On Thu, May 10, 2018 at 4:34 PM, Jeff King <p...@peff.net> wrote: > On Thu, May 10, 2018 at 03:58:52PM +0200, SZEDER Gábor wrote: >> Since testing bitmaps doesn't need a worktree in the first place, >> let's just create bare clones for the two JGit tests, so the cloned >> won't have an index, and these two tests can be executed even with >> split index enabled. > > Nice, this seems like a clever workaround. > > Reviewed-by: Jeff King <p...@peff.net> > > The more heavy-handed approach would be to just disable the JGIT prereq > when GIT_TEST_SPLIT_INDEX is in use, which would cover this and > potentially any other cases. This is nicer because it lets us continue > using the test. And it's not like we have a ton of jgit dependencies, > such that dealing with each individually would be a burden.
We could also 'sane_unset GIT_TEST_SPLIT_INDEX' in these two tests, but I think that we should do that only in tests that specifically check split index behavior (i.e. t1700).
