On Sat, Apr 27, 2013 at 03:23:18PM +0200, René Scharfe wrote: > How about squashing in this test? > > René > > --- > t/t4300-merge-tree.sh | 13 +++++++++++++ > 1 file changed, 13 insertions(+) > > diff --git a/t/t4300-merge-tree.sh b/t/t4300-merge-tree.sh > index bd43b3d..2defb42 100755 > --- a/t/t4300-merge-tree.sh > +++ b/t/t4300-merge-tree.sh > @@ -205,6 +205,19 @@ EXPECTED > test_cmp expected actual > ' > > +test_expect_success 'file remove A, B (same)' ' > + cat >expected <<\EXPECTED && > +EXPECTED > + > + git reset --hard initial && > + test_commit "rm-a-b-base" "ONE" "AAA" && > + git rm ONE && > + git commit -m "rm-a-b" && > + git tag "rm-a-b" && > + git merge-tree rm-a-b-base rm-a-b rm-a-b >actual && I'm not sure about using the same "our" and "their" refs here. The existing tests go out of their way to create separate commits - although since they contain identical trees I don't think that actually buys us anything. Since this test does fail without my patch, it clearly does trigger the affected code, so I think it's fine as is. > + test_cmp expected actual > +' > + > test_expect_success 'file change A, remove B' ' > cat >expected <<\EXPECTED && > removed in remote > -- > 1.8.2.1 -- 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