On 2013-05-01, Thomas de Grivel <billi...@gmail.com> wrote: > Le 04/29/13 06:03, Hugo Osvaldo Barrera a écrit : >> On 2013-04-20 12:15, Stuart Henderson wrote: >>> On 2013-04-20, Alokat MacMoneysack <mail...@alokat.org> wrote: >>>> Hi, >>>> >>>> first, I don't want to start a flame war about why is CVS better or not >> better than X - it's just a question. >>>> >>>> If you say, we use it because it just works - it's okay. :) >>>> >>>> So why does OpenBSD still uses CVS and don't migrate to SVN or something >> like git as other OSS projekts do? >>>> >>>> Regards, >>>> fritjof >>>> >>>> >>> >>> my 2p: like all version control software CVS has bugs, but between us, >>> developers have a reasonable idea of how to avoid them in CVS, there's >>> less knowledge about other version control systems. >>> >>> Also having the repository stored in human-readable (ish) files is an >>> advantage if there was ever any repo corruption. >> >> Some other CVS keeps checksums of every commit, and every commit contains >> the checksum of the last commit + this commits diff. This helps *prevent* >> corruption (or at least prevents it from spreading). >> I think that beats human-readable files to manually find corruptions >> (that may well spread). > > I ran into a hash collision once, using git rebase.
Like I said, "between us, developers have a reasonable idea of how to avoid them in CVS, there's less knowledge about other version control systems." ;) Out of interest, what happens in this case? does git notice and ask you to make some silly change to avoid the collision (tweak whitespace or something), or does it cause havoc?
