Re: Autoexpanding ZIP archives?

[Greg Hudson]
> Given that our solution already involves a change to the compression
> software used by OO, I don't see why it changes anything to also
> require a change to the archiving software. Just store the files in
> sorted order.

Not exactly, the proposed solution involves a change to a library used
by OO, and having the library sort its input filenames is not possible
- zlib just operates on streams, not whole zip files. The distinction
is that the OO people wouldn't need to get directly involved.

I'd agree though that the onus on making files friendly to binary diff
algorithms rests on the open source software producing the files. If
not sorted order, at least something deterministic, if this is not
already the case.