On Thu, Apr 09, 2009 at 06:04:22PM -0700, Mike Kupfer wrote: > Agreed, though if you know the area you're working in, you can do a > "make clobber" for part of the tree and then do "hg stat > <top-of-subtree>".
True. Still could be a massive amount of deletion. > I'm also thinking about changing nightly(1) so that it does "hg stat" > after it does "make clobber". This would have the advantage of flagging > clobber bugs as well as missing invocations of "hg add". Indeed, that would be an interesting thing to note. > Danek> With snapshots, you could make it work, but I'd rather see the > Danek> build objects go into a separate part of the tree, I think. > > Do you mean long-term, or do you mean we should do that instead of > fixing "make clobber"? Long term. Fixing make clobber is a good thing, too. But I'm not sure I'd "fix up" .hgignore until all the build objects are (or can be) placed in a tree of their own, or the performance issues worked out. Danek
