WriteHeader: improve writing bytebuffer

- Use backing byte[] buf for ByteBuffer instance creation
- Use backing byte array to write values and calc CRC

Signed-off-by: Hiroshi Miura <miurahr@linux.com>

diff --git a/dictzip-lib/src/main/java/org/dict/zip/DictZipHeader.java b/dictzip-lib/src/main/java/org/dict/zip/DictZipHeader.java
index 2e3988a..b1f7711 100644 (file)
--- a/dictzip-lib/src/main/java/org/dict/zip/DictZipHeader.java
+++ b/dictzip-lib/src/main/java/org/dict/zip/DictZipHeader.java
@@ -299,7 +299,8 @@ public class DictZipHeader {
         headerCrc.reset();
         // force Header CRC on
         h.setGzipFlag(FHCRC, true);
-        ByteBuffer bb = ByteBuffer.allocate(22).order(ByteOrder.LITTLE_ENDIAN);
+        byte[] buf = new byte[22];
+        ByteBuffer bb = ByteBuffer.wrap(buf).order(ByteOrder.LITTLE_ENDIAN);
         ByteBuffer chunkbb = ByteBuffer.allocate(h.chunkCount * 2).order(ByteOrder.LITTLE_ENDIAN);
         bb.putShort((short) GZIP_MAGIC);
         bb.put((byte) Deflater.DEFLATED);
@@ -314,13 +315,15 @@ public class DictZipHeader {
         bb.putShort((short) h.subfieldVersion);
         bb.putShort((short) h.chunkLength);
         bb.putShort((short) h.chunkCount);
-        out.write(bb.array());
+        out.write(buf);
+        headerCrc.update(buf);
         for (int i = 0; i < h.chunkCount; i++) {
             DictZipFileUtils.writeShort(out, h.chunks[i]);
             chunkbb.putShort((short) h.chunks[i]);
         }
-        headerCrc.update(bb.array());
-        headerCrc.update(chunkbb.array());
+        chunkbb.flip();
+        headerCrc.update(chunkbb);
+        //
         if (h.gzipFlag.get(FNAME)) {
             if (h.filename != null) {
                 out.write(h.filename.getBytes(CHARSET));