Re: [PATCH 1/7] test: use test_i18ncmp when checking --stat output
From: Junio C Hamano <hidden>
Date: 2016-06-15 22:53:17
Jonathan Nieder [off-list ref] writes:
quoted hunk
diff --git a/t/t3508-cherry-pick-many-commits.sh b/t/t3508-cherry-pick-many-commits.sh index 1b3a3441..d909e6db 100755 --- a/t/t3508-cherry-pick-many-commits.sh +++ b/t/t3508-cherry-pick-many-commits.sh@@ -55,7 +55,7 @@ test_expect_success 'cherry-pick first..fourth works' ' git diff --quiet HEAD other && sed -e "s/$_x05[0-9a-f][0-9a-f]/OBJID/" <actual >actual.fuzzy && - test_cmp expected actual.fuzzy && + test_i18ncmp expected actual.fuzzy && check_head_differs_from fourth '@@ -82,7 +82,7 @@ test_expect_success 'cherry-pick --strategy resolve first..fourth works' ' git diff --quiet other && git diff --quiet HEAD other && sed -e "s/$_x05[0-9a-f][0-9a-f]/OBJID/" <actual >actual.fuzzy && - test_cmp expected actual.fuzzy && + test_i18ncmp expected actual.fuzzy && check_head_differs_from fourth '
Honestly speaking, I am not sure if we should be making these comparisons to cast the informational output from cherry-pick that is meant for human concumption in stone to begin with. I think "cherry-pick first..fourth works" test should be comparing the output from something like "log --format=%s" over the range before and after the replaying. All the other changes in this patch I do agree with; in C locale these are expected to exactly match what we specify in these tests and I do not see any reason to update them, but it is OK that they are translated and attempting to limit the matches to untranslatable part is not worth it. test_i18ncmp is the tool for such comparison. But I do not think the progress output from range cherry-pick needs to be cast in stone even in C locale, hence the comments in the first two paragraphs from me.