This is an automated email from the ASF dual-hosted git repository.
ggregory pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/commons-compress.git
The following commit(s) were added to refs/heads/master by this push:
new 6c0ffcb32 [COMPRESS-695] Ability to use different InputStreams for
Zstd (#655)
6c0ffcb32 is described below
commit 6c0ffcb329c69c721b2104a2dd349643c80dd5b8
Author: mehmet-karaman <mehmet.kara...@advantest.com>
AuthorDate: Sun Mar 23 15:04:33 2025 +0100
[COMPRESS-695] Ability to use different InputStreams for Zstd (#655)
* Adjusted the ZipFile and the Builder. Added a Test for checking both.
* Implemented PR Comments.
* Fix Javadoc and checkstyle problem
---
.../commons/compress/archivers/zip/ZipFile.java | 43 ++++++++++++++--
.../harmony/unpack200/bytecode/Attribute.java | 2 +-
.../compress/archivers/zip/ZipFileTest.java | 57 ++++++++++++++++++++++
3 files changed, 97 insertions(+), 5 deletions(-)
diff --git
a/src/main/java/org/apache/commons/compress/archivers/zip/ZipFile.java
b/src/main/java/org/apache/commons/compress/archivers/zip/ZipFile.java
index cb5e482ab..c0bfd7c80 100644
--- a/src/main/java/org/apache/commons/compress/archivers/zip/ZipFile.java
+++ b/src/main/java/org/apache/commons/compress/archivers/zip/ZipFile.java
@@ -65,6 +65,7 @@
import org.apache.commons.io.FilenameUtils;
import org.apache.commons.io.build.AbstractOrigin.ByteArrayOrigin;
import org.apache.commons.io.build.AbstractStreamBuilder;
+import org.apache.commons.io.function.IOFunction;
import org.apache.commons.io.function.IOStream;
import org.apache.commons.io.input.BoundedInputStream;
@@ -138,6 +139,7 @@ public static class Builder extends
AbstractStreamBuilder<ZipFile, Builder> {
private boolean useUnicodeExtraFields = true;
private boolean ignoreLocalFileHeader;
private long maxNumberOfDisks = 1;
+ private IOFunction<InputStream, InputStream> zstdInputStream =
ZstdCompressorInputStream::new;
/**
* Constructs a new instance.
@@ -167,7 +169,24 @@ public ZipFile get() throws IOException {
actualDescription = path.toString();
}
final boolean closeOnError = seekableByteChannel != null;
- return new ZipFile(actualChannel, actualDescription, getCharset(),
useUnicodeExtraFields, closeOnError, ignoreLocalFileHeader);
+ return new ZipFile(actualChannel, actualDescription, getCharset(),
useUnicodeExtraFields, closeOnError, ignoreLocalFileHeader, zstdInputStream);
+ }
+
+ /**
+ * This method sets the {@link IOFunction} which will be used the
create an
+ * {@link InputStream}, which could be used to replace the officially
supported
+ * ZSTD compression library.
+ *
+ * @param zstdInpStreamFactory the IOFunction which gives the ability
to return
+ * a different InputStream for Zstd
compression, if
+ * parameter is null than it will be set
to default.
+ * @return {@code this} instance
+ * @since 1.28.0
+ */
+ public Builder setZstdInputStreamFactory(IOFunction<InputStream,
InputStream> zstdInpStreamFactory) {
+ this.zstdInputStream = zstdInpStreamFactory == null ?
ZstdCompressorInputStream::new : zstdInpStreamFactory;
+
+ return this;
}
/**
@@ -721,6 +740,8 @@ private static boolean tryToLocateSignature(final
SeekableByteChannel channel, f
private final ByteBuffer shortBbuf = ByteBuffer.wrap(shortBuf);
+ private final IOFunction<InputStream, InputStream> zstdInputStream;
+
private long centralDirectoryStartDiskNumber;
private long centralDirectoryStartRelativeOffset;
@@ -892,12 +913,13 @@ public ZipFile(final SeekableByteChannel channel, final
String encoding) throws
}
private ZipFile(final SeekableByteChannel channel, final String
channelDescription, final Charset encoding, final boolean useUnicodeExtraFields,
- final boolean closeOnError, final boolean ignoreLocalFileHeader)
throws IOException {
+ final boolean closeOnError, final boolean ignoreLocalFileHeader,
IOFunction<InputStream, InputStream> zstdInputStream) throws IOException {
this.isSplitZipArchive = channel instanceof
ZipSplitReadOnlySeekableByteChannel;
this.encoding = Charsets.toCharset(encoding, Builder.DEFAULT_CHARSET);
this.zipEncoding = ZipEncodingHelper.getZipEncoding(encoding);
this.useUnicodeExtraFields = useUnicodeExtraFields;
this.archive = channel;
+ this.zstdInputStream = zstdInputStream;
boolean success = false;
try {
final Map<ZipArchiveEntry, NameAndComment> entriesWithoutUTF8Flag
= populateFromCentralDirectory();
@@ -966,7 +988,8 @@ public ZipFile(final SeekableByteChannel channel, final
String channelDescriptio
private ZipFile(final SeekableByteChannel channel, final String
channelDescription, final String encoding, final boolean useUnicodeExtraFields,
final boolean closeOnError, final boolean ignoreLocalFileHeader)
throws IOException {
- this(channel, channelDescription, Charsets.toCharset(encoding),
useUnicodeExtraFields, closeOnError, ignoreLocalFileHeader);
+ this(channel, channelDescription, Charsets.toCharset(encoding),
useUnicodeExtraFields, closeOnError,
+ ignoreLocalFileHeader, ZstdCompressorInputStream::new);
}
/**
@@ -1234,7 +1257,7 @@ public void close() throws IOException {
return new Deflate64CompressorInputStream(is);
case ZSTD:
case ZSTD_DEPRECATED:
- return new ZstdCompressorInputStream(is);
+ return createZstdInputStream(is);
case XZ:
return new XZCompressorInputStream(is);
case AES_ENCRYPTED:
@@ -1254,6 +1277,18 @@ public void close() throws IOException {
}
}
+ /**
+ * Creates the appropriate InputStream for the ZSTD compression method.
