On Sam, 14 Nov 1998 Tom Lees wrote: >Initial indications with a couple of scripts I hacked up are slightly >disappointing (~20% savings as a diff from hamm to slink, tested for >2 or 3 packages). However, the script >doesn't do anything fancy yet (like compare individual files), so it may get >better. My scripts are attached anyway. I hacked a bit around with your scripts and made this attached script. it takes two debs and generates a diff by comparing every single file and generating lots of deltas. files which are new in the new package are just copied over, dont know if we should run an xdelta over it as it already gets compressed by tar/gzip later.... the compression is quite nice against the debs i tested it. about 40-50% gain. tell me what you think..... Kai -- Kai Duebbert e-Mail: kad@gmx.de , duebbert@informatik.tu-muenchen.de ICQ: 4884177
Attachment:
make-diff
Description: application/shellscript