Dan Pascu <[email protected]> writes: >> Argh! Could I confirm that this is with an external merge tool (and not >> just the standard darcs pull?) > > Yes it's with an external merge, but that's already stated in the bug > report. The internal merge is fine. This is also not the fault of the > external > merge tool. Darcs simply provides an empty file for one of the versions to > the > external merge tool, making it look like a major change was performed in that > version which leads to countless conflicts that create a mess. The result of > this is that I end up with a large chunk (read almost all of the file) > missing > (i.e all parts of the file that the pulled patch and the other conflicting > patch didn't touch).
Well, the issue is that no-one on the darcs team is using external merge, as far as I can tell. However, a testcase for the bug would certainly improve its changes of it being fixed. I am on a tight schedule, and converting informal bug descriptions into tests is something I currently cannot afford to spend time on. I can look at the (and probably fix) the bug if I get a test. This works for other bugs as well: I would say that on average, having an automated reproducer for a bug more than doubles its chances of a developer looking at it. So if people care about a certain bug, the best thing they can do (other than fixing it, of course) is providing a working test, ideally in form of a darcs patch. If you decide to make and post one, please use "Accept issueNNNN: short description." as the patch name, and store the testcase in tests/failing-issueNNNN-foo.sh. Yours, Petr. _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
