On Mon, Oct 27, 2008 at 10:10:05PM +0100, Benedikt Schmidt wrote: > One of the patches just simplifies a function in TouchesFile. > > The other one fixes (at least I hope so) a problem I notices when > testing the the filecache version on the ghc repo: > > Darcs changes FOO checks if FOO is known under another name in the > repo, i.e. there is an unrecorded patch that moves FOO. Previously, > darcs first computed all unrecorded changes, then computed the new > filename from these changes. Now it only checks pending which is of > course much faster.
I applied the first and last. The change to use read_pending I didn't apply, since that's a function that shouldn't have been exported in the first place, since it's usually incorrect to use it. In this case it would be safe, until someone decides to add the feature of automatically detecting renames. You're welcome to try to create a safe and efficient interface for examining a subset of the local changes, but until then I don't think it's the end of the world for darcs changes to be as slow as darcs whatsnew. Note that a very nice improvment for this would be to introduce a mechanism for obtaining the changes to just a subset of files, i.e. issue814. David _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
