[Zlib-devel] Google's version of zlib code

[glennrp at comcast.net]

I extracted the IDAT chunk data from 1750 files from
the Mozilla mercurial repository and recompressed them
in various ways.  The gzip, zopfli, bzip2, and xz
results are with the IDAT chunk data generated by
pngcrush -brute and then decompressed, including
only the PNG filter bytes and filtered pixel data.

      Method           time   IDAT bytes
    original files      n/a    5019516
    pngcrush           24.5    4299905
    pngcrush -brute   310.8    4281038
    gzip -9n            5.6    4339189
    zopfli --zlib     596.7    4035633
    bzip2 -9            3.8    4015342
    xz                 14.6    3813468

The bzip2 and xz results include whatever overhead
exists.  I subtracted 12 bytes per file from the
gzip results to predict the IDAT length.

The tools for extracting the IDAT data are
in the sourceforge pmt/pngzop project.

Glenn