result A reference to the string variable where the read + string is to be stored. + @return `true` if the \x00-byte indicating the end of the string was + encountered before the EOF; false` indicates an unexpected EOF. + */ + bool get_bson_cstr(string_t& result) + { + auto out = std::back_inserter(result); + while (true) + { + get(); + if (JSON_UNLIKELY(not unexpect_eof(input_format_t::bson, "cstring"))) + { + return false; + } + if (current == 0x00) + { + return true; + } + *out++ = static_cast<char>(current); + } + + return true; + } + + /*! + @brief Parses a zero-terminated string of length @a len from the BSON + input. + @param[in] len The length (including the zero-byte at the end) of the + string to be read. + @param[in, out] result A reference to the string variable where the read + string is to be stored. + @tparam NumberType The type of the length @a len + @pre len >= 1 + @return `true` if the string was successfully parsed + */ + template<typename NumberType> + bool get_bson_string(const NumberType len, string_t& result) + { + if (JSON_UNLIKELY(len < 1)) + { + auto last_token = get_token_string(); + return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, exception_message(input_format_t::bson, "string length must be at least 1, is " + std::to_string(len), "string"))); + } + + return get_string(input_format_t::bson, len - static_cast<NumberType>(1), result) and get() != std::char_traits<char>::eof(); + } + + /*! + @brief Read a BSON document element of the given @a element_type. + @param[in] element_type The BSON element type, c.f. http://bsonspec.org/spec.html + @param[in] element_type_parse_position The position in the input stream, + where the `element_type` was read. + @warning Not all BSON element types are supported yet. An unsupported + @a element_type will give rise to a parse_error.114: + Unsupported BSON record type 0x... + @return whether a valid BSON-object/array was passed to the SAX parser + */ + bool parse_bson_element_internal(const int element_type, + const std::size_t element_type_parse_position) + { + switch (element_type) + { + case 0x01: // double + { + double number; + return get_number<double, true>(input_format_t::bson, number) and sax->number_float(static_cast<number_float_t>(number), ""); + } + + case 0x02: // string + { + std::int32_t len; + string_t value; + return get_number<std::int32_t, true>(input_format_t::bson, len) and get_bson_string(len, value) and sax->string(value); + } + + case 0x03: // object + { + return parse_bson_internal(); + } + + case 0x04: // array + { + return parse_bson_array(); + } + + case 0x08: // boolean + { + return sax->boolean(get() != 0); + } + + case 0x0A: // null + { + return sax->null(); + } + + case 0x10: // int32 + { + std::int32_t value; + return get_number<std::int32_t, true>(input_format_t::bson, value) and sax->number_integer(value); + } + + case 0x12: // int64 + { + std::int64_t value; + return get_number<std::int64_t, true>(input_format_t::bson, value) and sax->number_integer(value); + } + + default: // anything else not supported (yet) + { + char cr[3]; + (std::snprintf)(cr, sizeof(cr), "%.2hhX", static_cast<unsigned char>(element_type)); + return sax->parse_error(element_type_parse_position, std::string(cr), parse_error::create(114, element_type_parse_position, "Unsupported BSON record type 0x" + std::string(cr))); + } + } + } + + /*! + @brief Read a BSON element list (as specified in the BSON-spec) + + The same binary layout is used for objects and arrays, hence it must be + indicated with the argument @a is_array which one is expected + (true --> array, false --> object). + + @param[in] is_array Determines if the element list being read is to be + treated as an object (@a is_array == false), or as an + array (@a is_array == true). + @return whether a valid BSON-object/array was passed to the SAX parser + */ + bool parse_bson_element_list(const bool is_array) + { + string_t key; + while (int element_type = get()) + { + if (JSON_UNLIKELY(not unexpect_eof(input_format_t::bson, "element list"))) + { + return false; + } + + const std::size_t element_type_parse_position = chars_read; + if (JSON_UNLIKELY(not get_bson_cstr(key))) + { + return false; + } + + if (not is_array) + { + if (not sax->key(key)) + { + return false; + } + } + + if (JSON_UNLIKELY(not parse_bson_element_internal(element_type, element_type_parse_position))) + { + return false; + } + + // get_bson_cstr only appends + key.clear(); + } + + return true; + } + + /*! + @brief Reads an array from the BSON input and passes it to the SAX-parser. + @return whether a valid BSON-array was passed to the SAX parser + */ + bool parse_bson_array() + { + std::int32_t document_size; + get_number<std::int32_t, true>(input_format_t::bson, document_size); + + if (JSON_UNLIKELY(not sax->start_array(std::size_t(-1)))) + { + return false; + } + + if (JSON_UNLIKELY(not parse_bson_element_list(/*is_array*/true))) + { + return false; + } + + return sax->end_array(); + } + + ////////// + // CBOR // + ////////// + /*! @param[in] get_char whether a new character should be retrieved from the input (true, default) or whether the last read character should be considered instead + + @return whether a valid CBOR value was passed to the SAX parser */ - BasicJsonType parse_cbor_internal(const bool get_char = true) + bool parse_cbor_internal(const bool get_char = true) { switch (get_char ? get() : current) { // EOF case std::char_traits<char>::eof(): - JSON_THROW(parse_error::create(110, chars_read, "unexpected end of input")); + return unexpect_eof(input_format_t::cbor, "value"); // Integer 0x00..0x17 (0..23) case 0x00: @@ -4585,19 +6837,31 @@ class binary_reader case 0x15: case 0x16: case 0x17: - return static_cast<number_unsigned_t>(current); + return sax->number_unsigned(static_cast<number_unsigned_t>(current)); case 0x18: // Unsigned integer (one-byte uint8_t follows) - return get_number<uint8_t>(); + { + uint8_t number; + return get_number(input_format_t::cbor, number) and sax->number_unsigned(number); + } case 0x19: // Unsigned integer (two-byte uint16_t follows) - return get_number<uint16_t>(); + { + uint16_t number; + return get_number(input_format_t::cbor, number) and sax->number_unsigned(number); + } case 0x1A: // Unsigned integer (four-byte uint32_t follows) - return get_number<uint32_t>(); + { + uint32_t number; + return get_number(input_format_t::cbor, number) and sax->number_unsigned(number); + } case 0x1B: // Unsigned integer (eight-byte uint64_t follows) - return get_number<uint64_t>(); + { + uint64_t number; + return get_number(input_format_t::cbor, number) and sax->number_unsigned(number); + } // Negative integer -1-0x00..-1-0x17 (-1..-24) case 0x20: @@ -4624,28 +6888,31 @@ class binary_reader case 0x35: case 0x36: case 0x37: - return static_cast<int8_t>(0x20 - 1 - current); + return sax->number_integer(static_cast<int8_t>(0x20 - 1 - current)); case 0x38: // Negative integer (one-byte uint8_t follows) { - // must be uint8_t ! - return static_cast<number_integer_t>(-1) - get_number<uint8_t>(); + uint8_t number; + return get_number(input_format_t::cbor, number) and sax->number_integer(static_cast<number_integer_t>(-1) - number); } case 0x39: // Negative integer -1-n (two-byte uint16_t follows) { - return static_cast<number_integer_t>(-1) - get_number<uint16_t>(); + uint16_t number; + return get_number(input_format_t::cbor, number) and sax->number_integer(static_cast<number_integer_t>(-1) - number); } case 0x3A: // Negative integer -1-n (four-byte uint32_t follows) { - return static_cast<number_integer_t>(-1) - get_number<uint32_t>(); + uint32_t number; + return get_number(input_format_t::cbor, number) and sax->number_integer(static_cast<number_integer_t>(-1) - number); } case 0x3B: // Negative integer -1-n (eight-byte uint64_t follows) { - return static_cast<number_integer_t>(-1) - - static_cast<number_integer_t>(get_number<uint64_t>()); + uint64_t number; + return get_number(input_format_t::cbor, number) and sax->number_integer(static_cast<number_integer_t>(-1) + - static_cast<number_integer_t>(number)); } // UTF-8 string (0x00..0x17 bytes follow) @@ -4679,7 +6946,8 @@ class binary_reader case 0x7B: // UTF-8 string (eight-byte uint64_t for n follow) case 0x7F: // UTF-8 string (indefinite length) { - return get_cbor_string(); + string_t s; + return get_cbor_string(s) and sax->string(s); } // array (0x00..0x17 data items follow) @@ -4707,39 +6975,34 @@ class binary_reader case 0x95: case 0x96: case 0x97: - { - return get_cbor_array(current & 0x1F); - } + return get_cbor_array(static_cast<std::size_t>(current & 0x1F)); case 0x98: // array (one-byte uint8_t for n follows) { - return get_cbor_array(get_number<uint8_t>()); + uint8_t len; + return get_number(input_format_t::cbor, len) and get_cbor_array(static_cast<std::size_t>(len)); } case 0x99: // array (two-byte uint16_t for n follow) { - return get_cbor_array(get_number<uint16_t>()); + uint16_t len; + return get_number(input_format_t::cbor, len) and get_cbor_array(static_cast<std::size_t>(len)); } case 0x9A: // array (four-byte uint32_t for n follow) { - return get_cbor_array(get_number<uint32_t>()); + uint32_t len; + return get_number(input_format_t::cbor, len) and get_cbor_array(static_cast<std::size_t>(len)); } case 0x9B: // array (eight-byte uint64_t for n follow) { - return get_cbor_array(get_number<uint64_t>()); + uint64_t len; + return get_number(input_format_t::cbor, len) and get_cbor_array(static_cast<std::size_t>(len)); } case 0x9F: // array (indefinite length) - { - BasicJsonType result = value_t::array; - while (get() != 0xFF) - { - result.push_back(parse_cbor_internal(false)); - } - return result; - } + return get_cbor_array(std::size_t(-1)); // map (0x00..0x17 pairs of data items follow) case 0xA0: @@ -4766,62 +7029,59 @@ class binary_reader case 0xB5: case 0xB6: case 0xB7: - { - return get_cbor_object(current & 0x1F); - } + return get_cbor_object(static_cast<std::size_t>(current & 0x1F)); case 0xB8: // map (one-byte uint8_t for n follows) { - return get_cbor_object(get_number<uint8_t>()); + uint8_t len; + return get_number(input_format_t::cbor, len) and get_cbor_object(static_cast<std::size_t>(len)); } case 0xB9: // map (two-byte uint16_t for n follow) { - return get_cbor_object(get_number<uint16_t>()); + uint16_t len; + return get_number(input_format_t::cbor, len) and get_cbor_object(static_cast<std::size_t>(len)); } case 0xBA: // map (four-byte uint32_t for n follow) { - return get_cbor_object(get_number<uint32_t>()); + uint32_t len; + return get_number(input_format_t::cbor, len) and get_cbor_object(static_cast<std::size_t>(len)); } case 0xBB: // map (eight-byte uint64_t for n follow) { - return get_cbor_object(get_number<uint64_t>()); + uint64_t len; + return get_number(input_format_t::cbor, len) and get_cbor_object(static_cast<std::size_t>(len)); } case 0xBF: // map (indefinite length) - { - BasicJsonType result = value_t::object; - while (get() != 0xFF) - { - auto key = get_cbor_string(); - result[key] = parse_cbor_internal(); - } - return result; - } + return get_cbor_object(std::size_t(-1)); case 0xF4: // false - { - return false; - } + return sax->boolean(false); case 0xF5: // true - { - return true; - } + return sax->boolean(true); case 0xF6: // null - { - return value_t::null; - } + return sax->null(); case 0xF9: // Half-Precision Float (two-byte IEEE 754) { - const int byte1 = get(); - check_eof(); - const int byte2 = get(); - check_eof(); + const int byte1_raw = get(); + if (JSON_UNLIKELY(not unexpect_eof(input_format_t::cbor, "number"))) + { + return false; + } + const int byte2_raw = get(); + if (JSON_UNLIKELY(not unexpect_eof(input_format_t::cbor, "number"))) + { + return false; + } + + const auto byte1 = static_cast<unsigned char>(byte1_raw); + const auto byte2 = static_cast<unsigned char>(byte2_raw); // code from RFC 7049, Appendix D, Figure 3: // As half-precision floating-point numbers were only added @@ -4832,51 +7092,244 @@ class binary_reader // half-precision floating-point numbers in the C language // is shown in Fig. 3. const int half = (byte1 << 8) + byte2; - const int exp = (half >> 10) & 0x1F; - const int mant = half & 0x3FF; - double val; - if (exp == 0) - { - val = std::ldexp(mant, -24); - } - else if (exp != 31) + const double val = [&half] { - val = std::ldexp(mant + 1024, exp - 25); - } - else - { - val = (mant == 0) ? std::numeric_limits<double>::infinity() - : std::numeric_limits<double>::quiet_NaN(); - } - return (half & 0x8000) != 0 ? -val : val; + const int exp = (half >> 10) & 0x1F; + const int mant = half & 0x3FF; + assert(0 <= exp and exp <= 32); + assert(0 <= mant and mant <= 1024); + switch (exp) + { + case 0: + return std::ldexp(mant, -24); + case 31: + return (mant == 0) + ? std::numeric_limits<double>::infinity() + : std::numeric_limits<double>::quiet_NaN(); + default: + return std::ldexp(mant + 1024, exp - 25); + } + }(); + return sax->number_float((half & 0x8000) != 0 + ? static_cast<number_float_t>(-val) + : static_cast<number_float_t>(val), ""); } case 0xFA: // Single-Precision Float (four-byte IEEE 754) { - return get_number<float>(); + float number; + return get_number(input_format_t::cbor, number) and sax->number_float(static_cast<number_float_t>(number), ""); } case 0xFB: // Double-Precision Float (eight-byte IEEE 754) { - return get_number<double>(); + double number; + return get_number(input_format_t::cbor, number) and sax->number_float(static_cast<number_float_t>(number), ""); } default: // anything else (0xFF is handled inside the other types) { - std::stringstream ss; - ss << std::setw(2) << std::uppercase << std::setfill('0') << std::hex << current; - JSON_THROW(parse_error::create(112, chars_read, "error reading CBOR; last byte: 0x" + ss.str())); + auto last_token = get_token_string(); + return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, exception_message(input_format_t::cbor, "invalid byte: 0x" + last_token, "value"))); } } } - BasicJsonType parse_msgpack_internal() + /*! + @brief reads a CBOR string + + This function first reads starting bytes to determine the expected + string length and then copies this number of bytes into a string. + Additionally, CBOR's strings with indefinite lengths are supported. + + @param[out] result created string + + @return whether string creation completed + */ + bool get_cbor_string(string_t& result) + { + if (JSON_UNLIKELY(not unexpect_eof(input_format_t::cbor, "string"))) + { + return false; + } + + switch (current) + { + // UTF-8 string (0x00..0x17 bytes follow) + case 0x60: + case 0x61: + case 0x62: + case 0x63: + case 0x64: + case 0x65: + case 0x66: + case 0x67: + case 0x68: + case 0x69: + case 0x6A: + case 0x6B: + case 0x6C: + case 0x6D: + case 0x6E: + case 0x6F: + case 0x70: + case 0x71: + case 0x72: + case 0x73: + case 0x74: + case 0x75: + case 0x76: + case 0x77: + { + return get_string(input_format_t::cbor, current & 0x1F, result); + } + + case 0x78: // UTF-8 string (one-byte uint8_t for n follows) + { + uint8_t len; + return get_number(input_format_t::cbor, len) and get_string(input_format_t::cbor, len, result); + } + + case 0x79: // UTF-8 string (two-byte uint16_t for n follow) + { + uint16_t len; + return get_number(input_format_t::cbor, len) and get_string(input_format_t::cbor, len, result); + } + + case 0x7A: // UTF-8 string (four-byte uint32_t for n follow) + { + uint32_t len; + return get_number(input_format_t::cbor, len) and get_string(input_format_t::cbor, len, result); + } + + case 0x7B: // UTF-8 string (eight-byte uint64_t for n follow) + { + uint64_t len; + return get_number(input_format_t::cbor, len) and get_string(input_format_t::cbor, len, result); + } + + case 0x7F: // UTF-8 string (indefinite length) + { + while (get() != 0xFF) + { + string_t chunk; + if (not get_cbor_string(chunk)) + { + return false; + } + result.append(chunk); + } + return true; + } + + default: + { + auto last_token = get_token_string(); + return sax->parse_error(chars_read, last_token, parse_error::create(113, chars_read, exception_message(input_format_t::cbor, "expected length specification (0x60-0x7B) or indefinite string type (0x7F); last byte: 0x" + last_token, "string"))); + } + } + } + + /*! + @param[in] len the length of the array or std::size_t(-1) for an + array of indefinite size + @return whether array creation completed + */ + bool get_cbor_array(const std::size_t len) + { + if (JSON_UNLIKELY(not sax->start_array(len))) + { + return false; + } + + if (len != std::size_t(-1)) + { + for (std::size_t i = 0; i < len; ++i) + { + if (JSON_UNLIKELY(not parse_cbor_internal())) + { + return false; + } + } + } + else + { + while (get() != 0xFF) + { + if (JSON_UNLIKELY(not parse_cbor_internal(false))) + { + return false; + } + } + } + + return sax->end_array(); + } + + /*! + @param[in] len the length of the object or std::size_t(-1) for an + object of indefinite size + @return whether object creation completed + */ + bool get_cbor_object(const std::size_t len) + { + if (not JSON_UNLIKELY(sax->start_object(len))) + { + return false; + } + + string_t key; + if (len != std::size_t(-1)) + { + for (std::size_t i = 0; i < len; ++i) + { + get(); + if (JSON_UNLIKELY(not get_cbor_string(key) or not sax->key(key))) + { + return false; + } + + if (JSON_UNLIKELY(not parse_cbor_internal())) + { + return false; + } + key.clear(); + } + } + else + { + while (get() != 0xFF) + { + if (JSON_UNLIKELY(not get_cbor_string(key) or not sax->key(key))) + { + return false; + } + + if (JSON_UNLIKELY(not parse_cbor_internal())) + { + return false; + } + key.clear(); + } + } + + return sax->end_object(); + } + + ///////////// + // MsgPack // + ///////////// + + /*! + @return whether a valid MessagePack value was passed to the SAX parser + */ + bool parse_msgpack_internal() { switch (get()) { // EOF case std::char_traits<char>::eof(): - JSON_THROW(parse_error::create(110, chars_read, "unexpected end of input")); + return unexpect_eof(input_format_t::msgpack, "value"); // positive fixint case 0x00: @@ -5007,7 +7460,7 @@ class binary_reader case 0x7D: case 0x7E: case 0x7F: - return static_cast<number_unsigned_t>(current); + return sax->number_unsigned(static_cast<number_unsigned_t>(current)); // fixmap case 0x80: @@ -5026,9 +7479,7 @@ class binary_reader case 0x8D: case 0x8E: case 0x8F: - { - return get_msgpack_object(current & 0x0F); - } + return get_msgpack_object(static_cast<std::size_t>(current & 0x0F)); // fixarray case 0x90: @@ -5047,9 +7498,7 @@ class binary_reader case 0x9D: case 0x9E: case 0x9F: - { - return get_msgpack_array(current & 0x0F); - } + return get_msgpack_array(static_cast<std::size_t>(current & 0x0F)); // fixstr case 0xA0: @@ -5084,73 +7533,113 @@ class binary_reader case 0xBD: case 0xBE: case 0xBF: - return get_msgpack_string(); + { + string_t s; + return get_msgpack_string(s) and sax->string(s); + } case 0xC0: // nil - return value_t::null; + return sax->null(); case 0xC2: // false - return false; + return sax->boolean(false); case 0xC3: // true - return true; + return sax->boolean(true); case 0xCA: // float 32 - return get_number<float>(); + { + float number; + return get_number(input_format_t::msgpack, number) and sax->number_float(static_cast<number_float_t>(number), ""); + } case 0xCB: // float 64 - return get_number<double>(); + { + double number; + return get_number(input_format_t::msgpack, number) and sax->number_float(static_cast<number_float_t>(number), ""); + } case 0xCC: // uint 8 - return get_number<uint8_t>(); + { + uint8_t number; + return get_number(input_format_t::msgpack, number) and sax->number_unsigned(number); + } case 0xCD: // uint 16 - return get_number<uint16_t>(); + { + uint16_t number; + return get_number(input_format_t::msgpack, number) and sax->number_unsigned(number); + } case 0xCE: // uint 32 - return get_number<uint32_t>(); + { + uint32_t number; + return get_number(input_format_t::msgpack, number) and sax->number_unsigned(number); + } case 0xCF: // uint 64 - return get_number<uint64_t>(); + { + uint64_t number; + return get_number(input_format_t::msgpack, number) and sax->number_unsigned(number); + } case 0xD0: // int 8 - return get_number<int8_t>(); + { + int8_t number; + return get_number(input_format_t::msgpack, number) and sax->number_integer(number); + } case 0xD1: // int 16 - return get_number<int16_t>(); + { + int16_t number; + return get_number(input_format_t::msgpack, number) and sax->number_integer(number); + } case 0xD2: // int 32 - return get_number<int32_t>(); + { + int32_t number; + return get_number(input_format_t::msgpack, number) and sax->number_integer(number); + } case 0xD3: // int 64 - return get_number<int64_t>(); + { + int64_t number; + return get_number(input_format_t::msgpack, number) and sax->number_integer(number); + } case 0xD9: // str 8 case 0xDA: // str 16 case 0xDB: // str 32 - return get_msgpack_string(); + { + string_t s; + return get_msgpack_string(s) and sax->string(s); + } case 0xDC: // array 16 { - return get_msgpack_array(get_number<uint16_t>()); + uint16_t len; + return get_number(input_format_t::msgpack, len) and get_msgpack_array(static_cast<std::size_t>(len)); } case 0xDD: // array 32 { - return get_msgpack_array(get_number<uint32_t>()); + uint32_t len; + return get_number(input_format_t::msgpack, len) and get_msgpack_array(static_cast<std::size_t>(len)); } case 0xDE: // map 16 { - return get_msgpack_object(get_number<uint16_t>()); + uint16_t len; + return get_number(input_format_t::msgpack, len) and get_msgpack_object(static_cast<std::size_t>(len)); } case 0xDF: // map 32 { - return get_msgpack_object(get_number<uint32_t>()); + uint32_t len; + return get_number(input_format_t::msgpack, len) and get_msgpack_object(static_cast<std::size_t>(len)); } - // positive fixint + // negative fixint case 0xE0: case 0xE1: case 0xE2: @@ -5183,338 +7672,749 @@ class binary_reader case 0xFD: case 0xFE: case 0xFF: - return static_cast<int8_t>(current); + return sax->number_integer(static_cast<int8_t>(current)); default: // anything else { - std::stringstream ss; - ss << std::setw(2) << std::uppercase << std::setfill('0') << std::hex << current; - JSON_THROW(parse_error::create(112, chars_read, - "error reading MessagePack; last byte: 0x" + ss.str())); + auto last_token = get_token_string(); + return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, exception_message(input_format_t::msgpack, "invalid byte: 0x" + last_token, "value"))); } } } /*! - @brief get next character from the input + @brief reads a MessagePack string - This function provides the interface to the used input adapter. It does - not throw in case the input reached EOF, but returns a -'ve valued - `std::char_traits<char>::eof()` in that case. + This function first reads starting bytes to determine the expected + string length and then copies this number of bytes into a string. - @return character read from the input + @param[out] result created string + + @return whether string creation completed */ - int get() + bool get_msgpack_string(string_t& result) { - ++chars_read; - return (current = ia->get_character()); - } + if (JSON_UNLIKELY(not unexpect_eof(input_format_t::msgpack, "string"))) + { + return false; + } - /* - @brief read a number from the input + switch (current) + { + // fixstr + case 0xA0: + case 0xA1: + case 0xA2: + case 0xA3: + case 0xA4: + case 0xA5: + case 0xA6: + case 0xA7: + case 0xA8: + case 0xA9: + case 0xAA: + case 0xAB: + case 0xAC: + case 0xAD: + case 0xAE: + case 0xAF: + case 0xB0: + case 0xB1: + case 0xB2: + case 0xB3: + case 0xB4: + case 0xB5: + case 0xB6: + case 0xB7: + case 0xB8: + case 0xB9: + case 0xBA: + case 0xBB: + case 0xBC: + case 0xBD: + case 0xBE: + case 0xBF: + { + return get_string(input_format_t::msgpack, current & 0x1F, result); + } - @tparam NumberType the type of the number + case 0xD9: // str 8 + { + uint8_t len; + return get_number(input_format_t::msgpack, len) and get_string(input_format_t::msgpack, len, result); + } - @return number of type @a NumberType + case 0xDA: // str 16 + { + uint16_t len; + return get_number(input_format_t::msgpack, len) and get_string(input_format_t::msgpack, len, result); + } - @note This function needs to respect the system's endianess, because - bytes in CBOR and MessagePack are stored in network order (big - endian) and therefore need reordering on little endian systems. + case 0xDB: // str 32 + { + uint32_t len; + return get_number(input_format_t::msgpack, len) and get_string(input_format_t::msgpack, len, result); + } + + default: + { + auto last_token = get_token_string(); + return sax->parse_error(chars_read, last_token, parse_error::create(113, chars_read, exception_message(input_format_t::msgpack, "expected length specification (0xA0-0xBF, 0xD9-0xDB); last byte: 0x" + last_token, "string"))); + } + } + } - @throw parse_error.110 if input has less than `sizeof(NumberType)` bytes + /*! + @param[in] len the length of the array + @return whether array creation completed */ - template<typename NumberType> NumberType get_number() + bool get_msgpack_array(const std::size_t len) { - // step 1: read input into array with system's byte order - std::array<uint8_t, sizeof(NumberType)> vec; - for (std::size_t i = 0; i < sizeof(NumberType); ++i) + if (JSON_UNLIKELY(not sax->start_array(len))) { - get(); - check_eof(); + return false; + } - // reverse byte order prior to conversion if necessary - if (is_little_endian) - { - vec[sizeof(NumberType) - i - 1] = static_cast<uint8_t>(current); - } - else + for (std::size_t i = 0; i < len; ++i) + { + if (JSON_UNLIKELY(not parse_msgpack_internal())) { - vec[i] = static_cast<uint8_t>(current); // LCOV_EXCL_LINE + return false; } } - // step 2: convert array into number of type T and return - NumberType result; - std::memcpy(&result, vec.data(), sizeof(NumberType)); - return result; + return sax->end_array(); } /*! - @brief create a string by reading characters from the input + @param[in] len the length of the object + @return whether object creation completed + */ + bool get_msgpack_object(const std::size_t len) + { + if (JSON_UNLIKELY(not sax->start_object(len))) + { + return false; + } - @param[in] len number of bytes to read + string_t key; + for (std::size_t i = 0; i < len; ++i) + { + get(); + if (JSON_UNLIKELY(not get_msgpack_string(key) or not sax->key(key))) + { + return false; + } - @note We can not reserve @a len bytes for the result, because @a len - may be too large. Usually, @ref check_eof() detects the end of - the input before we run out of string memory. + if (JSON_UNLIKELY(not parse_msgpack_internal())) + { + return false; + } + key.clear(); + } - @return string created by reading @a len bytes + return sax->end_object(); + } - @throw parse_error.110 if input has less than @a len bytes + //////////// + // UBJSON // + //////////// + + /*! + @param[in] get_char whether a new character should be retrieved from the + input (true, default) or whether the last read + character should be considered instead + + @return whether a valid UBJSON value was passed to the SAX parser */ - template<typename NumberType> - std::string get_string(const NumberType len) + bool parse_ubjson_internal(const bool get_char = true) { - std::string result; - std::generate_n(std::back_inserter(result), len, [this]() - { - get(); - check_eof(); - return static_cast<char>(current); - }); - return result; + return get_ubjson_value(get_char ? get_ignore_noop() : current); } /*! - @brief reads a CBOR string + @brief reads a UBJSON string - This function first reads starting bytes to determine the expected - string length and then copies this number of bytes into a string. - Additionally, CBOR's strings with indefinite lengths are supported. + This function is either called after reading the 'S' byte explicitly + indicating a string, or in case of an object key where the 'S' byte can be + left out. - @return string + @param[out] result created string + @param[in] get_char whether a new character should be retrieved from the + input (true, default) or whether the last read + character should be considered instead - @throw parse_error.110 if input ended - @throw parse_error.113 if an unexpected byte is read + @return whether string creation completed */ - std::string get_cbor_string() + bool get_ubjson_string(string_t& result, const bool get_char = true) { - check_eof(); + if (get_char) + { + get(); // TODO: may we ignore N here? + } + + if (JSON_UNLIKELY(not unexpect_eof(input_format_t::ubjson, "value"))) + { + return false; + } switch (current) { - // UTF-8 string (0x00..0x17 bytes follow) - case 0x60: - case 0x61: - case 0x62: - case 0x63: - case 0x64: - case 0x65: - case 0x66: - case 0x67: - case 0x68: - case 0x69: - case 0x6A: - case 0x6B: - case 0x6C: - case 0x6D: - case 0x6E: - case 0x6F: - case 0x70: - case 0x71: - case 0x72: - case 0x73: - case 0x74: - case 0x75: - case 0x76: - case 0x77: + case 'U': { - return get_string(current & 0x1F); + uint8_t len; + return get_number(input_format_t::ubjson, len) and