On Thu, Apr 22, 2010 at 01:27:12 -0400, Isaac Dupree wrote: > I discovered via Google that there was a tool that showed a Darcs repo > of patches into a Graphviz file showing the dependencies: "deps.pl", > used by Zooko in 2005, originally written by James Webb, and used > successfully by another person in 2007 (see those details later)
Ah good, I was wondering what became of that (especially since so many people want patch dependency viewing). > (Is there anyway to create a repo with no patches > in it that has the same patch-format, i.e. darcs1 vs. darcs2, as another > one? I didn't see any options in either `darcs get` or `darcs init` that > would let me do this. It's needed so that we can create a repo that > we're going to pull patches, one-at-a-time, into.) I don't think there's a way to do this in Darcs, but you could presumably parse the darcs show repo output [1]_ > And then... it *still* never finished running, on my quite small > hashed-format darcs repo (55 patches), with darcs 2.4, and I'm not > sure what's wrong. Should I give up and re-write it somehow if I want > graphviz results? (it's only 128 lines of code, and does hacky things > with darcs command-line and darcs interactive output) I have ingrained in my head that to do dependency viewing, we need a minimal context function <http://bugs.darcs.net/issue1514> and that a minimal context may not be so easy or in any case cheap. The minimal context is just the set of patches which cannot be commuted past a given patch. So intuitively for me, the dependencies of a patch are represented by doing a transitive reduction on the graph formed by pointing each patch to all the patches in its minimal context. But maybe I'm wrong. The list should shout me down if I'm just spreading misinformation. So how does this deps.pl work? Is it doing something which effectively works out the same as computing a minimal context? Is it reasonable to expect the idea to work (not to speak of the implementation?). I do remember that Camp has a patch viewing GUI, although I seem to recall Ian saying that his effectively computes the minimal context as well. .. [1] It may be a good idea to make darcs show repo output machine-readable. I may have some opinions on this for a future thread. -- Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow> PGP Key ID: 08AC04F9
signature.asc
Description: Digital signature
_______________________________________________ darcs-users mailing list darcs-users@darcs.net http://lists.osuosl.org/mailman/listinfo/darcs-users
