* Path hashing.

5 files changed, 178 insertions, 8 deletions

diff --git a/src/eval.cc b/src/eval.cc
index 14577c8738ae..c96cf6467ecc 100644
--- a/src/eval.cc
+++ b/src/eval.cc
@@ -148,10 +148,12 @@ static Hash computeDerived(Hash sourceHash, string targetName,
             throw Error("program " + progPath + 
                 " failed to create a result in " + targetPath);
 
+#if 0
         /* Remove write permission from the value. */
         int res = system(("chmod -R -w " + targetPath).c_str()); // !!! escaping
         if (WEXITSTATUS(res) != 0)
             throw Error("cannot remove write permission from " + targetPath);
+#endif
 
     } catch (exception &) {
 //         system(("rm -rf " + targetPath).c_str());
@@ -159,7 +161,7 @@ static Hash computeDerived(Hash sourceHash, string targetName,
     }
 
     /* Hash the result. */
-    Hash targetHash = hashFile(targetPath);
+    Hash targetHash = hashPath(targetPath);
 
     /* Register targetHash -> targetPath.  !!! this should be in
        values.cc. */
diff --git a/src/hash.cc b/src/hash.cc
index 9451ac8d80e4..9558d36705f3 100644
--- a/src/hash.cc
+++ b/src/hash.cc
@@ -1,9 +1,16 @@
+#include <iostream>
+
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <unistd.h>
+#include <dirent.h>
+#include <fcntl.h>
+
 extern "C" {
 #include "md5.h"
 }
 
 #include "hash.hh"
-#include <iostream>
 
 
 Hash::Hash()
@@ -88,3 +95,135 @@ Hash hashFile(const string & fileName)
     if (err) throw SysError("cannot hash file " + fileName);
     return hash;
 }
+
+
+struct HashSink : DumpSink
+{
+    struct md5_ctx ctx;
+    virtual void operator ()
+        (const unsigned char * data, unsigned int len)
+    {
+        md5_process_bytes(data, len, &ctx);
+    }
+};
+
+
+Hash hashPath(const string & path)
+{
+    Hash hash;
+    HashSink sink;
+    md5_init_ctx(&sink.ctx);
+    dumpPath(path, sink);
+    md5_finish_ctx(&sink.ctx, hash.hash);
+    return hash;
+}
+
+
+static void pad(unsigned int len, DumpSink & sink)
+{
+    if (len % 8) {
+        unsigned char zero[8];
+        memset(zero, 0, sizeof(zero));
+        sink(zero, 8 - (len % 8));
+    }
+}
+
+
+static void writeInt(unsigned int n, DumpSink & sink)
+{
+    unsigned char buf[8];
+    memset(buf, 0, sizeof(buf));
+    buf[0] = n & 0xff;
+    buf[1] = (n >> 8) & 0xff;
+    buf[2] = (n >> 16) & 0xff;
+    buf[3] = (n >> 24) & 0xff;
+    sink(buf, sizeof(buf));
+}
+
+
+static void writeString(const string & s, DumpSink & sink)
+{
+    unsigned int len = s.length();
+    writeInt(len, sink);
+    sink((const unsigned char *) s.c_str(), len);
+    pad(len, sink);
+}
+
+
+static void dumpEntries(const string & path, DumpSink & sink)
+{
+    DIR * dir = opendir(path.c_str());
+    if (!dir) throw SysError("opening directory " + path);
+    
+    struct dirent * dirent;
+
+    /* !!! sort entries */
+
+    while (errno = 0, dirent = readdir(dir)) {
+        string name = dirent->d_name;
+        if (name == "." || name == "..") continue;
+        writeString("entry", sink);
+        writeString("(", sink);
+        writeString("name", sink);
+        writeString(name, sink);
+        writeString("file", sink);
+        dumpPath(path + "/" + name, sink);
+        writeString(")", sink);
+    }
+
+    if (errno) throw SysError("reading directory " + path);
+    
+    closedir(dir); /* !!! close on exception */
+}
+
+
+static void dumpContents(const string & path, unsigned int size, 
+    DumpSink & sink)
+{
+    writeString("contents", sink);
+    writeInt(size, sink);
+
+    int fd = open(path.c_str(), O_RDONLY);
+    if (!fd) throw SysError("opening file " + path);
+    
+    unsigned char buf[16384];
+
+    unsigned int total = 0;
+    ssize_t n;
+    while ((n = read(fd, buf, sizeof(buf)))) {
+        if (n == -1) throw SysError("reading file " + path);
+        total += n;
+        sink(buf, n);
+    }
+
+    if (total != size)
+        throw SysError("file changed while reading it: " + path);
+
+    pad(size, sink);
+
+    close(fd); /* !!! close on exception */
+}
+
+
+void dumpPath(const string & path, DumpSink & sink)
+{
+    cerr << path << endl;
+
+    struct stat st;
+    if (lstat(path.c_str(), &st))
+        throw SysError("getting attributes of path " + path);
+
+    writeString("(", sink);
+
+    if (S_ISREG(st.st_mode)) {
+        writeString("type", sink);
+        writeString("regular", sink);
+        dumpContents(path, st.st_size, sink);
+    } else if (S_ISDIR(st.st_mode)) {
+        writeString("type", sink);
+        writeString("directory", sink);
+        dumpEntries(path, sink);
+    } else throw Error("unknown file type: " + path);
+
+    writeString(")", sink);
+}
diff --git a/src/hash.hh b/src/hash.hh
index 9d72e66dbdea..13c5275b468a 100644
--- a/src/hash.hh
+++ b/src/hash.hh
@@ -50,6 +50,11 @@ Hash hashFile(const string & fileName);
    follows:
 