get_string(input_format_t::ubjson, len, result); } - case 0x78: // UTF-8 string (one-byte uint8_t for n follows) + case 'i': { - return get_string(get_number<uint8_t>()); + int8_t len; + return get_number(input_format_t::ubjson, len) and get_string(input_format_t::ubjson, len, result); } - case 0x79: // UTF-8 string (two-byte uint16_t for n follow) + case 'I': { - return get_string(get_number<uint16_t>()); + int16_t len; + return get_number(input_format_t::ubjson, len) and get_string(input_format_t::ubjson, len, result); } - case 0x7A: // UTF-8 string (four-byte uint32_t for n follow) + case 'l': { - return get_string(get_number<uint32_t>()); + int32_t len; + return get_number(input_format_t::ubjson, len) and get_string(input_format_t::ubjson, len, result); } - case 0x7B: // UTF-8 string (eight-byte uint64_t for n follow) + case 'L': { - return get_string(get_number<uint64_t>()); + int64_t len; + return get_number(input_format_t::ubjson, len) and get_string(input_format_t::ubjson, len, result); } - case 0x7F: // UTF-8 string (indefinite length) + default: + auto last_token = get_token_string(); + return sax->parse_error(chars_read, last_token, parse_error::create(113, chars_read, exception_message(input_format_t::ubjson, "expected length type specification (U, i, I, l, L); last byte: 0x" + last_token, "string"))); + } + } + + /*! + @param[out] result determined size + @return whether size determination completed + */ + bool get_ubjson_size_value(std::size_t& result) + { + switch (get_ignore_noop()) + { + case 'U': { - std::string result; - while (get() != 0xFF) + uint8_t number; + if (JSON_UNLIKELY(not get_number(input_format_t::ubjson, number))) + { + return false; + } + result = static_cast<std::size_t>(number); + return true; + } + + case 'i': + { + int8_t number; + if (JSON_UNLIKELY(not get_number(input_format_t::ubjson, number))) { - check_eof(); - result.push_back(static_cast<char>(current)); + return false; } - return result; + result = static_cast<std::size_t>(number); + return true; + } + + case 'I': + { + int16_t number; + if (JSON_UNLIKELY(not get_number(input_format_t::ubjson, number))) + { + return false; + } + result = static_cast<std::size_t>(number); + return true; + } + + case 'l': + { + int32_t number; + if (JSON_UNLIKELY(not get_number(input_format_t::ubjson, number))) + { + return false; + } + result = static_cast<std::size_t>(number); + return true; + } + + case 'L': + { + int64_t number; + if (JSON_UNLIKELY(not get_number(input_format_t::ubjson, number))) + { + return false; + } + result = static_cast<std::size_t>(number); + return true; } default: { - std::stringstream ss; - ss << std::setw(2) << std::uppercase << std::setfill('0') << std::hex << current; - JSON_THROW(parse_error::create(113, chars_read, "expected a CBOR string; last byte: 0x" + ss.str())); + auto last_token = get_token_string(); + return sax->parse_error(chars_read, last_token, parse_error::create(113, chars_read, exception_message(input_format_t::ubjson, "expected length type specification (U, i, I, l, L) after '#'; last byte: 0x" + last_token, "size"))); } } } - template<typename NumberType> - BasicJsonType get_cbor_array(const NumberType len) + /*! + @brief determine the type and size for a container + + In the optimized UBJSON format, a type and a size can be provided to allow + for a more compact representation. + + @param[out] result pair of the size and the type + + @return whether pair creation completed + */ + bool get_ubjson_size_type(std::pair<std::size_t, int>& result) { - BasicJsonType result = value_t::array; - std::generate_n(std::back_inserter(*result.m_value.array), len, [this]() + result.first = string_t::npos; // size + result.second = 0; // type + + get_ignore_noop(); + + if (current == '$') { - return parse_cbor_internal(); - }); - return result; - } + result.second = get(); // must not ignore 'N', because 'N' maybe the type + if (JSON_UNLIKELY(not unexpect_eof(input_format_t::ubjson, "type"))) + { + return false; + } - template<typename NumberType> - BasicJsonType get_cbor_object(const NumberType len) - { - BasicJsonType result = value_t::object; - std::generate_n(std::inserter(*result.m_value.object, - result.m_value.object->end()), - len, [this]() + get_ignore_noop(); + if (JSON_UNLIKELY(current != '#')) + { + if (JSON_UNLIKELY(not unexpect_eof(input_format_t::ubjson, "value"))) + { + return false; + } + auto last_token = get_token_string(); + return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, exception_message(input_format_t::ubjson, "expected '#' after type information; last byte: 0x" + last_token, "size"))); + } + + return get_ubjson_size_value(result.first); + } + else if (current == '#') { - get(); - auto key = get_cbor_string(); - auto val = parse_cbor_internal(); - return std::make_pair(std::move(key), std::move(val)); - }); - return result; + return get_ubjson_size_value(result.first); + } + return true; } /*! - @brief reads a MessagePack string + @param prefix the previously read or set type prefix + @return whether value creation completed + */ + bool get_ubjson_value(const int prefix) + { + switch (prefix) + { + case std::char_traits<char>::eof(): // EOF + return unexpect_eof(input_format_t::ubjson, "value"); - This function first reads starting bytes to determine the expected - string length and then copies this number of bytes into a string. + case 'T': // true + return sax->boolean(true); + case 'F': // false + return sax->boolean(false); + + case 'Z': // null + return sax->null(); - @return string + case 'U': + { + uint8_t number; + return get_number(input_format_t::ubjson, number) and sax->number_unsigned(number); + } + + case 'i': + { + int8_t number; + return get_number(input_format_t::ubjson, number) and sax->number_integer(number); + } - @throw parse_error.110 if input ended - @throw parse_error.113 if an unexpected byte is read + case 'I': + { + int16_t number; + return get_number(input_format_t::ubjson, number) and sax->number_integer(number); + } + + case 'l': + { + int32_t number; + return get_number(input_format_t::ubjson, number) and sax->number_integer(number); + } + + case 'L': + { + int64_t number; + return get_number(input_format_t::ubjson, number) and sax->number_integer(number); + } + + case 'd': + { + float number; + return get_number(input_format_t::ubjson, number) and sax->number_float(static_cast<number_float_t>(number), ""); + } + + case 'D': + { + double number; + return get_number(input_format_t::ubjson, number) and sax->number_float(static_cast<number_float_t>(number), ""); + } + + case 'C': // char + { + get(); + if (JSON_UNLIKELY(not unexpect_eof(input_format_t::ubjson, "char"))) + { + return false; + } + if (JSON_UNLIKELY(current > 127)) + { + auto last_token = get_token_string(); + return sax->parse_error(chars_read, last_token, parse_error::create(113, chars_read, exception_message(input_format_t::ubjson, "byte after 'C' must be in range 0x00..0x7F; last byte: 0x" + last_token, "char"))); + } + string_t s(1, static_cast<char>(current)); + return sax->string(s); + } + + case 'S': // string + { + string_t s; + return get_ubjson_string(s) and sax->string(s); + } + + case '[': // array + return get_ubjson_array(); + + case '{': // object + return get_ubjson_object(); + + default: // anything else + { + auto last_token = get_token_string(); + return sax->parse_error(chars_read, last_token, parse_error::create(112, chars_read, exception_message(input_format_t::ubjson, "invalid byte: 0x" + last_token, "value"))); + } + } + } + + /*! + @return whether array creation completed */ - std::string get_msgpack_string() + bool get_ubjson_array() { - check_eof(); + std::pair<std::size_t, int> size_and_type; + if (JSON_UNLIKELY(not get_ubjson_size_type(size_and_type))) + { + return false; + } - switch (current) + if (size_and_type.first != string_t::npos) { - // fixstr - case 0xA0: - case 0xA1: - case 0xA2: - case 0xA3: - case 0xA4: - case 0xA5: - case 0xA6: - case 0xA7: - case 0xA8: - case 0xA9: - case 0xAA: - case 0xAB: - case 0xAC: - case 0xAD: - case 0xAE: - case 0xAF: - case 0xB0: - case 0xB1: - case 0xB2: - case 0xB3: - case 0xB4: - case 0xB5: - case 0xB6: - case 0xB7: - case 0xB8: - case 0xB9: - case 0xBA: - case 0xBB: - case 0xBC: - case 0xBD: - case 0xBE: - case 0xBF: + if (JSON_UNLIKELY(not sax->start_array(size_and_type.first))) { - return get_string(current & 0x1F); + return false; } - case 0xD9: // str 8 + if (size_and_type.second != 0) { - return get_string(get_number<uint8_t>()); + if (size_and_type.second != 'N') + { + for (std::size_t i = 0; i < size_and_type.first; ++i) + { + if (JSON_UNLIKELY(not get_ubjson_value(size_and_type.second))) + { + return false; + } + } + } + } + else + { + for (std::size_t i = 0; i < size_and_type.first; ++i) + { + if (JSON_UNLIKELY(not parse_ubjson_internal())) + { + return false; + } + } + } + } + else + { + if (JSON_UNLIKELY(not sax->start_array(std::size_t(-1)))) + { + return false; } - case 0xDA: // str 16 + while (current != ']') { - return get_string(get_number<uint16_t>()); + if (JSON_UNLIKELY(not parse_ubjson_internal(false))) + { + return false; + } + get_ignore_noop(); } + } - case 0xDB: // str 32 + return sax->end_array(); + } + + /*! + @return whether object creation completed + */ + bool get_ubjson_object() + { + std::pair<std::size_t, int> size_and_type; + if (JSON_UNLIKELY(not get_ubjson_size_type(size_and_type))) + { + return false; + } + + string_t key; + if (size_and_type.first != string_t::npos) + { + if (JSON_UNLIKELY(not sax->start_object(size_and_type.first))) { - return get_string(get_number<uint32_t>()); + return false; } - default: + if (size_and_type.second != 0) { - std::stringstream ss; - ss << std::setw(2) << std::uppercase << std::setfill('0') << std::hex << current; - JSON_THROW(parse_error::create(113, chars_read, - "expected a MessagePack string; last byte: 0x" + ss.str())); + for (std::size_t i = 0; i < size_and_type.first; ++i) + { + if (JSON_UNLIKELY(not get_ubjson_string(key) or not sax->key(key))) + { + return false; + } + if (JSON_UNLIKELY(not get_ubjson_value(size_and_type.second))) + { + return false; + } + key.clear(); + } + } + else + { + for (std::size_t i = 0; i < size_and_type.first; ++i) + { + if (JSON_UNLIKELY(not get_ubjson_string(key) or not sax->key(key))) + { + return false; + } + if (JSON_UNLIKELY(not parse_ubjson_internal())) + { + return false; + } + key.clear(); + } + } + } + else + { + if (JSON_UNLIKELY(not sax->start_object(std::size_t(-1)))) + { + return false; + } + + while (current != '}') + { + if (JSON_UNLIKELY(not get_ubjson_string(key, false) or not sax->key(key))) + { + return false; + } + if (JSON_UNLIKELY(not parse_ubjson_internal())) + { + return false; + } + get_ignore_noop(); + key.clear(); } } + + return sax->end_object(); } - template<typename NumberType> - BasicJsonType get_msgpack_array(const NumberType len) + /////////////////////// + // Utility functions // + /////////////////////// + + /*! + @brief get next character from the input + + This function provides the interface to the used input adapter. It does + not throw in case the input reached EOF, but returns a -'ve valued + `std::char_traits<char>::eof()` in that case. + + @return character read from the input + */ + int get() { - BasicJsonType result = value_t::array; - std::generate_n(std::back_inserter(*result.m_value.array), len, [this]() - { - return parse_msgpack_internal(); - }); - return result; + ++chars_read; + return (current = ia->get_character()); } - template<typename NumberType> - BasicJsonType get_msgpack_object(const NumberType len) + /*! + @return character read from the input after ignoring all 'N' entries + */ + int get_ignore_noop() { - BasicJsonType result = value_t::object; - std::generate_n(std::inserter(*result.m_value.object, - result.m_value.object->end()), - len, [this]() + do { get(); - auto key = get_msgpack_string(); - auto val = parse_msgpack_internal(); - return std::make_pair(std::move(key), std::move(val)); - }); - return result; + } + while (current == 'N'); + + return current; } - /*! - @brief check if input ended - @throw parse_error.110 if input ended + /* + @brief read a number from the input + + @tparam NumberType the type of the number + @param[in] format the current format (for diagnostics) + @param[out] result number of type @a NumberType + + @return whether conversion completed + + @note This function needs to respect the system's endianess, because + bytes in CBOR, MessagePack, and UBJSON are stored in network order + (big endian) and therefore need reordering on little endian systems. */ - void check_eof(const bool expect_eof = false) const + template<typename NumberType, bool InputIsLittleEndian = false> + bool get_number(const input_format_t format, NumberType& result) { - if (expect_eof) + // step 1: read input into array with system's byte order + std::array<uint8_t, sizeof(NumberType)> vec; + for (std::size_t i = 0; i < sizeof(NumberType); ++i) { - if (JSON_UNLIKELY(current != std::char_traits<char>::eof())) + get(); + if (JSON_UNLIKELY(not unexpect_eof(format, "number"))) { - JSON_THROW(parse_error::create(110, chars_read, "expected end of input")); + return false; + } + + // reverse byte order prior to conversion if necessary + if (is_little_endian && !InputIsLittleEndian) + { + vec[sizeof(NumberType) - i - 1] = static_cast<uint8_t>(current); + } + else + { + vec[i] = static_cast<uint8_t>(current); // LCOV_EXCL_LINE } } - else + + // step 2: convert array into number of type T and return + std::memcpy(&result, vec.data(), sizeof(NumberType)); + return true; + } + + /*! + @brief create a string by reading characters from the input + + @tparam NumberType the type of the number + @param[in] format the current format (for diagnostics) + @param[in] len number of characters to read + @param[out] result string created by reading @a len bytes + + @return whether string creation completed + + @note We can not reserve @a len bytes for the result, because @a len + may be too large. Usually, @ref unexpect_eof() detects the end of + the input before we run out of string memory. + */ + template<typename NumberType> + bool get_string(const input_format_t format, + const NumberType len, + string_t& result) + { + bool success = true; + std::generate_n(std::back_inserter(result), len, [this, &success, &format]() { - if (JSON_UNLIKELY(current == std::char_traits<char>::eof())) + get(); + if (JSON_UNLIKELY(not unexpect_eof(format, "string"))) { - JSON_THROW(parse_error::create(110, chars_read, "unexpected end of input")); + success = false; } + return static_cast<char>(current); + }); + return success; + } + + /*! + @param[in] format the current format (for diagnostics) + @param[in] context further context information (for diagnostics) + @return whether the last read character is not EOF + */ + bool unexpect_eof(const input_format_t format, const char* context) const + { + if (JSON_UNLIKELY(current == std::char_traits<char>::eof())) + { + return sax->parse_error(chars_read, "<end of file>", + parse_error::create(110, chars_read, exception_message(format, "unexpected end of input", context))); + } + return true; + } + + /*! + @return a string representation of the last read byte + */ + std::string get_token_string() const + { + char cr[3]; + (std::snprintf)(cr, 3, "%.2hhX", static_cast<unsigned char>(current)); + return std::string{cr}; + } + + /*! + @param[in] format the current format + @param[in] detail a detailed error message + @param[in] context further contect information + @return a message string to use in the parse_error exceptions + */ + std::string exception_message(const input_format_t format, + const std::string& detail, + const std::string& context) const + { + std::string error_msg = "syntax error while parsing "; + + switch (format) + { + case input_format_t::cbor: + error_msg += "CBOR"; + break; + + case input_format_t::msgpack: + error_msg += "MessagePack"; + break; + + case input_format_t::ubjson: + error_msg += "UBJSON"; + break; + + case input_format_t::bson: + error_msg += "BSON"; + break; + + // LCOV_EXCL_START + default: + assert(false); + // LCOV_EXCL_STOP } + + return error_msg + " " + context + ": " + detail; } private: @@ -5529,7 +8429,34 @@ class binary_reader /// whether we can assume little endianess const bool is_little_endian = little_endianess(); + + /// the SAX parser + json_sax_t* sax = nullptr; }; +} // namespace detail +} // namespace nlohmann + +// #include <nlohmann/detail/output/binary_writer.hpp> + + +#include <algorithm> // reverse +#include <array> // array +#include <cstdint> // uint8_t, uint16_t, uint32_t, uint64_t +#include <cstring> // memcpy +#include <limits> // numeric_limits + +// #include <nlohmann/detail/input/binary_reader.hpp> + +// #include <nlohmann/detail/output/output_adapters.hpp> + + +namespace nlohmann +{ +namespace detail +{ +/////////////////// +// binary writer // +/////////////////// /*! @brief serialization to CBOR and MessagePack values @@ -5537,6 +8464,8 @@ class binary_reader template<typename BasicJsonType, typename CharType> class binary_writer { + using string_t = typename BasicJsonType::string_t; + public: /*! @brief create a binary writer @@ -5549,7 +8478,28 @@ class binary_writer } /*! - @brief[in] j JSON value to serialize + @param[in] j JSON value to serialize + @pre j.type() == value_t::object + */ + void write_bson(const BasicJsonType& j) + { + switch (j.type()) + { + case value_t::object: + { + write_bson_object(*j.m_value.object); + break; + } + + default: + { + JSON_THROW(type_error::create(317, "to serialize to BSON, top-level type must be object, but is " + std::string(j.type_name()))); + } + } + } + + /*! + @param[in] j JSON value to serialize */ void write_cbor(const BasicJsonType& j) { @@ -5557,15 +8507,15 @@ class binary_writer { case value_t::null: { - oa->write_character(static_cast<CharType>(0xF6)); + oa->write_character(to_char_type(0xF6)); break; } case value_t::boolean: { oa->write_character(j.m_value.boolean - ? static_cast<CharType>(0xF5) - : static_cast<CharType>(0xF4)); + ? to_char_type(0xF5) + : to_char_type(0xF4)); break; } @@ -5582,22 +8532,22 @@ class binary_writer } else if (j.m_value.number_integer <= (std::numeric_limits<uint8_t>::max)()) { - oa->write_character(static_cast<CharType>(0x18)); + oa->write_character(to_char_type(0x18)); write_number(static_cast<uint8_t>(j.m_value.number_integer)); } else if (j.m_value.number_integer <= (std::numeric_limits<uint16_t>::max)()) { - oa->write_character(static_cast<CharType>(0x19)); + oa->write_character(to_char_type(0x19)); write_number(static_cast<uint16_t>(j.m_value.number_integer)); } else if (j.m_value.number_integer <= (std::numeric_limits<uint32_t>::max)()) { - oa->write_character(static_cast<CharType>(0x1A)); + oa->write_character(to_char_type(0x1A)); write_number(static_cast<uint32_t>(j.m_value.number_integer)); } else { - oa->write_character(static_cast<CharType>(0x1B)); + oa->write_character(to_char_type(0x1B)); write_number(static_cast<uint64_t>(j.m_value.number_integer)); } } @@ -5612,22 +8562,22 @@ class binary_writer } else if (positive_number <= (std::numeric_limits<uint8_t>::max)()) { - oa->write_character(static_cast<CharType>(0x38)); + oa->write_character(to_char_type(0x38)); write_number(static_cast<uint8_t>(positive_number)); } else if (positive_number <= (std::numeric_limits<uint16_t>::max)()) { - oa->write_character(static_cast<CharType>(0x39)); + oa->write_character(to_char_type(0x39)); write_number(static_cast<uint16_t>(positive_number)); } else if (positive_number <= (std::numeric_limits<uint32_t>::max)()) { - oa->write_character(static_cast<CharType>(0x3A)); + oa->write_character(to_char_type(0x3A)); write_number(static_cast<uint32_t>(positive_number)); } else { - oa->write_character(static_cast<CharType>(0x3B)); + oa->write_character(to_char_type(0x3B)); write_number(static_cast<uint64_t>(positive_number)); } } @@ -5642,30 +8592,30 @@ class binary_writer } else if (j.m_value.number_unsigned <= (std::numeric_limits<uint8_t>::max)()) { - oa->write_character(static_cast<CharType>(0x18)); + oa->write_character(to_char_type(0x18)); write_number(static_cast<uint8_t>(j.m_value.number_unsigned)); } else if (j.m_value.number_unsigned <= (std::numeric_limits<uint16_t>::max)()) { - oa->write_character(static_cast<CharType>(0x19)); + oa->write_character(to_char_type(0x19)); write_number(static_cast<uint16_t>(j.m_value.number_unsigned)); } else if (j.m_value.number_unsigned <= (std::numeric_limits<uint32_t>::max)()) { - oa->write_character(static_cast<CharType>(0x1A)); + oa->write_character(to_char_type(0x1A)); write_number(static_cast<uint32_t>(j.m_value.number_unsigned)); } else { - oa->write_character(static_cast<CharType>(0x1B)); + oa->write_character(to_char_type(0x1B)); write_number(static_cast<uint64_t>(j.m_value.number_unsigned)); } break; } - case value_t::number_float: // Double-Precision Float + case value_t::number_float: { - oa->write_character(static_cast<CharType>(0xFB)); + oa->write_character(get_cbor_float_prefix(j.m_value.number_float)); write_number(j.m_value.number_float); break; } @@ -5678,25 +8628,25 @@ class binary_writer { write_number(static_cast<uint8_t>(0x60 + N)); } - else if (N <= 0xFF) + else if (N <= (std::numeric_limits<uint8_t>::max)()) { - oa->write_character(static_cast<CharType>(0x78)); + oa->write_character(to_char_type(0x78)); write_number(static_cast<uint8_t>(N)); } - else if (N <= 0xFFFF) + else if (N <= (std::numeric_limits<uint16_t>::max)()) { - oa->write_character(static_cast<CharType>(0x79)); + oa->write_character(to_char_type(0x79)); write_number(static_cast<uint16_t>(N)); } - else if (N <= 0xFFFFFFFF) + else if (N <= (std::numeric_limits<uint32_t>::max)()) { - oa->write_character(static_cast<CharType>(0x7A)); + oa->write_character(to_char_type(0x7A)); write_number(static_cast<uint32_t>(N)); } // LCOV_EXCL_START - else if (N <= 0xFFFFFFFFFFFFFFFF) + else if (N <= (std::numeric_limits<uint64_t>::max)()) { - oa->write_character(static_cast<CharType>(0x7B)); + oa->write_character(to_char_type(0x7B)); write_number(static_cast<uint64_t>(N)); } // LCOV_EXCL_STOP @@ -5716,25 +8666,25 @@ class binary_writer { write_number(static_cast<uint8_t>(0x80 + N)); } - else if (N <= 0xFF) + else if (N <= (std::numeric_limits<uint8_t>::max)()) { - oa->write_character(static_cast<CharType>(0x98)); + oa->write_character(to_char_type(0x98)); write_number(static_cast<uint8_t>(N)); } - else if (N <= 0xFFFF) + else if (N <= (std::numeric_limits<uint16_t>::max)()) { - oa->write_character(static_cast<CharType>(0x99)); + oa->write_character(to_char_type(0x99)); write_number(static_cast<uint16_t>(N)); } - else if (N <= 0xFFFFFFFF) + else if (N <= (std::numeric_limits<uint32_t>::max)()) { - oa->write_character(static_cast<CharType>(0x9A)); + oa->write_character(to_char_type(0x9A)); write_number(static_cast<uint32_t>(N)); } // LCOV_EXCL_START - else if (N <= 0xFFFFFFFFFFFFFFFF) + else if (N <= (std::numeric_limits<uint64_t>::max)()) { - oa->write_character(static_cast<CharType>(0x9B)); + oa->write_character(to_char_type(0x9B)); write_number(static_cast<uint64_t>(N)); } // LCOV_EXCL_STOP @@ -5755,25 +8705,25 @@ class binary_writer { write_number(static_cast<uint8_t>(0xA0 + N)); } - else if (N <= 0xFF) + else if (N <= (std::numeric_limits<uint8_t>::max)()) { - oa->write_character(static_cast<CharType>(0xB8)); + oa->write_character(to_char_type(0xB8)); write_number(static_cast<uint8_t>(N)); } - else if (N <= 0xFFFF) + else if (N <= (std::numeric_limits<uint16_t>::max)()) { - oa->write_character(static_cast<CharType>(0xB9)); + oa->write_character(to_char_type(0xB9)); write_number(static_cast<uint16_t>(N)); } - else if (N <= 0xFFFFFFFF) + else if (N <= (std::numeric_limits<uint32_t>::max)()) { - oa->write_character(static_cast<CharType>(0xBA)); + oa->write_character(to_char_type(0xBA)); write_number(static_cast<uint32_t>(N)); } // LCOV_EXCL_START - else if (N <= 0xFFFFFFFFFFFFFFFF) + else if (N <= (std::numeric_limits<uint64_t>::max)()) { - oa->write_character(static_cast<CharType>(0xBB)); + oa->write_character(to_char_type(0xBB)); write_number(static_cast<uint64_t>(N)); } // LCOV_EXCL_STOP @@ -5793,7 +8743,7 @@ class binary_writer } /*! - @brief[in] j JSON value to serialize + @param[in] j JSON value to serialize */ void write_msgpack(const BasicJsonType& j) { @@ -5801,15 +8751,15 @@ class binary_writer { case value_t::null: // nil { - oa->write_character(static_cast<CharType>(0xC0)); + oa->write_character(to_char_type(0xC0)); break; } case value_t::boolean: // true and false { oa->write_character(j.m_value.boolean - ? static_cast<CharType>(0xC3) - : static_cast<CharType>(0xC2)); + ? to_char_type(0xC3) + : to_char_type(0xC2)); break; } @@ -5828,25 +8778,25 @@ class binary_writer else if (j.m_value.number_unsigned <= (std::numeric_limits<uint8_t>::max)()) { // uint 8 - oa->write_character(static_cast<CharType>(0xCC)); + oa->write_character(to_char_type(0xCC)); write_number(static_cast<uint8_t>(j.m_value.number_integer)); } else if (j.m_value.number_unsigned <= (std::numeric_limits<uint16_t>::max)()) { // uint 16 - oa->write_character(static_cast<CharType>(0xCD)); + oa->write_character(to_char_type(0xCD)); write_number(static_cast<uint16_t>(j.m_value.number_integer)); } else if (j.m_value.number_unsigned <= (std::numeric_limits<uint32_t>::max)()) { // uint 32 - oa->write_character(static_cast<CharType>(0xCE)); + oa->write_character(to_char_type(0xCE)); write_number(static_cast<uint32_t>(j.m_value.number_integer)); } else if (j.m_value.number_unsigned <= (std::numeric_limits<uint64_t>::max)()) { // uint 64 - oa->write_character(static_cast<CharType>(0xCF)); + oa->write_character(to_char_type(0xCF)); write_number(static_cast<uint64_t>(j.m_value.number_integer)); } } @@ -5861,28 +8811,28 @@ class binary_writer j.m_value.number_integer <= (std::numeric_limits<int8_t>::max)()) { // int 8 - oa->write_character(static_cast<CharType>(0xD0)); + oa->write_character(to_char_type(0xD0)); write_number(static_cast<int8_t>(j.m_value.number_integer)); } else if (j.m_value.number_integer >= (std::numeric_limits<int16_t>::min)() and j.m_value.number_integer <= (std::numeric_limits<int16_t>::max)()) { // int 16 - oa->write_character(static_cast<CharType>(0xD1)); + oa->write_character(to_char_type(0xD1)); write_number(static_cast<int16_t>(j.m_value.number_integer)); } else if (j.m_value.number_integer >= (std::numeric_limits<int32_t>::min)() and j.m_value.number_integer <= (std::numeric_limits<int32_t>::max)()) { // int 32 - oa->write_character(static_cast<CharType>(0xD2)); + oa->write_character(to_char_type(0xD2)); write_number(static_cast<int32_t>(j.m_value.number_integer)); } else if (j.m_value.number_integer >= (std::numeric_limits<int64_t>::min)() and j.m_value.number_integer <= (std::numeric_limits<int64_t>::max)()) { // int 64 - oa->write_character(static_cast<CharType>(0xD3)); + oa->write_character(to_char_type(0xD3)); write_number(static_cast<int64_t>(j.m_value.number_integer)); } } @@ -5899,33 +8849,33 @@ class binary_writer else if (j.m_value.number_unsigned <= (std::numeric_limits<uint8_t>::max)()) { // uint 8 - oa->write_character(static_cast<CharType>(0xCC)); + oa->write_character(to_char_type(0xCC)); write_number(static_cast<uint8_t>(j.m_value.number_integer)); } else if (j.m_value.number_unsigned <= (std::numeric_limits<uint16_t>::max)()) { // uint 16 - oa->write_character(static_cast<CharType>(0xCD)); + oa->write_character(to_char_type(0xCD)); write_number(static_cast<uint16_t>(j.m_value.number_integer)); } else if (j.m_value.number_unsigned <= (std::numeric_limits<uint32_t>::max)()) { // uint 32 - oa->write_character(static_cast<CharType>(0xCE)); + oa->write_character(to_char_type(0xCE)); write_number(static_cast<uint32_t>(j.m_value.number_integer)); } else if (j.m_value.number_unsigned <= (std::numeric_limits<uint64_t>::max)()) { // uint 64 - oa->write_character(static_cast<CharType>(0xCF)); + oa->write_character(to_char_type(0xCF)); write_number(static_cast<uint64_t>(j.m_value.number_integer)); } break; } - case value_t::number_float: // float 64 + case value_t::number_float: { - oa->write_character(static_cast<CharType>(0xCB)); + oa->write_character(get_msgpack_float_prefix(j.m_value.number_float)); write_number(j.m_value.number_float); break; } @@ -5939,22 +8889,22 @@ class binary_writer // fixstr write_number(static_cast<uint8_t>(0xA0 | N)); } - else if (N <= 255) + else if (N <= (std::numeric_limits<uint8_t>::max)()) { // str 8 - oa->write_character(static_cast<CharType>(0xD9)); + oa->write_character(to_char_type(0xD9)); write_number(static_cast<uint8_t>(N)); } - else if (N <= 65535) + else if (N <= (std::numeric_limits<uint16_t>::max)()) { // str 16 - oa->write_character(static_cast<CharType>(0xDA)); + oa->write_character(to_char_type(0xDA)); write_number(static_cast<uint16_t>(N)); } - else if (N <= 4294967295) + else if (N <= (std::numeric_limits<uint32_t>::max)()) { // str 