I think it is ahead because it was the first of the new VCS systems. The only 3 that really matter now (IMHO) are svn, git, and bzr. As svn was the first out there in the wild that actually worked, and made a huge difference against the older VCS's (SCCS, RCS, CVS), the switch was obvious.
Reasons for choosing either of the three may be well beyond the system itself. I chose git for a few not so obvious reasons:
- Subversion has way too many dependencies, and I didn't get it to work on HP-UX in 64bit mode in 5 working days, after which I gave up. (I was able to compile git to something that partly worked in about 4 hours).
- bzr is based on python, and I don't have python on HP-UX, and also don't want to enter the vortex of yak-shaving into getting that to run on HP-UX and maintain it in the future (in 64bit mode, as our main target for a new VCS was 64bit-only HP-UX). The base packages might be available from different sources, but they are all 32bit and won't cooperate in that 64bit environment.
- The future of perl5's VCS is git. We'll move from perforce to git after the release of perl-5.8.9, so knowing that VCS beforehand was a pre.
Enjoy, Have FUN! H.Merijn
| [reply] |
Because it's such a small step from CVS. It's certainly not the most advanced of the newer systems, or the fastest, or even the most reliable -- but it has the explicit goal of being "like CVS only with less suck". Just simple things like repo-wide revisions, rename tracking, file properties, and streamlining commandline operation. There are a lot of CVS users and a lot of former CVS users in the world and of course svn is going to appeal. | [reply] |
