* Made the garbage collector a lot faster. It no longer computes the

complete set of live and dead paths before starting the actual
  deletion, but determines liveness on demand.  I.e. for any path in
  the store, it first tries to delete all the referrers, and then the
  path itself.  This means that the collector can start deleting paths
  almost immediately.

src/libstore/gc.cc

@@ -401,7 +401,7 @@ static void addAdditionalRoots(PathSet & roots)
         if (isInStore(*i)) {
             Path path = toStorePath(*i);
             if (roots.find(path) == roots.end() && store->isValidPath(path)) {
-                debug(format("found additional root `%1%'") % path);
+                debug(format("got additional root `%1%'") % path);
                 roots.insert(path);
             }
         }
@@ -442,90 +442,158 @@ Paths topoSortPaths(const PathSet & paths)
 struct GCLimitReached { };
 
 
-void LocalStore::gcPath(const GCOptions & options, GCResults & results, 
+struct LocalStore::GCState
-    const Path & path)
 {
-    results.paths.insert(path);
+    GCOptions options;
-
+    GCResults & results;
-    if (!pathExists(path)) return;
+    PathSet roots;
-                
+    PathSet tempRoots;
-    /* Okay, it's safe to delete. */
+    PathSet deleted;
-    unsigned long long bytesFreed, blocksFreed;
+    PathSet live;
-    deleteFromStore(path, bytesFreed, blocksFreed);
+    PathSet busy;
-    results.bytesFreed += bytesFreed;
+    bool gcKeepOutputs;
-    results.blocksFreed += blocksFreed;
+    bool gcKeepDerivations;
-
+    GCState(GCResults & results_) : results(results_)
-    if (options.maxFreed && results.bytesFreed > options.maxFreed) {
-        printMsg(lvlInfo, format("deleted more than %1% bytes; stopping") % options.maxFreed);
-        throw GCLimitReached();
-    }
-
-    if (options.maxLinks) {
-        struct stat st;
-        if (stat(nixStore.c_str(), &st) == -1)
-            throw SysError(format("statting `%1%'") % nixStore);
-        if (st.st_nlink < options.maxLinks) {
-            printMsg(lvlInfo, format("link count on the store has dropped below %1%; stopping") % options.maxLinks);
-            throw GCLimitReached();
-        }
-    }
-}
-
-
-void LocalStore::gcPathRecursive(const GCOptions & options,
-    GCResults & results, PathSet & done, const Path & path)
     {
-    if (done.find(path) != done.end()) return;
+    }
-    done.insert(path);
+};
 
-    startNest(nest, lvlDebug, format("looking at `%1%'") % path);
 
-    /* Delete all the referrers first.  They must be garbage too,
+static bool doDelete(GCOptions::GCAction action)
-       since if they were live, then the current path would also be
+{
-       live.  Note that deleteFromStore() below still makes sure that
+    return action == GCOptions::gcDeleteDead
-       the referrer set has become empty, just in case.  (However that
+        || action == GCOptions::gcDeleteSpecific;
-       doesn't guard against deleting top-level paths that are only
+}
-       reachable from GC roots.) */
+
+
+bool LocalStore::isActiveTempFile(const GCState & state,
+    const Path & path, const string & suffix)
+{
+    return hasSuffix(path, suffix)
+        && state.tempRoots.find(string(path, 0, path.size() - suffix.size())) != state.tempRoots.end();
+}
+
+    
+bool LocalStore::tryToDelete(GCState & state, const Path & path)
+{
+    if (!pathExists(path)) return true;
+    if (state.deleted.find(path) != state.deleted.end()) return true;
+    if (state.live.find(path) != state.live.end()) return false;
+
+    startNest(nest, lvlDebug, format("considering whether to delete `%1%'") % path);
+
+    if (state.roots.find(path) != state.roots.end()) {
+        printMsg(lvlDebug, format("cannot delete `%1%' because it's a root") % path);
+        goto isLive;
+    }
+
+    if (isValidPath(path)) {
+
+        /* Recursively try to delete the referrers of this path.  If
+           any referrer can't be deleted, then this path can't be
+           deleted either. */
         PathSet referrers;
-    if (isValidPath(path))
         queryReferrers(path, referrers);
         foreach (PathSet::iterator, i, referrers)
-        if (*i != path) gcPathRecursive(options, results, done, *i);
+            if (*i != path && !tryToDelete(state, *i)) {
-
+                printMsg(lvlDebug, format("cannot delete `%1%' because it has live referrers") % path);
-    printMsg(lvlInfo, format("deleting `%1%'") % path);
+                goto isLive;
-            
-    gcPath(options, results, path);
             }
 
+        /* If gc-keep-derivations is set and this is a derivation,
+           then don't delete the derivation if any of the outputs are
+           live. */
+        if (state.gcKeepDerivations && isDerivation(path)) {
+            Derivation drv = derivationFromPath(path);
+            foreach (DerivationOutputs::iterator, i, drv.outputs)
+                if (!tryToDelete(state, i->second.path)) {
+                    printMsg(lvlDebug, format("cannot delete derivation `%1%' because its output is alive") % path);
+                    goto isLive;
+                }
+        }
 