32 - oa->write_character(static_cast<CharType>(0xDB)); + oa->write_character(to_char_type(0xDB)); write_number(static_cast<uint32_t>(N)); } @@ -5974,16 +8924,16 @@ class binary_writer // fixarray write_number(static_cast<uint8_t>(0x90 | N)); } - else if (N <= 0xFFFF) + else if (N <= (std::numeric_limits<uint16_t>::max)()) { // array 16 - oa->write_character(static_cast<CharType>(0xDC)); + oa->write_character(to_char_type(0xDC)); write_number(static_cast<uint16_t>(N)); } - else if (N <= 0xFFFFFFFF) + else if (N <= (std::numeric_limits<uint32_t>::max)()) { // array 32 - oa->write_character(static_cast<CharType>(0xDD)); + oa->write_character(to_char_type(0xDD)); write_number(static_cast<uint32_t>(N)); } @@ -6004,16 +8954,16 @@ class binary_writer // fixmap write_number(static_cast<uint8_t>(0x80 | (N & 0xF))); } - else if (N <= 65535) + else if (N <= (std::numeric_limits<uint16_t>::max)()) { // map 16 - oa->write_character(static_cast<CharType>(0xDE)); + oa->write_character(to_char_type(0xDE)); write_number(static_cast<uint16_t>(N)); } - else if (N <= 4294967295) + else if (N <= (std::numeric_limits<uint32_t>::max)()) { // map 32 - oa->write_character(static_cast<CharType>(0xDF)); + oa->write_character(to_char_type(0xDF)); write_number(static_cast<uint32_t>(N)); } @@ -6031,25 +8981,745 @@ class binary_writer } } + /*! + @param[in] j JSON value to serialize + @param[in] use_count whether to use '#' prefixes (optimized format) + @param[in] use_type whether to use '$' prefixes (optimized format) + @param[in] add_prefix whether prefixes need to be used for this value + */ + void write_ubjson(const BasicJsonType& j, const bool use_count, + const bool use_type, const bool add_prefix = true) + { + switch (j.type()) + { + case value_t::null: + { + if (add_prefix) + { + oa->write_character(to_char_type('Z')); + } + break; + } + + case value_t::boolean: + { + if (add_prefix) + { + oa->write_character(j.m_value.boolean + ? to_char_type('T') + : to_char_type('F')); + } + break; + } + + case value_t::number_integer: + { + write_number_with_ubjson_prefix(j.m_value.number_integer, add_prefix); + break; + } + + case value_t::number_unsigned: + { + write_number_with_ubjson_prefix(j.m_value.number_unsigned, add_prefix); + break; + } + + case value_t::number_float: + { + write_number_with_ubjson_prefix(j.m_value.number_float, add_prefix); + break; + } + + case value_t::string: + { + if (add_prefix) + { + oa->write_character(to_char_type('S')); + } + write_number_with_ubjson_prefix(j.m_value.string->size(), true); + oa->write_characters( + reinterpret_cast<const CharType*>(j.m_value.string->c_str()), + j.m_value.string->size()); + break; + } + + case value_t::array: + { + if (add_prefix) + { + oa->write_character(to_char_type('[')); + } + + bool prefix_required = true; + if (use_type and not j.m_value.array->empty()) + { + assert(use_count); + const CharType first_prefix = ubjson_prefix(j.front()); + const bool same_prefix = std::all_of(j.begin() + 1, j.end(), + [this, first_prefix](const BasicJsonType & v) + { + return ubjson_prefix(v) == first_prefix; + }); + + if (same_prefix) + { + prefix_required = false; + oa->write_character(to_char_type('$')); + oa->write_character(first_prefix); + } + } + + if (use_count) + { + oa->write_character(to_char_type('#')); + write_number_with_ubjson_prefix(j.m_value.array->size(), true); + } + + for (const auto& el : *j.m_value.array) + { + write_ubjson(el, use_count, use_type, prefix_required); + } + + if (not use_count) + { + oa->write_character(to_char_type(']')); + } + + break; + } + + case value_t::object: + { + if (add_prefix) + { + oa->write_character(to_char_type('{')); + } + + bool prefix_required = true; + if (use_type and not j.m_value.object->empty()) + { + assert(use_count); + const CharType first_prefix = ubjson_prefix(j.front()); + const bool same_prefix = std::all_of(j.begin(), j.end(), + [this, first_prefix](const BasicJsonType & v) + { + return ubjson_prefix(v) == first_prefix; + }); + + if (same_prefix) + { + prefix_required = false; + oa->write_character(to_char_type('$')); + oa->write_character(first_prefix); + } + } + + if (use_count) + { + oa->write_character(to_char_type('#')); + write_number_with_ubjson_prefix(j.m_value.object->size(), true); + } + + for (const auto& el : *j.m_value.object) + { + write_number_with_ubjson_prefix(el.first.size(), true); + oa->write_characters( + reinterpret_cast<const CharType*>(el.first.c_str()), + el.first.size()); + write_ubjson(el.second, use_count, use_type, prefix_required); + } + + if (not use_count) + { + oa->write_character(to_char_type('}')); + } + + break; + } + + default: + break; + } + } + private: + ////////// + // BSON // + ////////// + + /*! + @return The size of a BSON document entry header, including the id marker + and the entry name size (and its null-terminator). + */ + static std::size_t calc_bson_entry_header_size(const string_t& name) + { + const auto it = name.find(static_cast<typename string_t::value_type>(0)); + if (JSON_UNLIKELY(it != BasicJsonType::string_t::npos)) + { + JSON_THROW(out_of_range::create(409, + "BSON key cannot contain code point U+0000 (at byte " + std::to_string(it) + ")")); + } + + return /*id*/ 1ul + name.size() + /*zero-terminator*/1u; + } + + /*! + @brief Writes the given @a element_type and @a name to the output adapter + */ + void write_bson_entry_header(const string_t& name, + const std::uint8_t element_type) + { + oa->write_character(to_char_type(element_type)); // boolean + oa->write_characters( + reinterpret_cast<const CharType*>(name.c_str()), + name.size() + 1u); + } + + /*! + @brief Writes a BSON element with key @a name and boolean value @a value + */ + void write_bson_boolean(const string_t& name, + const bool value) + { + write_bson_entry_header(name, 0x08); + oa->write_character(value ? to_char_type(0x01) : to_char_type(0x00)); + } + + /*! + @brief Writes a BSON element with key @a name and double value @a value + */ + void write_bson_double(const string_t& name, + const double value) + { + write_bson_entry_header(name, 0x01); + write_number<double, true>(value); + } + + /*! + @return The size of the BSON-encoded string in @a value + */ + static std::size_t calc_bson_string_size(const string_t& value) + { + return sizeof(std::int32_t) + value.size() + 1ul; + } + + /*! + @brief Writes a BSON element with key @a name and string value @a value + */ + void write_bson_string(const string_t& name, + const string_t& value) + { + write_bson_entry_header(name, 0x02); + + write_number<std::int32_t, true>(static_cast<std::int32_t>(value.size() + 1ul)); + oa->write_characters( + reinterpret_cast<const CharType*>(value.c_str()), + value.size() + 1); + } + + /*! + @brief Writes a BSON element with key @a name and null value + */ + void write_bson_null(const string_t& name) + { + write_bson_entry_header(name, 0x0A); + } + + /*! + @return The size of the BSON-encoded integer @a value + */ + static std::size_t calc_bson_integer_size(const std::int64_t value) + { + if ((std::numeric_limits<std::int32_t>::min)() <= value and value <= (std::numeric_limits<std::int32_t>::max)()) + { + return sizeof(std::int32_t); + } + else + { + return sizeof(std::int64_t); + } + } + + /*! + @brief Writes a BSON element with key @a name and integer @a value + */ + void write_bson_integer(const string_t& name, + const std::int64_t value) + { + if ((std::numeric_limits<std::int32_t>::min)() <= value and value <= (std::numeric_limits<std::int32_t>::max)()) + { + write_bson_entry_header(name, 0x10); // int32 + write_number<std::int32_t, true>(static_cast<std::int32_t>(value)); + } + else + { + write_bson_entry_header(name, 0x12); // int64 + write_number<std::int64_t, true>(static_cast<std::int64_t>(value)); + } + } + + /*! + @return The size of the BSON-encoded unsigned integer in @a j + */ + static constexpr std::size_t calc_bson_unsigned_size(const std::uint64_t value) noexcept + { + return (value <= static_cast<std::uint64_t>((std::numeric_limits<std::int32_t>::max)())) + ? sizeof(std::int32_t) + : sizeof(std::int64_t); + } + + /*! + @brief Writes a BSON element with key @a name and unsigned @a value + */ + void write_bson_unsigned(const string_t& name, + const std::uint64_t value) + { + if (value <= static_cast<std::uint64_t>((std::numeric_limits<std::int32_t>::max)())) + { + write_bson_entry_header(name, 0x10 /* int32 */); + write_number<std::int32_t, true>(static_cast<std::int32_t>(value)); + } + else if (value <= static_cast<std::uint64_t>((std::numeric_limits<std::int64_t>::max)())) + { + write_bson_entry_header(name, 0x12 /* int64 */); + write_number<std::int64_t, true>(static_cast<std::int64_t>(value)); + } + else + { + JSON_THROW(out_of_range::create(407, "integer number " + std::to_string(value) + " cannot be represented by BSON as it does not fit int64")); + } + } + + /*! + @brief Writes a BSON element with key @a name and object @a value + */ + void write_bson_object_entry(const string_t& name, + const typename BasicJsonType::object_t& value) + { + write_bson_entry_header(name, 0x03); // object + write_bson_object(value); + } + + /*! + @return The size of the BSON-encoded array @a value + */ + static std::size_t calc_bson_array_size(const typename BasicJsonType::array_t& value) + { + std::size_t embedded_document_size = 0ul; + std::size_t array_index = 0ul; + + for (const auto& el : value) + { + embedded_document_size += calc_bson_element_size(std::to_string(array_index++), el); + } + + return sizeof(std::int32_t) + embedded_document_size + 1ul; + } + + /*! + @brief Writes a BSON element with key @a name and array @a value + */ + void write_bson_array(const string_t& name, + const typename BasicJsonType::array_t& value) + { + write_bson_entry_header(name, 0x04); // array + write_number<std::int32_t, true>(static_cast<std::int32_t>(calc_bson_array_size(value))); + + std::size_t array_index = 0ul; + + for (const auto& el : value) + { + write_bson_element(std::to_string(array_index++), el); + } + + oa->write_character(to_char_type(0x00)); + } + + /*! + @brief Calculates the size necessary to serialize the JSON value @a j with its @a name + @return The calculated size for the BSON document entry for @a j with the given @a name. + */ + static std::size_t calc_bson_element_size(const string_t& name, + const BasicJsonType& j) + { + const auto header_size = calc_bson_entry_header_size(name); + switch (j.type()) + { + case value_t::object: + return header_size + calc_bson_object_size(*j.m_value.object); + + case value_t::array: + return header_size + calc_bson_array_size(*j.m_value.array); + + case value_t::boolean: + return header_size + 1ul; + + case value_t::number_float: + return header_size + 8ul; + + case value_t::number_integer: + return header_size + calc_bson_integer_size(j.m_value.number_integer); + + case value_t::number_unsigned: + return header_size + calc_bson_unsigned_size(j.m_value.number_unsigned); + + case value_t::string: + return header_size + calc_bson_string_size(*j.m_value.string); + + case value_t::null: + return header_size + 0ul; + + // LCOV_EXCL_START + default: + assert(false); + return 0ul; + // LCOV_EXCL_STOP + }; + } + + /*! + @brief Serializes the JSON value @a j to BSON and associates it with the + key @a name. + @param name The name to associate with the JSON entity @a j within the + current BSON document + @return The size of the BSON entry + */ + void write_bson_element(const string_t& name, + const BasicJsonType& j) + { + switch (j.type()) + { + case value_t::object: + return write_bson_object_entry(name, *j.m_value.object); + + case value_t::array: + return write_bson_array(name, *j.m_value.array); + + case value_t::boolean: + return write_bson_boolean(name, j.m_value.boolean); + + case value_t::number_float: + return write_bson_double(name, j.m_value.number_float); + + case value_t::number_integer: + return write_bson_integer(name, j.m_value.number_integer); + + case value_t::number_unsigned: + return write_bson_unsigned(name, j.m_value.number_unsigned); + + case value_t::string: + return write_bson_string(name, *j.m_value.string); + + case value_t::null: + return write_bson_null(name); + + // LCOV_EXCL_START + default: + assert(false); + return; + // LCOV_EXCL_STOP + }; + } + + /*! + @brief Calculates the size of the BSON serialization of the given + JSON-object @a j. + @param[in] j JSON value to serialize + @pre j.type() == value_t::object + */ + static std::size_t calc_bson_object_size(const typename BasicJsonType::object_t& value) + { + std::size_t document_size = std::accumulate(value.begin(), value.end(), 0ul, + [](size_t result, const typename BasicJsonType::object_t::value_type & el) + { + return result += calc_bson_element_size(el.first, el.second); + }); + + return sizeof(std::int32_t) + document_size + 1ul; + } + + /*! + @param[in] j JSON value to serialize + @pre j.type() == value_t::object + */ + void write_bson_object(const typename BasicJsonType::object_t& value) + { + write_number<std::int32_t, true>(static_cast<std::int32_t>(calc_bson_object_size(value))); + + for (const auto& el : value) + { + write_bson_element(el.first, el.second); + } + + oa->write_character(to_char_type(0x00)); + } + + ////////// + // CBOR // + ////////// + + static constexpr CharType get_cbor_float_prefix(float /*unused*/) + { + return to_char_type(0xFA); // Single-Precision Float + } + + static constexpr CharType get_cbor_float_prefix(double /*unused*/) + { + return to_char_type(0xFB); // Double-Precision Float + } + + ///////////// + // MsgPack // + ///////////// + + static constexpr CharType get_msgpack_float_prefix(float /*unused*/) + { + return to_char_type(0xCA); // float 32 + } + + static constexpr CharType get_msgpack_float_prefix(double /*unused*/) + { + return to_char_type(0xCB); // float 64 + } + + //////////// + // UBJSON // + //////////// + + // UBJSON: write number (floating point) + template<typename NumberType, typename std::enable_if< + std::is_floating_point<NumberType>::value, int>::type = 0> + void write_number_with_ubjson_prefix(const NumberType n, + const bool add_prefix) + { + if (add_prefix) + { + oa->write_character(get_ubjson_float_prefix(n)); + } + write_number(n); + } + + // UBJSON: write number (unsigned integer) + template<typename NumberType, typename std::enable_if< + std::is_unsigned<NumberType>::value, int>::type = 0> + void write_number_with_ubjson_prefix(const NumberType n, + const bool add_prefix) + { + if (n <= static_cast<uint64_t>((std::numeric_limits<int8_t>::max)())) + { + if (add_prefix) + { + oa->write_character(to_char_type('i')); // int8 + } + write_number(static_cast<uint8_t>(n)); + } + else if (n <= (std::numeric_limits<uint8_t>::max)()) + { + if (add_prefix) + { + oa->write_character(to_char_type('U')); // uint8 + } + write_number(static_cast<uint8_t>(n)); + } + else if (n <= static_cast<uint64_t>((std::numeric_limits<int16_t>::max)())) + { + if (add_prefix) + { + oa->write_character(to_char_type('I')); // int16 + } + write_number(static_cast<int16_t>(n)); + } + else if (n <= static_cast<uint64_t>((std::numeric_limits<int32_t>::max)())) + { + if (add_prefix) + { + oa->write_character(to_char_type('l')); // int32 + } + write_number(static_cast<int32_t>(n)); + } + else if (n <= static_cast<uint64_t>((std::numeric_limits<int64_t>::max)())) + { + if (add_prefix) + { + oa->write_character(to_char_type('L')); // int64 + } + write_number(static_cast<int64_t>(n)); + } + else + { + JSON_THROW(out_of_range::create(407, "integer number " + std::to_string(n) + " cannot be represented by UBJSON as it does not fit int64")); + } + } + + // UBJSON: write number (signed integer) + template<typename NumberType, typename std::enable_if< + std::is_signed<NumberType>::value and + not std::is_floating_point<NumberType>::value, int>::type = 0> + void write_number_with_ubjson_prefix(const NumberType n, + const bool add_prefix) + { + if ((std::numeric_limits<int8_t>::min)() <= n and n <= (std::numeric_limits<int8_t>::max)()) + { + if (add_prefix) + { + oa->write_character(to_char_type('i')); // int8 + } + write_number(static_cast<int8_t>(n)); + } + else if (static_cast<int64_t>((std::numeric_limits<uint8_t>::min)()) <= n and n <= static_cast<int64_t>((std::numeric_limits<uint8_t>::max)())) + { + if (add_prefix) + { + oa->write_character(to_char_type('U')); // uint8 + } + write_number(static_cast<uint8_t>(n)); + } + else if ((std::numeric_limits<int16_t>::min)() <= n and n <= (std::numeric_limits<int16_t>::max)()) + { + if (add_prefix) + { + oa->write_character(to_char_type('I')); // int16 + } + write_number(static_cast<int16_t>(n)); + } + else if ((std::numeric_limits<int32_t>::min)() <= n and n <= (std::numeric_limits<int32_t>::max)()) + { + if (add_prefix) + { + oa->write_character(to_char_type('l')); // int32 + } + write_number(static_cast<int32_t>(n)); + } + else if ((std::numeric_limits<int64_t>::min)() <= n and n <= (std::numeric_limits<int64_t>::max)()) + { + if (add_prefix) + { + oa->write_character(to_char_type('L')); // int64 + } + write_number(static_cast<int64_t>(n)); + } + // LCOV_EXCL_START + else + { + JSON_THROW(out_of_range::create(407, "integer number " + std::to_string(n) + " cannot be represented by UBJSON as it does not fit int64")); + } + // LCOV_EXCL_STOP + } + + /*! + @brief determine the type prefix of container values + + @note This function does not need to be 100% accurate when it comes to + integer limits. In case a number exceeds the limits of int64_t, + this will be detected by a later call to function + write_number_with_ubjson_prefix. Therefore, we return 'L' for any + value that does not fit the previous limits. + */ + CharType ubjson_prefix(const BasicJsonType& j) const noexcept + { + switch (j.type()) + { + case value_t::null: + return 'Z'; + + case value_t::boolean: + return j.m_value.boolean ? 'T' : 'F'; + + case value_t::number_integer: + { + if ((std::numeric_limits<int8_t>::min)() <= j.m_value.number_integer and j.m_value.number_integer <= (std::numeric_limits<int8_t>::max)()) + { + return 'i'; + } + if ((std::numeric_limits<uint8_t>::min)() <= j.m_value.number_integer and j.m_value.number_integer <= (std::numeric_limits<uint8_t>::max)()) + { + return 'U'; + } + if ((std::numeric_limits<int16_t>::min)() <= j.m_value.number_integer and j.m_value.number_integer <= (std::numeric_limits<int16_t>::max)()) + { + return 'I'; + } + if ((std::numeric_limits<int32_t>::min)() <= j.m_value.number_integer and j.m_value.number_integer <= (std::numeric_limits<int32_t>::max)()) + { + return 'l'; + } + // no check and assume int64_t (see note above) + return 'L'; + } + + case value_t::number_unsigned: + { + if (j.m_value.number_unsigned <= (std::numeric_limits<int8_t>::max)()) + { + return 'i'; + } + if (j.m_value.number_unsigned <= (std::numeric_limits<uint8_t>::max)()) + { + return 'U'; + } + if (j.m_value.number_unsigned <= (std::numeric_limits<int16_t>::max)()) + { + return 'I'; + } + if (j.m_value.number_unsigned <= (std::numeric_limits<int32_t>::max)()) + { + return 'l'; + } + // no check and assume int64_t (see note above) + return 'L'; + } + + case value_t::number_float: + return get_ubjson_float_prefix(j.m_value.number_float); + + case value_t::string: + return 'S'; + + case value_t::array: + return '['; + + case value_t::object: + return '{'; + + default: // discarded values + return 'N'; + } + } + + static constexpr CharType get_ubjson_float_prefix(float /*unused*/) + { + return 'd'; // float 32 + } + + static constexpr CharType get_ubjson_float_prefix(double /*unused*/) + { + return 'D'; // float 64 + } + + /////////////////////// + // Utility functions // + /////////////////////// + /* @brief write a number to output input - @param[in] n number of type @a NumberType @tparam NumberType the type of the number + @tparam OutputIsLittleEndian Set to true if output data is + required to be little endian @note This function needs to respect the system's endianess, because bytes - in CBOR and MessagePack are stored in network order (big endian) and - therefore need reordering on little endian systems. + in CBOR, MessagePack, and UBJSON are stored in network order (big + endian) and therefore need reordering on little endian systems. */ - template<typename NumberType> void write_number(NumberType n) + template<typename NumberType, bool OutputIsLittleEndian = false> + void write_number(const NumberType n) { // step 1: write number to array of length NumberType std::array<CharType, sizeof(NumberType)> vec; std::memcpy(vec.data(), &n, sizeof(NumberType)); // step 2: write array to output (with possible reordering) - if (is_little_endian) + if (is_little_endian and not OutputIsLittleEndian) { // reverse byte order prior to conversion if necessary std::reverse(vec.begin(), vec.end()); @@ -6058,6 +9728,47 @@ class binary_writer oa->write_characters(vec.data(), sizeof(NumberType)); } + public: + // The following to_char_type functions are implement the conversion + // between uint8_t and CharType. In case CharType is not unsigned, + // such a conversion is required to allow values greater than 128. + // See <https://github.com/nlohmann/json/issues/1286> for a discussion. + template < typename C = CharType, + enable_if_t < std::is_signed<C>::value and std::is_signed<char>::value > * = nullptr > + static constexpr CharType to_char_type(std::uint8_t x) noexcept + { + return *reinterpret_cast<char*>(&x); + } + + template < typename C = CharType, + enable_if_t < std::is_signed<C>::value and std::is_unsigned<char>::value > * = nullptr > + static CharType to_char_type(std::uint8_t x) noexcept + { + static_assert(sizeof(std::uint8_t) == sizeof(CharType), "size of CharType must be equal to std::uint8_t"); + static_assert(std::is_pod<CharType>::value, "CharType must be POD"); + CharType result; + std::memcpy(&result, &x, sizeof(x)); + return result; + } + + template<typename C = CharType, + enable_if_t<std::is_unsigned<C>::value>* = nullptr> + static constexpr CharType to_char_type(std::uint8_t x) noexcept + { + return x; + } + + template < typename InputCharType, typename C = CharType, + enable_if_t < + std::is_signed<C>::value and + std::is_signed<char>::value and + std::is_same<char, typename std::remove_cv<InputCharType>::type>::value + > * = nullptr > + static constexpr CharType to_char_type(InputCharType x) noexcept + { + return x; + } + private: /// whether we can assume little endianess const bool is_little_endian = binary_reader<BasicJsonType>::little_endianess(); @@ -6065,11 +9776,1150 @@ class binary_writer /// the output output_adapter_t<CharType> oa = nullptr; }; +} // namespace detail +} // namespace nlohmann +// #include <nlohmann/detail/output/serializer.hpp> + + +#include <algorithm> // reverse, remove, fill, find, none_of +#include <array> // array +#include <cassert> // assert +#include <ciso646> // and, or +#include <clocale> // localeconv, lconv +#include <cmath> // labs, isfinite, isnan, signbit +#include <cstddef> // size_t, ptrdiff_t +#include <cstdint> // uint8_t +#include <cstdio> // snprintf +#include <limits> // numeric_limits +#include <string> // string +#include <type_traits> // is_same + +// #include <nlohmann/detail/exceptions.hpp> + +// #include <nlohmann/detail/conversions/to_chars.hpp> + + +#include <cassert> // assert +#include <ciso646> // or, and, not +#include <cmath> // signbit, isfinite +#include <cstdint> // intN_t, uintN_t +#include <cstring> // memcpy, memmove + +namespace nlohmann +{ +namespace detail +{ + +/*! +@brief implements the Grisu2 algorithm for binary to decimal floating-point +conversion. + +This implementation is a slightly modified version of the reference +implementation which may be obtained from +http://florian.loitsch.com/publications (bench.tar.gz). + +The code is distributed under the MIT license, Copyright (c) 2009 Florian Loitsch. + +For a detailed description of the algorithm see: + +[1] Loitsch, "Printing Floating-Point Numbers Quickly and Accurately with + Integers", Proceedings of the ACM SIGPLAN 2010 Conference on Programming + Language Design and Implementation, PLDI 2010 +[2] Burger, Dybvig, "Printing Floating-Point Numbers Quickly and Accurately", + Proceedings of the ACM SIGPLAN 1996 Conference on Programming Language + Design and Implementation, PLDI 1996 +*/ +namespace dtoa_impl +{ + +template <typename Target, typename Source> +Target reinterpret_bits(const Source source) +{ + static_assert(sizeof(Target) == sizeof(Source), "size mismatch"); + + Target target; + std::memcpy(&target, &source, sizeof(Source)); + return target; +} + +struct diyfp // f * 2^e +{ + static constexpr int kPrecision = 64; // = q + + uint64_t f = 0; + int e = 0; + + constexpr diyfp(uint64_t f_, int e_) noexcept : f(f_), e(e_) {} + + /*! + @brief returns x - y + @pre x.e == y.e and x.f >= y.f + */ + static diyfp sub(const diyfp& x, const diyfp& y) noexcept + { + assert(x.e == y.e); + assert(x.f >= y.f); + + return {x.f - y.f, x.e}; + } + + /*! + @brief returns x * y + @note The result is rounded. (Only the upper q bits are returned.) + */ + static diyfp mul(const diyfp& x, const diyfp& y) noexcept + { + static_assert(kPrecision == 64, "internal error"); + + // Computes: + // f = round((x.f * y.f) / 2^q) + // e = x.e + y.e + q + + // Emulate the 64-bit * 64-bit multiplication: + // + // p = u * v + // = (u_lo + 2^32 u_hi) (v_lo + 2^32 v_hi) + // = (u_lo v_lo ) + 2^32 ((u_lo v_hi ) + (u_hi v_lo )) + 2^64 (u_hi v_hi ) + // = (p0 ) + 2^32 ((p1 ) + (p2 )) + 2^64 (p3 ) + // = (p0_lo + 2^32 p0_hi) + 2^32 ((p1_lo + 2^32 p1_hi) + (p2_lo + 2^32 p2_hi)) + 2^64 (p3 ) + // = (p0_lo ) + 2^32 (p0_hi + p1_lo + p2_lo ) + 2^64 (p1_hi + p2_hi + p3) + // = (p0_lo ) + 2^32 (Q ) + 2^64 (H ) + // = (p0_lo ) + 2^32 (Q_lo + 2^32 Q_hi ) + 2^64 (H ) + // + // (Since Q might be larger than 2^32 - 1) + // + // = (p0_lo + 2^32 Q_lo) + 2^64 (Q_hi + H) + // + // (Q_hi + H does not overflow a 64-bit int) + // + // = p_lo + 2^64 p_hi + + const uint64_t u_lo = x.f & 0xFFFFFFFF; + const uint64_t u_hi = x.f >> 32; + const uint64_t v_lo = y.f & 0xFFFFFFFF; + const uint64_t v_hi = y.f >> 32; + + const uint64_t p0 = u_lo * v_lo; + const uint64_t p1 = u_lo * v_hi; + const uint64_t p2 = u_hi * v_lo; + const uint64_t p3 = u_hi * v_hi; + + const uint64_t p0_hi = p0 >> 32; + const uint64_t p1_lo = p1 & 0xFFFFFFFF; + const uint64_t p1_hi = p1 >> 32; + const uint64_t p2_lo = p2 & 0xFFFFFFFF; + const uint64_t p2_hi = p2 >> 32; + + uint64_t Q = p0_hi + p1_lo + p2_lo; + + // The full product might now be computed as + // + // p_hi = p3 + p2_hi + p1_hi + (Q >> 32) + // p_lo = p0_lo + (Q << 32) + // + // But in this particular case here, the full p_lo is not required. + // Effectively we only need to add the highest bit in p_lo to p_hi (and + // Q_hi + 1 does not overflow). + + Q += uint64_t{1} << (64 - 32 - 1); // round, ties up + + const uint64_t h = p3 + p2_hi + p1_hi + (Q >> 32); + + return {h, x.e + y.e + 64}; + } + + /*! + @brief normalize x such that the significand is >= 2^(q-1) + @pre x.f != 0 + */ + static diyfp normalize(diyfp x) noexcept + { + assert(x.f != 0); + + while ((x.f >> 63) == 0) + { + x.f <<= 1; + x.e--; + } + + return x; + } + + /*! + @brief normalize x such that the result has the exponent E + @pre e >= x.e and the upper e - x.e bits of x.f must be zero. + */ + static diyfp normalize_to(const diyfp& x, const int target_exponent) noexcept + { + const int delta = x.e - target_exponent; + + assert(delta >= 0); + assert(((x.f << delta) >> delta) == x.f); + + return {x.f << delta, target_exponent}; + } +}; + +struct boundaries +{ + diyfp w; + diyfp minus; + diyfp plus; +}; + +/*! +Compute the (normalized) diyfp representing the input number 'value' and its +boundaries. + +@pre value must be finite and positive +*/ +template <typename FloatType> +boundaries compute_boundaries(FloatType value) +{ + assert(std::isfinite(value)); + assert(value > 0); + + // Convert the IEEE representation into a diyfp. + // + // If v is denormal: + // value = 0.F * 2^(1 - bias) = ( F) * 2^(1 - bias - (p-1)) + // If v is normalized: + // value = 1.F * 2^(E - bias) = (2^(p-1) + F) * 2^(E - bias - (p-1)) + + static_assert(std::numeric_limits<FloatType>::is_iec559, + "internal error: dtoa_short requires an IEEE-754 floating-point implementation"); + + constexpr int kPrecision = std::numeric_limits<FloatType>::digits; // = p (includes the hidden bit) + constexpr int kBias = std::numeric_limits<FloatType>::max_exponent - 1 + (kPrecision - 1); + constexpr int kMinExp = 1 - kBias; + constexpr uint64_t kHiddenBit = uint64_t{1} << (kPrecision - 1); // = 2^(p-1) + + using bits_type = typename std::conditional< kPrecision == 24, uint32_t, uint64_t >::type; + + const uint64_t bits = reinterpret_bits<bits_type>(value); + const uint64_t E = bits >> (kPrecision - 1); + const uint64_t F = bits & (kHiddenBit - 1); + + const bool is_denormal = (E == 0); + const diyfp v = is_denormal + ? diyfp(F, kMinExp) + : diyfp(F + kHiddenBit, static_cast<int>(E) - kBias); + + // Compute the boundaries m- and m+ of the floating-point value + // v = f * 2^e. + // + // Determine v- and v+, the floating-point predecessor and successor if v, + // respectively. + // + // v- = v - 2^e if f != 2^(p-1) or e == e_min (A) + // = v - 2^(e-1) if f == 2^(p-1) and e > e_min (B) + // + // v+ = v + 2^e + // + // Let m- = (v- + v) / 2 and m+ = (v + v+) / 2. All real numbers _strictly_ + // between m- and m+ round to v, regardless of how the input rounding + // algorithm breaks ties. + // + // ---+-------------+-------------+-------------+-------------+--- (A) + // v- m- v m+ v+ + // + // -----------------+------+------+-------------+-------------+--- (B) + // v- m- v m+ v+ + + const bool lower_boundary_is_closer = (F == 0 and E > 1); + const diyfp m_plus = diyfp(2 * v.f + 1, v.e - 1); + const diyfp m_minus = lower_boundary_is_closer + ? diyfp(4 * v.f - 1, v.e - 2) // (B) + : diyfp(2 * v.f - 1, v.e - 1); // (A) + + // Determine the normalized w+ = m+. + const diyfp w_plus = diyfp::normalize(m_plus); + + // Determine w- = m- such that e_(w-) = e_(w+). + const diyfp w_minus = diyfp::normalize_to(m_minus, w_plus.e); + + return {diyfp::normalize(v), w_minus, w_plus}; +} + +// Given normalized diyfp w, Grisu needs to find a (normalized) cached +// power-of-ten c, such that the exponent of the product c * w = f * 2^e lies +// within a certain range [alpha, gamma] (Definition 3.2 from [1]) +// +// alpha <= e = e_c + e_w + q <= gamma +// +// or +// +// f_c * f_w * 2^alpha <= f_c 2^(e_c) * f_w 2^(e_w) * 2^q +// <= f_c * f_w * 2^gamma +// +// Since c and w are normalized, i.e. 2^(q-1) <= f < 2^q, this implies +// +// 2^(q-1) * 2^(q-1) * 2^alpha <= c * w * 2^q < 2^q * 2^q * 2^gamma +// +// or +// +// 2^(q - 2 + alpha) <= c * w < 2^(q + gamma) +// +// The choice of (alpha,gamma) determines the size of the table and the form of +// the digit generation procedure. Using (alpha,gamma)=(-60,-32) works out well +// in practice: +// +// The idea is to cut the number c * w = f * 2^e into two parts, which can be +// processed independently: An integral part p1, and a fractional part p2: +// +// f * 2^e = ( (f div 2^-e) * 2^-e + (f mod 2^-e) ) * 2^e +// = (f div 2^-e) + (f mod 2^-e) * 2^e +// = p1 + p2 * 2^e +// +// The conversion of p1 into decimal form requires a series of divisions and +// modulos by (a power of) 10. These operations are faster for 32-bit than for +// 64-bit integers, so p1 should ideally fit into a 32-bit integer. This can be +// achieved by choosing +// +// -e >= 32 or e <= -32 := gamma +// +// In order to convert the fractional part +// +// p2 * 2^e = p2 / 2^-e = d[-1] / 10^1 + d[-2] / 10^2 + ... +// +// into decimal form, the fraction is repeatedly multiplied by 10 and the digits +// d[-i] are extracted in order: +// +// (10 * p2) div 2^-e = d[-1] +// (10 * p2) mod 2^-e = d[-2] / 10^1 + ... +// +// The multiplication by 10 must not overflow. It is sufficient to choose +// +// 10 * p2 < 16 * p2 = 2^4 * p2 <= 2^64. +// +// Since p2 = f mod 2^-e < 2^-e, +// +// -e <= 60 or e >= -60 := alpha + +constexpr int kAlpha = -60; +constexpr int kGamma = -32; + +struct cached_power // c = f * 2^e ~= 10^k +{ + uint64_t f; + int e; + int k; +}; + +/*! +For a normalized diyfp w = f * 2^e, this function returns a (normalized) cached +power-of-ten c = f_c * 2^e_c, such that the exponent of the product w * c +satisfies (Definition 3.2 from [1]) + + alpha <= e_c + e + q <= gamma. +*/ +inline cached_power get_cached_power_for_binary_exponent(int e) +{ + // Now + // + // alpha <= e_c + e + q <= gamma (1) + // ==> f_c * 2^alpha <= c * 2^e * 2^q + // + // and since the c's are normalized, 2^(q-1) <= f_c, + // + // ==> 2^(q - 1 + alpha) <= c * 2^(e + q) + // ==> 2^(alpha - e - 1) <= c + // + // If c were an exakt power of ten, i.e. c = 10^k, one may determine k as + // + // k = ceil( log_10( 2^(alpha - e - 1) ) ) + // = ceil( (alpha - e - 1) * log_10(2) ) + // + // From the paper: + // "In theory the result of the procedure could be wrong since c is rounded, + // and the computation itself is approximated [...]