* Improved Nix. Resources (package descriptors and other source

  files) are now referenced using their cryptographic hashes.  

  This ensures that if two package descriptors have the same contents,
  then they describe the same package.  This property is not as
  trivial as it sounds: generally import relations cause this property
  not to hold w.r.t. temporality.  But since imports also use hashes
  to reference other packages, equality follows by induction.

1 files changed, 267 insertions, 74 deletions

diff --git a/src/nix.cc b/src/nix.cc
index 0c786527082c..8108c2fca9fc 100644
--- a/src/nix.cc
+++ b/src/nix.cc
@@ -1,6 +1,10 @@
 #include <iostream>
+#include <fstream>
 #include <memory>
 #include <string>
+#include <sstream>
+#include <list>
+#include <cstdio>
 
 #include <unistd.h>
 #include <sys/stat.h>
@@ -15,10 +19,16 @@ using namespace std;
 #define PKGINFO_PATH "/pkg/sys/var/pkginfo"
 
 
+static string dbRefs = "refs";
+static string dbInstPkgs = "pkginst";
+
+
 static string prog;
 static string dbfile = PKGINFO_PATH;
 
 
+/* Wrapper class that ensures that the database is closed upon
+   object destruction. */
 class Db2 : public Db 
 {
 public:
@@ -81,85 +91,250 @@ void delDB(const string & dbname, const string & key)
 }
 
 
-void getPkg(int argc, char * * argv)
+/* Verify that a reference is valid (that is, is a MD5 hash code). */
+void checkRef(const string & s)
 {
-    string pkg;
-    string src;
-    string inst;
-    string cmd;
-    int res;
+    string err = "invalid reference: " + s;
+    if (s.length() != 32)
+        throw err;
+    for (int i = 0; i < 32; i++) {
+        char c = s[i];
+        if (!((c >= '0' && c <= '9') ||
+              (c >= 'a' && c <= 'f')))
+            throw err;
+    }
+}
 
-    if (argc != 1)
-        throw string("arguments missing in get-pkg");
 
-    pkg = argv[0];
+/* Compute the MD5 hash of a file. */
+string makeRef(string filename)
+{
+    char hash[33];
+
+    FILE * pipe = popen(("md5sum " + filename).c_str(), "r");
+    if (!pipe) throw string("cannot execute md5sum");
+
+    if (fread(hash, 32, 1, pipe) != 1)
+        throw string("cannot read hash from md5sum");
+    hash[32] = 0;
+
+    pclose(pipe);
+
+    checkRef(hash);
+    return hash;
+}
 
-    if (queryDB("pkginst", pkg, inst)) {
-        cout << inst << endl;
-        return;
+
+struct Dep
+{
+    string name;
+    string ref;
+    Dep(string _name, string _ref)
+    {
+        name = _name;
+        ref = _ref;
     }
+};
+
+typedef list<Dep> DepList;
+
+
+void readPkgDescr(const string & pkgfile,
+    DepList & pkgImports, DepList & fileImports)
+{
+    ifstream file;
+    file.exceptions(ios::badbit);
+    file.open(pkgfile.c_str());
     
-    cerr << "package " << pkg << " is not yet installed\n";
-        
-    if (!queryDB("pkgsrc", pkg, src)) 
-        throw string("source of package " + string(pkg) + " is not known");
+    while (!file.eof()) {
+        string line;
+        getline(file, line);
 
-    inst = "/pkg/" + pkg;
+        int n = line.find('#');
+        if (n >= 0) line = line.erase(n);
 
-    cmd = "rsync -a \"" + src + "\"/ \"" + inst + "\"";
+        if ((int) line.find_first_not_of(" ") < 0) continue;
         
