[Nut-upsdev] Time for a distributed VCS?

Eric S. Raymond esr at thyrsus.com
Wed Nov 30 18:41:52 UTC 2011


Arnaud Quette <aquette.dev at gmail.com>:
> > It is still working fine.
> 
> ok. if you one day feel a replacement is needed, you know where to
> knock... (not on the head please ;)

Noted.
 
> but I'm all for a doc review from you, if you're interested in.
> The docs (user manual, developer guide and website) are stable now,
> though still evolving.

Going on my work queue, but not likely to happen for a while.

You mentioned earlier that the weirdness in commit 7 was a cvs2svn 
artifact. I know how to deal with that now.  It may result in a parent 
pointer for that commit that is slightly off, but given that those
changesets were imported from CVS the 'correct' attachment point is not
very well defined in the first place.

Large, old Subversion repos with CVS in their ancient history, like
yours, are the exact reason I created reposurgeon in the first place -
I had to git-lift one of these (the GPSD project) and the random
cvs2svn cruft offended me. I originally built reposurgeon to remove
that scar tissue, though it has other uses now.

Now I'm taking another whack at the problem.  By the time I'm done,
repostreamer is going to end up doing better-quality Subversion lifts
with less fiddling than git-svn or either version of svn2git.  The
nut repo makes a pretty good test for repostreamer's armor-plating.
-- 
		<a href="http://www.catb.org/~esr/">Eric S. Raymond</a>



More information about the Nut-upsdev mailing list