. In practice, however, + // this simple function is sufficient." + // + // For IEEE double precision floating-point numbers converted into + // normalized diyfp's w = f * 2^e, with q = 64, + // + // e >= -1022 (min IEEE exponent) + // -52 (p - 1) + // -52 (p - 1, possibly normalize denormal IEEE numbers) + // -11 (normalize the diyfp) + // = -1137 + // + // and + // + // e <= +1023 (max IEEE exponent) + // -52 (p - 1) + // -11 (normalize the diyfp) + // = 960 + // + // This binary exponent range [-1137,960] results in a decimal exponent + // range [-307,324]. One does not need to store a cached power for each + // k in this range. For each such k it suffices to find a cached power + // such that the exponent of the product lies in [alpha,gamma]. + // This implies that the difference of the decimal exponents of adjacent + // table entries must be less than or equal to + // + // floor( (gamma - alpha) * log_10(2) ) = 8. + // + // (A smaller distance gamma-alpha would require a larger table.) + + // NB: + // Actually this function returns c, such that -60 <= e_c + e + 64 <= -34. + + constexpr int kCachedPowersSize = 79; + constexpr int kCachedPowersMinDecExp = -300; + constexpr int kCachedPowersDecStep = 8; + + static constexpr cached_power kCachedPowers[] = + { + { 0xAB70FE17C79AC6CA, -1060, -300 }, + { 0xFF77B1FCBEBCDC4F, -1034, -292 }, + { 0xBE5691EF416BD60C, -1007, -284 }, + { 0x8DD01FAD907FFC3C, -980, -276 }, + { 0xD3515C2831559A83, -954, -268 }, + { 0x9D71AC8FADA6C9B5, -927, -260 }, + { 0xEA9C227723EE8BCB, -901, -252 }, + { 0xAECC49914078536D, -874, -244 }, + { 0x823C12795DB6CE57, -847, -236 }, + { 0xC21094364DFB5637, -821, -228 }, + { 0x9096EA6F3848984F, -794, -220 }, + { 0xD77485CB25823AC7, -768, -212 }, + { 0xA086CFCD97BF97F4, -741, -204 }, + { 0xEF340A98172AACE5, -715, -196 }, + { 0xB23867FB2A35B28E, -688, -188 }, + { 0x84C8D4DFD2C63F3B, -661, -180 }, + { 0xC5DD44271AD3CDBA, -635, -172 }, + { 0x936B9FCEBB25C996, -608, -164 }, + { 0xDBAC6C247D62A584, -582, -156 }, + { 0xA3AB66580D5FDAF6, -555, -148 }, + { 0xF3E2F893DEC3F126, -529, -140 }, + { 0xB5B5ADA8AAFF80B8, -502, -132 }, + { 0x87625F056C7C4A8B, -475, -124 }, + { 0xC9BCFF6034C13053, -449, -116 }, + { 0x964E858C91BA2655, -422, -108 }, + { 0xDFF9772470297EBD, -396, -100 }, + { 0xA6DFBD9FB8E5B88F, -369, -92 }, + { 0xF8A95FCF88747D94, -343, -84 }, + { 0xB94470938FA89BCF, -316, -76 }, + { 0x8A08F0F8BF0F156B, -289, -68 }, + { 0xCDB02555653131B6, -263, -60 }, + { 0x993FE2C6D07B7FAC, -236, -52 }, + { 0xE45C10C42A2B3B06, -210, -44 }, + { 0xAA242499697392D3, -183, -36 }, + { 0xFD87B5F28300CA0E, -157, -28 }, + { 0xBCE5086492111AEB, -130, -20 }, + { 0x8CBCCC096F5088CC, -103, -12 }, + { 0xD1B71758E219652C, -77, -4 }, + { 0x9C40000000000000, -50, 4 }, + { 0xE8D4A51000000000, -24, 12 }, + { 0xAD78EBC5AC620000, 3, 20 }, + { 0x813F3978F8940984, 30, 28 }, + { 0xC097CE7BC90715B3, 56, 36 }, + { 0x8F7E32CE7BEA5C70, 83, 44 }, + { 0xD5D238A4ABE98068, 109, 52 }, + { 0x9F4F2726179A2245, 136, 60 }, + { 0xED63A231D4C4FB27, 162, 68 }, + { 0xB0DE65388CC8ADA8, 189, 76 }, + { 0x83C7088E1AAB65DB, 216, 84 }, + { 0xC45D1DF942711D9A, 242, 92 }, + { 0x924D692CA61BE758, 269, 100 }, + { 0xDA01EE641A708DEA, 295, 108 }, + { 0xA26DA3999AEF774A, 322, 116 }, + { 0xF209787BB47D6B85, 348, 124 }, + { 0xB454E4A179DD1877, 375, 132 }, + { 0x865B86925B9BC5C2, 402, 140 }, + { 0xC83553C5C8965D3D, 428, 148 }, + { 0x952AB45CFA97A0B3, 455, 156 }, + { 0xDE469FBD99A05FE3, 481, 164 }, + { 0xA59BC234DB398C25, 508, 172 }, + { 0xF6C69A72A3989F5C, 534, 180 }, + { 0xB7DCBF5354E9BECE, 561, 188 }, + { 0x88FCF317F22241E2, 588, 196 }, + { 0xCC20CE9BD35C78A5, 614, 204 }, + { 0x98165AF37B2153DF, 641, 212 }, + { 0xE2A0B5DC971F303A, 667, 220 }, + { 0xA8D9D1535CE3B396, 694, 228 }, + { 0xFB9B7CD9A4A7443C, 720, 236 }, + { 0xBB764C4CA7A44410, 747, 244 }, + { 0x8BAB8EEFB6409C1A, 774, 252 }, + { 0xD01FEF10A657842C, 800, 260 }, + { 0x9B10A4E5E9913129, 827, 268 }, + { 0xE7109BFBA19C0C9D, 853, 276 }, + { 0xAC2820D9623BF429, 880, 284 }, + { 0x80444B5E7AA7CF85, 907, 292 }, + { 0xBF21E44003ACDD2D, 933, 300 }, + { 0x8E679C2F5E44FF8F, 960, 308 }, + { 0xD433179D9C8CB841, 986, 316 }, + { 0x9E19DB92B4E31BA9, 1013, 324 }, + }; + + // This computation gives exactly the same results for k as + // k = ceil((kAlpha - e - 1) * 0.30102999566398114) + // for |e| <= 1500, but doesn't require floating-point operations. + // NB: log_10(2) ~= 78913 / 2^18 + assert(e >= -1500); + assert(e <= 1500); + const int f = kAlpha - e - 1; + const int k = (f * 78913) / (1 << 18) + static_cast<int>(f > 0); + + const int index = (-kCachedPowersMinDecExp + k + (kCachedPowersDecStep - 1)) / kCachedPowersDecStep; + assert(index >= 0); + assert(index < kCachedPowersSize); + static_cast<void>(kCachedPowersSize); // Fix warning. + + const cached_power cached = kCachedPowers[index]; + assert(kAlpha <= cached.e + e + 64); + assert(kGamma >= cached.e + e + 64); + + return cached; +} + +/*! +For n != 0, returns k, such that pow10 := 10^(k-1) <= n < 10^k. +For n == 0, returns 1 and sets pow10 := 1. +*/ +inline int find_largest_pow10(const uint32_t n, uint32_t& pow10) +{ + // LCOV_EXCL_START + if (n >= 1000000000) + { + pow10 = 1000000000; + return 10; + } + // LCOV_EXCL_STOP + else if (n >= 100000000) + { + pow10 = 100000000; + return 9; + } + else if (n >= 10000000) + { + pow10 = 10000000; + return 8; + } + else if (n >= 1000000) + { + pow10 = 1000000; + return 7; + } + else if (n >= 100000) + { + pow10 = 100000; + return 6; + } + else if (n >= 10000) + { + pow10 = 10000; + return 5; + } + else if (n >= 1000) + { + pow10 = 1000; + return 4; + } + else if (n >= 100) + { + pow10 = 100; + return 3; + } + else if (n >= 10) + { + pow10 = 10; + return 2; + } + else + { + pow10 = 1; + return 1; + } +} + +inline void grisu2_round(char* buf, int len, uint64_t dist, uint64_t delta, + uint64_t rest, uint64_t ten_k) +{ + assert(len >= 1); + assert(dist <= delta); + assert(rest <= delta); + assert(ten_k > 0); + + // <--------------------------- delta ----> + // <---- dist ---------> + // --------------[------------------+-------------------]-------------- + // M- w M+ + // + // ten_k + // <------> + // <---- rest ----> + // --------------[------------------+----+--------------]-------------- + // w V + // = buf * 10^k + // + // ten_k represents a unit-in-the-last-place in the decimal representation + // stored in buf. + // Decrement buf by ten_k while this takes buf closer to w. + + // The tests are written in this order to avoid overflow in unsigned + // integer arithmetic. + + while (rest < dist + and delta - rest >= ten_k + and (rest + ten_k < dist or dist - rest > rest + ten_k - dist)) + { + assert(buf[len - 1] != '0'); + buf[len - 1]--; + rest += ten_k; + } +} + +/*! +Generates V = buffer * 10^decimal_exponent, such that M- <= V <= M+. +M- and M+ must be normalized and share the same exponent -60 <= e <= -32. +*/ +inline void grisu2_digit_gen(char* buffer, int& length, int& decimal_exponent, + diyfp M_minus, diyfp w, diyfp M_plus) +{ + static_assert(kAlpha >= -60, "internal error"); + static_assert(kGamma <= -32, "internal error"); + + // Generates the digits (and the exponent) of a decimal floating-point + // number V = buffer * 10^decimal_exponent in the range [M-, M+]. The diyfp's + // w, M- and M+ share the same exponent e, which satisfies alpha <= e <= gamma. + // + // <--------------------------- delta ----> + // <---- dist ---------> + // --------------[------------------+-------------------]-------------- + // M- w M+ + // + // Grisu2 generates the digits of M+ from left to right and stops as soon as + // V is in [M-,M+]. + + assert(M_plus.e >= kAlpha); + assert(M_plus.e <= kGamma); + + uint64_t delta = diyfp::sub(M_plus, M_minus).f; // (significand of (M+ - M-), implicit exponent is e) + uint64_t dist = diyfp::sub(M_plus, w ).f; // (significand of (M+ - w ), implicit exponent is e) + + // Split M+ = f * 2^e into two parts p1 and p2 (note: e < 0): + // + // M+ = f * 2^e + // = ((f div 2^-e) * 2^-e + (f mod 2^-e)) * 2^e + // = ((p1 ) * 2^-e + (p2 )) * 2^e + // = p1 + p2 * 2^e + + const diyfp one(uint64_t{1} << -M_plus.e, M_plus.e); + + auto p1 = static_cast<uint32_t>(M_plus.f >> -one.e); // p1 = f div 2^-e (Since -e >= 32, p1 fits into a 32-bit int.) + uint64_t p2 = M_plus.f & (one.f - 1); // p2 = f mod 2^-e + + // 1) + // + // Generate the digits of the integral part p1 = d[n-1]...d[1]d[0] + + assert(p1 > 0); + + uint32_t pow10; + const int k = find_largest_pow10(p1, pow10); + + // 10^(k-1) <= p1 < 10^k, pow10 = 10^(k-1) + // + // p1 = (p1 div 10^(k-1)) * 10^(k-1) + (p1 mod 10^(k-1)) + // = (d[k-1] ) * 10^(k-1) + (p1 mod 10^(k-1)) + // + // M+ = p1 + p2 * 2^e + // = d[k-1] * 10^(k-1) + (p1 mod 10^(k-1)) + p2 * 2^e + // = d[k-1] * 10^(k-1) + ((p1 mod 10^(k-1)) * 2^-e + p2) * 2^e + // = d[k-1] * 10^(k-1) + ( rest) * 2^e + // + // Now generate the digits d[n] of p1 from left to right (n = k-1,...,0) + // + // p1 = d[k-1]...d[n] * 10^n + d[n-1]...d[0] + // + // but stop as soon as + // + // rest * 2^e = (d[n-1]...d[0] * 2^-e + p2) * 2^e <= delta * 2^e + + int n = k; + while (n > 0) + { + // Invariants: + // M+ = buffer * 10^n + (p1 + p2 * 2^e) (buffer = 0 for n = k) + // pow10 = 10^(n-1) <= p1 < 10^n + // + const uint32_t d = p1 / pow10; // d = p1 div 10^(n-1) + const uint32_t r = p1 % pow10; // r = p1 mod 10^(n-1) + // + // M+ = buffer * 10^n + (d * 10^(n-1) + r) + p2 * 2^e + // = (buffer * 10 + d) * 10^(n-1) + (r + p2 * 2^e) + // + assert(d <= 9); + buffer[length++] = static_cast<char>('0' + d); // buffer := buffer * 10 + d + // + // M+ = buffer * 10^(n-1) + (r + p2 * 2^e) + // + p1 = r; + n--; + // + // M+ = buffer * 10^n + (p1 + p2 * 2^e) + // pow10 = 10^n + // + + // Now check if enough digits have been generated. + // Compute + // + // p1 + p2 * 2^e = (p1 * 2^-e + p2) * 2^e = rest * 2^e + // + // Note: + // Since rest and delta share the same exponent e, it suffices to + // compare the significands. + const uint64_t rest = (uint64_t{p1} << -one.e) + p2; + if (rest <= delta) + { + // V = buffer * 10^n, with M- <= V <= M+. + + decimal_exponent += n; + + // We may now just stop. But instead look if the buffer could be + // decremented to bring V closer to w. + // + // pow10 = 10^n is now 1 ulp in the decimal representation V. + // The rounding procedure works with diyfp's with an implicit + // exponent of e. + // + // 10^n = (10^n * 2^-e) * 2^e = ulp * 2^e + // + const uint64_t ten_n = uint64_t{pow10} << -one.e; + grisu2_round(buffer, length, dist, delta, rest, ten_n); + + return; + } + + pow10 /= 10; + // + // pow10 = 10^(n-1) <= p1 < 10^n + // Invariants restored. + } + + // 2) + // + // The digits of the integral part have been generated: + // + // M+ = d[k-1]...d[1]d[0] + p2 * 2^e + // = buffer + p2 * 2^e + // + // Now generate the digits of the fractional part p2 * 2^e. + // + // Note: + // No decimal point is generated: the exponent is adjusted instead. + // + // p2 actually represents the fraction + // + // p2 * 2^e + // = p2 / 2^-e + // = d[-1] / 10^1 + d[-2] / 10^2 + ... + // + // Now generate the digits d[-m] of p1 from left to right (m = 1,2,...) + // + // p2 * 2^e = d[-1]d[-2]...d[-m] * 10^-m + // + 10^-m * (d[-m-1] / 10^1 + d[-m-2] / 10^2 + ...) + // + // using + // + // 10^m * p2 = ((10^m * p2) div 2^-e) * 2^-e + ((10^m * p2) mod 2^-e) + // = ( d) * 2^-e + ( r) + // + // or + // 10^m * p2 * 2^e = d + r * 2^e + // + // i.e. + // + // M+ = buffer + p2 * 2^e + // = buffer + 10^-m * (d + r * 2^e) + // = (buffer * 10^m + d) * 10^-m + 10^-m * r * 2^e + // + // and stop as soon as 10^-m * r * 2^e <= delta * 2^e + + assert(p2 > delta); + + int m = 0; + for (;;) + { + // Invariant: + // M+ = buffer * 10^-m + 10^-m * (d[-m-1] / 10 + d[-m-2] / 10^2 + ...) * 2^e + // = buffer * 10^-m + 10^-m * (p2 ) * 2^e + // = buffer * 10^-m + 10^-m * (1/10 * (10 * p2) ) * 2^e + // = buffer * 10^-m + 10^-m * (1/10 * ((10*p2 div 2^-e) * 2^-e + (10*p2 mod 2^-e)) * 2^e + // + assert(p2 <= UINT64_MAX / 10); + p2 *= 10; + const uint64_t d = p2 >> -one.e; // d = (10 * p2) div 2^-e + const uint64_t r = p2 & (one.f - 1); // r = (10 * p2) mod 2^-e + // + // M+ = buffer * 10^-m + 10^-m * (1/10 * (d * 2^-e + r) * 2^e + // = buffer * 10^-m + 10^-m * (1/10 * (d + r * 2^e)) + // = (buffer * 10 + d) * 10^(-m-1) + 10^(-m-1) * r * 2^e + // + assert(d <= 9); + buffer[length++] = static_cast<char>('0' + d); // buffer := buffer * 10 + d + // + // M+ = buffer * 10^(-m-1) + 10^(-m-1) * r * 2^e + // + p2 = r; + m++; + // + // M+ = buffer * 10^-m + 10^-m * p2 * 2^e + // Invariant restored. + + // Check if enough digits have been generated. + // + // 10^-m * p2 * 2^e <= delta * 2^e + // p2 * 2^e <= 10^m * delta * 2^e + // p2 <= 10^m * delta + delta *= 10; + dist *= 10; + if (p2 <= delta) + { + break; + } + } + + // V = buffer * 10^-m, with M- <= V <= M+. + + decimal_exponent -= m; + + // 1 ulp in the decimal representation is now 10^-m. + // Since delta and dist are now scaled by 10^m, we need to do the + // same with ulp in order to keep the units in sync. + // + // 10^m * 10^-m = 1 = 2^-e * 2^e = ten_m * 2^e + // + const uint64_t ten_m = one.f; + grisu2_round(buffer, length, dist, delta, p2, ten_m); + + // By construction this algorithm generates the shortest possible decimal + // number (Loitsch, Theorem 6.2) which rounds back to w. + // For an input number of precision p, at least + // + // N = 1 + ceil(p * log_10(2)) + // + // decimal digits are sufficient to identify all binary floating-point + // numbers (Matula, "In-and-Out conversions"). + // This implies that the algorithm does not produce more than N decimal + // digits. + // + // N = 17 for p = 53 (IEEE double precision) + // N = 9 for p = 24 (IEEE single precision) +} + +/*! +v = buf * 10^decimal_exponent +len is the length of the buffer (number of decimal digits) +The buffer must be large enough, i.e. >= max_digits10. +*/ +inline void grisu2(char* buf, int& len, int& decimal_exponent, + diyfp m_minus, diyfp v, diyfp m_plus) +{ + assert(m_plus.e == m_minus.e); + assert(m_plus.e == v.e); + + // --------(-----------------------+-----------------------)-------- (A) + // m- v m+ + // + // --------------------(-----------+-----------------------)-------- (B) + // m- v m+ + // + // First scale v (and m- and m+) such that the exponent is in the range + // [alpha, gamma]. + + const cached_power cached = get_cached_power_for_binary_exponent(m_plus.e); + + const diyfp c_minus_k(cached.f, cached.e); // = c ~= 10^-k + + // The exponent of the products is = v.e + c_minus_k.e + q and is in the range [alpha,gamma] + const diyfp w = diyfp::mul(v, c_minus_k); + const diyfp w_minus = diyfp::mul(m_minus, c_minus_k); + const diyfp w_plus = diyfp::mul(m_plus, c_minus_k); + + // ----(---+---)---------------(---+---)---------------(---+---)---- + // w- w w+ + // = c*m- = c*v = c*m+ + // + // diyfp::mul rounds its result and c_minus_k is approximated too. w, w- and + // w+ are now off by a small amount. + // In fact: + // + // w - v * 10^k < 1 ulp + // + // To account for this inaccuracy, add resp. subtract 1 ulp. + // + // --------+---[---------------(---+---)---------------]---+-------- + // w- M- w M+ w+ + // + // Now any number in [M-, M+] (bounds included) will round to w when input, + // regardless of how the input rounding algorithm breaks ties. + // + // And digit_gen generates the shortest possible such number in [M-, M+]. + // Note that this does not mean that Grisu2 always generates the shortest + // possible number in the interval (m-, m+). + const diyfp M_minus(w_minus.f + 1, w_minus.e); + const diyfp M_plus (w_plus.f - 1, w_plus.e ); + + decimal_exponent = -cached.k; // = -(-k) = k + + grisu2_digit_gen(buf, len, decimal_exponent, M_minus, w, M_plus); +} + +/*! +v = buf * 10^decimal_exponent +len is the length of the buffer (number of decimal digits) +The buffer must be large enough, i.e. >= max_digits10. +*/ +template <typename FloatType> +void grisu2(char* buf, int& len, int& decimal_exponent, FloatType value) +{ + static_assert(diyfp::kPrecision >= std::numeric_limits<FloatType>::digits + 3, + "internal error: not enough precision"); + + assert(std::isfinite(value)); + assert(value > 0); + + // If the neighbors (and boundaries) of 'value' are always computed for double-precision + // numbers, all float's can be recovered using strtod (and strtof). However, the resulting + // decimal representations are not exactly "short". + // + // The documentation for 'std::to_chars' (https://en.cppreference.com/w/cpp/utility/to_chars) + // says "value is converted to a string as if by std::sprintf in the default ("C") locale" + // and since sprintf promotes float's to double's, I think this is exactly what 'std::to_chars' + // does. + // On the other hand, the documentation for 'std::to_chars' requires that "parsing the + // representation using the corresponding std::from_chars function recovers value exactly". That + // indicates that single precision floating-point numbers should be recovered using + // 'std::strtof'. + // + // NB: If the neighbors are computed for single-precision numbers, there is a single float + // (7.0385307e-26f) which can't be recovered using strtod. The resulting double precision + // value is off by 1 ulp. +#if 0 + const boundaries w = compute_boundaries(static_cast<double>(value)); +#else + const boundaries w = compute_boundaries(value); +#endif + + grisu2(buf, len, decimal_exponent, w.minus, w.w, w.plus); +} + +/*! +@brief appends a decimal representation of e to buf +@return a pointer to the element following the exponent. +@pre -1000 < e < 1000 +*/ +inline char* append_exponent(char* buf, int e) +{ + assert(e > -1000); + assert(e < 1000); + + if (e < 0) + { + e = -e; + *buf++ = '-'; + } + else + { + *buf++ = '+'; + } + + auto k = static_cast<uint32_t>(e); + if (k < 10) + { + // Always print at least two digits in the exponent. + // This is for compatibility with printf("%g"). + *buf++ = '0'; + *buf++ = static_cast<char>('0' + k); + } + else if (k < 100) + { + *buf++ = static_cast<char>('0' + k / 10); + k %= 10; + *buf++ = static_cast<char>('0' + k); + } + else + { + *buf++ = static_cast<char>('0' + k / 100); + k %= 100; + *buf++ = static_cast<char>('0' + k / 10); + k %= 10; + *buf++ = static_cast<char>('0' + k); + } + + return buf; +} + +/*! +@brief prettify v = buf * 10^decimal_exponent + +If v is in the range [10^min_exp, 10^max_exp) it will be printed in fixed-point +notation. Otherwise it will be printed in exponential notation. + +@pre min_exp < 0 +@pre max_exp > 0 +*/ +inline char* format_buffer(char* buf, int len, int decimal_exponent, + int min_exp, int max_exp) +{ + assert(min_exp < 0); + assert(max_exp > 0); + + const int k = len; + const int n = len + decimal_exponent; + + // v = buf * 10^(n-k) + // k is the length of the buffer (number of decimal digits) + // n is the position of the decimal point relative to the start of the buffer. + + if (k <= n and n <= max_exp) + { + // digits[000] + // len <= max_exp + 2 + + std::memset(buf + k, '0', static_cast<size_t>(n - k)); + // Make it look like a floating-point number (#362, #378) + buf[n + 0] = '.'; + buf[n + 1] = '0'; + return buf + (n + 2); + } + + if (0 < n and n <= max_exp) + { + // dig.its + // len <= max_digits10 + 1 + + assert(k > n); + + std::memmove(buf + (n + 1), buf + n, static_cast<size_t>(k - n)); + buf[n] = '.'; + return buf + (k + 1); + } + + if (min_exp < n and n <= 0) + { + // 0.[000]digits + // len <= 2 + (-min_exp - 1) + max_digits10 + + std::memmove(buf + (2 + -n), buf, static_cast<size_t>(k)); + buf[0] = '0'; + buf[1] = '.'; + std::memset(buf + 2, '0', static_cast<size_t>(-n)); + return buf + (2 + (-n) + k); + } + + if (k == 1) + { + // dE+123 + // len <= 1 + 5 + + buf += 1; + } + else + { + // d.igitsE+123 + // len <= max_digits10 + 1 + 5 + + std::memmove(buf + 2, buf + 1, static_cast<size_t>(k - 1)); + buf[1] = '.'; + buf += 1 + k; + } + + *buf++ = 'e'; + return append_exponent(buf, n - 1); +} + +} // namespace dtoa_impl + +/*! +@brief generates a decimal representation of the floating-point number value in [first, last). + +The format of the resulting decimal representation is similar to printf's %g +format. Returns an iterator pointing past-the-end of the decimal representation. + +@note The input number must be finite, i.e. NaN's and Inf's are not supported. +@note The buffer must be large enough. +@note The result is NOT null-terminated. +*/ +template <typename FloatType> +char* to_chars(char* first, const char* last, FloatType value) +{ + static_cast<void>(last); // maybe unused - fix warning + assert(std::isfinite(value)); + + // Use signbit(value) instead of (value < 0) since signbit works for -0. + if (std::signbit(value)) + { + value = -value; + *first++ = '-'; + } + + if (value == 0) // +-0 + { + *first++ = '0'; + // Make it look like a floating-point number (#362, #378) + *first++ = '.'; + *first++ = '0'; + return first; + } + + assert(last - first >= std::numeric_limits<FloatType>::max_digits10); + + // Compute v = buffer * 10^decimal_exponent. + // The decimal digits are stored in the buffer, which needs to be interpreted + // as an unsigned decimal integer. + // len is the length of the buffer, i.e. the number of decimal digits. + int len = 0; + int decimal_exponent = 0; + dtoa_impl::grisu2(first, len, decimal_exponent, value); + + assert(len <= std::numeric_limits<FloatType>::max_digits10); + + // Format the buffer like printf("%.*g", prec, value) + constexpr int kMinExp = -4; + // Use digits10 here to increase compatibility with version 2. + constexpr int kMaxExp = std::numeric_limits<FloatType>::digits10; + + assert(last - first >= kMaxExp + 2); + assert(last - first >= 2 + (-kMinExp - 1) + std::numeric_limits<FloatType>::max_digits10); + assert(last - first >= std::numeric_limits<FloatType>::max_digits10 + 6); + + return dtoa_impl::format_buffer(first, len, decimal_exponent, kMinExp, kMaxExp); +} + +} // namespace detail +} // namespace nlohmann + +// #include <nlohmann/detail/macro_scope.hpp> + +// #include <nlohmann/detail/meta/cpp_future.hpp> + +// #include <nlohmann/detail/output/binary_writer.hpp> + +// #include <nlohmann/detail/output/output_adapters.hpp> + +// #include <nlohmann/detail/value_t.hpp> + + +namespace nlohmann +{ +namespace detail +{ /////////////////// // serialization // /////////////////// +/// how to treat decoding errors +enum class error_handler_t +{ + strict, ///< throw a type_error exception in case of invalid UTF-8 + replace, ///< replace invalid UTF-8 sequences with U+FFFD + ignore ///< ignore invalid UTF-8 sequences +}; + template<typename BasicJsonType> class serializer { @@ -6077,20 +10927,32 @@ class serializer using number_float_t = typename BasicJsonType::number_float_t; using number_integer_t = typename BasicJsonType::number_integer_t; using number_unsigned_t = typename BasicJsonType::number_unsigned_t; + static constexpr uint8_t UTF8_ACCEPT = 0; + static constexpr uint8_t UTF8_REJECT = 1; + public: /*! @param[in] s output stream to serialize to @param[in] ichar indentation character to use + @param[in] error_handler_ how to react on decoding errors */ - serializer(output_adapter_t<char> s, const char ichar) - : o(std::move(s)), loc(std::localeconv()), - thousands_sep(loc->thousands_sep == nullptr ? '\0' : * (loc->thousands_sep)), - decimal_point(loc->decimal_point == nullptr ? '\0' : * (loc->decimal_point)), - indent_char(ichar), indent_string(512, indent_char) {} + serializer(output_adapter_t<char> s, const char ichar, + error_handler_t error_handler_ = error_handler_t::strict) + : o(std::move(s)) + , loc(std::localeconv()) + , thousands_sep(loc->thousands_sep == nullptr ? '\0' : * (loc->thousands_sep)) + , decimal_point(loc->decimal_point == nullptr ? '\0' : * (loc->decimal_point)) + , indent_char(ichar) + , indent_string(512, indent_char) + , error_handler(error_handler_) + {} // delete because of pointer members serializer(const serializer&) = delete; serializer& operator=(const serializer&) = delete; + serializer(serializer&&) = delete; + serializer& operator=(serializer&&) = delete; + ~serializer() = default; /*! @brief internal implementation of the serialization function @@ -6303,323 +11165,249 @@ class serializer private: /*! - @brief returns the number of expected bytes following in UTF-8 string - - @param[in] u the first byte of a UTF-8 string - @return the number of expected bytes following - */ - static constexpr std::size_t bytes_following(const uint8_t u) - { - return ((u <= 127) ? 0 - : ((192 <= u and u <= 223) ? 1 - : ((224 <= u and u <= 239) ? 2 - : ((240 <= u and u <= 247) ? 