Add zip writer strategy to convert ZipArchive to via std ZipOutputStream

Author: Col-E

src/main/java/software/coley/llzip/ZipCompressions.java

@@ -1,6 +1,9 @@
 package software.coley.llzip;
 
 import software.coley.llzip.part.LocalFileHeader;
+import software.coley.llzip.strategy.DeflateDecompressor;
+
+import java.io.IOException;
 
 /**
  * Constants for {@link LocalFileHeader#getCompressionMethod()}.
@@ -120,4 +123,96 @@ public interface ZipCompressions {
 	 * AE-x encryption marker.
 	 */
 	int AE_x = 99;
+
+	/**
+	 * @param method
+	 * 		Compression method value.
+	 *
+	 * @return Name of method.
+	 */
+	static String getName(int method) {
+		switch (method) {
+			case STRORED:
+				return "STRORED";
+			case SHRUNK:
+				return "SHRUNK";
+			case REDUCED_F1:
+				return "REDUCED_F1";
+			case REDUCED_F2:
+				return "REDUCED_F2";
+			case REDUCED_F3:
+				return "REDUCED_F3";
+			case REDUCED_F4:
+				return "REDUCED_F4";
+			case IMPLODED:
+				return "IMPLODED";
+			case RESERVED_TOKENIZING:
+				return "RESERVED_TOKENIZING";
+			case DEFLATED:
+				return "DEFLATED";
+			case DEFLATED_64:
+				return "DEFLATED_64";
+			case PKWARE_IMPLODING:
+				return "PKWARE_IMPLODING";
+			case PKWARE_RESERVED_11:
+				return "PKWARE_RESERVED_11";
+			case BZIP2:
+				return "BZIP2";
+			case PKWARE_RESERVED_13:
+				return "PKWARE_RESERVED_13";
+			case LZMA:
+				return "LZMA";
+			case PKWARE_RESERVED_15:
+				return "PKWARE_RESERVED_15";
+			case CMPSC:
+				return "CMPSC";
+			case PKWARE_RESERVED_17:
+				return "PKWARE_RESERVED_17";
+			case IBM_TERSE:
+				return "IBM_TERSE";
+			case IBM_LZ77:
+				return "IBM_LZ77";
+			case DEPRECATED_ZSTD:
+				return "DEPRECATED_ZSTD";
+			case ZSTANDARD:
+				return "ZSTANDARD";
+			case MP3:
+				return "MP3";
+			case XZ:
+				return "XZ";
+			case JPEG:
+				return "JPEG";
+			case WAVPACK:
+				return "WAVPACK";
+			case PPMD:
+				return "PPMD";
+			case AE_x:
+				return "AE_x";
+			default:
+				return "Unknown[" + method + "]";
+		}
+	}
+
+	/**
+	 * @param header
+	 * 		Header with {@link LocalFileHeader#getFileData()} to decompress.
+	 *
+	 * @return Decompressed {@code byte[]}.
+	 *
+	 * @throws IOException
+	 * 		When the decompression failed.
+	 */
+	static byte[] decompress(LocalFileHeader header) throws IOException {
+		int method = header.getCompressionMethod();
+		switch (method) {
+			case STRORED:
+				return header.getFileData();
+			case DEFLATED:
+				return header.decompress(new DeflateDecompressor());
+			default:
+				// TODO: Support other decompressing techniques
+				String methodName = getName(method);
+				throw new IOException("Unsupported compression method: " + methodName);
+		}
+	}
 }

src/main/java/software/coley/llzip/part/CentralDirectoryFileHeader.java

@@ -14,7 +14,7 @@
  */
 public class CentralDirectoryFileHeader implements ZipPart, ZipRead {
 	private transient int offset = -1;
-	private transient LocalFileHeader linked;
+	private transient LocalFileHeader linkedFileHeader;
 	// Zip spec elements
 	private int versionMadeBy;
 	private int versionNeededToExtract;
@@ -78,16 +78,16 @@ public int offset() {
 	/**
 	 * @return The file header associated with {@link #getRelativeOffsetOfLocalHeader()}. May be {@code null}.
 	 */
-	public LocalFileHeader getLinked() {
-		return linked;
+	public LocalFileHeader getLinkedFileHeader() {
+		return linkedFileHeader;
 	}
 
