jade wip

Change-Id: I4d5a0bebeeefc35dac0bbc065c104c70c24aae17

lix/libutil/tarfile.cc

@@ -1,9 +1,11 @@
 #include <archive.h>
 #include <archive_entry.h>
 #include <kj/async.h>
+#include <map>
 
 #include "async-io.hh"
 #include "file-descriptor.hh"
+#include "lix/libstore/temporary-dir.hh"
 #include "lix/libutil/c-calls.hh"
 #include "lix/libutil/charptr-cast.hh"
 #include "lix/libutil/file-system.hh"
@@ -86,13 +88,57 @@ void TarArchive::close()
     check(archive_read_close(this->archive.get()), "Failed to close archive (%s)");
 }
 
+namespace {
+
+struct IndexEntry
+{
+    struct IndexDirectory
+    {
+        std::map<std::string, IndexEntry> entries;
+    };
+
+    struct IndexFile
+    {
+        uint64_t start;
+        uint64_t size;
+    };
+
+    std::variant<IndexDirectory, IndexFile> innerData;
+    int mode;
+};
+
+/** Extract a TAR archive into an in-memory index with data stored on disk,
+ * because fundamentally all we have to do when importing archives to the store
+ * is to sort the entries and then send it into a NAR serializer.
+ *
+ * This deals with macOS performance problems by not materializing the files to
+ * disk in the first place.
+ *
+ * https://git.lix.systems/lix-project/lix/issues/1072
+ */
+struct InMemoryIndex
+{
+    Path tempDir;
+    AutoDelete tempDirDeleter;
+    AutoCloseFD backingFile;
+    std::ofstream writer;
+
+    IndexEntry topLevel;
+
+    InMemoryIndex() : tempDir(createTempDir()), tempDirDeleter(tempDir, true)
+    {
+        backingFile = sys::open(tempDir + "/extracted", O_RDWR | O_TRUNC | O_CLOEXEC);
+    }
+};
+
+}
+
 static void extract_archive(TarArchive & archive, const Path & destDir)
 {
     requireCString(destDir);
 
-    int flags = ARCHIVE_EXTRACT_TIME
-        | ARCHIVE_EXTRACT_SECURE_SYMLINKS
-        | ARCHIVE_EXTRACT_SECURE_NODOTDOT;
+    int flags = ARCHIVE_EXTRACT_TIME | ARCHIVE_EXTRACT_SECURE_SYMLINKS
+        | ARCHIVE_EXTRACT_SECURE_NODOTDOT | ARCHIVE_EXTRACT_NO_HFS_COMPRESSION;
 
     for (;;) {
         struct archive_entry * entry;