> > <snip> > >> >> While I was working on the examples for this email reply I realized that the >> problem is only present for paths given in a client-spec. I added a test >> case to prove that. That means I don’t need to request *all* paths. I >> *think* it is sufficient to request the paths mentioned in the client spec >> which would usually be really fast. Stay tuned for PATCH v5. > > I've just tried a small experiment with stock unaltered git-p4. > > - Started p4d with -C1 option to case-fold. > - Add some files with different cases of directory (Path/File1, > PATH/File2, pATH/File3). > - git-p4 clone. > > The result of the clone is separate directories if I do nothing > special (PATH/File1, Path/File2, etc). But if I set core.ignorecase, I > get a single case-folded directory, Path/File1, Path/File2, etc. I'm > still failing to get how that isn't what you need (other than being a > bit obscure to get to the right invocation). > > I've put a script that shows this here: > > https://github.com/luked99/quick-git-p4-case-folding-test As mentioned I realized that the problem occurs only if you use client specs. Can you take a look at this test case / run it? https://github.com/larsxschneider/git/blob/d3a191cb5fb4d8f5f48ca9314c772169d5dbf65b/t/t9821-git-p4-path-variations.sh#L112-L127
Does this make sense to you? If you want to I could also modify “quick-git-p4-case-folding-test” to show the problem. Cheers, Lars -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