    hash(path) = md5(dump(path))
+*/
+Hash hashPath(const string & path);
+
+
+/* Dump a path as follows:
 
    IF path points to a REGULAR FILE:
      dump(path) = attrs(
@@ -60,7 +65,7 @@ Hash hashFile(const string & fileName);
    IF path points to a DIRECTORY:
      dump(path) = attrs(
        [ ("type", "directory")
-       , ("entries", concat(map(f, entries(path))))
+       , ("entries", concat(map(f, sort(entries(path)))))
        ])
        where f(fn) = attrs(
          [ ("name", fn)
@@ -72,17 +77,25 @@ Hash hashFile(const string & fileName);
      attrs(as) = concat(map(attr, as)) + encN(0) 
      attrs((a, b)) = encS(a) + encS(b)
 
-     encS(s) = encN(len(s)) + s
+     encS(s) = encN(len(s)) + s + (padding until next 64-bit boundary)
 
      encN(n) = 64-bit little-endian encoding of n.
 
      contents(path) = the contents of a regular file.
 
+     sort(strings) = lexicographic sort by 8-bit value (strcmp).
+
      entries(path) = the entries of a directory, without `.' and
      `..'.
 
      `+' denotes string concatenation. */
-Hash hashPath(const string & path);
+
+struct DumpSink 
+{
+    virtual void operator () (const unsigned char * data, unsigned int len) = 0;
+};
+
+void dumpPath(const string & path, DumpSink & sink);
 
 
 #endif /* !__HASH_H */
diff --git a/src/test.cc b/src/test.cc
index 79468182ea3c..b37a16a1fe0c 100644
--- a/src/test.cc
+++ b/src/test.cc
@@ -19,6 +19,15 @@ void evalTest(Expr e)
 }
 
 
+struct MySink : DumpSink
+{
+    virtual void operator () (const unsigned char * data, unsigned int len)
+    {
+        cout.write((char *) data, len);
+    }
+};
+
+
 void runTests()
 {
     /* Hashing. */
@@ -36,6 +45,13 @@ void runTests()
         abort();
     } catch (BadRefError err) { };
 
+    /* Dumping. */
+
+#if 0
+    MySink sink;
+    dumpPath("scratch", sink);
+    cout << (string) hashPath("scratch") << endl;
+#endif
 
     /* Set up the test environment. */
 
diff --git a/src/values.cc b/src/values.cc
index 064203ae2836..77a6f928e1a2 100644
--- a/src/values.cc
+++ b/src/values.cc
@@ -19,7 +19,7 @@ static string absValuePath(string s)
 
 Hash addValue(string path)
 {
-    Hash hash = hashFile(path);
+    Hash hash = hashPath(path);
 
     string name;
     if (queryDB(nixDB, dbRefs, hash, name)) {
@@ -72,8 +72,8 @@ string queryValuePath(Hash hash)
         if (queryDB(nixDB, dbRefs, hash, name)) {
             string fn = absValuePath(name);
 
-            /* Verify that the file hasn't changed. !!! race */
-            if (hashFile(fn) != hash)
+            /* Verify that the file hasn't changed. !!! race !!! slow */
+            if (hashPath(fn) != hash)
                 throw Error("file " + fn + " is stale");
 
             return fn;