From 2eba4e5b41c299d82e5aa0b974a1f039997ecf6e Mon Sep 17 00:00:00 2001 From: Anna Papitto Date: Fri, 14 Jul 2023 12:19:27 -0700 Subject: [PATCH] PackReverseIndex: open file if present otherwise compute The existing #read and #computeFromIndex static builder methods require the caller to choose whether to supply an input stream of a reverse index file or a forward index to compute the reverse index from, which is slower. Allow a caller to provide a file path where the pack's reverse index might be and the pack's forward index index and simply get some reverse index instance back. Prefer opening and parsing the file if it is present, to save computation time. Otherwise, fall back onto computing the reverse index from the pack's forward index. Change-Id: I09bdd4b813ad62c86add586417b2ab86e9331aec Signed-off-by: Anna Papitto --- ...9ae0e6e789088ea0f51f164f489d14.corrupt.rev | Bin 0 -> 76 bytes ...deda40019ae0e6e789088ea0f51f164f489d14.rev | Bin 0 -> 76 bytes .../storage/file/PackReverseIndexTest.java | 44 ++++++++++++++++++ .../storage/file/PackReverseIndexFactory.java | 33 +++++++++++++ 4 files changed, 77 insertions(+) create mode 100644 org.eclipse.jgit.test/tst-rsrc/org/eclipse/jgit/test/resources/pack-cbdeda40019ae0e6e789088ea0f51f164f489d14.corrupt.rev create mode 100644 org.eclipse.jgit.test/tst-rsrc/org/eclipse/jgit/test/resources/pack-cbdeda40019ae0e6e789088ea0f51f164f489d14.rev diff --git a/org.eclipse.jgit.test/tst-rsrc/org/eclipse/jgit/test/resources/pack-cbdeda40019ae0e6e789088ea0f51f164f489d14.corrupt.rev b/org.eclipse.jgit.test/tst-rsrc/org/eclipse/jgit/test/resources/pack-cbdeda40019ae0e6e789088ea0f51f164f489d14.corrupt.rev new file mode 100644 index 0000000000000000000000000000000000000000..74283a2f9aa54045e24d9fb4e21131e4b70cb703 GIT binary patch literal 76 zcmWIYbctYKU|@t|79h<8#LQ4WE0AW`EvssK`2$Dk36=}_#V;9}ZgZXx{*cRjJ%9c| QGqpXZOU)LlWxWFe0OW-g=l}o! literal 0 HcmV?d00001 diff --git a/org.eclipse.jgit.test/tst-rsrc/org/eclipse/jgit/test/resources/pack-cbdeda40019ae0e6e789088ea0f51f164f489d14.rev b/org.eclipse.jgit.test/tst-rsrc/org/eclipse/jgit/test/resources/pack-cbdeda40019ae0e6e789088ea0f51f164f489d14.rev new file mode 100644 index 0000000000000000000000000000000000000000..6ac7d65f674cb2a4317149c01f29d0e641201d73 GIT binary patch literal 76 zcmWIYbctYKU|@t|79h<8#LQ4WE0AW`EvssK`2$Dk36=}_#V;9}ZgZXx{*cRjJ%9c| SGqpXZOU)LlWofsFH3I PackIndex.open( + getResourceFileFor(noRevFilePrefix, PackExt.INDEX))); + + assertTrue(computed instanceof PackReverseIndexComputed); + } + + @Test + public void open_readGoodFile() throws IOException { + String hasRevFilePrefix = "pack-cbdeda40019ae0e6e789088ea0f51f164f489d14."; + PackReverseIndex version1 = PackReverseIndexFactory.openOrCompute( + getResourceFileFor(hasRevFilePrefix, PackExt.REVERSE_INDEX), 6, + () -> PackIndex.open( + getResourceFileFor(hasRevFilePrefix, PackExt.INDEX))); + + assertTrue(version1 instanceof PackReverseIndexV1); + } + + @Test + public void open_readCorruptFile() { + String hasRevFilePrefix = "pack-cbdeda40019ae0e6e789088ea0f51f164f489d14."; + + assertThrows(IOException.class, + () -> PackReverseIndexFactory.openOrCompute( + getResourceFileFor(hasRevFilePrefix + "corrupt.", + PackExt.REVERSE_INDEX), + 6, () -> PackIndex.open(getResourceFileFor( + hasRevFilePrefix, PackExt.INDEX)))); + } + @Test public void read_badMagic() { byte[] badMagic = new byte[] { 'R', 'B', 'A', 'D', // magic @@ -53,4 +92,9 @@ public void read_unsupportedVersion2() { assertThrows(IOException.class, () -> PackReverseIndexFactory.readFromFile(in, 0, () -> null)); } + + private File getResourceFileFor(String packFilePrefix, PackExt ext) { + return JGitTestUtil + .getTestResourceFile(packFilePrefix + ext.getExtension()); + } } diff --git a/org.eclipse.jgit/src/org/eclipse/jgit/internal/storage/file/PackReverseIndexFactory.java b/org.eclipse.jgit/src/org/eclipse/jgit/internal/storage/file/PackReverseIndexFactory.java index b16da5ae8..32830c3cf 100644 --- a/org.eclipse.jgit/src/org/eclipse/jgit/internal/storage/file/PackReverseIndexFactory.java +++ b/org.eclipse.jgit/src/org/eclipse/jgit/internal/storage/file/PackReverseIndexFactory.java @@ -14,6 +14,8 @@ import static org.eclipse.jgit.internal.storage.file.PackReverseIndex.VERSION_1; import java.io.DataInput; +import java.io.File; +import java.io.FileNotFoundException; import java.io.IOException; import java.io.InputStream; import java.security.DigestInputStream; @@ -23,11 +25,42 @@ import org.eclipse.jgit.internal.JGitText; import org.eclipse.jgit.lib.Constants; import org.eclipse.jgit.util.IO; +import org.eclipse.jgit.util.io.SilentFileInputStream; /** * Factory for creating instances of {@link PackReverseIndex}. */ public final class PackReverseIndexFactory { + /** + * Create an in-memory pack reverse index by reading it from the given file + * if the file exists, or computing it from the given pack index if the file + * doesn't exist. + * + * @param idxFile + * the file to read the pack file from, if it exists + * @param objectCount + * the number of objects in the corresponding pack + * @param packIndexSupplier + * a function to lazily get the corresponding forward index + * @return the reverse index instance + * @throws IOException + * if reading from the file fails + */ + static PackReverseIndex openOrCompute(File idxFile, long objectCount, + PackBitmapIndex.SupplierWithIOException packIndexSupplier) + throws IOException { + try (SilentFileInputStream fd = new SilentFileInputStream(idxFile)) { + return readFromFile(fd, objectCount, packIndexSupplier); + } catch (FileNotFoundException e) { + return computeFromIndex(packIndexSupplier.get()); + } catch (IOException e) { + throw new IOException( + MessageFormat.format(JGitText.get().unreadablePackIndex, + idxFile.getAbsolutePath()), + e); + } + } + /** * Compute an in-memory pack reverse index from the in-memory pack forward * index. This computation uses insertion sort, which has a quadratic