3 : std::string::npos)))); - } + @brief dump escaped string - /*! - @brief calculates the extra space to escape a JSON string + Escape a string by replacing certain special characters by a sequence of an + escape character (backslash) and another character and other control + characters by a sequence of "\u" followed by a four-digit hex + representation. The escaped string is written to output stream @a o. @param[in] s the string to escape @param[in] ensure_ascii whether to escape non-ASCII characters with \uXXXX sequences - @return the number of characters required to escape string @a s @complexity Linear in the length of string @a s. */ - static std::size_t extra_space(const string_t& s, - const bool ensure_ascii) noexcept + void dump_escaped(const string_t& s, const bool ensure_ascii) { - std::size_t res = 0; + uint32_t codepoint; + uint8_t state = UTF8_ACCEPT; + std::size_t bytes = 0; // number of bytes written to string_buffer + + // number of bytes written at the point of the last valid byte + std::size_t bytes_after_last_accept = 0; + std::size_t undumped_chars = 0; for (std::size_t i = 0; i < s.size(); ++i) { - switch (s[i]) - { - // control characters that can be escaped with a backslash - case '"': - case '\\': - case '\b': - case '\f': - case '\n': - case '\r': - case '\t': - { - // from c (1 byte) to \x (2 bytes) - res += 1; - break; - } + const auto byte = static_cast<uint8_t>(s[i]); - // control characters that need \uxxxx escaping - case 0x00: - case 0x01: - case 0x02: - case 0x03: - case 0x04: - case 0x05: - case 0x06: - case 0x07: - case 0x0B: - case 0x0E: - case 0x0F: - case 0x10: - case 0x11: - case 0x12: - case 0x13: - case 0x14: - case 0x15: - case 0x16: - case 0x17: - case 0x18: - case 0x19: - case 0x1A: - case 0x1B: - case 0x1C: - case 0x1D: - case 0x1E: - case 0x1F: - { - // from c (1 byte) to \uxxxx (6 bytes) - res += 5; - break; - } - - default: + switch (decode(state, codepoint, byte)) + { + case UTF8_ACCEPT: // decode found a new code point { - if (ensure_ascii and (s[i] & 0x80 or s[i] == 0x7F)) + switch (codepoint) { - const auto bytes = bytes_following(static_cast<uint8_t>(s[i])); - // invalid characters will be detected by throw_if_invalid_utf8 - assert (bytes != std::string::npos); - - if (bytes == 3) + case 0x08: // backspace { - // codepoints that need 4 bytes (i.e., 3 additional - // bytes) in UTF-8 need a surrogate pair when \u - // escaping is used: from 4 bytes to \uxxxx\uxxxx - // (12 bytes) - res += (12 - bytes - 1); + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = 'b'; + break; } - else + + case 0x09: // horizontal tab { - // from x bytes to \uxxxx (6 bytes) - res += (6 - bytes - 1); + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = 't'; + break; } - // skip the additional bytes - i += bytes; - } - break; - } - } - } - - return res; - } - - static void escape_codepoint(int codepoint, string_t& result, std::size_t& pos) - { - // expecting a proper codepoint - assert(0x00 <= codepoint and codepoint <= 0x10FFFF); - - // the last written character was the backslash before the 'u' - assert(result[pos] == '\\'); - - // write the 'u' - result[++pos] = 'u'; - - // convert a number 0..15 to its hex representation (0..f) - static const std::array<char, 16> hexify = - { - { - '0', '1', '2', '3', '4', '5', '6', '7', - '8', '9', 'a', 'b', 'c', 'd', 'e', 'f' - } - }; - - if (codepoint < 0x10000) - { - // codepoints U+0000..U+FFFF can be represented as \uxxxx. - result[++pos] = hexify[(codepoint >> 12) & 0x0F]; - result[++pos] = hexify[(codepoint >> 8) & 0x0F]; - result[++pos] = hexify[(codepoint >> 4) & 0x0F]; - result[++pos] = hexify[codepoint & 0x0F]; - } - else - { - // codepoints U+10000..U+10FFFF need a surrogate pair to be - // represented as \uxxxx\uxxxx. - // http://www.unicode.org/faq/utf_bom.html#utf16-4 - codepoint -= 0x10000; - const int high_surrogate = 0xD800 | ((codepoint >> 10) & 0x3FF); - const int low_surrogate = 0xDC00 | (codepoint & 0x3FF); - result[++pos] = hexify[(high_surrogate >> 12) & 0x0F]; - result[++pos] = hexify[(high_surrogate >> 8) & 0x0F]; - result[++pos] = hexify[(high_surrogate >> 4) & 0x0F]; - result[++pos] = hexify[high_surrogate & 0x0F]; - ++pos; // backslash is already in output - result[++pos] = 'u'; - result[++pos] = hexify[(low_surrogate >> 12) & 0x0F]; - result[++pos] = hexify[(low_surrogate >> 8) & 0x0F]; - result[++pos] = hexify[(low_surrogate >> 4) & 0x0F]; - result[++pos] = hexify[low_surrogate & 0x0F]; - } - - ++pos; - } + case 0x0A: // newline + { + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = 'n'; + break; + } - /*! - @brief dump escaped string + case 0x0C: // formfeed + { + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = 'f'; + break; + } - Escape a string by replacing certain special characters by a sequence of an - escape character (backslash) and another character and other control - characters by a sequence of "\u" followed by a four-digit hex - representation. The escaped string is written to output stream @a o. + case 0x0D: // carriage return + { + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = 'r'; + break; + } - @param[in] s the string to escape - @param[in] ensure_ascii whether to escape non-ASCII characters with - \uXXXX sequences + case 0x22: // quotation mark + { + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = '\"'; + break; + } - @complexity Linear in the length of string @a s. - */ - void dump_escaped(const string_t& s, const bool ensure_ascii) const - { - throw_if_invalid_utf8(s); + case 0x5C: // reverse solidus + { + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = '\\'; + break; + } - const auto space = extra_space(s, ensure_ascii); - if (space == 0) - { - o->write_characters(s.c_str(), s.size()); - return; - } + default: + { + // escape control characters (0x00..0x1F) or, if + // ensure_ascii parameter is used, non-ASCII characters + if ((codepoint <= 0x1F) or (ensure_ascii and (codepoint >= 0x7F))) + { + if (codepoint <= 0xFFFF) + { + (std::snprintf)(string_buffer.data() + bytes, 7, "\\u%04x", + static_cast<uint16_t>(codepoint)); + bytes += 6; + } + else + { + (std::snprintf)(string_buffer.data() + bytes, 13, "\\u%04x\\u%04x", + static_cast<uint16_t>(0xD7C0 + (codepoint >> 10)), + static_cast<uint16_t>(0xDC00 + (codepoint & 0x3FF))); + bytes += 12; + } + } + else + { + // copy byte to buffer (all previous bytes + // been copied have in default case above) + string_buffer[bytes++] = s[i]; + } + break; + } + } - // create a result string of necessary size - string_t result(s.size() + space, '\\'); - std::size_t pos = 0; + // write buffer and reset index; there must be 13 bytes + // left, as this is the maximal number of bytes to be + // written ("\uxxxx\uxxxx\0") for one code point + if (string_buffer.size() - bytes < 13) + { + o->write_characters(string_buffer.data(), bytes); + bytes = 0; + } - for (std::size_t i = 0; i < s.size(); ++i) - { - switch (s[i]) - { - case '"': // quotation mark (0x22) - { - result[pos + 1] = '"'; - pos += 2; + // remember the byte position of this accept + bytes_after_last_accept = bytes; + undumped_chars = 0; break; } - case '\\': // reverse solidus (0x5C) + case UTF8_REJECT: // decode found invalid UTF-8 byte { - // nothing to change - pos += 2; - break; - } + switch (error_handler) + { + case error_handler_t::strict: + { + std::string sn(3, '\0'); + (std::snprintf)(&sn[0], sn.size(), "%.2X", byte); + JSON_THROW(type_error::create(316, "invalid UTF-8 byte at index " + std::to_string(i) + ": 0x" + sn)); + } - case '\b': // backspace (0x08) - { - result[pos + 1] = 'b'; - pos += 2; - break; - } + case error_handler_t::ignore: + case error_handler_t::replace: + { + // in case we saw this character the first time, we + // would like to read it again, because the byte + // may be OK for itself, but just not OK for the + // previous sequence + if (undumped_chars > 0) + { + --i; + } - case '\f': // formfeed (0x0C) - { - result[pos + 1] = 'f'; - pos += 2; + // reset length buffer to the last accepted index; + // thus removing/ignoring the invalid characters + bytes = bytes_after_last_accept; + + if (error_handler == error_handler_t::replace) + { + // add a replacement character + if (ensure_ascii) + { + string_buffer[bytes++] = '\\'; + string_buffer[bytes++] = 'u'; + string_buffer[bytes++] = 'f'; + string_buffer[bytes++] = 'f'; + string_buffer[bytes++] = 'f'; + string_buffer[bytes++] = 'd'; + } + else + { + string_buffer[bytes++] = detail::binary_writer<BasicJsonType, char>::to_char_type('\xEF'); + string_buffer[bytes++] = detail::binary_writer<BasicJsonType, char>::to_char_type('\xBF'); + string_buffer[bytes++] = detail::binary_writer<BasicJsonType, char>::to_char_type('\xBD'); + } + bytes_after_last_accept = bytes; + } + + undumped_chars = 0; + + // continue processing the string + state = UTF8_ACCEPT; + break; + } + } break; } - case '\n': // newline (0x0A) + default: // decode found yet incomplete multi-byte code point { - result[pos + 1] = 'n'; - pos += 2; + if (not ensure_ascii) + { + // code point will not be escaped - copy byte to buffer + string_buffer[bytes++] = s[i]; + } + ++undumped_chars; break; } + } + } - case '\r': // carriage return (0x0D) + // we finished processing the string + if (JSON_LIKELY(state == UTF8_ACCEPT)) + { + // write buffer + if (bytes > 0) + { + o->write_characters(string_buffer.data(), bytes); + } + } + else + { + // we finish reading, but do not accept: string was incomplete + switch (error_handler) + { + case error_handler_t::strict: { - result[pos + 1] = 'r'; - pos += 2; - break; + std::string sn(3, '\0'); + (std::snprintf)(&sn[0], sn.size(), "%.2X", static_cast<uint8_t>(s.back())); + JSON_THROW(type_error::create(316, "incomplete UTF-8 string; last byte: 0x" + sn)); } - case '\t': // horizontal tab (0x09) + case error_handler_t::ignore: { - result[pos + 1] = 't'; - pos += 2; + // write all accepted bytes + o->write_characters(string_buffer.data(), bytes_after_last_accept); break; } - default: + case error_handler_t::replace: { - // escape control characters (0x00..0x1F) or, if - // ensure_ascii parameter is used, non-ASCII characters - if ((0x00 <= s[i] and s[i] <= 0x1F) or - (ensure_ascii and (s[i] & 0x80 or s[i] == 0x7F))) + // write all accepted bytes + o->write_characters(string_buffer.data(), bytes_after_last_accept); + // add a replacement character + if (ensure_ascii) { - const auto bytes = bytes_following(static_cast<uint8_t>(s[i])); - // invalid characters will be detected by throw_if_invalid_utf8 - assert (bytes != std::string::npos); - - // check that the additional bytes are present - assert(i + bytes < s.size()); - - // to use \uxxxx escaping, we first need to calculate - // the codepoint from the UTF-8 bytes - int codepoint = 0; - - // bytes is unsigned type: - assert(bytes <= 3); - switch (bytes) - { - case 0: - { - codepoint = s[i] & 0xFF; - break; - } - - case 1: - { - codepoint = ((s[i] & 0x3F) << 6) - + (s[i + 1] & 0x7F); - break; - } - - case 2: - { - codepoint = ((s[i] & 0x1F) << 12) - + ((s[i + 1] & 0x7F) << 6) - + (s[i + 2] & 0x7F); - break; - } - - case 3: - { - codepoint = ((s[i] & 0xF) << 18) - + ((s[i + 1] & 0x7F) << 12) - + ((s[i + 2] & 0x7F) << 6) - + (s[i + 3] & 0x7F); - break; - } - - default: - break; // LCOV_EXCL_LINE - } - - escape_codepoint(codepoint, result, pos); - i += bytes; + o->write_characters("\\ufffd", 6); } else { - // all other characters are added as-is - result[pos++] = s[i]; + o->write_characters("\xEF\xBF\xBD", 3); } break; } } } - - assert(pos == result.size()); - o->write_characters(result.c_str(), result.size()); } /*! @@ -6644,7 +11432,7 @@ class serializer return; } - const bool is_negative = (x <= 0) and (x != 0); // see issue #755 + const bool is_negative = std::is_same<NumberType, number_integer_t>::value and not (x >= 0); // see issue #755 std::size_t i = 0; while (x != 0) @@ -6679,17 +11467,39 @@ class serializer void dump_float(number_float_t x) { // NaN / inf - if (not std::isfinite(x) or std::isnan(x)) + if (not std::isfinite(x)) { o->write_characters("null", 4); return; } - // get number of digits for a text -> float -> text round-trip - static constexpr auto d = std::numeric_limits<number_float_t>::digits10; + // If number_float_t is an IEEE-754 single or double precision number, + // use the Grisu2 algorithm to produce short numbers which are + // guaranteed to round-trip, using strtof and strtod, resp. + // + // NB: The test below works if <long double> == <double>. + static constexpr bool is_ieee_single_or_double + = (std::numeric_limits<number_float_t>::is_iec559 and std::numeric_limits<number_float_t>::digits == 24 and std::numeric_limits<number_float_t>::max_exponent == 128) or + (std::numeric_limits<number_float_t>::is_iec559 and std::numeric_limits<number_float_t>::digits == 53 and std::numeric_limits<number_float_t>::max_exponent == 1024); + + dump_float(x, std::integral_constant<bool, is_ieee_single_or_double>()); + } + + void dump_float(number_float_t x, std::true_type /*is_ieee_single_or_double*/) + { + char* begin = number_buffer.data(); + char* end = ::nlohmann::detail::to_chars(begin, begin + number_buffer.size(), x); + + o->write_characters(begin, static_cast<size_t>(end - begin)); + } + + void dump_float(number_float_t x, std::false_type /*is_ieee_single_or_double*/) + { + // get number of digits for a float -> text -> float round-trip + static constexpr auto d = std::numeric_limits<number_float_t>::max_digits10; // the actual conversion - std::ptrdiff_t len = snprintf(number_buffer.data(), number_buffer.size(), "%.*g", d, x); + std::ptrdiff_t len = (std::snprintf)(number_buffer.data(), number_buffer.size(), "%.*g", d, x); // negative value indicates an error assert(len > 0); @@ -6744,15 +11554,16 @@ class serializer followed. @param[in,out] state the state of the decoding + @param[in,out] codep codepoint (valid only if resulting state is UTF8_ACCEPT) @param[in] byte next byte to decode + @return new state - @note The function has been edited: a std::array is used and the code - point is not calculated. + @note The function has been edited: a std::array is used. @copyright Copyright (c) 2008-2009 Bjoern Hoehrmann <bjoern@hoehrmann.de> @sa http://bjoern.hoehrmann.de/utf-8/decoder/dfa/ */ - static void decode(uint8_t& state, const uint8_t byte) + static uint8_t decode(uint8_t& state, uint32_t& codep, const uint8_t byte) noexcept { static const std::array<uint8_t, 400> utf8d = { @@ -6775,42 +11586,13 @@ class serializer }; const uint8_t type = utf8d[byte]; - state = utf8d[256u + state * 16u + type]; - } - /*! - @brief throw an exception if a string is not UTF-8 encoded - - @param[in] str UTF-8 string to check - @throw type_error.316 if passed string is not UTF-8 encoded + codep = (state != UTF8_ACCEPT) + ? (byte & 0x3fu) | (codep << 6) + : static_cast<uint32_t>(0xff >> type) & (byte); - @since version 3.0.0 - */ - static void throw_if_invalid_utf8(const std::string& str) - { - // start with state 0 (= accept) - uint8_t state = 0; - - for (size_t i = 0; i < str.size(); ++i) - { - const auto byte = static_cast<uint8_t>(str[i]); - decode(state, byte); - if (state == 1) - { - // state 1 means reject - std::stringstream ss; - ss << std::setw(2) << std::uppercase << std::setfill('0') << std::hex << static_cast<int>(byte); - JSON_THROW(type_error::create(316, "invalid UTF-8 byte at index " + std::to_string(i) + ": 0x" + ss.str())); - } - } - - if (state != 0) - { - // we finish reading, but do not accept: string was incomplete - std::stringstream ss; - ss << std::setw(2) << std::uppercase << std::setfill('0') << std::hex << static_cast<int>(static_cast<uint8_t>(str.back())); - JSON_THROW(type_error::create(316, "incomplete UTF-8 string; last byte: 0x" + ss.str())); - } + state = utf8d[256u + state * 16u + type]; + return state; } private: @@ -6827,13 +11609,33 @@ class serializer /// the locale's decimal point character const char decimal_point = '\0'; + /// string buffer + std::array<char, 512> string_buffer{{}}; + /// the indentation character const char indent_char; - /// the indentation string string_t indent_string; + + /// error_handler how to react on decoding errors + const error_handler_t error_handler; }; +} // namespace detail +} // namespace nlohmann + +// #include <nlohmann/detail/json_ref.hpp> + +#include <initializer_list> +#include <utility> + +// #include <nlohmann/detail/meta/type_traits.hpp> + + +namespace nlohmann +{ +namespace detail +{ template<typename BasicJsonType> class json_ref { @@ -6852,15 +11654,19 @@ class json_ref : owned_value(init), value_ref(&owned_value), is_rvalue(true) {} - template<class... Args> - json_ref(Args&& ... args) - : owned_value(std::forward<Args>(args)...), value_ref(&owned_value), is_rvalue(true) - {} + template < + class... Args, + enable_if_t<std::is_constructible<value_type, Args...>::value, int> = 0 > + json_ref(Args && ... args) + : owned_value(std::forward<Args>(args)...), value_ref(&owned_value), + is_rvalue(true) {} // class should be movable only json_ref(json_ref&&) = default; json_ref(const json_ref&) = delete; json_ref& operator=(const json_ref&) = delete; + json_ref& operator=(json_ref&&) = delete; + ~json_ref() = default; value_type moved_or_copied() const { @@ -6886,74 +11692,30 @@ class json_ref value_type* value_ref = nullptr; const bool is_rvalue; }; +} // namespace detail +} // namespace nlohmann -} // namespace detail - -/// namespace to hold default `to_json` / `from_json` functions -namespace -{ -constexpr const auto& to_json = detail::static_const<detail::to_json_fn>::value; -constexpr const auto& from_json = detail::static_const<detail::from_json_fn>::value; -} - - -/*! -@brief default JSONSerializer template argument - -This serializer ignores the template arguments and uses ADL -([argument-dependent lookup](http://en.cppreference.com/w/cpp/language/adl)) -for serialization. -*/ -template<typename, typename> -struct adl_serializer -{ - /*! - @brief convert a JSON value to any value type - - This function is usually called by the `get()` function of the - @ref basic_json class (either explicit or via conversion operators). - - @param[in] j JSON value to read from - @param[in,out] val value to write to - */ - template<typename BasicJsonType, typename ValueType> - static void from_json(BasicJsonType&& j, ValueType& val) noexcept( - noexcept(::nlohmann::from_json(std::forward<BasicJsonType>(j), val))) - { - ::nlohmann::from_json(std::forward<BasicJsonType>(j), val); - } +// #include <nlohmann/detail/json_pointer.hpp> - /*! - @brief convert any value type to a JSON value - This function is usually called by the constructors of the @ref basic_json - class. +#include <cassert> // assert +#include <numeric> // accumulate +#include <string> // string +#include <vector> // vector - @param[in,out] j JSON value to write to - @param[in] val value to read from - */ - template<typename BasicJsonType, typename ValueType> - static void to_json(BasicJsonType& j, ValueType&& val) noexcept( - noexcept(::nlohmann::to_json(j, std::forward<ValueType>(val)))) - { - ::nlohmann::to_json(j, std::forward<ValueType>(val)); - } -}; +// #include <nlohmann/detail/macro_scope.hpp> -/*! -@brief JSON Pointer +// #include <nlohmann/detail/exceptions.hpp> -A JSON pointer defines a string syntax for identifying a specific value -within a JSON document. It can be used with functions `at` and -`operator[]`. Furthermore, JSON pointers are the base for JSON patches. +// #include <nlohmann/detail/value_t.hpp> -@sa [RFC 6901](https://tools.ietf.org/html/rfc6901) -@since version 2.0.0 -*/ +namespace nlohmann +{ +template<typename BasicJsonType> class json_pointer { - /// allow basic_json to access private members + // allow basic_json to access private members NLOHMANN_BASIC_JSON_TPL_DECLARATION friend class basic_json; @@ -6967,19 +11729,21 @@ class json_pointer @param[in] s string representing the JSON pointer; if omitted, the empty string is assumed which references the whole JSON value - @throw parse_error.107 if the given JSON pointer @a s is nonempty and - does not begin with a slash (`/`); see example below + @throw parse_error.107 if the given JSON pointer @a s is nonempty and does + not begin with a slash (`/`); see example below - @throw parse_error.108 if a tilde (`~`) in the given JSON pointer @a s - is not followed by `0` (representing `~`) or `1` (representing `/`); - see example below + @throw parse_error.108 if a tilde (`~`) in the given JSON pointer @a s is + not followed by `0` (representing `~`) or `1` (representing `/`); see + example below - @liveexample{The example shows the construction several valid JSON - pointers as well as the exceptional behavior.,json_pointer} + @liveexample{The example shows the construction several valid JSON pointers + as well as the exceptional behavior.,json_pointer} @since version 2.0.0 */ - explicit json_pointer(const std::string& s = "") : reference_tokens(split(s)) {} + explicit json_pointer(const std::string& s = "") + : reference_tokens(split(s)) + {} /*! @brief return a string representation of the JSON pointer @@ -6996,7 +11760,7 @@ class json_pointer @since version 2.0.0 */ - std::string to_string() const noexcept + std::string to_string() const { return std::accumulate(reference_tokens.begin(), reference_tokens.end(), std::string{}, @@ -7021,7 +11785,7 @@ class json_pointer */ static int array_index(const std::string& s) { - size_t processed_chars = 0; + std::size_t processed_chars = 0; const int res = std::stoi(s, &processed_chars); // check if the string was completely read @@ -7051,7 +11815,7 @@ class json_pointer } /// return whether pointer points to the root document - bool is_root() const + bool is_root() const noexcept { return reference_tokens.empty(); } @@ -7076,8 +11840,66 @@ class json_pointer @throw parse_error.109 if array index is not a number @throw type_error.313 if value cannot be unflattened */ - NLOHMANN_BASIC_JSON_TPL_DECLARATION - NLOHMANN_BASIC_JSON_TPL& get_and_create(NLOHMANN_BASIC_JSON_TPL& j) const; + BasicJsonType& get_and_create(BasicJsonType& j) const + { + using size_type = typename BasicJsonType::size_type; + auto result = &j; + + // in case no reference tokens exist, return a reference to the JSON value + // j which will be overwritten by a primitive value + for (const auto& reference_token : reference_tokens) + { + switch (result->m_type) + { + case detail::value_t::null: + { + if (reference_token == "0") + { + // start a new array if reference token is 0 + result = &result->operator[](0); + } + else + { + // start a new object otherwise + result = &result->operator[](reference_token); + } + break; + } + + case detail::value_t::object: + { + // create an entry in the object + result = &result->operator[](reference_token); + break; + } + + case detail::value_t::array: + { + // create an entry in the array + JSON_TRY + { + result = &result->operator[](static_cast<size_type>(array_index(reference_token))); + } + JSON_CATCH(std::invalid_argument&) + { + JSON_THROW(detail::parse_error::create(109, 0, "array index '" + reference_token + "' is not a number")); + } + break; + } + + /* + The following code is only reached if there exists a reference + token _and_ the current value is primitive. In this case, we have + an error situation, because primitive values may only occur as + single value; that is, with an empty list of reference tokens. + */ + default: + JSON_THROW(detail::type_error::create(313, "invalid value to unflatten")); + } + } + + return *result; + } /*! @brief return a reference to the pointed to value @@ -7098,8 +11920,75 @@ class json_pointer @throw parse_error.109 if an array index was not a number @throw out_of_range.404 if the JSON pointer can not be resolved */ - NLOHMANN_BASIC_JSON_TPL_DECLARATION - NLOHMANN_BASIC_JSON_TPL& get_unchecked(NLOHMANN_BASIC_JSON_TPL* ptr) const; + BasicJsonType& get_unchecked(BasicJsonType* ptr) const + { + using size_type = typename BasicJsonType::size_type; + for (const auto& reference_token : reference_tokens) + { + // convert null values to arrays or objects before continuing + if (ptr->m_type == detail::value_t::null) + { + // check if reference token is a number + const bool nums = + std::all_of(reference_token.begin(), reference_token.end(), + [](const char x) + { + return (x >= '0' and x <= '9'); + }); + + // change value to array for numbers or "-" or to object otherwise + *ptr = (nums or reference_token == "-") + ? detail::value_t::array + : detail::value_t::object; + } + + switch (ptr->m_type) + { + case detail::value_t::object: + { + // use unchecked object access + ptr = &ptr->operator[](reference_token); + break; + } + + case detail::value_t::array: + { + // error condition (cf. RFC 6901, Sect. 4) + if (JSON_UNLIKELY(reference_token.size() > 1 and reference_token[0] == '0')) + { + JSON_THROW(detail::parse_error::create(106, 0, + "array index '" + reference_token + + "' must not begin with '0'")); + } + + if (reference_token == "-") + { + // explicitly treat "-" as index beyond the end + ptr = &ptr->operator[](ptr->m_value.array->size()); + } + else + { + // convert array index to number; unchecked access + JSON_TRY + { + ptr = &ptr->operator[]( + static_cast<size_type>(array_index(reference_token))); + } + JSON_CATCH(std::invalid_argument&) + { + JSON_THROW(detail::parse_error::create(109, 0, "array index '" + reference_token + "' is not a number")); + } + } + break; + } + + default: + JSON_THROW(detail::out_of_range::create(404, "unresolved reference token '" + reference_token + "'")); + } + } + + return *ptr; + } /*! @throw parse_error.106 if an array index begins with '0' @@ -7107,8 +11996,57 @@ class json_pointer @throw out_of_range.402 if the array index '-' is used @throw out_of_range.404 if the JSON pointer can not be resolved */ - NLOHMANN_BASIC_JSON_TPL_DECLARATION - NLOHMANN_BASIC_JSON_TPL& get_checked(NLOHMANN_BASIC_JSON_TPL* ptr) const; + BasicJsonType& get_checked(BasicJsonType* ptr) const + { + using size_type = typename BasicJsonType::size_type; + for (const auto& reference_token : reference_tokens) + { + switch (ptr->m_type) + { + case detail::value_t::object: + { + // note: at performs range check + ptr = &ptr->at(reference_token); + break; + } + + case detail::value_t::array: + { + if (JSON_UNLIKELY(reference_token == "-")) + { + // "-" always fails the range check + JSON_THROW(detail::out_of_range::create(402, + "array index '-' (" + std::to_string(ptr->m_value.array->size()) + + ") is out of range")); + } + + // error condition (cf. RFC 6901, Sect. 4) + if (JSON_UNLIKELY(reference_token.size() > 1 and reference_token[0] == '0')) + { + JSON_THROW(detail::parse_error::create(106, 0, + "array index '" + reference_token + + "' must not begin with '0'")); + } + + // note: at performs range check + JSON_TRY + { + ptr = &ptr->at(static_cast<size_type>(array_index(reference_token))); + } + JSON_CATCH(std::invalid_argument&) + { + JSON_THROW(detail::parse_error::create(109, 0, "array index '" + reference_token + "' is not a number")); + } + break; + } + + default: + JSON_THROW(detail::out_of_range::create(404, "unresolved reference token '" + reference_token + "'")); + } + } + + return *ptr; + } /*! @brief return a const reference to the pointed to value @@ -7123,8 +12061,58 @@ class json_pointer @throw out_of_range.402 if the array index '-' is used @throw out_of_range.404 if the JSON pointer can not be resolved */ - NLOHMANN_BASIC_JSON_TPL_DECLARATION - const NLOHMANN_BASIC_JSON_TPL& get_unchecked(const NLOHMANN_BASIC_JSON_TPL* ptr) const; + const BasicJsonType& get_unchecked(const BasicJsonType* ptr) const + { + using size_type = typename BasicJsonType::size_type; + for (const auto& reference_token : reference_tokens) + { + switch (ptr->m_type) + { + case detail::value_t::object: + { + // use unchecked object access + ptr = &ptr->operator[](reference_token); + break; + } + + case detail::value_t::array: + { + if (JSON_UNLIKELY(reference_token == "-")) + { + // "-" cannot be used for const access + JSON_THROW(detail::out_of_range::create(402, + "array index '-' (" + std::to_string(ptr->m_value.array->size()) + + ") is out of range")); + } + + // error condition (cf. RFC 6901, Sect. 4) + if (JSON_UNLIKELY(reference_token.size() > 1 and reference_token[0] == '0')) + { + JSON_THROW(detail::parse_error::create(106, 0, + "array index '" + reference_token + + "' must not begin with '0'")); + } + + // use unchecked array access + JSON_TRY + { + ptr = &ptr->operator[]( + static_cast<size_type>(array_index(reference_token))); + } + JSON_CATCH(std::invalid_argument&) + { + JSON_THROW(detail::parse_error::create(109, 0, "array index '" + reference_token + "' is not a number")); + } + break; + } + + default: + JSON_THROW(detail::out_of_range::create(404, "unresolved reference token '" + reference_token + "'")); + } + } + + return *ptr; + } /*! @throw parse_error.106 if an array index begins with '0' @@ -7132,8 +12120,57 @@ class json_pointer @throw out_of_range.402 if the array index '-' is used @throw out_of_range.404 if the JSON pointer can not be resolved */ - NLOHMANN_BASIC_JSON_TPL_DECLARATION - const NLOHMANN_BASIC_JSON_TPL& get_checked(const NLOHMANN_BASIC_JSON_TPL* ptr) const; + const BasicJsonType& get_checked(const BasicJsonType* ptr) const + { + using size_type = typename BasicJsonType::size_type; + for (const auto& reference_token : reference_tokens) + { + switch (ptr->m_type) + { + case detail::value_t::object: + { + // note: at performs range check + ptr = &ptr->at(reference_token); + break; + } + + case detail::value_t::array: + { + if (JSON_UNLIKELY(reference_token == "-")) + { + // "-" always fails the range check + JSON_THROW(detail::out_of_range::create(402, + "array index '-' (" + std::to_string(ptr->m_value.array->size()) + + ") is out of range")); + } + + // error condition (cf. RFC 6901, Sect. 4) + if (JSON_UNLIKELY(reference_token.size() > 1 and reference_token[0] == '0')) + { + JSON_THROW(detail::parse_error::create(106, 0, + "array index '" + reference_token + + "' must not begin with '0'")); + } + + // note: at performs range check + JSON_TRY + { + ptr = &ptr->at(static_cast<size_type>(array_index(reference_token))); + } + JSON_CATCH(std::invalid_argument&) + { + JSON_THROW(detail::parse_error::create(109, 0, "array index '" + reference_token + "' is not a number")); + } + break; + } + + default: + JSON_THROW(detail::out_of_range::create(404, "unresolved reference token '" + reference_token + "'")); + } + } + + return *ptr; + } /*! @brief split the string input to reference tokens @@ -7170,11 +12207,11 @@ class json_pointer std::size_t slash = reference_string.find_first_of('/', 1), // set the beginning of the first reference token start = 1; - // we can stop if start == string::npos+1 = 0 + // we can stop if start == 0 (if slash == std::string::npos) start != 0; // set the beginning of the next reference token // (will eventually be 0 if slash == std::string::npos) - start = slash + 1, + start = (slash == std::string::npos) ? 