 	/**
-	 * @param linked
+	 * @param header
 	 * 		The file header associated with {@link #getRelativeOffsetOfLocalHeader()}. May be {@code null}.
 	 */
-	public void link(LocalFileHeader linked) {
-		this.linked = linked;
+	public void link(LocalFileHeader header) {
+		this.linkedFileHeader = header;
 	}
 
 	/**
@@ -453,15 +453,15 @@ public boolean equals(Object o) {
 				internalFileAttributes == that.internalFileAttributes &&
 				externalFileAttributes == that.externalFileAttributes &&
 				relativeOffsetOfLocalHeader == that.relativeOffsetOfLocalHeader &&
-				Objects.equals(linked, that.linked) &&
+				Objects.equals(linkedFileHeader, that.linkedFileHeader) &&
 				fileName.equals(that.fileName) &&
 				Arrays.equals(extraField, that.extraField) &&
 				fileComment.equals(that.fileComment);
 	}
 
 	@Override
 	public int hashCode() {
-		int result = Objects.hash(offset, linked, versionMadeBy, versionNeededToExtract, generalPurposeBitFlag,
+		int result = Objects.hash(offset, linkedFileHeader, versionMadeBy, versionNeededToExtract, generalPurposeBitFlag,
 				compressionMethod, lastModFileTime, lastModFileDate, crc32, compressedSize, uncompressedSize,
 				fileNameLength, extraFieldLength, fileCommentLength, diskNumberStart, internalFileAttributes,
 				externalFileAttributes, relativeOffsetOfLocalHeader, fileName, fileComment);

src/main/java/software/coley/llzip/part/LocalFileHeader.java

@@ -13,6 +13,7 @@
  */
 public class LocalFileHeader implements ZipPart, ZipRead {
 	private transient int offset = -1;
+	private transient CentralDirectoryFileHeader linkedDirectoryFileHeader;
 	// Zip spec elements
 	private int versionNeededToExtract;
 	private int generalPurposeBitFlag;
@@ -74,6 +75,21 @@ public byte[] decompress(Decompressor decompressor) throws IOException {
 		return decompressor.decompress(this, fileData);
 	}
 
+	/**
+	 * @return The central directory file header this file is associated with.
+	 */
+	public CentralDirectoryFileHeader getLinkedDirectoryFileHeader() {
+		return linkedDirectoryFileHeader;
+	}
+
+	/**
+	 * @param directoryFileHeader
+	 * 		The central directory file header this file is associated with.
+	 */
+	public void link(CentralDirectoryFileHeader directoryFileHeader) {
+		this.linkedDirectoryFileHeader = directoryFileHeader;
+	}
+
 	/**
 	 * @return Zip software version.
 	 */

src/main/java/software/coley/llzip/part/ZipRead.java

@@ -14,6 +14,8 @@ public interface ZipRead {
 	 */
 	void read(byte[] data, int offset);
 
-	// TODO: Write conventions, then rename to 'ZipReadWrite' and tie into ZipWriterStrategy
-	//        - maybe transform primitive 'offset' to data type references and auto-compute offsets in output.
+	// TODO: Write conventions, then rename to 'ZipReadWrite' and tie into a new ZipWriterStrategy
+	//        - similar to read, but with data-output-stream writes bytes to stream
+	//        - obviously requires offset data to be correct for a 'valid' output
+	//            - but still allows invalid output to facilitate crafting intentional malformed zips/jars
 }

src/main/java/software/coley/llzip/strategy/DefaultZipReaderStrategy.java

@@ -51,6 +51,7 @@ public void read(ZipArchive zip, byte[] data) throws IOException {
 				file.read(data, offset);
 				zip.getParts().add(file);
 				directory.link(file);
+				file.link(directory);
 				offsets.add(offset);
 			} else {
 				logger.warn("Central-Directory-File-Header's offset[{}] to Local-File-Header does not match the Local-File-Header magic!", offset);

