On Tue, Jul 28, 2020 at 7:35 PM brian m. carlson <sandals@xxxxxxxxxxxxxxxxxxxx> wrote: > Compute the length of an object ID instead of hard-coding 40-based > values. > > Signed-off-by: brian m. carlson <sandals@xxxxxxxxxxxxxxxxxxxx> > --- > diff --git a/t/t8002-blame.sh b/t/t8002-blame.sh > @@ -6,6 +6,10 @@ test_description='git blame' > +test_expect_success 'setup' ' > + hexsz=$(test_oid hexsz) > +' In the previous version of this series, the "setup" test invoked test_oid_init() as its very first step. This version doesn't. As a reviewer, I was caught off-guard by this unexpected and unexplained difference between versions. The script works fine without test_oid_init() anyhow since test-lib.sh invokes test_oid_init(), so the test_oid_init() call introduced here by the previous version was redundant. Some of the patches in this series add test_oid_init() calls to their "setup" tests, while others don't, which makes for a somewhat confusing impression as one reads the series. In general, it would be nice for the patches to paint a consistent picture (i.e either uniformly employ test_oid_init() or don't), however, I would not want to see a re-roll just for that. Also, since the final patch of the series ends up removing all those test_oid_init() calls anyhow, it's all straightened out in the end.
