On 10/14/05, Miles Bader <[EMAIL PROTECTED]> wrote: > 2005/10/14, Mark Flacy <[EMAIL PROTECTED]>: > > I (and presumably the list) would really be interested in reading what > > those problems with tagline inventory mode were. > > I think it's pretty simple... "That's not what CVS did!!!"
Well, we are still running some cvs->arch imports regularly, and all the approaches for that expect "names" inventory. Even if I add shell scripts to use taglines automagically, if I create a file and contribute it upstream, the cvs->arch gateway will assign it an arbitrary tag, and it'll conflict with my original file, even if they are one and the same. I'll happily accept that perhaps I didn't find the one true way of doing it. If anyone is using Arch successfully to track upstream projects that don't use arch, and have it automagically figure out that the file/patch coming from the other side has been seen locally, please, PLEASE, tell me how. I mean, in a simple way that doesn't expect Arch to be everywhere, or me to attach bits of magic metadata here and there. At the end of the day I actually don't care if git does it with bubblegum and papier mache. It does it. Game over. I bow to the proof that a smart idea works even if it's not the canonical textbook answer (no pun intended). Hey, that's the FOSS/hacker spirit, and what makes us strong. Whining (!) about theoretical "flaws" is what Oracle DBAs do when someone talks about PostgreSQL or some other FOSS alternative. So I'd appreciate someone downloading and building git and tricking it to do something *really* stupid. C'mon, show that the penguin has no clothes. Seems to be easy to explain it, shouldn't be hard to prove it to the unwashed masses. martin
_______________________________________________ Gnu-arch-users mailing list Gnu-arch-users@gnu.org http://lists.gnu.org/mailman/listinfo/gnu-arch-users GNU arch home page: http://savannah.gnu.org/projects/gnu-arch/