src/main/java/software/coley/llzip/strategy/JavaZipWriterStategy.java

@@ -0,0 +1,35 @@
+package software.coley.llzip.strategy;
+
+import software.coley.llzip.ZipArchive;
+import software.coley.llzip.ZipCompressions;
+import software.coley.llzip.part.CentralDirectoryFileHeader;
+import software.coley.llzip.part.LocalFileHeader;
+
+import java.io.IOException;
+import java.io.OutputStream;
+import java.util.Optional;
+import java.util.zip.ZipEntry;
+import java.util.zip.ZipOutputStream;
+
+/**
+ * Uses the Java {@link ZipOutputStream} to recompute the zip file format.
+ * The only used data in this case is {@link LocalFileHeader#getFileData()} and the file name
+ * which can be either {@link CentralDirectoryFileHeader#getFileName()} or {@link LocalFileHeader#getFileName()}.
+ *
+ * @author Matt Coley
+ */
+public class JavaZipWriterStategy implements ZipWriterStrategy {
+	@Override
+	public void write(ZipArchive archive, OutputStream os) throws IOException {
+		try (ZipOutputStream zos = new ZipOutputStream(os)) {
+			for (LocalFileHeader fileHeader : archive.getLocalFiles()) {
+				String name = Optional.ofNullable(fileHeader.getLinkedDirectoryFileHeader())
+						.map(CentralDirectoryFileHeader::getFileName)
+						.orElse(fileHeader.getFileName());
+				zos.putNextEntry(new ZipEntry(name));
+				zos.write(ZipCompressions.decompress(fileHeader));
+				zos.closeEntry();
+			}
+		}
+	}
+}

src/main/java/software/coley/llzip/strategy/JvmZipReaderStrategy.java

@@ -90,6 +90,7 @@ else if (Array.startsWith(data, jvmBaseOffset, ZipPatterns.CENTRAL_DIRECTORY_FIL
 					file.read(data, offset);
 					zip.getParts().add(file);
 					directory.link(file);
+					file.link(directory);
 					offsets.add(offset);
 				} catch (Exception ex) {
 					logger.warn("Failed to read 'local file header' at offset[{}]", offset);

src/main/java/software/coley/llzip/util/NameComparator.java

@@ -0,0 +1,49 @@
+package software.coley.llzip.util;
+
+import software.coley.llzip.part.CentralDirectoryFileHeader;
+import software.coley.llzip.part.LocalFileHeader;
+import software.coley.llzip.part.ZipPart;
+import software.coley.llzip.strategy.JavaZipWriterStategy;
+
+import java.util.Comparator;
+import java.util.Optional;
+
+/**
+ * For sorting by {@link CentralDirectoryFileHeader#getFileName()} and {@link LocalFileHeader#getFileName()}.
+ * This is intended to be used in cases like {@link JavaZipWriterStategy} where offset information is ignored anyways.
+ *
+ * @author Matt Coley
+ */
+public class NameComparator implements Comparator<ZipPart> {
+	private final Comparator<ZipPart> fallback;
+
+	/**
+	 * @param fallback
+	 * 		Fallback comparator for
+	 */
+	public NameComparator(Comparator<ZipPart> fallback) {
+		this.fallback = fallback;
+	}
+
+	@Override
+	public int compare(ZipPart o1, ZipPart o2) {
+		if (o1 instanceof LocalFileHeader && o2 instanceof LocalFileHeader) {
+			LocalFileHeader header1 = (LocalFileHeader) o1;
+			LocalFileHeader header2 = (LocalFileHeader) o2;
+			String name1 = Optional.ofNullable(header1.getLinkedDirectoryFileHeader())
+					.map(CentralDirectoryFileHeader::getFileName)
+					.orElse(header1.getFileName());
+			String name2 = Optional.ofNullable(header2.getLinkedDirectoryFileHeader())
+					.map(CentralDirectoryFileHeader::getFileName)
+					.orElse(header2.getFileName());
+			return name1.compareTo(name2);
+		} else if (o1 instanceof CentralDirectoryFileHeader && o2 instanceof CentralDirectoryFileHeader) {
+			CentralDirectoryFileHeader header1 = (CentralDirectoryFileHeader) o1;
+			CentralDirectoryFileHeader header2 = (CentralDirectoryFileHeader) o2;
+			String name1 = header1.getFileName();
+			String name2 = header2.getFileName();
+			return name1.compareTo(name2);
+		}
+		return fallback.compare(o1, o2);
+	}
+}

