NarAccessor: Run in constant memory

src/libstore/binary-cache-store.cc

@@ -155,13 +155,17 @@ void BinaryCacheStore::addToStore(const ValidPathInfo & info, Source & narSource
 
     auto now1 = std::chrono::steady_clock::now();
 
+    /* Read the NAR simultaneously into a CompressionSink+FileSink (to
+       write the compressed NAR to disk), into a HashSink (to get the
+       NAR hash), and into a NarAccessor (to get the NAR listing). */
     HashSink fileHashSink(htSHA256);
-
+    std::shared_ptr<FSAccessor> narAccessor;
     {
     FdSink fileSink(fdTemp.get());
     TeeSink teeSink(fileSink, fileHashSink);
     auto compressionSink = makeCompressionSink(compression, teeSink);
-    copyNAR(narSource, *compressionSink);
+    TeeSource teeSource(narSource, *compressionSink);
+    narAccessor = makeNarAccessor(teeSource);
     compressionSink->finish();
     }
 
@@ -200,10 +204,9 @@ void BinaryCacheStore::addToStore(const ValidPathInfo & info, Source & narSource
     #if 0
     auto accessor_ = std::dynamic_pointer_cast<RemoteFSAccessor>(accessor);
 
-    auto narAccessor = makeNarAccessor(nar);
-
     if (accessor_)
         accessor_->addToCache(printStorePath(info.path), *nar, narAccessor);
+    #endif
 
     /* Optionally write a JSON file containing a listing of the
        contents of the NAR. */
@@ -216,15 +219,13 @@ void BinaryCacheStore::addToStore(const ValidPathInfo & info, Source & narSource
 
             {
                 auto res = jsonRoot.placeholder("root");
-                listNar(res, narAccessor, "", true);
+                listNar(res, ref<FSAccessor>(narAccessor), "", true);
             }
         }
 
         upsertFile(std::string(info.path.to_string()) + ".ls", jsonOut.str(), "application/json");
     }
-    #endif
 
-    #if 0
     /* Optionally maintain an index of DWARF debug info files
        consisting of JSON files named 'debuginfo/<build-id>' that
        specify the NAR file and member containing the debug info. */
@@ -281,7 +282,6 @@ void BinaryCacheStore::addToStore(const ValidPathInfo & info, Source & narSource
             threadPool.process();
         }
     }
-    #endif
 
     /* Atomically write the NAR file. */
     if (repair || !fileExists(narInfo->url)) {

src/libstore/daemon.cc

@@ -391,7 +391,8 @@ static void performOp(TunnelLogger * logger, ref<Store> store,
         }
         HashType hashAlgo = parseHashType(s);
 
-        TeeSource savedNAR(from);
+        StringSink savedNAR;
+        TeeSource savedNARSource(from, savedNAR);
         RetrieveRegularNARSink savedRegular;
 
         if (method == FileIngestionMethod::Recursive) {
@@ -399,7 +400,7 @@ static void performOp(TunnelLogger * logger, ref<Store> store,
                a string so that we can pass it to
                addToStoreFromDump(). */
             ParseSink sink; /* null sink; just parse the NAR */
-            parseDump(sink, savedNAR);
+            parseDump(sink, savedNARSource);
         } else
             parseDump(savedRegular, from);
 
@@ -407,7 +408,7 @@ static void performOp(TunnelLogger * logger, ref<Store> store,
         if (!savedRegular.regular) throw Error("regular file expected");
 
         auto path = store->addToStoreFromDump(
-            method == FileIngestionMethod::Recursive ? *savedNAR.data : savedRegular.s,
+            method == FileIngestionMethod::Recursive ? *savedNAR.s : savedRegular.s,
             baseName,
             method,
             hashAlgo);
@@ -733,7 +734,7 @@ static void performOp(TunnelLogger * logger, ref<Store> store,
         else {
             TeeParseSink tee(from);
             parseDump(tee, tee.source);
-            saved = std::move(*tee.source.data);
+            saved = std::move(*tee.saved.s);
             source = std::make_unique<StringSource>(saved);
         }
 

src/libstore/export-import.cc

@@ -77,15 +77,15 @@ StorePaths Store::importPaths(Source & source, std::shared_ptr<FSAccessor> acces
         if (deriver != "")
             info.deriver = parseStorePath(deriver);
 
-        info.narHash = hashString(htSHA256, *tee.source.data);
-        info.narSize = tee.source.data->size();
+        info.narHash = hashString(htSHA256, *tee.saved.s);
+        info.narSize = tee.saved.s->size();
 
         // Ignore optional legacy signature.
         if (readInt(source) == 1)
             readString(source);
 
