Optimize buffer reuse and improve NIO usage in ZIP archiver

- Refine ByteArrayOutputStream.reset() to safely handle empty buffers.
- Use Path and Files API in DeferredScatterOutputStream and OffloadingOutputStream for better performance and reliability.
- Expose Path in OffloadingOutputStream to avoid unnecessary File object conversions.

Author: slachiewicz

src/main/java/org/codehaus/plexus/archiver/zip/ByteArrayOutputStream.java

@@ -262,7 +262,7 @@ public synchronized void reset() {
         } else {
             // Throw away old buffers
             currentBuffer = null;
-            int size = buffers.get(0).length;
+            int size = buffers.isEmpty() ? 1024 : buffers.get(0).length;
             buffers.clear();
             needNewBuffer(size);
             reuseBuffers = true;

src/main/java/org/codehaus/plexus/archiver/zip/DeferredScatterOutputStream.java

@@ -17,9 +17,10 @@
  */
 package org.codehaus.plexus.archiver.zip;
 
-import java.io.File;
 import java.io.IOException;
 import java.io.InputStream;
+import java.nio.file.Files;
+import java.nio.file.Path;
 
 import org.apache.commons.compress.parallel.ScatterGatherBackingStore;
 
@@ -48,9 +49,9 @@ public void closeForWriting() throws IOException {
 
     @Override
     public void close() throws IOException {
-        File file = dfos.getFile();
+        Path file = dfos.getOutputPath();
         if (file != null) {
-            file.delete();
+            Files.deleteIfExists(file);
         }
     }
 }

src/main/java/org/codehaus/plexus/archiver/zip/OffloadingOutputStream.java

@@ -142,6 +142,10 @@ public byte[] getData() {
         return null;
     }
 
+    public Path getOutputPath() {
+        return outputPath;
+    }
+
     /**
      * Returns either the output file specified in the constructor or
      * the temporary file created or null.