Mark Mentovai <mark@xxxxxxxxxxxx> writes: > Forcing the entire test suite to run from a realpath-canonicalized > path seems like a blunt instrument, though. I view it differently. It is a very effective way to make sure that the tests work for one person does not mysteriously break for another, only due to where they kept the git repository and where in their filesystem they arranged the tests to run. That is not the kind of breakage we want our developers to be scratching their heads about. The part of the test-lib I showed already was trying to make sure that our idea of $PWD matches the physical world. It did so only for PWD, by doing 'cd -P "$TRASH_DIRECTORY"' early in the test setup sequence, but it did so after setting HOME, leaving it inconsistent between physical and logical world. Which is the bug that my suggested fix addresses. Besides the problem we saw in this discussion thread is *not* a bug in Git and its normalization of its working directory, is it? It was a bug in the test script that assumed that $HOME already was normalized, right? If we want to make sure Git (not its tests, but Git itself) works when (pwd) and (pwd -P) are different, we should have a dedicated test to arrange that for everybody, regardless of where the tester checks out their git repository. We can only do so on symlink capable filesystems, but we do have enough support in the test framework to do so if you want to write such a test.
