Merge remote-tracking branch 'upstream/master' into optional-derivation-output-storepath

6 files changed, 180 insertions, 14 deletions

diff --git a/src/libstore/content-address.cc b/src/libstore/content-address.cc
index 3d753836f..6cb69d0a9 100644
--- a/src/libstore/content-address.cc
+++ b/src/libstore/content-address.cc
@@ -82,4 +82,16 @@ std::string renderContentAddress(std::optional<ContentAddress> ca) {
     return ca ? renderContentAddress(*ca) : "";
 }
 
+Hash getContentAddressHash(const ContentAddress & ca)
+{
+    return std::visit(overloaded {
+        [](TextHash th) {
+            return th.hash;
+        },
+        [](FixedOutputHash fsh) {
+            return fsh.hash;
+        }
+    }, ca);
+}
+
 }
diff --git a/src/libstore/content-address.hh b/src/libstore/content-address.hh
index ba4797f5b..22a039242 100644
--- a/src/libstore/content-address.hh
+++ b/src/libstore/content-address.hh
@@ -53,4 +53,6 @@ ContentAddress parseContentAddress(std::string_view rawCa);
 
 std::optional<ContentAddress> parseContentAddressOpt(std::string_view rawCaOpt);
 
+Hash getContentAddressHash(const ContentAddress & ca);
+
 }
diff --git a/src/libstore/globals.hh b/src/libstore/globals.hh
index 58cf08763..4d5eec7bf 100644
--- a/src/libstore/globals.hh
+++ b/src/libstore/globals.hh
@@ -365,6 +365,9 @@ public:
 
     Setting<bool> warnDirty{this, true, "warn-dirty",
         "Whether to warn about dirty Git/Mercurial trees."};
+
+    Setting<size_t> narBufferSize{this, 32 * 1024 * 1024, "nar-buffer-size",
+        "Maximum size of NARs before spilling them to disk."};
 };
 
 
diff --git a/src/libstore/local-store.cc b/src/libstore/local-store.cc
index 7ecaa8743..d1aca9790 100644
--- a/src/libstore/local-store.cc
+++ b/src/libstore/local-store.cc
@@ -969,7 +969,7 @@ void LocalStore::addToStore(const ValidPathInfo & info, Source & source,
 
         PathLocks outputLock;
 
-        Path realPath = realStoreDir + "/" + std::string(info.path.to_string());
+        auto realPath = Store::toRealPath(info.path);
 
         /* Lock the output path.  But don't lock if we're being called
            from a build hook (whose parent process already acquired a
@@ -1040,8 +1040,7 @@ StorePath LocalStore::addToStoreFromDump(const string & dump, const string & nam
         /* The first check above is an optimisation to prevent
            unnecessary lock acquisition. */
 
-        Path realPath = realStoreDir + "/";
-        realPath += dstPath.to_string();
+        auto realPath = Store::toRealPath(dstPath);
 
         PathLocks outputLock({realPath});
 
@@ -1091,16 +1090,125 @@ StorePath LocalStore::addToStore(const string & name, const Path & _srcPath,
 {
     Path srcPath(absPath(_srcPath));
 
-    /* Read the whole path into memory. This is not a very scalable
-       method for very large paths, but `copyPath' is mainly used for
-       small files. */
-    StringSink sink;
-    if (method == FileIngestionMethod::Recursive)
-        dumpPath(srcPath, sink, filter);
-    else
-        sink.s = make_ref<std::string>(readFile(srcPath));
+    /* For computing the NAR hash. */
+    auto sha256Sink = std::make_unique<HashSink>(htSHA256);
+
+    /* For computing the store path. In recursive SHA-256 mode, this
+       is the same as the NAR hash, so no need to do it again. */
+    std::unique_ptr<HashSink> hashSink =
+        method == FileIngestionMethod::Recursive && hashAlgo == htSHA256
+        ? nullptr
+        : std::make_unique<HashSink>(hashAlgo);
+
+    /* Read the source path into memory, but only if it's up to
+       narBufferSize bytes. If it's larger, write it to a temporary
+       location in the Nix store. If the subsequently computed
+       destination store path is already valid, we just delete the
+       temporary path. Otherwise, we move it to the destination store
+       path. */
+    bool inMemory = true;
+    std::string nar;
+
+    auto source = sinkToSource([&](Sink & sink) {
+
+        LambdaSink sink2([&](const unsigned char * buf, size_t len) {
+            (*sha256Sink)(buf, len);
+            if (hashSink) (*hashSink)(buf, len);
+
+            if (inMemory) {
+                if (nar.size() + len > settings.narBufferSize) {
+                    inMemory = false;
+                    sink << 1;
+                    sink((const unsigned char *) nar.data(), nar.size());
+                    nar.clear();
+                } else {
+                    nar.append((const char *) buf, len);
+                }
+            }
+
+            if (!inMemory) sink(buf, len);
+        });
+
+        if (method == FileIngestionMethod::Recursive)
+            dumpPath(srcPath, sink2, filter);
+        else
+            readFile(srcPath, sink2);
+    });
+
+    std::unique_ptr<AutoDelete> delTempDir;
+    Path tempPath;
+
+    try {
+        /* Wait for the source coroutine to give us some dummy
+           data. This is so that we don't create the temporary
+           directory if the NAR fits in memory. */
+        readInt(*source);
+
+        auto tempDir = createTempDir(realStoreDir, "add");
+        delTempDir = std::make_unique<AutoDelete>(tempDir);
+        tempPath = tempDir + "/x";
+
+        if (method == FileIngestionMethod::Recursive)
+            restorePath(tempPath, *source);
+        else
+            writeFile(tempPath, *source);
+
+    } catch (EndOfFile &) {
+        if (!inMemory) throw;
+        /* The NAR fits in memory, so we didn't do restorePath(). */
+    }
+
+    auto sha256 = sha256Sink->finish();
+
+    Hash hash = hashSink ? hashSink->finish().first : sha256.first;
+
+    auto dstPath = makeFixedOutputPath(method, hash, name);
+
+    addTempRoot(dstPath);
+
+    if (repair || !isValidPath(dstPath)) {
+
+        /* The first check above is an optimisation to prevent
+           unnecessary lock acquisition. */
+
+        auto realPath = Store::toRealPath(dstPath);
 