0 : slash + 1, // find next slash slash = reference_string.find_first_of('/', start)) { @@ -7230,7 +12267,7 @@ class json_pointer {} } - /// escape "~"" to "~0" and "/" to "~1" + /// escape "~" to "~0" and "/" to "~1" static std::string escape(std::string s) { replace_substring(s, "~", "~0"); @@ -7252,10 +12289,57 @@ class json_pointer @note Empty objects or arrays are flattened to `null`. */ - NLOHMANN_BASIC_JSON_TPL_DECLARATION static void flatten(const std::string& reference_string, - const NLOHMANN_BASIC_JSON_TPL& value, - NLOHMANN_BASIC_JSON_TPL& result); + const BasicJsonType& value, + BasicJsonType& result) + { + switch (value.m_type) + { + case detail::value_t::array: + { + if (value.m_value.array->empty()) + { + // flatten empty array as null + result[reference_string] = nullptr; + } + else + { + // iterate array and use index as reference string + for (std::size_t i = 0; i < value.m_value.array->size(); ++i) + { + flatten(reference_string + "/" + std::to_string(i), + value.m_value.array->operator[](i), result); + } + } + break; + } + + case detail::value_t::object: + { + if (value.m_value.object->empty()) + { + // flatten empty object as null + result[reference_string] = nullptr; + } + else + { + // iterate object and use keys as reference string + for (const auto& element : *value.m_value.object) + { + flatten(reference_string + "/" + escape(element.first), element.second, result); + } + } + break; + } + + default: + { + // add primitive value with its reference string + result[reference_string] = value; + break; + } + } + } /*! @param[in] value flattened JSON @@ -7267,19 +12351,112 @@ class json_pointer @throw type_error.315 if object values are not primitive @throw type_error.313 if value cannot be unflattened */ - NLOHMANN_BASIC_JSON_TPL_DECLARATION - static NLOHMANN_BASIC_JSON_TPL - unflatten(const NLOHMANN_BASIC_JSON_TPL& value); + static BasicJsonType + unflatten(const BasicJsonType& value) + { + if (JSON_UNLIKELY(not value.is_object())) + { + JSON_THROW(detail::type_error::create(314, "only objects can be unflattened")); + } + + BasicJsonType result; + + // iterate the JSON object values + for (const auto& element : *value.m_value.object) + { + if (JSON_UNLIKELY(not element.second.is_primitive())) + { + JSON_THROW(detail::type_error::create(315, "values in object must be primitive")); + } + + // assign value to reference pointed to by JSON pointer; Note that if + // the JSON pointer is "" (i.e., points to the whole value), function + // get_and_create returns a reference to result itself. An assignment + // will then create a primitive value. + json_pointer(element.first).get_and_create(result) = element.second; + } + + return result; + } friend bool operator==(json_pointer const& lhs, - json_pointer const& rhs) noexcept; + json_pointer const& rhs) noexcept + { + return (lhs.reference_tokens == rhs.reference_tokens); + } friend bool operator!=(json_pointer const& lhs, - json_pointer const& rhs) noexcept; + json_pointer const& rhs) noexcept + { + return not (lhs == rhs); + } /// the reference tokens std::vector<std::string> reference_tokens; }; +} // namespace nlohmann + +// #include <nlohmann/adl_serializer.hpp> + + +#include <utility> + +// #include <nlohmann/detail/conversions/from_json.hpp> + +// #include <nlohmann/detail/conversions/to_json.hpp> + + +namespace nlohmann +{ + +template<typename, typename> +struct adl_serializer +{ + /*! + @brief convert a JSON value to any value type + + This function is usually called by the `get()` function of the + @ref basic_json class (either explicit or via conversion operators). + + @param[in] j JSON value to read from + @param[in,out] val value to write to + */ + template<typename BasicJsonType, typename ValueType> + static auto from_json(BasicJsonType&& j, ValueType& val) noexcept( + noexcept(::nlohmann::from_json(std::forward<BasicJsonType>(j), val))) + -> decltype(::nlohmann::from_json(std::forward<BasicJsonType>(j), val), void()) + { + ::nlohmann::from_json(std::forward<BasicJsonType>(j), val); + } + + /*! + @brief convert any value type to a JSON value + + This function is usually called by the constructors of the @ref basic_json + class. + + @param[in,out] j JSON value to write to + @param[in] val value to read from + */ + template <typename BasicJsonType, typename ValueType> + static auto to_json(BasicJsonType& j, ValueType&& val) noexcept( + noexcept(::nlohmann::to_json(j, std::forward<ValueType>(val)))) + -> decltype(::nlohmann::to_json(j, std::forward<ValueType>(val)), void()) + { + ::nlohmann::to_json(j, std::forward<ValueType>(val)); + } +}; + +} // namespace nlohmann + + +/*! +@brief namespace for Niels Lohmann +@see https://github.com/nlohmann +@since version 1.0.0 +*/ +namespace nlohmann +{ /*! @brief a class to store JSON values @@ -7305,42 +12482,42 @@ and `from_json()` (@ref adl_serializer by default) @requirement The class satisfies the following concept requirements: - Basic - - [DefaultConstructible](http://en.cppreference.com/w/cpp/concept/DefaultConstructible): + - [DefaultConstructible](https://en.cppreference.com/w/cpp/named_req/DefaultConstructible): JSON values can be default constructed. The result will be a JSON null value. - - [MoveConstructible](http://en.cppreference.com/w/cpp/concept/MoveConstructible): + - [MoveConstructible](https://en.cppreference.com/w/cpp/named_req/MoveConstructible): A JSON value can be constructed from an rvalue argument. - - [CopyConstructible](http://en.cppreference.com/w/cpp/concept/CopyConstructible): + - [CopyConstructible](https://en.cppreference.com/w/cpp/named_req/CopyConstructible): A JSON value can be copy-constructed from an lvalue expression. - - [MoveAssignable](http://en.cppreference.com/w/cpp/concept/MoveAssignable): + - [MoveAssignable](https://en.cppreference.com/w/cpp/named_req/MoveAssignable): A JSON value van be assigned from an rvalue argument. - - [CopyAssignable](http://en.cppreference.com/w/cpp/concept/CopyAssignable): + - [CopyAssignable](https://en.cppreference.com/w/cpp/named_req/CopyAssignable): A JSON value can be copy-assigned from an lvalue expression. - - [Destructible](http://en.cppreference.com/w/cpp/concept/Destructible): + - [Destructible](https://en.cppreference.com/w/cpp/named_req/Destructible): JSON values can be destructed. - Layout - - [StandardLayoutType](http://en.cppreference.com/w/cpp/concept/StandardLayoutType): + - [StandardLayoutType](https://en.cppreference.com/w/cpp/named_req/StandardLayoutType): JSON values have - [standard layout](http://en.cppreference.com/w/cpp/language/data_members#Standard_layout): + [standard layout](https://en.cppreference.com/w/cpp/language/data_members#Standard_layout): All non-static data members are private and standard layout types, the class has no virtual functions or (virtual) base classes. - Library-wide - - [EqualityComparable](http://en.cppreference.com/w/cpp/concept/EqualityComparable): + - [EqualityComparable](https://en.cppreference.com/w/cpp/named_req/EqualityComparable): JSON values can be compared with `==`, see @ref operator==(const_reference,const_reference). - - [LessThanComparable](http://en.cppreference.com/w/cpp/concept/LessThanComparable): + - [LessThanComparable](https://en.cppreference.com/w/cpp/named_req/LessThanComparable): JSON values can be compared with `<`, see @ref operator<(const_reference,const_reference). - - [Swappable](http://en.cppreference.com/w/cpp/concept/Swappable): + - [Swappable](https://en.cppreference.com/w/cpp/named_req/Swappable): Any JSON lvalue or rvalue of can be swapped with any lvalue or rvalue of other compatible types, using unqualified function call @ref swap(). - - [NullablePointer](http://en.cppreference.com/w/cpp/concept/NullablePointer): + - [NullablePointer](https://en.cppreference.com/w/cpp/named_req/NullablePointer): JSON values can be compared against `std::nullptr_t` objects which are used to model the `null` value. - Container - - [Container](http://en.cppreference.com/w/cpp/concept/Container): + - [Container](https://en.cppreference.com/w/cpp/named_req/Container): JSON values can be used like STL containers and provide iterator access. - - [ReversibleContainer](http://en.cppreference.com/w/cpp/concept/ReversibleContainer); + - [ReversibleContainer](https://en.cppreference.com/w/cpp/named_req/ReversibleContainer); JSON values can be used like STL containers and provide reverse iterator access. @@ -7367,15 +12544,19 @@ class basic_json { private: template<detail::value_t> friend struct detail::external_constructor; - friend ::nlohmann::json_pointer; + friend ::nlohmann::json_pointer<basic_json>; friend ::nlohmann::detail::parser<basic_json>; friend ::nlohmann::detail::serializer<basic_json>; template<typename BasicJsonType> friend class ::nlohmann::detail::iter_impl; template<typename BasicJsonType, typename CharType> friend class ::nlohmann::detail::binary_writer; - template<typename BasicJsonType> + template<typename BasicJsonType, typename SAX> friend class ::nlohmann::detail::binary_reader; + template<typename BasicJsonType> + friend class ::nlohmann::detail::json_sax_dom_parser; + template<typename BasicJsonType> + friend class ::nlohmann::detail::json_sax_dom_callback_parser; /// workaround type for MSVC using basic_json_t = NLOHMANN_BASIC_JSON_TPL; @@ -7403,13 +12584,19 @@ class basic_json public: using value_t = detail::value_t; - /// @copydoc nlohmann::json_pointer - using json_pointer = ::nlohmann::json_pointer; + /// JSON Pointer, see @ref nlohmann::json_pointer + using json_pointer = ::nlohmann::json_pointer<basic_json>; template<typename T, typename SFINAE> using json_serializer = JSONSerializer<T, SFINAE>; + /// how to treat decoding errors + using error_handler_t = detail::error_handler_t; /// helper type for initializer lists of basic_json values using initializer_list_t = std::initializer_list<detail::json_ref<basic_json>>; + using input_format_t = detail::input_format_t; + /// SAX interface type, see @ref nlohmann::json_sax + using json_sax_t = json_sax<basic_json>; + //////////////// // exceptions // //////////////// @@ -7517,10 +12704,13 @@ class basic_json result["copyright"] = "(C) 2013-2017 Niels Lohmann"; result["name"] = "JSON for Modern C++"; result["url"] = "https://github.com/nlohmann/json"; - result["version"] = - { - {"string", "3.0.1"}, {"major", 3}, {"minor", 0}, {"patch", 1} - }; + result["version"]["string"] = + std::to_string(NLOHMANN_JSON_VERSION_MAJOR) + "." + + std::to_string(NLOHMANN_JSON_VERSION_MINOR) + "." + + std::to_string(NLOHMANN_JSON_VERSION_PATCH); + result["version"]["major"] = NLOHMANN_JSON_VERSION_MAJOR; + result["version"]["minor"] = NLOHMANN_JSON_VERSION_MINOR; + result["version"]["patch"] = NLOHMANN_JSON_VERSION_PATCH; #ifdef _WIN32 result["platform"] = "win32"; @@ -7622,10 +12812,10 @@ class basic_json - When all names are unique, objects will be interoperable in the sense that all software implementations receiving that object will agree on the name-value mappings. - - When the names within an object are not unique, later stored name/value - pairs overwrite previously stored name/value pairs, leaving the used - names unique. For instance, `{"key": 1}` and `{"key": 2, "key": 1}` will - be treated as equal and both stored as `{"key": 1}`. + - When the names within an object are not unique, it is unspecified which + one of the values for a given key will be chosen. For instance, + `{"key": 2, "key": 1}` could be equal to either `{"key": 1}` or + `{"key": 2}`. - Internally, name/value pairs are stored in lexicographical order of the names. Objects will also be serialized (see @ref dump) in this order. For instance, `{"b": 1, "a": 2}` and `{"a": 2, "b": 1}` will be stored @@ -8139,7 +13329,7 @@ class basic_json object = nullptr; // silence warning, see #821 if (JSON_UNLIKELY(t == value_t::null)) { - JSON_THROW(other_error::create(500, "961c151d2e87f2686a955a9be24d316f1362bf21 3.0.1")); // LCOV_EXCL_LINE + JSON_THROW(other_error::create(500, "961c151d2e87f2686a955a9be24d316f1362bf21 3.5.0")); // LCOV_EXCL_LINE } break; } @@ -8182,7 +13372,7 @@ class basic_json array = create<array_t>(std::move(value)); } - void destroy(value_t t) + void destroy(value_t t) noexcept { switch (t) { @@ -8227,7 +13417,7 @@ class basic_json value is changed, because the invariant expresses a relationship between @a m_type and @a m_value. */ - void assert_invariant() const + void assert_invariant() const noexcept { assert(m_type != value_t::object or m_value.object != nullptr); assert(m_type != value_t::array or m_value.array != nullptr); @@ -8307,7 +13497,6 @@ class basic_json */ using parser_callback_t = typename parser::parser_callback_t; - ////////////////// // constructors // ////////////////// @@ -8409,6 +13598,7 @@ class basic_json - @a CompatibleType is not derived from `std::istream`, - @a CompatibleType is not @ref basic_json (to avoid hijacking copy/move constructors), + - @a CompatibleType is not a different @ref basic_json type (i.e. with different template arguments) - @a CompatibleType is not a @ref basic_json nested type (e.g., @ref json_pointer, @ref iterator, etc ...) - @ref @ref json_serializer<U> has a @@ -8432,21 +13622,91 @@ class basic_json @since version 2.1.0 */ - template<typename CompatibleType, typename U = detail::uncvref_t<CompatibleType>, - detail::enable_if_t<not std::is_base_of<std::istream, U>::value and - not std::is_same<U, basic_json_t>::value and - not detail::is_basic_json_nested_type< - basic_json_t, U>::value and - detail::has_to_json<basic_json, U>::value, - int> = 0> - basic_json(CompatibleType && val) noexcept(noexcept(JSONSerializer<U>::to_json( - std::declval<basic_json_t&>(), std::forward<CompatibleType>(val)))) + template <typename CompatibleType, + typename U = detail::uncvref_t<CompatibleType>, + detail::enable_if_t< + not detail::is_basic_json<U>::value and detail::is_compatible_type<basic_json_t, U>::value, int> = 0> + basic_json(CompatibleType && val) noexcept(noexcept( + JSONSerializer<U>::to_json(std::declval<basic_json_t&>(), + std::forward<CompatibleType>(val)))) { JSONSerializer<U>::to_json(*this, std::forward<CompatibleType>(val)); assert_invariant(); } /*! + @brief create a JSON value from an existing one + + This is a constructor for existing @ref basic_json types. + It does not hijack copy/move constructors, since the parameter has different + template arguments than the current ones. + + The constructor tries to convert the internal @ref m_value of the parameter. + + @tparam BasicJsonType a type such that: + - @a BasicJsonType is a @ref basic_json type. + - @a BasicJsonType has different template arguments than @ref basic_json_t. + + @param[in] val the @ref basic_json value to be converted. + + @complexity Usually linear in the size of the passed @a val, also + depending on the implementation of the called `to_json()` + method. + + @exceptionsafety Depends on the called constructor. For types directly + supported by the library (i.e., all types for which no `to_json()` function + was provided), strong guarantee holds: if an exception is thrown, there are + no changes to any JSON value. + + @since version 3.2.0 + */ + template <typename BasicJsonType, + detail::enable_if_t< + detail::is_basic_json<BasicJsonType>::value and not std::is_same<basic_json, BasicJsonType>::value, int> = 0> + basic_json(const BasicJsonType& val) + { + using other_boolean_t = typename BasicJsonType::boolean_t; + using other_number_float_t = typename BasicJsonType::number_float_t; + using other_number_integer_t = typename BasicJsonType::number_integer_t; + using other_number_unsigned_t = typename BasicJsonType::number_unsigned_t; + using other_string_t = typename BasicJsonType::string_t; + using other_object_t = typename BasicJsonType::object_t; + using other_array_t = typename BasicJsonType::array_t; + + switch (val.type()) + { + case value_t::boolean: + JSONSerializer<other_boolean_t>::to_json(*this, val.template get<other_boolean_t>()); + break; + case value_t::number_float: + JSONSerializer<other_number_float_t>::to_json(*this, val.template get<other_number_float_t>()); + break; + case value_t::number_integer: + JSONSerializer<other_number_integer_t>::to_json(*this, val.template get<other_number_integer_t>()); + break; + case value_t::number_unsigned: + JSONSerializer<other_number_unsigned_t>::to_json(*this, val.template get<other_number_unsigned_t>()); + break; + case value_t::string: + JSONSerializer<other_string_t>::to_json(*this, val.template get_ref<const other_string_t&>()); + break; + case value_t::object: + JSONSerializer<other_object_t>::to_json(*this, val.template get_ref<const other_object_t&>()); + break; + case value_t::array: + JSONSerializer<other_array_t>::to_json(*this, val.template get_ref<const other_array_t&>()); + break; + case value_t::null: + *this = nullptr; + break; + case value_t::discarded: + m_type = value_t::discarded; + break; + } + assert_invariant(); + } + + /*! @brief create a container (array or object) from an initializer list Creates a JSON value of type array or object from the passed initializer @@ -8718,7 +13978,7 @@ class basic_json @warning A precondition is enforced with a runtime assertion that will result in calling `std::abort` if this precondition is not met. Assertions can be disabled by defining `NDEBUG` at compile time. - See http://en.cppreference.com/w/cpp/error/assert for more + See https://en.cppreference.com/w/cpp/error/assert for more information. @throw invalid_iterator.201 if iterators @a first and @a last are not @@ -8858,7 +14118,7 @@ class basic_json changes to any JSON value. @requirement This function helps `basic_json` satisfying the - [Container](http://en.cppreference.com/w/cpp/concept/Container) + [Container](https://en.cppreference.com/w/cpp/named_req/Container) requirements: - The complexity is linear. - As postcondition, it holds: `other == basic_json(other)`. @@ -8943,7 +14203,7 @@ class basic_json exceptions. @requirement This function helps `basic_json` satisfying the - [MoveConstructible](http://en.cppreference.com/w/cpp/concept/MoveConstructible) + [MoveConstructible](https://en.cppreference.com/w/cpp/named_req/MoveConstructible) requirements. @liveexample{The code below shows the move constructor explicitly called @@ -8977,7 +14237,7 @@ class basic_json @complexity Linear. @requirement This function helps `basic_json` satisfying the - [Container](http://en.cppreference.com/w/cpp/concept/Container) + [Container](https://en.cppreference.com/w/cpp/named_req/Container) requirements: - The complexity is linear. @@ -8988,7 +14248,7 @@ class basic_json @since version 1.0.0 */ - reference& operator=(basic_json other) noexcept ( + basic_json& operator=(basic_json other) noexcept ( std::is_nothrow_move_constructible<value_t>::value and std::is_nothrow_move_assignable<value_t>::value and std::is_nothrow_move_constructible<json_value>::value and @@ -9014,14 +14274,14 @@ class basic_json @complexity Linear. @requirement This function helps `basic_json` satisfying the - [Container](http://en.cppreference.com/w/cpp/concept/Container) + [Container](https://en.cppreference.com/w/cpp/named_req/Container) requirements: - The complexity is linear. - All stored elements are destroyed and all memory is freed. @since version 1.0.0 */ - ~basic_json() + ~basic_json() noexcept { assert_invariant(); m_value.destroy(m_type); @@ -9054,6 +14314,10 @@ class basic_json @param[in] ensure_ascii If @a ensure_ascii is true, all non-ASCII characters in the output are escaped with `\uXXXX` sequences, and the result consists of ASCII characters only. + @param[in] error_handler how to react on decoding errors; there are three + possible values: `strict` (throws and exception in case a decoding error + occurs; default), `replace` (replace invalid UTF-8 sequences with U+FFFD), + and `ignore` (ignore invalid UTF-8 sequences during serialization). @return string containing the serialization of the JSON value @@ -9072,13 +14336,16 @@ class basic_json @see https://docs.python.org/2/library/json.html#json.dump @since version 1.0.0; indentation character @a indent_char, option - @a ensure_ascii and exceptions added in version 3.0.0 + @a ensure_ascii and exceptions added in version 3.0.0; error + handlers added in version 3.4.0. */ - string_t dump(const int indent = -1, const char indent_char = ' ', - const bool ensure_ascii = false) const + string_t dump(const int indent = -1, + const char indent_char = ' ', + const bool ensure_ascii = false, + const error_handler_t error_handler = error_handler_t::strict) const { string_t result; - serializer s(detail::output_adapter<char>(result), indent_char); + serializer s(detail::output_adapter<char, string_t>(result), indent_char, error_handler); if (indent >= 0) { @@ -9619,11 +14886,34 @@ class basic_json } /*! + @brief get special-case overload + + This overloads converts the current @ref basic_json in a different + @ref basic_json type + + @tparam BasicJsonType == @ref basic_json + + @return a copy of *this, converted into @tparam BasicJsonType + + @complexity Depending on the implementation of the called `from_json()` + method. + + @since version 3.2.0 + */ + template<typename BasicJsonType, detail::enable_if_t< + not std::is_same<BasicJsonType, basic_json>::value and + detail::is_basic_json<BasicJsonType>::value, int> = 0> + BasicJsonType get() const + { + return *this; + } + + /*! @brief get a value (explicit) Explicit type conversion between the JSON value and a compatible value - which is [CopyConstructible](http://en.cppreference.com/w/cpp/concept/CopyConstructible) - and [DefaultConstructible](http://en.cppreference.com/w/cpp/concept/DefaultConstructible). + which is [CopyConstructible](https://en.cppreference.com/w/cpp/named_req/CopyConstructible) + and [DefaultConstructible](https://en.cppreference.com/w/cpp/named_req/DefaultConstructible). The value is converted by calling the @ref json_serializer<ValueType> `from_json()` method. @@ -9659,7 +14949,7 @@ class basic_json */ template<typename ValueTypeCV, typename ValueType = detail::uncvref_t<ValueTypeCV>, detail::enable_if_t < - not std::is_same<basic_json_t, ValueType>::value and + not detail::is_basic_json<ValueType>::value and detail::has_from_json<basic_json_t, ValueType>::value and not detail::has_non_default_from_json<basic_json_t, ValueType>::value, int> = 0> @@ -9683,8 +14973,8 @@ class basic_json @brief get a value (explicit); special case Explicit type conversion between the JSON value and a compatible value - which is **not** [CopyConstructible](http://en.cppreference.com/w/cpp/concept/CopyConstructible) - and **not** [DefaultConstructible](http://en.cppreference.com/w/cpp/concept/DefaultConstructible). + which is **not** [CopyConstructible](https://en.cppreference.com/w/cpp/named_req/CopyConstructible) + and **not** [DefaultConstructible](https://en.cppreference.com/w/cpp/named_req/DefaultConstructible). The value is converted by calling the @ref json_serializer<ValueType> `from_json()` method. @@ -9723,17 +15013,64 @@ class basic_json } /*! - @brief get a pointer value (explicit) + @brief get a value (explicit) - Explicit pointer access to the internally stored JSON value. No copies are + Explicit type conversion between the JSON value and a compatible value. + The value is filled into the input parameter by calling the @ref json_serializer<ValueType> + `from_json()` method. + + The function is equivalent to executing + @code {.cpp} + ValueType v; + JSONSerializer<ValueType>::from_json(*this, v); + @endcode + + This overloads is chosen if: + - @a ValueType is not @ref basic_json, + - @ref json_serializer<ValueType> has a `from_json()` method of the form + `void from_json(const basic_json&, ValueType&)`, and + + @tparam ValueType the input parameter type. + + @return the input parameter, allowing chaining calls. + + @throw what @ref json_serializer<ValueType> `from_json()` method throws + + @liveexample{The example below shows several conversions from JSON values + to other types. There a few things to note: (1) Floating-point numbers can + be converted to integers\, (2) A JSON array can be converted to a standard + `std::vector<short>`\, (3) A JSON object can be converted to C++ + associative containers such as `std::unordered_map<std::string\, + json>`.,get_to} + + @since version 3.3.0 + */ + template<typename ValueType, + detail::enable_if_t < + not detail::is_basic_json<ValueType>::value and + detail::has_from_json<basic_json_t, ValueType>::value, + int> = 0> + ValueType & get_to(ValueType& v) const noexcept(noexcept( + JSONSerializer<ValueType>::from_json(std::declval<const basic_json_t&>(), v))) + { + JSONSerializer<ValueType>::from_json(*this, v); + return v; + } + + + /*! + @brief get a pointer value (implicit) + + Implicit pointer access to the internally stored JSON value. No copies are made. - @warning The pointer becomes invalid if the underlying JSON object - changes. + @warning Writing data to the pointee of the result yields an undefined + state. @tparam PointerType pointer type; must be a pointer to @ref array_t, @ref object_t, @ref string_t, @ref boolean_t, @ref number_integer_t, - @ref number_unsigned_t, or @ref number_float_t. + @ref number_unsigned_t, or @ref number_float_t. Enforced by a static + assertion. @return pointer to the internally stored JSON value if the requested pointer type @a PointerType fits to the JSON value; `nullptr` otherwise @@ -9743,45 +15080,43 @@ class basic_json @liveexample{The example below shows how pointers to internal values of a JSON value can be requested. Note that no type conversions are made and a `nullptr` is returned if the value and the requested pointer type does not - match.,get__PointerType} - - @sa @ref get_ptr() for explicit pointer-member access + match.,get_ptr} @since version 1.0.0 */ template<typename PointerType, typename std::enable_if< std::is_pointer<PointerType>::value, int>::type = 0> - PointerType get() noexcept + auto get_ptr() noexcept -> decltype(std::declval<basic_json_t&>().get_impl_ptr(std::declval<PointerType>())) { - // delegate the call to get_ptr - return get_ptr<PointerType>(); + // delegate the call to get_impl_ptr<>() + return get_impl_ptr(static_cast<PointerType>(nullptr)); } /*! - @brief get a pointer value (explicit) - @copydoc get() + @brief get a pointer value (implicit) + @copydoc get_ptr() */ template<typename PointerType, typename std::enable_if< - std::is_pointer<PointerType>::value, int>::type = 0> - constexpr const PointerType get() const noexcept + std::is_pointer<PointerType>::value and + std::is_const<typename std::remove_pointer<PointerType>::type>::value, int>::type = 0> + constexpr auto get_ptr() const noexcept -> decltype(std::declval<const basic_json_t&>().get_impl_ptr(std::declval<PointerType>())) { - // delegate the call to get_ptr - return get_ptr<PointerType>(); + // delegate the call to get_impl_ptr<>() const + return get_impl_ptr(static_cast<PointerType>(nullptr)); } /*! - @brief get a pointer value (implicit) + @brief get a pointer value (explicit) - Implicit pointer access to the internally stored JSON value. No copies are + Explicit pointer access to the internally stored JSON value. No copies are made. - @warning Writing data to the pointee