Thanks, Aevar for looking into this. I haven't looked into detail, but: - s/where/var/ in your patch 1 - I think that variables should be unset with env --unset instead of sane_unset, because (as far as I can tell) the effects of sane_unset "leak" from test block to test block. - On my computer, some tests still fail. I have included a patch (patch 1) that makes these tests succeed. - I noticed that some of the tests fail because fetch-pack doesn't support protocol v2. It turns out that adding support for protocol v2 in fetch-pack wasn't too difficult, so I have done it. This is patch 2. - With that support added, some of the other tests no longer need the unsetting of the environment variable. This is patch 3.
If you agree with the general direction I'm going in, when you send out another version, I would add patch 2 somewhere near the beginning of the set, and then squash both patch 1 and patch 3 in the G_T_P_V=2 patch. Jonathan Tan (3): squash this into your patch builtin/fetch-pack: support protocol version 2 also squash this into your patch builtin/fetch-pack.c | 9 ++++++--- t/t5500-fetch-pack.sh | 13 +++++++------ t/t5539-fetch-http-shallow.sh | 12 ++++++++---- t/t5541-http-push-smart.sh | 10 ++++++++-- t/t5551-http-fetch-smart.sh | 23 +++++++++++++++-------- t/t5570-git-daemon.sh | 2 +- t/t5616-partial-clone.sh | 3 +-- 7 files changed, 46 insertions(+), 26 deletions(-) -- 2.19.0.271.gfe8321ec05.dirty
