this reminds me. I tried using trackdown recently, and one of the first
patches it unpulled was the one that sets up the test case :(
it seems a much better algorithm would be to iterate through the patches
and test the whole repo without just that patch. (and any that depend on
it), I find that very commonly, it gets past a point where it unpulls
something that is actually needed and the whole trackdown fails back to
stage one.
I would be willing to do the coding, but don't know much about darcs
internals (but know haskell quite well).
looking at the trackdown code, it appears to just get a list of patches,
and applys the inverse of each one in order running the test case each
time.
it seems like it would be simple to invert a patch, apply it, test, then
apply the original patch and then invert the next, test it, reapply the patch
and so forth... the only thing I am not sure about is how to "follow
through" to get at all the dependencies of a patch so they can be
temporarily undone too, and when it is a group of patches, how do I
invert and apply the whole group?
does this sound like it would work? does it seem like a good idea? any
pointers?
John
--
John Meacham - ⑆repetae.net⑆john⑈
_______________________________________________
darcs-users mailing list
[email protected]
http://www.abridgegame.org/mailman/listinfo/darcs-users
