| Message ID | 20190422044812.GA28493@sigill.intra.peff.net (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | t5516: fix mismerge in 'next' | expand |
Jeff King <peff@peff.net> writes: > The merge of jk/fetch-reachability-error-fix conflicts with > jt/test-protocol-version, but the conflict resolution done by > b4ce8375c018 has a typo (looks like an editor mistake): > ... Indeed it was. Thanks for spotting. > The same problem is present in the merges to pu, which is not surprising > since the 'next' merge was probably just resolved by rerere. Yup, here is what I did: $ git checkout b4ce8375c018 $ git am jk-t5516-fix.patch $ CG=$(git describe) $ git reset --hard b4ce8375c018^ $ git merge jk/fetch-reachability-error-fix $ git rerere forget t/t5516-fetch-push.sh $ git diff $CG ;# shows the typo $ git checkout $CG t/t5516-fetch-push.sh # I should learn to use 'git restore' one of these days. $ git diff ;# shows --cc $ git rerere $ git reset --hard $ git merge jk/fetch-reachability-error-fix $ git diff $CG ;# shows no difference That should teach rerere about the right resolution. And to carry the fix in 'next': $ git checkout next $ git am -s3 jk-t5516-fix.patch Thanks.
diff --git a/t/t5516-fetch-push.sh b/t/t5516-fetch-push.sh index 09e7178987..c81ca360ac 100755 --- a/t/t5516-fetch-push.sh +++ b/t/t5516-fetch-push.sh @@ -1242,7 +1242,7 @@ do # Some protocol versions (e.g. 2) support fetching # unadvertised objects, so restrict this test to v0. test_must_fail env GIT_TEST_PROTOCOL_VERSION= \ - git fetepo/.git $SHA1_3 && + git fetch ../testrepo/.git $SHA1_3 && test_must_fail env GIT_TEST_PROTOCOL_VERSION= \ git fetch ../testrepo/.git $SHA1_1 && git --git-dir=../testrepo/.git config uploadpack.allowreachablesha1inwant true &&
The merge of jk/fetch-reachability-error-fix conflicts with jt/test-protocol-version, but the conflict resolution done by b4ce8375c018 has a typo (looks like an editor mistake): $ git show b4ce8375c018 [...] - test_must_fail git fetch ../testrepo/.git $SHA1_3 && - test_must_fail git fetch ../testrepo/.git $SHA1_1 && + # Some protocol versions (e.g. 2) support fetching + # unadvertised objects, so restrict this test to v0. - test_must_fail ok=sigpipe env GIT_TEST_PROTOCOL_VERSION= \ - git fetch ../testrepo/.git $SHA1_3 && - test_must_fail ok=sigpipe env GIT_TEST_PROTOCOL_VERSION= \ ++ test_must_fail env GIT_TEST_PROTOCOL_VERSION= \ ++ git fetepo/.git $SHA1_3 && ++ test_must_fail env GIT_TEST_PROTOCOL_VERSION= \ + git fetch ../testrepo/.git $SHA1_1 && The tests don't notice the bogus command because we expect the command to fail. Signed-off-by: Jeff King <peff@peff.net> --- The same problem is present in the merges to pu, which is not surprising since the 'next' merge was probably just resolved by rerere. We'd presumably want a fixup like this for 'next' until the next rewind, but the more important thing is to adjust the rerere cache so the eventual merge to 'master' is correct. t/t5516-fetch-push.sh | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)