collection-server: Zip files that contain non-ASCII characters seem to be unusable

[Chris Wagner]

As the summary says. It seems like all files that come "after" a file that has "fun characters" (non-ASCII characters, I believe) will be un-extractable.

The problem may be a simple matter of flipping a bit when writing the zip file... Of course, that would likely require editing the Haskell-based Codec.Archive.Zip library that we're making use of.

It may also be a matter of simply passing the filenames to the 'addFilesToArchive' function as UTF8-encoded text, though I suspect that is not the case.

According to the official zip "specification", it is now possible to store filenames (within a zip archive) in a UTF-8 character encoding: http://www.pkware.com/documents/casestudies/APPNOTE.TXT

I'm not sure that most/many/any applications properly support this zip extension though (but I have not tested it). If attempting to store UTF8-encoded filenames causes problems with more applications than it doesn't, then we may want to consider another approach. But with that said, support for these Unicode-based filenames is more likely to improve over time, than to get worse.