Ant's code has been written against the 970311 spec with an eye on InfoZIP's C implementation for zip 2.x. The latest version is 20040528 with six or seven revisions between the two. zip 3.x introduces a different support mechanism for UTF-8 encoded filenames that probably should be detected by Ant and taken as a hint to not use native encoding. It also has support for files bigger than 2 GByte and some other stuff. Since an archive using that features will require a recent client to extract, we should probably only create archives using them when specifically asked for it - but the archive reading code should detect and support them.
*** Bug 45548 has been marked as a duplicate of this bug. ***
bug 45548 contains an (not yet reviewed) patch for the UTF-8 flag
https://issues.apache.org/jira/browse/SANDBOX-176 is commons-compress' part of this issue
ZIP64 is slowly but steadily coming to Commons Compress https://issues.apache.org/jira/browse/COMPRESS-36 Once that is released the Compress Antlib will work for big archives (the encoding stuff has been covered already anyway). I don't expect the changes required to get ported back to Ant. I'm re-routing this issue to the Compress Antlib since it may need a new option to enable ZIP64 support in special cases.
The just release Compress Antlib 1.1 supports Zip64 extensions.