of the result yields an undefined - state. + @warning The pointer becomes invalid if the underlying JSON object + changes. @tparam PointerType pointer type; must be a pointer to @ref array_t, @ref object_t, @ref string_t, @ref boolean_t, @ref number_integer_t, - @ref number_unsigned_t, or @ref number_float_t. Enforced by a static - assertion. + @ref number_unsigned_t, or @ref number_float_t. @return pointer to the internally stored JSON value if the requested pointer type @a PointerType fits to the JSON value; `nullptr` otherwise @@ -9791,59 +15126,30 @@ class basic_json @liveexample{The example below shows how pointers to internal values of a JSON value can be requested. Note that no type conversions are made and a `nullptr` is returned if the value and the requested pointer type does not - match.,get_ptr} + match.,get__PointerType} + + @sa @ref get_ptr() for explicit pointer-member access @since version 1.0.0 */ template<typename PointerType, typename std::enable_if< std::is_pointer<PointerType>::value, int>::type = 0> - PointerType get_ptr() noexcept + auto get() noexcept -> decltype(std::declval<basic_json_t&>().template get_ptr<PointerType>()) { - // get the type of the PointerType (remove pointer and const) - using pointee_t = typename std::remove_const<typename - std::remove_pointer<typename - std::remove_const<PointerType>::type>::type>::type; - // make sure the type matches the allowed types - static_assert( - std::is_same<object_t, pointee_t>::value - or std::is_same<array_t, pointee_t>::value - or std::is_same<string_t, pointee_t>::value - or std::is_same<boolean_t, pointee_t>::value - or std::is_same<number_integer_t, pointee_t>::value - or std::is_same<number_unsigned_t, pointee_t>::value - or std::is_same<number_float_t, pointee_t>::value - , "incompatible pointer type"); - - // delegate the call to get_impl_ptr<>() - return get_impl_ptr(static_cast<PointerType>(nullptr)); + // delegate the call to get_ptr + return get_ptr<PointerType>(); } /*! - @brief get a pointer value (implicit) - @copydoc get_ptr() + @brief get a pointer value (explicit) + @copydoc get() */ template<typename PointerType, typename std::enable_if< - std::is_pointer<PointerType>::value and - std::is_const<typename std::remove_pointer<PointerType>::type>::value, int>::type = 0> - constexpr const PointerType get_ptr() const noexcept + std::is_pointer<PointerType>::value, int>::type = 0> + constexpr auto get() const noexcept -> decltype(std::declval<const basic_json_t&>().template get_ptr<PointerType>()) { - // get the type of the PointerType (remove pointer and const) - using pointee_t = typename std::remove_const<typename - std::remove_pointer<typename - std::remove_const<PointerType>::type>::type>::type; - // make sure the type matches the allowed types - static_assert( - std::is_same<object_t, pointee_t>::value - or std::is_same<array_t, pointee_t>::value - or std::is_same<string_t, pointee_t>::value - or std::is_same<boolean_t, pointee_t>::value - or std::is_same<number_integer_t, pointee_t>::value - or std::is_same<number_unsigned_t, pointee_t>::value - or std::is_same<number_float_t, pointee_t>::value - , "incompatible pointer type"); - - // delegate the call to get_impl_ptr<>() const - return get_impl_ptr(static_cast<PointerType>(nullptr)); + // delegate the call to get_ptr + return get_ptr<PointerType>(); } /*! @@ -9925,13 +15231,16 @@ class basic_json template < typename ValueType, typename std::enable_if < not std::is_pointer<ValueType>::value and not std::is_same<ValueType, detail::json_ref<basic_json>>::value and - not std::is_same<ValueType, typename string_t::value_type>::value + not std::is_same<ValueType, typename string_t::value_type>::value and + not detail::is_basic_json<ValueType>::value + #ifndef _MSC_VER // fix for issue #167 operator<< ambiguity under VS2015 and not std::is_same<ValueType, std::initializer_list<typename string_t::value_type>>::value -#endif -#if defined(JSON_HAS_CPP_17) +#if defined(JSON_HAS_CPP_17) && defined(_MSC_VER) and _MSC_VER <= 1914 and not std::is_same<ValueType, typename std::string_view>::value #endif +#endif + and detail::is_detected<detail::get_template_function, const basic_json_t&, ValueType>::value , int >::type = 0 > operator ValueType() const { @@ -10195,7 +15504,7 @@ class basic_json return m_value.array->operator[](idx); } - JSON_THROW(type_error::create(305, "cannot use operator[] with " + std::string(type_name()))); + JSON_THROW(type_error::create(305, "cannot use operator[] with a numeric argument with " + std::string(type_name()))); } /*! @@ -10225,7 +15534,7 @@ class basic_json return m_value.array->operator[](idx); } - JSON_THROW(type_error::create(305, "cannot use operator[] with " + std::string(type_name()))); + JSON_THROW(type_error::create(305, "cannot use operator[] with a numeric argument with " + std::string(type_name()))); } /*! @@ -10271,7 +15580,7 @@ class basic_json return m_value.object->operator[](key); } - JSON_THROW(type_error::create(305, "cannot use operator[] with " + std::string(type_name()))); + JSON_THROW(type_error::create(305, "cannot use operator[] with a string argument with " + std::string(type_name()))); } /*! @@ -10313,7 +15622,7 @@ class basic_json return m_value.object->find(key)->second; } - JSON_THROW(type_error::create(305, "cannot use operator[] with " + std::string(type_name()))); + JSON_THROW(type_error::create(305, "cannot use operator[] with a string argument with " + std::string(type_name()))); } /*! @@ -10360,7 +15669,7 @@ class basic_json return m_value.object->operator[](key); } - JSON_THROW(type_error::create(305, "cannot use operator[] with " + std::string(type_name()))); + JSON_THROW(type_error::create(305, "cannot use operator[] with a string argument with " + std::string(type_name()))); } /*! @@ -10403,7 +15712,7 @@ class basic_json return m_value.object->find(key)->second; } - JSON_THROW(type_error::create(305, "cannot use operator[] with " + std::string(type_name()))); + JSON_THROW(type_error::create(305, "cannot use operator[] with a string argument with " + std::string(type_name()))); } /*! @@ -10476,7 +15785,7 @@ class basic_json /*! @brief overload for a default value of type const char* - @copydoc basic_json::value(const typename object_t::key_type&, ValueType) const + @copydoc basic_json::value(const typename object_t::key_type&, const ValueType&) const */ string_t value(const typename object_t::key_type& key, const char* default_value) const { @@ -10512,7 +15821,7 @@ class basic_json @return copy of the element at key @a key or @a default_value if @a key is not found - @throw type_error.306 if the JSON value is not an objec; in that case, + @throw type_error.306 if the JSON value is not an object; in that case, using `value()` with a key makes no sense. @complexity Logarithmic in the size of the container. @@ -10536,7 +15845,7 @@ class basic_json { return ptr.get_checked(this); } - JSON_CATCH (out_of_range&) + JSON_INTERNAL_CATCH (out_of_range&) { return default_value; } @@ -11047,7 +16356,7 @@ class basic_json @complexity Constant. @requirement This function helps `basic_json` satisfying the - [Container](http://en.cppreference.com/w/cpp/concept/Container) + [Container](https://en.cppreference.com/w/cpp/named_req/Container) requirements: - The complexity is constant. @@ -11086,7 +16395,7 @@ class basic_json @complexity Constant. @requirement This function helps `basic_json` satisfying the - [Container](http://en.cppreference.com/w/cpp/concept/Container) + [Container](https://en.cppreference.com/w/cpp/named_req/Container) requirements: - The complexity is constant. - Has the semantics of `const_cast<const basic_json&>(*this).begin()`. @@ -11118,7 +16427,7 @@ class basic_json @complexity Constant. @requirement This function helps `basic_json` satisfying the - [Container](http://en.cppreference.com/w/cpp/concept/Container) + [Container](https://en.cppreference.com/w/cpp/named_req/Container) requirements: - The complexity is constant. @@ -11157,7 +16466,7 @@ class basic_json @complexity Constant. @requirement This function helps `basic_json` satisfying the - [Container](http://en.cppreference.com/w/cpp/concept/Container) + [Container](https://en.cppreference.com/w/cpp/named_req/Container) requirements: - The complexity is constant. - Has the semantics of `const_cast<const basic_json&>(*this).end()`. @@ -11187,7 +16496,7 @@ class basic_json @complexity Constant. @requirement This function helps `basic_json` satisfying the - [ReversibleContainer](http://en.cppreference.com/w/cpp/concept/ReversibleContainer) + [ReversibleContainer](https://en.cppreference.com/w/cpp/named_req/ReversibleContainer) requirements: - The complexity is constant. - Has the semantics of `reverse_iterator(end())`. @@ -11224,7 +16533,7 @@ class basic_json @complexity Constant. @requirement This function helps `basic_json` satisfying the - [ReversibleContainer](http://en.cppreference.com/w/cpp/concept/ReversibleContainer) + [ReversibleContainer](https://en.cppreference.com/w/cpp/named_req/ReversibleContainer) requirements: - The complexity is constant. - Has the semantics of `reverse_iterator(begin())`. @@ -11261,7 +16570,7 @@ class basic_json @complexity Constant. @requirement This function helps `basic_json` satisfying the - [ReversibleContainer](http://en.cppreference.com/w/cpp/concept/ReversibleContainer) + [ReversibleContainer](https://en.cppreference.com/w/cpp/named_req/ReversibleContainer) requirements: - The complexity is constant. - Has the semantics of `const_cast<const basic_json&>(*this).rbegin()`. @@ -11290,7 +16599,7 @@ class basic_json @complexity Constant. @requirement This function helps `basic_json` satisfying the - [ReversibleContainer](http://en.cppreference.com/w/cpp/concept/ReversibleContainer) + [ReversibleContainer](https://en.cppreference.com/w/cpp/named_req/ReversibleContainer) requirements: - The complexity is constant. - Has the semantics of `const_cast<const basic_json&>(*this).rend()`. @@ -11361,18 +16670,100 @@ class basic_json @note The name of this function is not yet final and may change in the future. + + @deprecated This stream operator is deprecated and will be removed in + future 4.0.0 of the library. Please use @ref items() instead; + that is, replace `json::iterator_wrapper(j)` with `j.items()`. */ - static iteration_proxy<iterator> iterator_wrapper(reference ref) + JSON_DEPRECATED + static iteration_proxy<iterator> iterator_wrapper(reference ref) noexcept { - return iteration_proxy<iterator>(ref); + return ref.items(); } /*! @copydoc iterator_wrapper(reference) */ - static iteration_proxy<const_iterator> iterator_wrapper(const_reference ref) + JSON_DEPRECATED + static iteration_proxy<const_iterator> iterator_wrapper(const_reference ref) noexcept { - return iteration_proxy<const_iterator>(ref); + return ref.items(); + } + + /*! + @brief helper to access iterator member functions in range-based for + + This function allows to access @ref iterator::key() and @ref + iterator::value() during range-based for loops. In these loops, a + reference to the JSON values is returned, so there is no access to the + underlying iterator. + + For loop without `items()` function: + + @code{cpp} + for (auto it = j_object.begin(); it != j_object.end(); ++it) + { + std::cout << "key: " << it.key() << ", value:" << it.value() << '\n'; + } + @endcode + + Range-based for loop without `items()` function: + + @code{cpp} + for (auto it : j_object) + { + // "it" is of type json::reference and has no key() member + std::cout << "value: " << it << '\n'; + } + @endcode + + Range-based for loop with `items()` function: + + @code{cpp} + for (auto& el : j_object.items()) + { + std::cout << "key: " << el.key() << ", value:" << el.value() << '\n'; + } + @endcode + + The `items()` function also allows to use + [structured bindings](https://en.cppreference.com/w/cpp/language/structured_binding) + (C++17): + + @code{cpp} + for (auto& [key, val] : j_object.items()) + { + std::cout << "key: " << key << ", value:" << val << '\n'; + } + @endcode + + @note When iterating over an array, `key()` will return the index of the + element as string (see example). For primitive types (e.g., numbers), + `key()` returns an empty string. + + @return iteration proxy object wrapping @a ref with an interface to use in + range-based for loops + + @liveexample{The following code shows how the function is used.,items} + + @exceptionsafety Strong guarantee: if an exception is thrown, there are no + changes in the JSON value. + + @complexity Constant. + + @since version 3.1.0, structured bindings support since 3.5.0. + */ + iteration_proxy<iterator> items() noexcept + { + return iteration_proxy<iterator>(*this); + } + + /*! + @copydoc items() + */ + iteration_proxy<const_iterator> items() const noexcept + { + return iteration_proxy<const_iterator>(*this); } /// @} @@ -11417,7 +16808,7 @@ class basic_json false in the case of a string. @requirement This function helps `basic_json` satisfying the - [Container](http://en.cppreference.com/w/cpp/concept/Container) + [Container](https://en.cppreference.com/w/cpp/named_req/Container) requirements: - The complexity is constant. - Has the semantics of `begin() == end()`. @@ -11488,7 +16879,7 @@ class basic_json the case of a string. @requirement This function helps `basic_json` satisfying the - [Container](http://en.cppreference.com/w/cpp/concept/Container) + [Container](https://en.cppreference.com/w/cpp/named_req/Container) requirements: - The complexity is constant. - Has the semantics of `std::distance(begin(), end())`. @@ -11558,7 +16949,7 @@ class basic_json @exceptionsafety No-throw guarantee: this function never throws exceptions. @requirement This function helps `basic_json` satisfying the - [Container](http://en.cppreference.com/w/cpp/concept/Container) + [Container](https://en.cppreference.com/w/cpp/named_req/Container) requirements: - The complexity is constant. - Has the semantics of returning `b.size()` where `b` is the largest @@ -11970,6 +17361,26 @@ class basic_json return {it, res.second}; } + /// Helper for insertion of an iterator + /// @note: This uses std::distance to support GCC 4.8, + /// see https://github.com/nlohmann/json/pull/1257 + template<typename... Args> + iterator insert_iterator(const_iterator pos, Args&& ... args) + { + iterator result(this); + assert(m_value.array != nullptr); + + auto insert_pos = std::distance(m_value.array->begin(), pos.m_it.array_iterator); + m_value.array->insert(pos.m_it.array_iterator, std::forward<Args>(args)...); + result.m_it.array_iterator = m_value.array->begin() + insert_pos; + + // This could have been written as: + // result.m_it.array_iterator = m_value.array->insert(pos.m_it.array_iterator, cnt, val); + // but the return value of insert is missing in GCC 4.8, so it is written this way instead. + + return result; + } + /*! @brief inserts element @@ -12004,9 +17415,7 @@ class basic_json } // insert to array and return iterator - iterator result(this); - result.m_it.array_iterator = m_value.array->insert(pos.m_it.array_iterator, val); - return result; + return insert_iterator(pos, val); } JSON_THROW(type_error::create(309, "cannot use insert() with " + std::string(type_name()))); @@ -12057,9 +17466,7 @@ class basic_json } // insert to array and return iterator - iterator result(this); - result.m_it.array_iterator = m_value.array->insert(pos.m_it.array_iterator, cnt, val); - return result; + return insert_iterator(pos, cnt, val); } JSON_THROW(type_error::create(309, "cannot use insert() with " + std::string(type_name()))); @@ -12121,12 +17528,7 @@ class basic_json } // insert to array and return iterator - iterator result(this); - result.m_it.array_iterator = m_value.array->insert( - pos.m_it.array_iterator, - first.m_it.array_iterator, - last.m_it.array_iterator); - return result; + return insert_iterator(pos, first.m_it.array_iterator, last.m_it.array_iterator); } /*! @@ -12168,9 +17570,7 @@ class basic_json } // insert to array and return iterator - iterator result(this); - result.m_it.array_iterator = m_value.array->insert(pos.m_it.array_iterator, ilist.begin(), ilist.end()); - return result; + return insert_iterator(pos, ilist.begin(), ilist.end()); } /*! @@ -12312,7 +17712,7 @@ class basic_json // passed iterators must belong to objects if (JSON_UNLIKELY(not first.m_object->is_object() - or not first.m_object->is_object())) + or not last.m_object->is_object())) { JSON_THROW(invalid_iterator::create(202, "iterators first and last must point to objects")); } @@ -12943,8 +18343,8 @@ class basic_json /*! @brief serialize to stream - @deprecated This stream operator is deprecated and will be removed in a - future version of the library. Please use + @deprecated This stream operator is deprecated and will be removed in + future 4.0.0 of the library. Please use @ref operator<<(std::ostream&, const basic_json&) instead; that is, replace calls like `j >> o;` with `o << j;`. @since version 1.0.0; deprecated since version 3.0.0 @@ -12999,6 +18399,8 @@ class basic_json @param[in] cb a parser callback function of type @ref parser_callback_t which is used to control the deserialization by filtering unwanted values (optional) + @param[in] allow_exceptions whether to throw exceptions in case of a + parse error (optional, true by default) @return result of the deserialization @@ -13027,19 +18429,7 @@ class basic_json @since version 2.0.3 (contiguous containers) */ - static basic_json parse(detail::input_adapter i, - const parser_callback_t cb = nullptr, - const bool allow_exceptions = true) - { - basic_json result; - parser(i, cb, allow_exceptions).parse(true, result); - return result; - } - - /*! - @copydoc basic_json parse(detail::input_adapter, const parser_callback_t) - */ - static basic_json parse(detail::input_adapter& i, + static basic_json parse(detail::input_adapter&& i, const parser_callback_t cb = nullptr, const bool allow_exceptions = true) { @@ -13048,14 +18438,80 @@ class basic_json return result; } - static bool accept(detail::input_adapter i) + static bool accept(detail::input_adapter&& i) { return parser(i).accept(true); } - static bool accept(detail::input_adapter& i) + /*! + @brief generate SAX events + + The SAX event lister must follow the interface of @ref json_sax. + + This function reads from a compatible input. Examples are: + - an array of 1-byte values + - strings with character/literal type with size of 1 byte + - input streams + - container with contiguous storage of 1-byte values. Compatible container + types include `std::vector`, `std::string`, `std::array`, + `std::valarray`, and `std::initializer_list`. Furthermore, C-style + arrays can be used with `std::begin()`/`std::end()`. User-defined + containers can be used as long as they implement random-access iterators + and a contiguous storage. + + @pre Each element of the container has a size of 1 byte. Violating this + precondition yields undefined behavior. **This precondition is enforced + with a static assertion.** + + @pre The container storage is contiguous. Violating this precondition + yields undefined behavior. **This precondition is enforced with an + assertion.** + @pre Each element of the container has a size of 1 byte. Violating this + precondition yields undefined behavior. **This precondition is enforced + with a static assertion.** + + @warning There is no way to enforce all preconditions at compile-time. If + the function is called with a noncompliant container and with + assertions switched off, the behavior is undefined and will most + likely yield segmentation violation. + + @param[in] i input to read from + @param[in,out] sax SAX event listener + @param[in] format the format to parse (JSON, CBOR, MessagePack, or UBJSON) + @param[in] strict whether the input has to be consumed completely + + @return return value of the last processed SAX event + + @throw parse_error.101 if a parse error occurs; example: `""unexpected end + of input; expected string literal""` + @throw parse_error.102 if to_unicode fails or surrogate error + @throw parse_error.103 if to_unicode fails + + @complexity Linear in the length of the input. The parser is a predictive + LL(1) parser. The complexity can be higher if the SAX consumer @a sax has + a super-linear complexity. + + @note A UTF-8 byte order mark is silently ignored. + + @liveexample{The example below demonstrates the `sax_parse()` function + reading from string and processing the events with a user-defined SAX + event consumer.,sax_parse} + + @since version 3.2.0 + */ + template <typename SAX> + static bool sax_parse(detail::input_adapter&& i, SAX* sax, + input_format_t format = input_format_t::json, + const bool strict = true) { - return parser(i).accept(true); + assert(sax); + switch (format) + { + case input_format_t::json: + return parser(std::move(i)).sax_parse(sax, strict); + default: + return detail::binary_reader<basic_json, SAX>(std::move(i)).sax_parse(format, sax, strict); + } } /*! @@ -13127,10 +18583,19 @@ class basic_json return parser(detail::input_adapter(first, last)).accept(true); } + template<class IteratorType, class SAX, typename std::enable_if< + std::is_base_of< + std::random_access_iterator_tag, + typename std::iterator_traits<IteratorType>::iterator_category>::value, int>::type = 0> + static bool sax_parse(IteratorType first, IteratorType last, SAX* sax) + { + return parser(detail::input_adapter(first, last)).sax_parse(sax); + } + /*! @brief deserialize from stream - @deprecated This stream operator is deprecated and will be removed in a - future version of the library. Please use + @deprecated This stream operator is deprecated and will be removed in + version 4.0.0 of the library. Please use @ref operator>>(std::istream&, basic_json&) instead; that is, replace calls like `j << i;` with `i >> j;`. @since version 1.0.0; deprecated since version 3.0.0 @@ -13331,9 +18796,11 @@ class basic_json vector in CBOR format.,to_cbor} @sa http://cbor.io - @sa @ref from_cbor(const std::vector<uint8_t>&, const size_t) for the + @sa @ref from_cbor(detail::input_adapter&&, const bool, const bool) for the analogous deserialization @sa @ref to_msgpack(const basic_json&) for the related MessagePack format + @sa @ref to_ubjson(const basic_json&, const bool, const bool) for the + related UBJSON format @since version 2.0.9 */ @@ -13426,9 +18893,10 @@ class basic_json vector in MessagePack format.,to_msgpack} @sa http://msgpack.org - @sa @ref from_msgpack(const std::vector<uint8_t>&, const size_t) for the - analogous deserialization + @sa @ref from_msgpack for the analogous deserialization @sa @ref to_cbor(const basic_json& for the related CBOR format + @sa @ref to_ubjson(const basic_json&, const bool, const bool) for the + related UBJSON format @since version 2.0.9 */ @@ -13450,6 +18918,192 @@ class basic_json } /*! + @brief create a UBJSON serialization of a given JSON value + + Serializes a given JSON value @a j to a byte vector using the UBJSON + (Universal Binary JSON) serialization format. UBJSON aims to be more compact + than JSON itself, yet more efficient to parse. + + The library uses the following mapping from JSON values types to + UBJSON types according to the UBJSON specification: + + JSON value type | value/range | UBJSON type | marker + --------------- | --------------------------------- | ----------- | ------ + null | `null` | null | `Z` + boolean | `true` | true | `T` + boolean | `false` | false | `F` + number_integer | -9223372036854775808..-2147483649 | int64 | `L` + number_integer | -2147483648..-32769 | int32 | `l` + number_integer | -32768..-129 | int16 | `I` + number_integer | -128..127 | int8 | `i` + number_integer | 128..255 | uint8 | `U` + number_integer | 256..32767 | int16 | `I` + number_integer | 32768..2147483647 | int32 | `l` + number_integer | 2147483648..9223372036854775807 | int64 | `L` + number_unsigned | 0..127 | int8 | `i` + number_unsigned | 128..255 | uint8 | `U` + number_unsigned | 256..32767 | int16 | `I` + number_unsigned | 32768..2147483647 | int32 | `l` + number_unsigned | 2147483648..9223372036854775807 | int64 | `L` + number_float | *any value* | float64 | `D` + string | *with shortest length indicator* | string | `S` + array | *see notes on optimized format* | array | `[` + object | *see notes on optimized format* | map | `{` + + @note The mapping is **complete** in the sense that any JSON value type + can be converted to a UBJSON value. + + @note The following values can **not** be converted to a UBJSON value: + - strings with more than 9223372036854775807 bytes (theoretical) + - unsigned integer numbers above 9223372036854775807 + + @note The following markers are not used in the conversion: + - `Z`: no-op values are not created. + - `C`: single-byte strings are serialized with `S` markers. + + @note Any UBJSON output created @ref to_ubjson can be successfully parsed + by @ref from_ubjson. + + @note If NaN or Infinity are stored inside a JSON number, they are + serialized properly. This behavior differs from the @ref dump() + function which serializes NaN or Infinity to `null`. + + @note The optimized formats for containers are supported: Parameter + @a use_size adds size information to the beginning of a container and + removes the closing marker. Parameter @a use_type further checks + whether all elements of a container have the same type and adds the + type marker to the beginning of the container. The @a use_type + parameter must only be used together with @a use_size = true. Note + that @a use_size = true alone may result in larger representations - + the benefit of this parameter is that the receiving side is + immediately informed on the number of elements of the container. + + @param[in] j JSON value to serialize + @param[in] use_size whether to add size annotations to container types + @param[in] use_type whether to add type annotations to container types + (must be combined with @a use_size = true) + @return UBJSON serialization as byte vector + + @complexity Linear in the size of the JSON value @a j. + + @liveexample{The example shows the serialization of a JSON value to a byte + vector in UBJSON format.,to_ubjson} + + @sa http://ubjson.org + @sa @ref from_ubjson(detail::input_adapter&&, const bool, const bool) for the + analogous deserialization + @sa @ref to_cbor(const basic_json& for the related CBOR format + @sa @ref to_msgpack(const basic_json&) for the related MessagePack format + + @since version 3.1.0 + */ + static std::vector<uint8_t> to_ubjson(const basic_json& j, + const bool use_size = false, + const bool use_type = false) + { + std::vector<uint8_t> result; + to_ubjson(j, result, use_size, use_type); + return result; + } + + static void to_ubjson(const basic_json& j, detail::output_adapter<uint8_t> o, + const bool use_size = false, const bool use_type = false) + { + binary_writer<uint8_t>(o).write_ubjson(j, use_size, use_type); + } + + static void to_ubjson(const basic_json& j, detail::output_adapter<char> o, + const bool use_size = false, const bool use_type = false) + { + binary_writer<char>(o).write_ubjson(j, use_size, use_type); + } + + + /*! + @brief Serializes the given JSON object `j` to BSON and returns a vector + containing the corresponding BSON-representation. + + BSON (Binary JSON) is a binary format in which zero or more ordered key/value pairs are + stored as a single entity (a so-called document). + + The library uses the following mapping from JSON values types to BSON types: + + JSON value type | value/range | BSON type | marker + --------------- | --------------------------------- | ----------- | ------ + null | `null` | null | 0x0A + boolean | `true`, `false` | boolean | 0x08 + number_integer | -9223372036854775808..