-    res = system(cmd.c_str());
-    if (!WIFEXITED(res) || WEXITSTATUS(res) != 0)
-        throw string("unable to copy sources");
+        istringstream str(line);
+
+        string name, op, ref;
+        str >> name >> op >> ref;
+
+        checkRef(ref);
+
+        if (op == "<-") 
+            pkgImports.push_back(Dep(name, ref));
+        else if (op == "=")
+            fileImports.push_back(Dep(name, ref));
+        else throw string("invalid operator " + op);
+    }
+}
+
+
+string getPkg(string pkgref);
+
+
+typedef pair<string, string> EnvPair;
+typedef list<EnvPair> Environment;
+
+
+void installPkg(string pkgref)
+{
+    string pkgfile;
+    string src;
+    string path;
+    string cmd;
+    string builder;
+
+    if (!queryDB("refs", pkgref, pkgfile))
+        throw string("unknown package " + pkgref);
+
+    cerr << "installing package " + pkgref + " from " + pkgfile + "\n";
+
+    /* Verify that the file hasn't changed. !!! race */
+    if (makeRef(pkgfile) != pkgref)
+        throw string("file " + pkgfile + " is stale");
+
+    /* Read the package description file. */
+    DepList pkgImports, fileImports;
+    readPkgDescr(pkgfile, pkgImports, fileImports);
+
+    /* Recursively fetch all the dependencies, filling in the
+       environment as we go along. */
+    Environment env;
+
+    for (DepList::iterator it = pkgImports.begin();
+         it != pkgImports.end(); it++)
+    {
+        cerr << "fetching package dependency "
+             << it->name << " <- " << it->ref
+             << endl;
+        env.push_back(EnvPair(it->name, getPkg(it->ref)));
+    }
+
+    for (DepList::iterator it = fileImports.begin();
+         it != fileImports.end(); it++)
+    {
+        cerr << "fetching file dependency "
+             << it->name << " = " << it->ref
+             << endl;
 
-    if (chdir(inst.c_str()))
-        throw string("unable to chdir to package directory");
+        string file;
 
-    /* Prepare for building.  Clean the environment so that the
-       build process does not inherit things it shouldn't. */
-    setenv("PATH", "/pkg/sys/bin", 1);
+        if (!queryDB("refs", it->ref, file))
+            throw string("unknown file " + it->ref);
 
-    res = system("./buildme");
-    if (!WIFEXITED(res) || WEXITSTATUS(res) != 0)
+        if (makeRef(file) != it->ref)
+            throw string("file " + file + " is stale");
+
+        if (it->name == "build")
+            builder = file;
+        else
+            env.push_back(EnvPair(it->name, file));
+    }
+
+    if (builder == "")
+        throw string("no builder specified");
+
+    /* Construct a path for the installed package. */
+    path = "/pkg/" + pkgref;
+
+    /* Create the path. */
+    if (mkdir(path.c_str(), 0777))
+        throw string("unable to create directory " + path);
+
+    /* Fork a child to build the package. */
+    pid_t pid;
+    switch (pid = fork()) {
+
+    case -1:
+        throw string("unable to fork");
+
+    case 0: /* child */
+
+        /* Go to the build directory. */
+        if (chdir(path.c_str())) {
+            cout << "unable to chdir to package directory\n";
+            _exit(1);
+        }
+
+        /* Fill in the environment.  We don't bother freeing the
+           strings, since we'll exec or die soon anyway. */
+        const char * env2[env.size() + 1];
+        int i = 0;
+        for (Environment::iterator it = env.begin();
+             it != env.end(); it++, i++)
+            env2[i] = (new string(it->first + "=" + it->second))->c_str();
+        env2[i] = 0;
+
+        /* Execute the builder.  This should not return. */
+        execle(builder.c_str(), builder.c_str(), 0, env2);
+
+        cout << strerror(errno) << endl;
+
+        cout << "unable to execute builder\n";
+        _exit(1);
+    }
+
+    /* parent */
+
+    /* Wait for the child to finish. */
+    int status;
+    if (waitpid(pid, &status, 0) != pid)
+        throw string("unable to wait for child");
+    
+    if (!WIFEXITED(status) || WEXITSTATUS(status) != 0)
         throw string("unable to build package");
 
-    setDB("pkginst", pkg, inst);
+    setDB(dbInstPkgs, pkgref, path);
+}
+
 
-    cout << inst << endl;
+string getPkg(string pkgref)
+{
+    string path;
+    checkRef(pkgref);
+    while (!queryDB(dbInstPkgs, pkgref, path))
+        installPkg(pkgref);
+    return path;
 }
 