src/test/java/software/coley/llzip/CompressionTests.java

@@ -9,7 +9,6 @@
 import software.coley.llzip.part.LocalFileHeader;
 import software.coley.llzip.strategy.Decompressor;
 import software.coley.llzip.strategy.DeflateDecompressor;
-import software.coley.llzip.strategy.JvmZipReaderStrategy;
 
 import java.io.IOException;
 import java.nio.file.Files;
@@ -60,14 +59,14 @@ public void testDeflateJvmJar() {
 			// The red herring class that most zip tools see
 			CentralDirectoryFileHeader redHerringCentralDir = zip.getCentralDirectories().get(0);
 			assertEquals("Hello.class", redHerringCentralDir.getFileName());
-			assertNull( redHerringCentralDir.getLinked(), "The red herring central directory got linked");
+			assertNull( redHerringCentralDir.getLinkedFileHeader(), "The red herring central directory got linked");
 			byte[] redHerringClassData = zip.getLocalFiles().get(1).decompress(new DeflateDecompressor());
 			assertDefinesString(redHerringClassData, "Hello world!");
 			// The real class that gets run by the JVM
 			CentralDirectoryFileHeader jvmCentralDir = zip.getCentralDirectories().get(3);
 			assertEquals("Hello.class/", jvmCentralDir.getFileName());
-			assertNotEquals("Hello.class/", jvmCentralDir.getLinked().getFileName());
-			byte[] classData = jvmCentralDir.getLinked().decompress(new DeflateDecompressor());
+			assertNotEquals("Hello.class/", jvmCentralDir.getLinkedFileHeader().getFileName());
+			byte[] classData = jvmCentralDir.getLinkedFileHeader().decompress(new DeflateDecompressor());
 			assertDefinesString(classData, "The secret code is: ROSE");
 		} catch (IOException ex) {
 			fail(ex);
@@ -84,13 +83,13 @@ public void testDeflateJvmJarWithGarbageHeader() {
 			// The red herring class that most zip tools see
 			CentralDirectoryFileHeader redHerringCentralDir = zip.getCentralDirectories().get(1);
 			assertEquals("Hello\t.class", redHerringCentralDir.getFileName());
-			byte[] redHerringClassData = redHerringCentralDir.getLinked().decompress(new DeflateDecompressor());
+			byte[] redHerringClassData = redHerringCentralDir.getLinkedFileHeader().decompress(new DeflateDecompressor());
 			assertDefinesString(redHerringClassData, "Hello world!");
 			// The real class that gets run by the JVM
 			CentralDirectoryFileHeader jvmCentralDir = zip.getCentralDirectories().get(0);
 			assertEquals("Hello.class/", jvmCentralDir.getFileName());
-			assertNotEquals("Hello.class/", jvmCentralDir.getLinked().getFileName());
-			byte[] classData = jvmCentralDir.getLinked().decompress(new DeflateDecompressor());
+			assertNotEquals("Hello.class/", jvmCentralDir.getLinkedFileHeader().getFileName());
+			byte[] classData = jvmCentralDir.getLinkedFileHeader().decompress(new DeflateDecompressor());
 			assertDefinesString(classData, "The secret code is: ROSE");
 		} catch (IOException ex) {
 			fail(ex);

src/test/java/software/coley/llzip/PartParseTests.java

@@ -89,7 +89,7 @@ public void testJvmTrickJar() {
 
 	private static boolean hasFile(ZipArchive zip, String name) {
 		return zip.getCentralDirectories().stream()
-				.anyMatch(cdfh -> cdfh.getLinked() != null &&
+				.anyMatch(cdfh -> cdfh.getLinkedFileHeader() != null &&
 						cdfh.getFileName().equals(name));
 	}
 }