-2147483649 | int64 | 0x12 + number_integer | -2147483648..2147483647 | int32 | 0x10 + number_integer | 2147483648..9223372036854775807 | int64 | 0x12 + number_unsigned | 0..2147483647 | int32 | 0x10 + number_unsigned | 2147483648..9223372036854775807 | int64 | 0x12 + number_unsigned | 9223372036854775808..18446744073709551615| -- | -- + number_float | *any value* | double | 0x01 + string | *any value* | string | 0x02 + array | *any value* | document | 0x04 + object | *any value* | document | 0x03 + + @warning The mapping is **incomplete**, since only JSON-objects (and things + contained therein) can be serialized to BSON. + Also, integers larger than 9223372036854775807 cannot be serialized to BSON, + and the keys may not contain U+0000, since they are serialized a + zero-terminated c-strings. + + @throw out_of_range.407 if `j.is_number_unsigned() && j.get<std::uint64_t>() > 9223372036854775807` + @throw out_of_range.409 if a key in `j` contains a NULL (U+0000) + @throw type_error.317 if `!j.is_object()` + + @pre The input `j` is required to be an object: `j.is_object() == true`. + + @note Any BSON output created via @ref to_bson can be successfully parsed + by @ref from_bson. + + @param[in] j JSON value to serialize + @return BSON serialization as byte vector + + @complexity Linear in the size of the JSON value @a j. + + @liveexample{The example shows the serialization of a JSON value to a byte + vector in BSON format.,to_bson} + + @sa http://bsonspec.org/spec.html + @sa @ref from_bson(detail::input_adapter&&, const bool strict) for the + analogous deserialization + @sa @ref to_ubjson(const basic_json&, const bool, const bool) for the + related UBJSON format + @sa @ref to_cbor(const basic_json&) for the related CBOR format + @sa @ref to_msgpack(const basic_json&) for the related MessagePack format + */ + static std::vector<uint8_t> to_bson(const basic_json& j) + { + std::vector<uint8_t> result; + to_bson(j, result); + return result; + } + + /*! + @brief Serializes the given JSON object `j` to BSON and forwards the + corresponding BSON-representation to the given output_adapter `o`. + @param j The JSON object to convert to BSON. + @param o The output adapter that receives the binary BSON representation. + @pre The input `j` shall be an object: `j.is_object() == true` + @sa @ref to_bson(const basic_json&) + */ + static void to_bson(const basic_json& j, detail::output_adapter<uint8_t> o) + { + binary_writer<uint8_t>(o).write_bson(j); + } + + /*! + @copydoc to_bson(const basic_json&, detail::output_adapter<uint8_t>) + */ + static void to_bson(const basic_json& j, detail::output_adapter<char> o) + { + binary_writer<char>(o).write_bson(j); + } + + + /*! @brief create a JSON value from an input in CBOR format Deserializes a given input @a i to a JSON value using the CBOR (Concise @@ -13490,7 +19144,7 @@ class basic_json map | object | 0xBF False | `false` | 0xF4 True | `true` | 0xF5 - Nill | `null` | 0xF6 + Null | `null` | 0xF6 Half-Precision Float | number_float | 0xF9 Single-Precision Float | number_float | 0xFA Double-Precision Float | number_float | 0xFB @@ -13518,6 +19172,9 @@ class basic_json @param[in] i an input in CBOR format convertible to an input adapter @param[in] strict whether to expect the input to be consumed until EOF (true by default) + @param[in] allow_exceptions whether to throw exceptions in case of a + parse error (optional, true by default) + @return deserialized JSON value @throw parse_error.110 if the given input ends prematurely or the end of @@ -13533,27 +19190,39 @@ class basic_json @sa http://cbor.io @sa @ref to_cbor(const basic_json&) for the analogous serialization - @sa @ref from_msgpack(detail::input_adapter, const bool) for the + @sa @ref from_msgpack(detail::input_adapter&&, const bool, const bool) for the related MessagePack format + @sa @ref from_ubjson(detail::input_adapter&&, const bool, const bool) for the + related UBJSON format @since version 2.0.9; parameter @a start_index since 2.1.1; changed to consume input adapters, removed start_index parameter, and added - @a strict parameter since 3.0.0 + @a strict parameter since 3.0.0; added @a allow_exceptions parameter + since 3.2.0 */ - static basic_json from_cbor(detail::input_adapter i, - const bool strict = true) + static basic_json from_cbor(detail::input_adapter&& i, + const bool strict = true, + const bool allow_exceptions = true) { - return binary_reader(i).parse_cbor(strict); + basic_json result; + detail::json_sax_dom_parser<basic_json> sdp(result, allow_exceptions); + const bool res = binary_reader(detail::input_adapter(i)).sax_parse(input_format_t::cbor, &sdp, strict); + return res ? result : basic_json(value_t::discarded); } /*! - @copydoc from_cbor(detail::input_adapter, const bool) + @copydoc from_cbor(detail::input_adapter&&, const bool, const bool) */ template<typename A1, typename A2, detail::enable_if_t<std::is_constructible<detail::input_adapter, A1, A2>::value, int> = 0> - static basic_json from_cbor(A1 && a1, A2 && a2, const bool strict = true) + static basic_json from_cbor(A1 && a1, A2 && a2, + const bool strict = true, + const bool allow_exceptions = true) { - return binary_reader(detail::input_adapter(std::forward<A1>(a1), std::forward<A2>(a2))).parse_cbor(strict); + basic_json result; + detail::json_sax_dom_parser<basic_json> sdp(result, allow_exceptions); + const bool res = binary_reader(detail::input_adapter(std::forward<A1>(a1), std::forward<A2>(a2))).sax_parse(input_format_t::cbor, &sdp, strict); + return res ? result : basic_json(value_t::discarded); } /*! @@ -13606,6 +19275,10 @@ class basic_json adapter @param[in] strict whether to expect the input to be consumed until EOF (true by default) + @param[in] allow_exceptions whether to throw exceptions in case of a + parse error (optional, true by default) + + @return deserialized JSON value @throw parse_error.110 if the given input ends prematurely or the end of file was not reached when @a strict was set to true @@ -13620,29 +19293,212 @@ class basic_json @sa http://msgpack.org @sa @ref to_msgpack(const basic_json&) for the analogous serialization - @sa @ref from_cbor(detail::input_adapter, const bool) for the related CBOR - format + @sa @ref from_cbor(detail::input_adapter&&, const bool, const bool) for the + related CBOR format + @sa @ref from_ubjson(detail::input_adapter&&, const bool, const bool) for + the related UBJSON format + @sa @ref from_bson(detail::input_adapter&&, const bool, const bool) for + the related BSON format @since version 2.0.9; parameter @a start_index since 2.1.1; changed to consume input adapters, removed start_index parameter, and added - @a strict parameter since 3.0.0 + @a strict parameter since 3.0.0; added @a allow_exceptions parameter + since 3.2.0 + */ + static basic_json from_msgpack(detail::input_adapter&& i, + const bool strict = true, + const bool allow_exceptions = true) + { + basic_json result; + detail::json_sax_dom_parser<basic_json> sdp(result, allow_exceptions); + const bool res = binary_reader(detail::input_adapter(i)).sax_parse(input_format_t::msgpack, &sdp, strict); + return res ? result : basic_json(value_t::discarded); + } + + /*! + @copydoc from_msgpack(detail::input_adapter&&, const bool, const bool) + */ + template<typename A1, typename A2, + detail::enable_if_t<std::is_constructible<detail::input_adapter, A1, A2>::value, int> = 0> + static basic_json from_msgpack(A1 && a1, A2 && a2, + const bool strict = true, + const bool allow_exceptions = true) + { + basic_json result; + detail::json_sax_dom_parser<basic_json> sdp(result, allow_exceptions); + const bool res = binary_reader(detail::input_adapter(std::forward<A1>(a1), std::forward<A2>(a2))).sax_parse(input_format_t::msgpack, &sdp, strict); + return res ? result : basic_json(value_t::discarded); + } + + /*! + @brief create a JSON value from an input in UBJSON format + + Deserializes a given input @a i to a JSON value using the UBJSON (Universal + Binary JSON) serialization format. + + The library maps UBJSON types to JSON value types as follows: + + UBJSON type | JSON value type | marker + ----------- | --------------------------------------- | ------ + no-op | *no value, next value is read* | `N` + null | `null` | `Z` + false | `false` | `F` + true | `true` | `T` + float32 | number_float | `d` + float64 | number_float | `D` + uint8 | number_unsigned | `U` + int8 | number_integer | `i` + int16 | number_integer | `I` + int32 | number_integer | `l` + int64 | number_integer | `L` + string | string | `S` + char | string | `C` + array | array (optimized values are supported) | `[` + object | object (optimized values are supported) | `{` + + @note The mapping is **complete** in the sense that any UBJSON value can + be converted to a JSON value. + + @param[in] i an input in UBJSON format convertible to an input adapter + @param[in] strict whether to expect the input to be consumed until EOF + (true by default) + @param[in] allow_exceptions whether to throw exceptions in case of a + parse error (optional, true by default) + + @return deserialized JSON value + + @throw parse_error.110 if the given input ends prematurely or the end of + file was not reached when @a strict was set to true + @throw parse_error.112 if a parse error occurs + @throw parse_error.113 if a string could not be parsed successfully + + @complexity Linear in the size of the input @a i. + + @liveexample{The example shows the deserialization of a byte vector in + UBJSON format to a JSON value.,from_ubjson} + + @sa http://ubjson.org + @sa @ref to_ubjson(const basic_json&, const bool, const bool) for the + analogous serialization + @sa @ref from_cbor(detail::input_adapter&&, const bool, const bool) for the + related CBOR format + @sa @ref from_msgpack(detail::input_adapter&&, const bool, const bool) for + the related MessagePack format + @sa @ref from_bson(detail::input_adapter&&, const bool, const bool) for + the related BSON format + + @since version 3.1.0; added @a allow_exceptions parameter since 3.2.0 */ - static basic_json from_msgpack(detail::input_adapter i, - const bool strict = true) + static basic_json from_ubjson(detail::input_adapter&& i, + const bool strict = true, + const bool allow_exceptions = true) { - return binary_reader(i).parse_msgpack(strict); + basic_json result; + detail::json_sax_dom_parser<basic_json> sdp(result, allow_exceptions); + const bool res = binary_reader(detail::input_adapter(i)).sax_parse(input_format_t::ubjson, &sdp, strict); + return res ? result : basic_json(value_t::discarded); } /*! - @copydoc from_msgpack(detail::input_adapter, const bool) + @copydoc from_ubjson(detail::input_adapter&&, const bool, const bool) */ template<typename A1, typename A2, detail::enable_if_t<std::is_constructible<detail::input_adapter, A1, A2>::value, int> = 0> - static basic_json from_msgpack(A1 && a1, A2 && a2, const bool strict = true) + static basic_json from_ubjson(A1 && a1, A2 && a2, + const bool strict = true, + const bool allow_exceptions = true) { - return binary_reader(detail::input_adapter(std::forward<A1>(a1), std::forward<A2>(a2))).parse_msgpack(strict); + basic_json result; + detail::json_sax_dom_parser<basic_json> sdp(result, allow_exceptions); + const bool res = binary_reader(detail::input_adapter(std::forward<A1>(a1), std::forward<A2>(a2))).sax_parse(input_format_t::ubjson, &sdp, strict); + return res ? result : basic_json(value_t::discarded); } + /*! + @brief Create a JSON value from an input in BSON format + + Deserializes a given input @a i to a JSON value using the BSON (Binary JSON) + serialization format. + + The library maps BSON record types to JSON value types as follows: + + BSON type | BSON marker byte | JSON value type + --------------- | ---------------- | --------------------------- + double | 0x01 | number_float + string | 0x02 | string + document | 0x03 | object + array | 0x04 | array + binary | 0x05 | still unsupported + undefined | 0x06 | still unsupported + ObjectId | 0x07 | still unsupported + boolean | 0x08 | boolean + UTC Date-Time | 0x09 | still unsupported + null | 0x0A | null + Regular Expr. | 0x0B | still unsupported + DB Pointer | 0x0C | still unsupported + JavaScript Code | 0x0D | still unsupported + Symbol | 0x0E | still unsupported + JavaScript Code | 0x0F | still unsupported + int32 | 0x10 | number_integer + Timestamp | 0x11 | still unsupported + 128-bit decimal float | 0x13 | still unsupported + Max Key | 0x7F | still unsupported + Min Key | 0xFF | still unsupported + + @warning The mapping is **incomplete**. The unsupported mappings + are indicated in the table above. + + @param[in] i an input in BSON format convertible to an input adapter + @param[in] strict whether to expect the input to be consumed until EOF + (true by default) + @param[in] allow_exceptions whether to throw exceptions in case of a + parse error (optional, true by default) + + @return deserialized JSON value + + @throw parse_error.114 if an unsupported BSON record type is encountered + + @complexity Linear in the size of the input @a i. + + @liveexample{The example shows the deserialization of a byte vector in + BSON format to a JSON value.,from_bson} + + @sa http://bsonspec.org/spec.html + @sa @ref to_bson(const basic_json&) for the analogous serialization + @sa @ref from_cbor(detail::input_adapter&&, const bool, const bool) for the + related CBOR format + @sa @ref from_msgpack(detail::input_adapter&&, const bool, const bool) for + the related MessagePack format + @sa @ref from_ubjson(detail::input_adapter&&, const bool, const bool) for the + related UBJSON format + */ + static basic_json from_bson(detail::input_adapter&& i, + const bool strict = true, + const bool allow_exceptions = true) + { + basic_json result; + detail::json_sax_dom_parser<basic_json> sdp(result, allow_exceptions); + const bool res = binary_reader(detail::input_adapter(i)).sax_parse(input_format_t::bson, &sdp, strict); + return res ? result : basic_json(value_t::discarded); + } + + /*! + @copydoc from_bson(detail::input_adapter&&, const bool, const bool) + */ + template<typename A1, typename A2, + detail::enable_if_t<std::is_constructible<detail::input_adapter, A1, A2>::value, int> = 0> + static basic_json from_bson(A1 && a1, A2 && a2, + const bool strict = true, + const bool allow_exceptions = true) + { + basic_json result; + detail::json_sax_dom_parser<basic_json> sdp(result, allow_exceptions); + const bool res = binary_reader(detail::input_adapter(std::forward<A1>(a1), std::forward<A2>(a2))).sax_parse(input_format_t::bson, &sdp, strict); + return res ? result : basic_json(value_t::discarded); + } + + + /// @} ////////////////////////// @@ -14008,20 +19864,20 @@ class basic_json // avoid undefined behavior JSON_THROW(out_of_range::create(401, "array index " + std::to_string(idx) + " is out of range")); } - else - { - // default case: insert add offset - parent.insert(parent.begin() + static_cast<difference_type>(idx), val); - } + + // default case: insert add offset + parent.insert(parent.begin() + static_cast<difference_type>(idx), val); } break; } + // LCOV_EXCL_START default: { // if there exists a parent it cannot be primitive - assert(false); // LCOV_EXCL_LINE + assert(false); } + // LCOV_EXCL_STOP } } }; @@ -14066,7 +19922,7 @@ class basic_json // wrapper to get a value for an operation const auto get_value = [&val](const std::string & op, const std::string & member, - bool string_type) -> basic_json& + bool string_type) -> basic_json & { // find value auto it = val.m_value.object->find(member); @@ -14163,7 +20019,7 @@ class basic_json // the "path" location must exist - use at() success = (result.at(ptr) == get_value("test", "value", false)); } - JSON_CATCH (out_of_range&) + JSON_INTERNAL_CATCH (out_of_range&) { // ignore out of range errors: success remains false } @@ -14216,6 +20072,7 @@ class basic_json diff for two JSON values.,diff} @sa @ref patch -- apply a JSON patch + @sa @ref merge_patch -- apply a JSON Merge Patch @sa [RFC 6902 (JSON Patch)](https://tools.ietf.org/html/rfc6902) @@ -14347,418 +20204,86 @@ class basic_json } /// @} -}; -///////////// -// presets // -///////////// + //////////////////////////////// + // JSON Merge Patch functions // + //////////////////////////////// -/*! -@brief default JSON class - -This type is the default specialization of the @ref basic_json class which -uses the standard template types. - -@since version 1.0.0 -*/ -using json = basic_json<>; - -////////////////// -// json_pointer // -////////////////// - -NLOHMANN_BASIC_JSON_TPL_DECLARATION -NLOHMANN_BASIC_JSON_TPL& -json_pointer::get_and_create(NLOHMANN_BASIC_JSON_TPL& j) const -{ - using size_type = typename NLOHMANN_BASIC_JSON_TPL::size_type; - auto result = &j; - - // in case no reference tokens exist, return a reference to the JSON value - // j which will be overwritten by a primitive value - for (const auto& reference_token : reference_tokens) - { - switch (result->m_type) - { - case detail::value_t::null: - { - if (reference_token == "0") - { - // start a new array if reference token is 0 - result = &result->operator[](0); - } - else - { - // start a new object otherwise - result = &result->operator[](reference_token); - } - break; - } - - case detail::value_t::object: - { - // create an entry in the object - result = &result->operator[](reference_token); - break; - } - - case detail::value_t::array: - { - // create an entry in the array - JSON_TRY - { - result = &result->operator[](static_cast<size_type>(array_index(reference_token))); - } - JSON_CATCH(std::invalid_argument&) - { - JSON_THROW(detail::parse_error::create(109, 0, "array index '" + reference_token + "' is not a number")); - } - break; - } + /// @name JSON Merge Patch functions + /// @{ - /* - The following code is only reached if there exists a reference - token _and_ the current value is primitive. In this case, we have - an error situation, because primitive values may only occur as - single value; that is, with an empty list of reference tokens. - */ - default: - JSON_THROW(detail::type_error::create(313, "invalid value to unflatten")); - } - } + /*! + @brief applies a JSON Merge Patch + + The merge patch format is primarily intended for use with the HTTP PATCH + method as a means of describing a set of modifications to a target + resource's content. This function applies a merge patch to the current + JSON value. + + The function implements the following algorithm from Section 2 of + [RFC 7396 (JSON Merge Patch)](https://tools.ietf.org/html/rfc7396): + + ``` + define MergePatch(Target, Patch): + if Patch is an Object: + if Target is not an Object: + Target = {} // Ignore the contents and set it to an empty Object + for each Name/Value pair in Patch: + if Value is null: + if Name exists in Target: + remove the Name/Value pair from Target + else: + Target[Name] = MergePatch(Target[Name], Value) + return Target + else: + return Patch + ``` + + Thereby, `Target` is the current object; that is, the patch is applied to + the current value. + + @param[in] apply_patch the patch to apply + + @complexity Linear in the lengths of @a patch. + + @liveexample{The following code shows how a JSON Merge Patch is applied to + a JSON document.,merge_patch} - return *result; -} + @sa @ref patch -- apply a JSON patch + @sa [RFC 7396 (JSON Merge Patch)](https://tools.ietf.org/html/rfc7396) -NLOHMANN_BASIC_JSON_TPL_DECLARATION -NLOHMANN_BASIC_JSON_TPL& -json_pointer::get_unchecked(NLOHMANN_BASIC_JSON_TPL* ptr) const -{ - using size_type = typename NLOHMANN_BASIC_JSON_TPL::size_type; - for (const auto& reference_token : reference_tokens) + @since version 3.0.0 + */ + void merge_patch(const basic_json& apply_patch) { - // convert null values to arrays or objects before continuing - if (ptr->m_type == detail::value_t::null) - { - // check if reference token is a number - const bool nums = - std::all_of(reference_token.begin(), reference_token.end(), - [](const char x) - { - return (x >= '0' and x <= '9'); - }); - - // change value to array for numbers or "-" or to object otherwise - *ptr = (nums or reference_token == "-") - ? detail::value_t::array - : detail::value_t::object; - } - - switch (ptr->m_type) + if (apply_patch.is_object()) { - case detail::value_t::object: + if (not is_object()) { - // use unchecked object access - ptr = &ptr->operator[](reference_token); - break; + *this = object(); } - - case detail::value_t::array: + for (auto it = apply_patch.begin(); it != apply_patch.end(); ++it) { - // error condition (cf. RFC 6901, Sect. 4) - if (JSON_UNLIKELY(reference_token.size() > 1 and reference_token[0] == '0')) - { - JSON_THROW(detail::parse_error::create(106, 0, - "array index '" + reference_token + - "' must not begin with '0'")); - } - - if (reference_token == "-") + if (it.value().is_null()) { - // explicitly treat "-" as index beyond the end - ptr = &ptr->operator[](ptr->m_value.array->size()); + erase(it.key()); } else { - // convert array index to number; unchecked access - JSON_TRY - { - ptr = &ptr->operator[]( - static_cast<size_type>(array_index(reference_token))); - } - JSON_CATCH(std::invalid_argument&) - { - JSON_THROW(detail::parse_error::create(109, 0, "array index '" + reference_token + "' is not a number")); - } - } - break; - } - - default: - JSON_THROW(detail::out_of_range::create(404, "unresolved reference token '" + reference_token + "'")); - } - } - - return *ptr; -} - -NLOHMANN_BASIC_JSON_TPL_DECLARATION -NLOHMANN_BASIC_JSON_TPL& -json_pointer::get_checked(NLOHMANN_BASIC_JSON_TPL* ptr) const -{ - using size_type = typename NLOHMANN_BASIC_JSON_TPL::size_type; - for (const auto& reference_token : reference_tokens) - { - switch (ptr->m_type) - { - case detail::value_t::object: - { - // note: at performs range check - ptr = &ptr->at(reference_token); - break; - } - - case detail::value_t::array: - { - if (JSON_UNLIKELY(reference_token == "-")) - { - // "-" always fails the range check - JSON_THROW(detail::out_of_range::create(402, - "array index '-' (" + std::to_string(ptr->m_value.array->size()) + - ") is out of range")); - } - - // error condition (cf. RFC 6901, Sect. 4) - if (JSON_UNLIKELY(reference_token.size() > 1 and reference_token[0] == '0')) - { - JSON_THROW(detail::parse_error::create(106, 0, - "array index '" + reference_token + - "' must not begin with '0'")); - } - - // note: at performs range check - JSON_TRY - { - ptr = &ptr->at(static_cast<size_type>(array_index(reference_token))); + operator[](it.key()).merge_patch(it.value()); } - JSON_CATCH(std::invalid_argument&) - { - JSON_THROW(detail::parse_error::create(109, 0, "array index '" + reference_token + "' is not a number")); - } - break; } - - default: - JSON_THROW(detail::out_of_range::create(404, "unresolved reference token '" + reference_token + "'")); } - } - - return *ptr; -} - -NLOHMANN_BASIC_JSON_TPL_DECLARATION -const NLOHMANN_BASIC_JSON_TPL& -json_pointer::get_unchecked(const NLOHMANN_BASIC_JSON_TPL* ptr) const -{ - using size_type = typename NLOHMANN_BASIC_JSON_TPL::size_type; - for (const auto& reference_token : reference_tokens) - { - switch (ptr->m_type) - { - case detail::value_t::object: - { - // use unchecked object access - ptr = &ptr->operator[](reference_token); - break; - } - - case detail::value_t::array: - { - if (JSON_UNLIKELY(reference_token == "-")) - { - // "-" cannot be used for const access - JSON_THROW(detail::out_of_range::create(402, - "array index '-' (" + std::to_string(ptr->m_value.array->size()) + - ") is out of range")); - } - - // error condition (cf. RFC 6901, Sect. 4) - if (JSON_UNLIKELY(reference_token.size() > 1 and reference_token[0] == '0')) - { - JSON_THROW(detail::parse_error::create(106, 0, - "array index '" + reference_token + - "' must not begin with '0'")); - } - - // use unchecked array access - JSON_TRY - { - ptr = &ptr->operator[]( - static_cast<size_type>(array_index(reference_token))); - } - JSON_CATCH(std::invalid_argument&) - { - JSON_THROW(detail::parse_error::create(109, 0, "array index '" + reference_token + "' is not a number")); - } - break; - } - - default: - JSON_THROW(detail::out_of_range::create(404, "unresolved reference token '" + reference_token + "'")); - } - } - - return *ptr; -} - -NLOHMANN_BASIC_JSON_TPL_DECLARATION -const NLOHMANN_BASIC_JSON_TPL& -json_pointer::get_checked(const NLOHMANN_BASIC_JSON_TPL* ptr) const -{ - using size_type = typename NLOHMANN_BASIC_JSON_TPL::size_type; - for (const auto& reference_token : reference_tokens) - { - switch (ptr->m_type) - { - case detail::value_t::object: - { - // note: at performs range check - ptr = &ptr->at(reference_token); - break; - } - - case detail::value_t::array: - { - if (JSON_UNLIKELY(reference_token == "-")) - { - // "-" always fails the range check - JSON_THROW(detail::out_of_range::create(402, - "array index '-' (" + std::to_string(ptr->m_value.array->size()) + - ") is out of range")); - } - - // error condition (cf. RFC 6901, Sect. 4) - if (JSON_UNLIKELY(reference_token.size() > 1 and reference_token[0] == '0')) - { - JSON_THROW(detail::parse_error::create(106, 0, - "array index '" + reference_token + - "' must not begin with '0'")); - } - - // note: at performs range check - JSON_TRY - { - ptr = &ptr->at(static_cast<size_type>(array_index(reference_token))); - } - JSON_CATCH(std::invalid_argument&) - { - JSON_THROW(detail::parse_error::create(109, 0, "array index '" + reference_token + "' is not a number")); - } - break; - } - - default: - JSON_THROW(detail::out_of_range::create(404, "unresolved reference token '" + reference_token + "'")); - } - } - - return *ptr; -} - -NLOHMANN_BASIC_JSON_TPL_DECLARATION -void json_pointer::flatten(const std::string& reference_string, - const NLOHMANN_BASIC_JSON_TPL& value, - NLOHMANN_BASIC_JSON_TPL& result) -{ - switch (value.m_type) - { - case detail::value_t::array: - { - if (value.m_value.array->empty()) - { - // flatten empty array as null - result[reference_string] = nullptr; - } - else - { - // iterate array and use index as reference string - for (std::size_t i = 0; i < value.m_value.array->size(); ++i) - { - flatten(reference_string + "/" + std::to_string(i), - value.m_value.array->operator[](i), result); - } - } - break; - } - - case detail::value_t::object: - { - if (value.m_value.object->empty()) - { - // flatten empty object as null - result[reference_string] = nullptr; - } - else - { - // iterate object and use keys as reference string - for (const auto& element : *value.m_value.object) - { - flatten(reference_string + "/" + escape(element.first), element.second, result); - } - } - break; - } - - default: - { - // add primitive value with its reference string - result[reference_string] = value; - break; - } - } -} - -NLOHMANN_BASIC_JSON_TPL_DECLARATION -NLOHMANN_BASIC_JSON_TPL -json_pointer::unflatten(const NLOHMANN_BASIC_JSON_TPL& value) -{ - if (JSON_UNLIKELY(not value.is_object())) - { - JSON_THROW(detail::type_error::create(314, "only objects can be unflattened")); - } - - NLOHMANN_BASIC_JSON_TPL result; - - // iterate the JSON object values - for (const auto& element : *value.m_value.object) - { - if (JSON_UNLIKELY(not element.second.is_primitive())) + else { - JSON_THROW(detail::type_error::create(315, "values in object must be primitive")); + *this = apply_patch; } - - // assign value to reference pointed to by JSON pointer; Note that if - // the JSON pointer is "" (i.e., points to the whole value), function - // get_and_create returns a reference to result itself. An assignment - // will then create a primitive value. - json_pointer(element.first).get_and_create(result) = element.second; } - return result; -} - -inline bool operator==(json_pointer const& lhs, json_pointer const& rhs) noexcept -{ - return (lhs.reference_tokens == rhs.reference_tokens); -} - -inline bool operator!=(json_pointer const& lhs, json_pointer const& rhs) noexcept -{ - return not (lhs == rhs); -} + /// @} +}; } // namespace nlohmann - /////////////////////// // nonmember support // /////////////////////// @@ -14766,20 +20291,6 @@ inline bool operator!=(json_pointer const& lhs, json_pointer const& rhs) noexcep // specialization of std::swap, and std::hash namespace std { -/*! -@brief exchanges the values of two JSON objects - -@since version 1.0.0 -*/ -template<> -inline void swap(nlohmann::json& j1, - nlohmann::json& j2) noexcept( - is_nothrow_move_constructible<nlohmann::json>::value and - is_nothrow_move_assignable<nlohmann::json>::value - ) -{ - j1.swap(j2); -} /// hash value for JSON objects template<> @@ -14815,6 +20326,20 @@ struct less< ::nlohmann::detail::value_t> } }; +/*! +@brief exchanges the values of two JSON objects + +@since version 1.0.0 +*/ +template<> +inline void swap<nlohmann::json>(nlohmann::json& j1, nlohmann::json& j2) noexcept( + is_nothrow_move_constructible<nlohmann::json>::value and + is_nothrow_move_assignable<nlohmann::json>::value +) +{ + j1.swap(j2); +} + } // namespace std /*! @@ -14853,6 +20378,9 @@ inline nlohmann::json::json_pointer operator "" _json_pointer(const char* s, std return nlohmann::json::json_pointer(std::string(s, n)); } +// #include <nlohmann/detail/macro_unscope.hpp> + + // restore GCC/clang diagnostic settings #if defined(__clang__) || defined(__GNUC__) || defined(__GNUG__) #pragma GCC diagnostic pop @@ -14862,13 +20390,17 @@ inline nlohmann::json::json_pointer operator "" _json_pointer(const char* s, std #endif // clean up +#undef JSON_INTERNAL_CATCH #undef JSON_CATCH #undef JSON_THROW #undef JSON_TRY #undef JSON_LIKELY #undef JSON_UNLIKELY #undef JSON_DEPRECATED +#undef JSON_HAS_CPP_14 +#undef JSON_HAS_CPP_17 #undef NLOHMANN_BASIC_JSON_TPL_DECLARATION #undef NLOHMANN_BASIC_JSON_TPL + #endif diff --git a/src/resolve-system-dependencies/local.mk b/src/resolve-system-dependencies/local.mk index 8792a4a252fa..f9db16268be0 100644 --- a/src/resolve-system-dependencies/local.mk +++ b/src/resolve-system-dependencies/local.mk @@ -6,6 +6,6 @@ resolve-system-dependencies_DIR := $(d) resolve-system-dependencies_INSTALL_DIR := $(libexecdir)/nix -resolve-system-dependencies_LIBS := libstore libmain libutil libformat +resolve-system-dependencies_LIBS := libstore libmain libutil resolve-system-dependencies_SOURCES := $(d)/resolve-system-dependencies.cc |