Name: rm29839 Date: 10/13/97


In working with JAR files using java.util.zip
classes, one glaring thing comes to mind. When
making a new zip file, it would be nice (and faster)
to have a copy method on the ZipOutputStream that
uses a ZipEntry as a reference. From an API standpoint,
the application can reference the entry, make a determination
as to whether or not to output the entry from one zip file
to another and simply use the copyEntry() to move it from one
file to another. This would also be faster than
the application decompressing the data (if it's ZipEntry.DEFLATED)
and subsequently recompressing it via a write method. The following
outlines what I'm talking about..

You have to do this now...

ZipInputStream foo = new ZipInputStream(..);
ZipOutputStream out = new ZipOutputStream(..);
ZipEntry entry;

while(true)
{
  try {entry = foo.getNextEntry();} catch (IOException e) {break;}
  if(entry == null)
     break;
  if(keepEntry(entry))
  {
    byte [] bytes = new byte[(int)entry.getSize()] // broken
    foo.read(bytes,0,bytes.length); // decompress the data
    out.putNextEntry(entry);
    out.write(bytes,0,bytes.length); // subsequenly recompress it.
    out.closeEntry();
  }
}

So, if you're selectively going from one zip to
another you're constantly decompressing and
recompressing data. To simplify this... The
if statement above now becomes..

  if(keepEntry(entry))
    out.copyEntry(entry); // move it over

Hope you approve.
company - Park City Group , email - ###@###.###
======================================================================