> > > I'd like to see an optional logging feature that keeps track of which of the
> > > used compressors was most successful in a given round of compressing.
> > >
> >
> > And another idea...
> >
> > Let zipmax use the original zipped contents when it does the comparison
> between
> > the resulting zips. That way, if you already have a bunch of zip files that
> > were optimized using zipmax and 7 compressors and you discover a new
> compressor,
> > you can just run it against that one compressor instead of all eight (original
> 7
> > + new 1).
> >
>
> Where should zipmax see that it was already compressed by zipmax without using
> flags inside the zip which increases the size? ;)
>

Maybe I'm confused about how zipmax works. A zip file is a collection of compressed chunks. Each chunk being a file, possibly compressed with a different method supported by the zip format. I thought that zipmax took the best compressed chunks from multiple zipped archives and combined them into a new zip archive. If I'm correct, and this is how it works, then wouldn't it be possible to do the following...

1. unzip an archive
2. re-zip the archive with a given compressor
3. compare each of the compressed chunks in the new archive with the same compressed chunk in the original archive. when a chunk in the new archive is smaller, replace that chunk in the original archive with the smaller chunk.