-static bool isLive(const Path & path, const PathSet & livePaths,
+        /* If gc-keep-outputs is set, then don't delete this path if
-    const PathSet & tempRoots, const PathSet & tempRootsClosed)
+           its deriver is not garbage.  !!! This is somewhat buggy,
-{
+           since there might be multiple derivers, but the database
-    if (livePaths.find(path) != livePaths.end() ||
+           only stores one. */
-        tempRootsClosed.find(path) != tempRootsClosed.end()) return true;
+        if (state.gcKeepOutputs) {
+            Path deriver = queryDeriver(path);
+            /* Break an infinite recursion if gc-keep-derivations and
+               gc-keep-outputs are both set by tentatively assuming
+               that this path is garbage.  This is a safe assumption
+               because at this point, the only thing that can prevent
+               it from being garbage is the deriver.  Since
+               tryToDelete() works "upwards" through the dependency
+               graph, it won't encouter this path except in the call
+               to tryToDelete() in the gc-keep-derivation branch. */
+            state.deleted.insert(path);
+            if (deriver != "" && !tryToDelete(state, deriver)) {
+                state.deleted.erase(path);
+                printMsg(lvlDebug, format("cannot delete `%1%' because its deriver is alive") % path);
+                goto isLive;
+            }
+        }
+    }
+
+    else {
 
         /* A lock file belonging to a path that we're building right
            now isn't garbage. */
-    if (hasSuffix(path, ".lock") && tempRoots.find(string(path, 0, path.size() - 5)) != tempRoots.end())
+        if (isActiveTempFile(state, path, ".lock")) return false;
-        return true;
 
         /* Don't delete .chroot directories for derivations that are
            currently being built. */
-    if (hasSuffix(path, ".chroot") && tempRoots.find(string(path, 0, path.size() - 7)) != tempRoots.end())
+        if (isActiveTempFile(state, path, ".chroot")) return false;
+
+    }
+
+    /* The path is garbage, so delete it. */
+    if (doDelete(state.options.action)) {
+        printMsg(lvlInfo, format("deleting `%1%'") % path);
+
+        unsigned long long bytesFreed, blocksFreed;
+        deleteFromStore(path, bytesFreed, blocksFreed);
+        state.results.bytesFreed += bytesFreed;
+        state.results.blocksFreed += blocksFreed;
+
+        if (state.options.maxFreed && state.results.bytesFreed > state.options.maxFreed) {
+            printMsg(lvlInfo, format("deleted more than %1% bytes; stopping") % state.options.maxFreed);
+            throw GCLimitReached();
+        }
+
+        if (state.options.maxLinks) {
+            struct stat st;
+            if (stat(nixStore.c_str(), &st) == -1)
+                throw SysError(format("statting `%1%'") % nixStore);
+            if (st.st_nlink < state.options.maxLinks) {
+                printMsg(lvlInfo, format("link count on the store has dropped below %1%; stopping") % state.options.maxLinks);
+                throw GCLimitReached();
+            }
+        }
+        
+    } else
+        printMsg(lvlTalkative, format("would delete `%1%'") % path);
+    
+    state.deleted.insert(path);
+    if (state.options.action != GCOptions::gcReturnLive)
+        state.results.paths.insert(path);
     return true;
 
+ isLive:
+    state.live.insert(path);
+    if (state.options.action == GCOptions::gcReturnLive)
+        state.results.paths.insert(path);
     return false;
 }
     
 
 void LocalStore::collectGarbage(const GCOptions & options, GCResults & results)
 {
-    bool gcKeepOutputs =
+    GCState state(results);
-        queryBoolSetting("gc-keep-outputs", false);
+    state.options = options;    
-    bool gcKeepDerivations =
+    
-        queryBoolSetting("gc-keep-derivations", true);
+    state.gcKeepOutputs = queryBoolSetting("gc-keep-outputs", false);
-    int gcKeepOutputsThreshold = 
+    state.gcKeepDerivations = queryBoolSetting("gc-keep-derivations", true);
-        queryIntSetting ("gc-keep-outputs-threshold", defaultGcLevel);
     
     /* Acquire the global GC root.  This prevents
        a) New roots from being added.
@@ -537,126 +605,59 @@ void LocalStore::collectGarbage(const GCOptions & options, GCResults & results)
     printMsg(lvlError, format("finding garbage collector roots..."));
     Roots rootMap = options.ignoreLiveness ? Roots() : nix::findRoots(true);
 
-    PathSet roots;
+    foreach (Roots::iterator, i, rootMap) state.roots.insert(i->second);
-    foreach (Roots::iterator, i, rootMap) roots.insert(i->second);
 