+ *
+ * @param is the input stream which should be used for compression.
+ * @return the {@link InputStream} for handling the Zstd compression.
+ * @throws IOException if an I/O error occurs.
+ * @since 1.28.0
+ */
+ protected InputStream createZstdInputStream(final InputStream is) throws
IOException {
+ return zstdInputStream.apply(is);
+ }
+
/**
* Gets the raw stream of the archive entry (compressed form).
* <p>
diff --git
a/src/main/java/org/apache/commons/compress/harmony/unpack200/bytecode/Attribute.java
b/src/main/java/org/apache/commons/compress/harmony/unpack200/bytecode/Attribute.java
index 1ba085b0a..d18ed20a6 100644
---
a/src/main/java/org/apache/commons/compress/harmony/unpack200/bytecode/Attribute.java
+++
b/src/main/java/org/apache/commons/compress/harmony/unpack200/bytecode/Attribute.java
@@ -126,7 +126,7 @@ protected void resolve(final ClassConstantPool pool) {
* Writes this body to the given output stream.
*
* @param out the output.
- * @throws IOException
+ * @throws IOException if an I/O error occurs.
*/
protected abstract void writeBody(DataOutputStream out) throws IOException;
diff --git
a/src/test/java/org/apache/commons/compress/archivers/zip/ZipFileTest.java
b/src/test/java/org/apache/commons/compress/archivers/zip/ZipFileTest.java
index 8f4e013f4..ebc4b7c0d 100644
--- a/src/test/java/org/apache/commons/compress/archivers/zip/ZipFileTest.java
+++ b/src/test/java/org/apache/commons/compress/archivers/zip/ZipFileTest.java
@@ -66,8 +66,36 @@
import org.junit.jupiter.api.Assumptions;
import org.junit.jupiter.api.Test;
+import io.airlift.compress.zstd.ZstdInputStream;
+
public class ZipFileTest extends AbstractTest {
+ /**
+ * This Class simulates the case where the Zip File uses the
aircompressors {@link ZstdInputStream}
+ */
+ private final class AirliftZstdZipFile extends ZipFile {
+ private boolean used;
+
+ private AirliftZstdZipFile(final File file) throws IOException {
+ super(file);
+ }
+
+ protected InputStream createZstdInputStream(final InputStream is)
throws IOException {
+ return new ZstdInputStream(is) {
+
+ @Override
+ public int read(final byte[] outputBuffer, final int
outputOffset, final int outputLength) throws IOException {
+ used = true;
+ return super.read(outputBuffer, outputOffset,
outputLength);
+ }
+ };
+ }
+
+ public boolean isUsed() {
+ return used;
+ }
+ }
+
private static final int OUT_OF_MEMORY = 137;
private static void assertEntryName(final ArrayList<ZipArchiveEntry>
entries, final int index, final String expectedName) {
@@ -382,6 +410,35 @@ public void testDuplicateEntry() throws Exception {
}
}
+ @Test
+ public void testAlternativeZstdInputStream() throws Exception {
+ final File archive = getFile("COMPRESS-692/compress-692.zip");
+ try (AirliftZstdZipFile zf = new AirliftZstdZipFile(archive)) {
+ final byte[] buffer = new byte[7000];
+ final ZipArchiveEntry ze = zf.getEntry("dolor.txt");
+ assertNotNull(ze);
+ try (InputStream inputStream = zf.getInputStream(ze)) {
+ assertNotNull(inputStream);
+ assertFalse(zf.isUsed());
+ final int bytesRead =
org.apache.commons.compress.utils.IOUtils.readFully(inputStream, buffer);
+ assertEquals(6066, bytesRead);
+ assertTrue(zf.isUsed());
+ }
+ }
+
+ try (ZipFile builtZipFile =
ZipFile.builder().setPath(archive.getAbsolutePath()).setZstdInputStreamFactory(ZstdInputStream::new).get())
{
+ final byte[] buffer = new byte[7000];
+ final ZipArchiveEntry ze = builtZipFile.getEntry("dolor.txt");
+ assertNotNull(ze);
+ try (InputStream inputStream = builtZipFile.getInputStream(ze)) {
+ assertTrue(inputStream instanceof ZstdInputStream);
+ assertNotNull(inputStream);
+ final int bytesRead =
org.apache.commons.compress.utils.IOUtils.readFully(inputStream, buffer);
+ assertEquals(6066, bytesRead);
+ }
+ }
+ }
+
/**
* Test entries alignment.
*/
