On Wed, Feb 24, 2016 at 10:14:03AM +0000, John Keeping wrote: > On Wed, Feb 24, 2016 at 03:26:57AM -0500, Jeff King wrote: > > On Wed, Feb 24, 2016 at 08:47:00AM +0100, Johannes Schindelin wrote: > > > > > I cannot think of a way how to test this: all of the regression > > > tests run inside Git's own worktree, and we cannot even assume > > > that /tmp/ is outside of a worktree (or that it exists). > > > > We could make the test conditional on whether we are in a repo. Anybody > > who builds from a tarball, or who uses --root would then run the test. > > Could we use GIT_CEILING_DIRECTORIES for this? If it's set to > TEST_OUTPUT_DIRECTORY won't that cover the in-tree and out-of-tree test > cases? Oh, right. That's much less nasty than my suggestion. > We probably do still want Peff's REPOLESS prereq just in case someone is > collecting test results in a repository, We can create arbitrary hierarchies within the trash directory. So even without removing the trash-dir .git, we could probably do: mkdir -p non-repo/foo && ( cd non-repo && GIT_CEILING_DIRECTORIES=$(pwd) && export GIT_CEILING_DIRECTORIES && cd foo && git config foo.bar baz ) or something. That should work everywhere. > but I think that will see much > better coverage than relying on people running tests from the tarball. With mine I think we'd see coverage from git devs, as it works with --root, too (and if you're not using --root with a RAM disk, I highly recommend it for the speedup). But your suggestion is way better. -Peff -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html