 
-void registerPkg(int argc, char * * argv)
+string absPath(string filename)
 {
-    char * pkg;
-    char * src;
-    
-    if (argc != 2)
-        throw string("arguments missing in register-pkg");
+    if (filename[0] != '/') {
+        char buf[PATH_MAX];
+        if (!getcwd(buf, sizeof(buf)))
+            throw string("cannot get cwd");
+        filename = string(buf) + "/" + filename;
+        /* !!! canonicalise */
+    }
+    return filename;
+}
 
-    pkg = argv[0];
-    src = argv[1];
 
-    setDB("pkgsrc", pkg, src);
+void registerFile(string filename)
+{
+    filename = absPath(filename);
+    setDB(dbRefs, makeRef(filename), filename);
 }
 
 
 /* This is primarily used for bootstrapping. */
-void registerInstalledPkg(int argc, char * * argv)
+void registerInstalledPkg(string pkgref, string path)
 {
-    string pkg;
-    string inst;
-    
-    if (argc != 2)
-        throw string("arguments missing in register-installed-pkg");
-
-    pkg = argv[0];
-    inst = argv[1];
-    
-    if (inst == "")
-        delDB("pkginst", pkg);
+    checkRef(pkgref);
+    if (path == "")
+        delDB(dbInstPkgs, pkgref);
     else
-        setDB("pkginst", pkg, inst);
+        setDB(dbInstPkgs, pkgref, path);
+}
+
+
+void initDB()
+{
+    openDB(dbRefs, false);
+    openDB(dbInstPkgs, false);
 }
 
 
@@ -168,18 +343,29 @@ void run(int argc, char * * argv)
     string cmd;
 
     if (argc < 1)
-        throw string("command not specified\n");
+        throw string("command not specified");
 
     cmd = argv[0];
     argc--, argv++;
 
-    if (cmd == "get-pkg")
-        getPkg(argc, argv);
-    else if (cmd == "register-pkg")
-        registerPkg(argc, argv);
-    else if (cmd == "register-installed-pkg")
-        registerInstalledPkg(argc, argv);
-    else
+    if (cmd == "init") {
+        if (argc != 0)
+            throw string("init doesn't have arguments");
+        initDB();
+    } else if (cmd == "getpkg") {
+        if (argc != 1)
+            throw string("arguments missing in getpkg");
+        string path = getPkg(argv[0]);
+        cout << path << endl;
+    } else if (cmd == "reg") {
+        if (argc != 1)
+            throw string("arguments missing in reg");
+        registerFile(argv[0]);
+    } else if (cmd == "regpkg") {
+        if (argc != 2)
+            throw string("arguments missing in regpkg");
+        registerInstalledPkg(argv[0], argv[1]);
+    } else
         throw string("unknown command: " + string(cmd));
 }
     
@@ -190,31 +376,38 @@ int main(int argc, char * * argv)
 
     prog = argv[0];
 
-    while ((c = getopt(argc, argv, "d:")) != EOF) {
-        
-        switch (c) {
+    umask(0022);
 
-        case 'd':
-            dbfile = optarg;
-            break;
+    try {
 
-        default:
-            cerr << "unknown option\n";
-            break;
-        }
+        while ((c = getopt(argc, argv, "d:")) != EOF) {
+        
+            switch (c) {
 
-    }
+            case 'd':
+                dbfile = optarg;
+                break;
 
-    argc -= optind, argv += optind;
+            default:
+                throw string("unknown option");
+                break;
 
-    try {
+            }
+        }
+
+        argc -= optind, argv += optind;
         run(argc, argv);
-        return 0;
+
     } catch (DbException e) {
         cerr << "db exception: " << e.what() << endl;
         return 1;
+    } catch (exception e) {
+        cerr << e.what() << endl;
+        return 1;
     } catch (string s) {
         cerr << s << endl;
         return 1;
     }
+
+    return 0;
 }