It appears to me that the ZipInputStream and ZipOutputStream should
not know about the Compressors. Instead the ZipEntry should handle
this bit. The ZipInputStream, ZipOutputStream and ZipFile understands that the zip file consists of ZipEntries, the ZipEntry class understands everything about a particular ZipEntry including the compressor it needs to use.



* Implement ZipFsDir.glob
* ZipFile.checkIntegrity method
* non-MSDOS permission attributes
** See mail from Ned Konz to ruby-talk subj. "Re: SV: [ANN] Archive 0.2"
* Packager version, required unpacker version in zip headers
** See mail from Ned Konz to ruby-talk subj. "Re: SV: [ANN] Archive 0.2"
* implement small gui app on top, to exercise library and to discover
what features should be added to rubyzip next.
* implement storing attributes and ownership information