     /* 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). */
     if (!options.ignoreLiveness)
-        addAdditionalRoots(roots);
+        addAdditionalRoots(state.roots);
 
     if (options.action == GCOptions::gcReturnRoots) {
-        results.paths = roots;
+        results.paths = state.roots;
-        return;
-    }
-
-    /* Determine the live paths which is just the closure of the
-       roots under the `references' relation. */
-    printMsg(lvlError, format("computing live paths..."));
-    PathSet livePaths;
-    foreach (PathSet::const_iterator, i, roots)
-        computeFSClosure(canonPath(*i), livePaths);
-
-    if (gcKeepDerivations) {
-        foreach (PathSet::iterator, i, livePaths) {
-            /* Note that the deriver need not be valid (e.g., if we
-               previously ran the collector with `gcKeepDerivations'
-               turned off). */
-            Path deriver = queryDeriver(*i);
-            if (deriver != "" && isValidPath(deriver))
-                computeFSClosure(deriver, livePaths);
-        }
-    }
-
-    if (gcKeepOutputs) {
-        /* Hmz, identical to storePathRequisites in nix-store. */
-        foreach (PathSet::iterator, i, livePaths)
-            if (isDerivation(*i)) {
-                Derivation drv = derivationFromPath(*i);
-
-		string gcLevelStr = drv.env["__gcLevel"];
-		int gcLevel;
-		if (!string2Int(gcLevelStr, gcLevel))
-		    gcLevel = defaultGcLevel;
-		
-		if (gcLevel >= gcKeepOutputsThreshold)    
-		    foreach (DerivationOutputs::iterator, j, drv.outputs)
-			if (isValidPath(j->second.path))
-			    computeFSClosure(j->second.path, livePaths);
-            }
-    }
-
-    if (options.action == GCOptions::gcReturnLive) {
-        results.paths = livePaths;
         return;
     }
 
     /* 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. */
-    PathSet tempRoots;
     FDs fds;
-    readTempRoots(tempRoots, fds);
+    readTempRoots(state.tempRoots, fds);
-
+    state.roots.insert(state.tempRoots.begin(), state.tempRoots.end());
-    /* Close the temporary roots.  Note that we *cannot* do this in
-       readTempRoots(), because there we may not have all locks yet,
-       meaning that an invalid path can become valid (and thus add to
-       the references graph) after we have added it to the closure
-       (and computeFSClosure() assumes that the presence of a path
-       means that it has already been closed). */
-    PathSet tempRootsClosed;
-    foreach (PathSet::iterator, i, tempRoots)
-        if (isValidPath(*i))
-            computeFSClosure(*i, tempRootsClosed);
-        else
-            tempRootsClosed.insert(*i);
 
     /* After this point the set of roots or temporary roots cannot
        increase, since we hold locks on everything.  So everything
-       that is not currently in in `livePaths' or `tempRootsClosed'
+       that is not reachable from `roots'. */
-       can be deleted. */
 
-    /* Read the Nix store directory to find all currently existing
+    /* Now either delete all garbage paths, or just the specified
-       paths and filter out all live paths. */
+       paths (for gcDeleteSpecific). */
-    printMsg(lvlError, format("reading the Nix store..."));
-    PathSet storePaths;
 
-    if (options.action != GCOptions::gcDeleteSpecific) {
+    if (options.action == GCOptions::gcDeleteSpecific) {
-        Paths entries = readDirectory(nixStore);
-        foreach (Paths::iterator, i, entries) {
-            Path path = canonPath(nixStore + "/" + *i);
-            if (!isLive(path, livePaths, tempRoots, tempRootsClosed)) storePaths.insert(path);
-        }
-    }
 
-    else {
         foreach (PathSet::iterator, i, options.pathsToDelete) {
             assertStorePath(*i);
-            storePaths.insert(*i);
+            if (!tryToDelete(state, *i))
-            if (isLive(*i, livePaths, tempRoots, tempRootsClosed))
                 throw Error(format("cannot delete path `%1%' since it is still alive") % *i);
         }
-    }
         
-    if (options.action == GCOptions::gcReturnDead) {
+    } else {
-        results.paths.insert(storePaths.begin(), storePaths.end());
-        return;
-    }
         
-    /* Delete all dead store paths (or until one of the stop
+        printMsg(lvlError, format("reading the Nix store..."));
-       conditions is reached), respecting the partial ordering
+        Paths entries = readDirectory(nixStore);
-       determined by the references graph. */
 
-    PathSet done;
+        if (doDelete(state.options.action))
-    try {
             printMsg(lvlError, format("deleting garbage..."));
-        foreach (PathSet::iterator, i, storePaths)
+        else
-            gcPathRecursive(options, results, done, *i);
+            printMsg(lvlError, format("determining live/dead paths..."));
+    
+        try {
+            foreach (Paths::iterator, i, entries)
+                tryToDelete(state, canonPath(nixStore + "/" + *i));
         } catch (GCLimitReached & e) {
         }
     }        
+}
 
 
 }

src/libstore/local-store.hh

@@ -174,11 +174,12 @@ private:
     
     void upgradeStore12();
 
-    void gcPath(const GCOptions & options, GCResults & results,
+    struct GCState;
-        const Path & path);
 
-    void gcPathRecursive(const GCOptions & options,
+    bool tryToDelete(GCState & state, const Path & path);
-        GCResults & results, PathSet & done, const Path & path);
+    
+    bool isActiveTempFile(const GCState & state,
+        const Path & path, const string & suffix);
         
     void startSubstituter(const Path & substituter,
         RunningSubstituter & runningSubstituter);