Affects Version/s: None
Fix Version/s: 1.0
I've recently applied three patches to the Ant code base that may be of interest to commons-compress as well.
use more modern collections (since Ant trunk now officially is JDK 1.4+). This results in some minor performance improvements because of less synchronization and a theoretical bigger improvement for entries with many extra fields.
make sure the archive is closed in the presence of a RuntimeExcepetion which may happen when LFH parsing fails.
skip extra fields that don't follow the recommended HEADER-ID, LENGTH structure at the end of the extra field (this is what unzip does as well).
There are two more issues open in Ant's bugzilla I may be working on over the next few weeks (low prio for me):
https://issues.apache.org/bugzilla/show_bug.cgi?id=46637 - which may be impossible to do in a backwards compatible way (not a concern for a sandbox component, I guess) and the more important https://issues.apache.org/bugzilla/show_bug.cgi?id=45429 which would allow archives to contain files bigger than 2 GB and support UTF8 filenames in a more portable way.
Technically I am a commons committer, maybe we can collaborate on the required changes via the commons dev list (or Ant's).