On Thu, Aug 14, 2003 at 06:03:29PM -0500, Karl Bongers wrote: > If your actual source is > 100M you need to find your way out of the > twilight zone :) I would second this observation. GIFs, JPEGs, and other binary data is a poor match for CVS. They cannot be version controlled anyway, and must be specified as binary data. I believe CVS simply compares timestamps and uploads any newer files it finds. If these binary files need to be under version control, then subversion or bitkeeper is probably what you need. If migrating is out of the question, then may be time to refactor the CVS repository and pull out image/binary files into a distributed tarball. Put a target in your Makefile that uses curl or wget to download and extract the data to the correct directories. If that is not an option, perhaps using the file command and some shell scripting to selectively update files might be the answer. You can use "cvs update [file[ file[ ...]]]". So, maybe something like: sourcefiles:=$(wildcard *.c) datafiles:=$(shell ...) update_source: cvs update $(sourcefiles) update_data: cvs update -- Chad Walstrom <chewie at wookimus.net> http://www.wookimus.net/ assert(expired(knowledge)); /* core dump */ -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 240 bytes Desc: not available Url : http://shadowknight.real-time.com/pipermail/tclug-list/attachments/20030815/df6fa7ae/attachment.pgp