-    return addToStoreFromDump(*sink.s, name, method, hashAlgo, repair);
+        PathLocks outputLock({realPath});
+
+        if (repair || !isValidPath(dstPath)) {
+
+            deletePath(realPath);
+
+            autoGC();
+
+            if (inMemory) {
+                /* Restore from the NAR in memory. */
+                StringSource source(nar);
+                if (method == FileIngestionMethod::Recursive)
+                    restorePath(realPath, source);
+                else
+                    writeFile(realPath, source);
+            } else {
+                /* Move the temporary path we restored above. */
+                if (rename(tempPath.c_str(), realPath.c_str()))
+                    throw Error("renaming '%s' to '%s'", tempPath, realPath);
+            }
+
+            canonicalisePathMetaData(realPath, -1); // FIXME: merge into restorePath
+
+            optimisePath(realPath);
+
+            ValidPathInfo info(dstPath);
+            info.narHash = sha256.first;
+            info.narSize = sha256.second;
+            info.ca = FixedOutputHash { .method = method, .hash = hash };
+            registerValidPath(info);
+        }
+
+        outputLock.setDeletion(true);
+    }
+
+    return dstPath;
 }
 
 
@@ -1114,8 +1222,7 @@ StorePath LocalStore::addTextToStore(const string & name, const string & s,
 
     if (repair || !isValidPath(dstPath)) {
 
-        Path realPath = realStoreDir + "/";
-        realPath += dstPath.to_string();
+        auto realPath = Store::toRealPath(dstPath);
 
         PathLocks outputLock({realPath});
 
diff --git a/src/libstore/store-api.cc b/src/libstore/store-api.cc
index c0a8bc9f6..8d46bb436 100644
--- a/src/libstore/store-api.cc
+++ b/src/libstore/store-api.cc
@@ -7,6 +7,7 @@
 #include "json.hh"
 #include "derivations.hh"
 #include "url.hh"
+#include "archive.hh"
 
 #include <future>
 
@@ -221,6 +222,40 @@ StorePath Store::computeStorePathForText(const string & name, const string & s,
 }
 
 
+ValidPathInfo Store::addToStoreSlow(std::string_view name, const Path & srcPath,
+    FileIngestionMethod method, HashType hashAlgo,
+    std::optional<Hash> expectedCAHash)
+{
+    /* FIXME: inefficient: we're reading/hashing 'tmpFile' three
+       times. */
+
+    auto [narHash, narSize] = hashPath(htSHA256, srcPath);
+
+    auto hash = method == FileIngestionMethod::Recursive
+        ? hashAlgo == htSHA256
+          ? narHash
+          : hashPath(hashAlgo, srcPath).first
+        : hashFile(hashAlgo, srcPath);
+
+    if (expectedCAHash && expectedCAHash != hash)
+        throw Error("hash mismatch for '%s'", srcPath);
+
+    ValidPathInfo info(makeFixedOutputPath(method, hash, name));
+    info.narHash = narHash;
+    info.narSize = narSize;
+    info.ca = FixedOutputHash { .method = method, .hash = hash };
+
+    if (!isValidPath(info.path)) {
+        auto source = sinkToSource([&](Sink & sink) {
+            dumpPath(srcPath, sink);
+        });
+        addToStore(info, *source);
+    }
+
+    return info;
+}
+
+
 Store::Store(const Params & params)
     : Config(params)
     , state({(size_t) pathInfoCacheSize})
diff --git a/src/libstore/store-api.hh b/src/libstore/store-api.hh
index b122e05d6..b1dd1f478 100644
--- a/src/libstore/store-api.hh
+++ b/src/libstore/store-api.hh
@@ -450,6 +450,13 @@ public:
         FileIngestionMethod method = FileIngestionMethod::Recursive, HashType hashAlgo = htSHA256,
         PathFilter & filter = defaultPathFilter, RepairFlag repair = NoRepair) = 0;
 
+    /* Copy the contents of a path to the store and register the
+       validity the resulting path, using a constant amount of
+       memory. */
+    ValidPathInfo addToStoreSlow(std::string_view name, const Path & srcPath,
+        FileIngestionMethod method = FileIngestionMethod::Recursive, HashType hashAlgo = htSHA256,
+        std::optional<Hash> expectedCAHash = {});
+
     // FIXME: remove?
     virtual StorePath addToStoreFromDump(const string & dump, const string & name,
         FileIngestionMethod method = FileIngestionMethod::Recursive, HashType hashAlgo = htSHA256, RepairFlag repair = NoRepair)