On Thu, 6 May 2010, Trey Harris wrote: > In a message dated Thu, 6 May 2010, John BORIS writes: > >> Since I started this thread let me chime back in. I know the need for >> VCS but using it for binaries I think , for my setup, isn't a good fit. >> What I have done have always kept copies of ecery compiled COBOL code >> and C code with an extention to match the date it was taken out of >> service and also a directory tree by version number. Now I am a small >> shop. 1 Manager and three programmers. I don't count myself as a >> programmer, just the keeper of the "keys". >> >> I agree with Yves about pulling a ton of data when you pull the >> repository. > > Again--not trying to defend git, it just happens to be what I'm familiar > with. Pulling the entire Linux source tree, including all history, takes > just a very few minutes on a fast link and just a few megabytes of space, > and refreshes even after weeks since the prior refresh, take seconds. > Most people have the reaction, "whoah, did something go wrong? It couldn't > possibly have pulled that fast" the first time they do a git pull. > > The git protocol is fantastically compressed and optimized.
The problem arises when you have files that cannot be compressed. Think mp3 or video files that may get a minor tweak to them, but are now a completely different binary. Allthat git can do is to store both versions. If you have something like this that changes a lot, the git repository can end up being very large (in a pathalogical situation it can actually make git unusable) David Lang _______________________________________________ Discuss mailing list Discuss@lopsa.org http://lopsa.org/cgi-bin/mailman/listinfo/discuss This list provided by the League of Professional System Administrators http://lopsa.org/