         // Can't use underlying source, which would have been exhausted
-        auto source = StringSource { *tee.source.data };
+        auto source = StringSource { *tee.saved.s };
         addToStore(info, source, NoRepair, checkSigs, accessor);
 
         res.push_back(info.path);

src/libstore/nar-accessor.cc

@@ -18,7 +18,7 @@ struct NarMember
 
     /* If this is a regular file, position of the contents of this
        file in the NAR. */
-    size_t start = 0, size = 0;
+    uint64_t start = 0, size = 0;
 
     std::string target;
 
@@ -34,17 +34,19 @@ struct NarAccessor : public FSAccessor
 
     NarMember root;
 
-    struct NarIndexer : ParseSink, StringSource
+    struct NarIndexer : ParseSink, Source
     {
         NarAccessor & acc;
+        Source & source;
 
         std::stack<NarMember *> parents;
 
-        std::string currentStart;
         bool isExec = false;
 
-        NarIndexer(NarAccessor & acc, const std::string & nar)
-            : StringSource(nar), acc(acc)
+        uint64_t pos = 0;
+
+        NarIndexer(NarAccessor & acc, Source & source)
+            : acc(acc), source(source)
         { }
 
         void createMember(const Path & path, NarMember member) {
@@ -79,31 +81,38 @@ struct NarAccessor : public FSAccessor
 
         void preallocateContents(unsigned long long size) override
         {
-            currentStart = string(s, pos, 16);
-            assert(size <= std::numeric_limits<size_t>::max());
-            parents.top()->size = (size_t)size;
+            assert(size <= std::numeric_limits<uint64_t>::max());
+            parents.top()->size = (uint64_t) size;
             parents.top()->start = pos;
         }
 
         void receiveContents(unsigned char * data, unsigned int len) override
-        {
-            // Sanity check
-            if (!currentStart.empty()) {
-                assert(len < 16 || currentStart == string((char *) data, 16));
-                currentStart.clear();
-            }
-        }
+        { }
 
         void createSymlink(const Path & path, const string & target) override
         {
             createMember(path,
                 NarMember{FSAccessor::Type::tSymlink, false, 0, 0, target});
         }
+
+        size_t read(unsigned char * data, size_t len) override
+        {
+            auto n = source.read(data, len);
+            pos += n;
+            return n;
+        }
     };
 
     NarAccessor(ref<const std::string> nar) : nar(nar)
     {
-        NarIndexer indexer(*this, *nar);
+        StringSource source(*nar);
+        NarIndexer indexer(*this, source);
+        parseDump(indexer, indexer);
+    }
+
+    NarAccessor(Source & source)
+    {
+        NarIndexer indexer(*this, source);
         parseDump(indexer, indexer);
     }
 
@@ -219,6 +228,11 @@ ref<FSAccessor> makeNarAccessor(ref<const std::string> nar)
     return make_ref<NarAccessor>(nar);
 }
 
+ref<FSAccessor> makeNarAccessor(Source & source)
+{
+    return make_ref<NarAccessor>(source);
+}
+
 ref<FSAccessor> makeLazyNarAccessor(const std::string & listing,
     GetNarBytes getNarBytes)
 {

src/libstore/nar-accessor.hh

@@ -6,10 +6,14 @@
 
 namespace nix {
 
+struct Source;
+
 /* Return an object that provides access to the contents of a NAR
    file. */
 ref<FSAccessor> makeNarAccessor(ref<const std::string> nar);
 
+ref<FSAccessor> makeNarAccessor(Source & source);
+
 /* Create a NAR accessor from a NAR listing (in the format produced by
    listNar()). The callback getNarBytes(offset, length) is used by the
    readFile() method of the accessor to get the contents of files

src/libutil/archive.hh

@@ -65,9 +65,10 @@ struct ParseSink
 
 struct TeeParseSink : ParseSink
 {
+    StringSink saved;
     TeeSource source;
 
-    TeeParseSink(Source & source) : source(source) { }
+    TeeParseSink(Source & source) : source(source, saved) { }
 };
 
 void parseDump(ParseSink & sink, Source & source);

src/libutil/serialise.hh

@@ -179,17 +179,17 @@ struct TeeSink : Sink
 };
 
 
-/* Adapter class of a Source that saves all data read to `s'. */
+/* Adapter class of a Source that saves all data read to a sink. */
 struct TeeSource : Source
 {
     Source & orig;
-    ref<std::string> data;
-    TeeSource(Source & orig)
-        : orig(orig), data(make_ref<std::string>()) { }
+    Sink & sink;
+    TeeSource(Source & orig, Sink & sink)
+        : orig(orig), sink(sink) { }
     size_t read(unsigned char * data, size_t len)
     {
         size_t n = orig.read(data, len);
-        this->data->append((const char *) data, n);
+        sink(data, len);
         return n;
     }
 };