Support tracking prefix data in front of archives

Author: Col-E

pom.xml

@@ -6,7 +6,7 @@
 
     <groupId>software.coley</groupId>
     <artifactId>lljzip</artifactId>
-    <version>2.4.1</version>
+    <version>2.5.0</version>
 
     <name>LL Java ZIP</name>
     <description>Lower level ZIP support for Java</description>

src/main/java/software/coley/lljzip/format/model/ZipArchive.java

@@ -1,12 +1,14 @@
 package software.coley.lljzip.format.model;
 
+import software.coley.lljzip.format.read.ZipReader;
 import software.coley.lljzip.format.transform.ZipPartMapper;
 import software.coley.lljzip.util.OffsetComparator;
 
 import javax.annotation.Nonnull;
 import javax.annotation.Nullable;
 import java.io.Closeable;
 import java.io.IOException;
+import java.lang.foreign.MemorySegment;
 import java.util.*;
 import java.util.function.Predicate;
 import java.util.stream.Collectors;
@@ -19,6 +21,7 @@
 public class ZipArchive implements AutoCloseable, Iterable<ZipPart> {
 	private final List<ZipPart> parts = new ArrayList<>();
 	private final Closeable closableBackingResource;
+	private MemorySegment prefixData;
 
 	/**
 	 * New zip archive without any backing resource.
@@ -37,6 +40,28 @@ public ZipArchive(@Nonnull Closeable closableBackingResource) {
 		this.closableBackingResource = closableBackingResource;
 	}
 
+	/**
+	 * If the {@link ZipReader} used to read this archive supports tracking such information this will return
+	 * any data not contained in the archive that was found at the front of the input content.
+	 * <p/>
+	 * Some applications like Jar2Exe will prepend a loader executable in front of the jar, in which case this content
+	 * would be that executable. It can be any kind of arbitrary data though.
+	 *
+	 * @return Data at the front of the archive.
+	 */
+	@Nullable
+	public MemorySegment getPrefixData() {
+		return prefixData;
+	}
+
+	/**
+	 * @param prefixData
+	 * 		Data at the front of the archive.
+	 */
+	public void setPrefixData(@Nullable MemorySegment prefixData) {
+		this.prefixData = prefixData;
+	}
+
 	/**
 	 * @param mapper
 	 * 		Part mapper to manipulate contained zip parts.

src/main/java/software/coley/lljzip/format/read/ForwardScanZipReader.java

@@ -13,8 +13,7 @@
 import javax.annotation.Nonnull;
 import java.io.IOException;
 import java.lang.foreign.MemorySegment;
-import java.util.HashSet;
-import java.util.Set;
+import java.util.*;
 
 /**
  * The standard read strategy that should work with standard zip archives.
@@ -67,7 +66,7 @@ public void read(@Nonnull ZipArchive zip, @Nonnull MemorySegment data) throws IO
 
 		// Read local files
 		// - Set to prevent duplicate file header entries for the same offset
-		Set<Long> offsets = new HashSet<>();
+		NavigableSet<Long> offsets = new TreeSet<>();
 		for (CentralDirectoryFileHeader directory : zip.getCentralDirectories()) {
 			long offset = zipStart + directory.getRelativeOffsetOfLocalHeader();
 			if (!offsets.contains(offset) && MemorySegmentUtil.readQuad(data, offset) == ZipPatterns.LOCAL_FILE_HEADER_QUAD) {
@@ -83,6 +82,12 @@ public void read(@Nonnull ZipArchive zip, @Nonnull MemorySegment data) throws IO
 			}
 		}
 
+		// Record any data appearing at the front of the file not associated with the ZIP file contents.
+		if (!offsets.isEmpty()) {
+			long firstOffset = offsets.first();
+			zip.setPrefixData(data.asSlice(0, firstOffset));
+		}
+
 		// Sort based on order
 		zip.sortParts(new OffsetComparator());
 	}

src/main/java/software/coley/lljzip/format/read/JvmZipReader.java

@@ -40,7 +40,7 @@ public JvmZipReader() {
 	 *
 	 * @param skipRevisitedCenToLocalLinks
 	 * 		Flag to skip creating duplicate {@link LocalFileHeader} entries if multiple
-	 * 		{@link CentralDirectoryFileHeader} point to the same location.
+	 *        {@link CentralDirectoryFileHeader} point to the same location.
 	 */
 	public JvmZipReader(boolean skipRevisitedCenToLocalLinks) {
 		this(new JvmZipPartAllocator(), skipRevisitedCenToLocalLinks);
@@ -53,7 +53,7 @@ public JvmZipReader(boolean skipRevisitedCenToLocalLinks) {
 	 * 		Allocator to use.
 	 * @param skipRevisitedCenToLocalLinks
 	 * 		Flag to skip creating duplicate {@link LocalFileHeader} entries if multiple
-	 * 		{@link CentralDirectoryFileHeader} point to the same location.
+	 *        {@link CentralDirectoryFileHeader} point to the same location.
 	 */
 	public JvmZipReader(@Nonnull ZipPartAllocator allocator, boolean skipRevisitedCenToLocalLinks) {
 		super(allocator);
@@ -203,6 +203,12 @@ else if (MemorySegmentUtil.readQuad(data, jvmBaseFileOffset) == ZipPatterns.CENT
 			}
 		}
 
+		// Record any data appearing at the front of the file not associated with the ZIP file contents.
+		if (!entryOffsets.isEmpty()) {
+			long firstOffset = entryOffsets.first();
+			zip.setPrefixData(data.asSlice(0, firstOffset));
+		}
+
 		// Sort based on order
 		zip.sortParts(new OffsetComparator());
 	}

src/main/java/software/coley/lljzip/format/read/NaiveLocalFileZipReader.java

@@ -34,12 +34,18 @@ public NaiveLocalFileZipReader(@Nonnull ZipPartAllocator allocator) {
 
 	@Override
 	public void read(@Nonnull ZipArchive zip, @Nonnull MemorySegment data) throws IOException {
-		long localFileOffset = -1;
-		while ((localFileOffset = MemorySegmentUtil.indexOfQuad(data, localFileOffset + 1, ZipPatterns.LOCAL_FILE_HEADER_QUAD)) >= 0) {
+		long localFileOffset = MemorySegmentUtil.indexOfQuad(data, 0, ZipPatterns.LOCAL_FILE_HEADER_QUAD);
+		if (localFileOffset < 0) return;
+		if (localFileOffset > 0) {
+			// The first offset containing archive data is not at the first byte.
+			// Record whatever content is at the front.
+			zip.setPrefixData(data.asSlice(0, localFileOffset));
+		}
+		do {
 			LocalFileHeader file = newLocalFileHeader();
 			file.read(data, localFileOffset);
 			zip.addPart(file);
 			postProcessLocalFileHeader(file);
-		}
+		} while ((localFileOffset = MemorySegmentUtil.indexOfQuad(data, localFileOffset + 1, ZipPatterns.LOCAL_FILE_HEADER_QUAD)) >= 0);
 	}
 }