Jeff King <peff@xxxxxxxx> writes:
> Commit f0cbe742f4 (blame: add a test to cover blame_coalesce(),
> 2019-06-20) added a test case where blame can usefully coalesce two
> groups of lines. But since it relies on the normal blame output, it only
> exercises the code and can't tell whether the lines were actually
> joined into a single group.
>
> However, by using --porcelain output, we can see how git-blame considers
> the groupings (and likewise how the coalescing might have a real
> user-visible impact for a tool that uses the porcelain-output
> groupings). This lets us confirm that we are indeed coalescing correctly
> (and the fact that this test case requires coalescing can be verified by
> dropping the call to blame_coalesce(), causing the test to fail).
>
> Signed-off-by: Jeff King <peff@xxxxxxxx>
> ---
> t/t8003-blame-corner-cases.sh | 11 ++++-------
> 1 file changed, 4 insertions(+), 7 deletions(-)
>
> diff --git a/t/t8003-blame-corner-cases.sh b/t/t8003-blame-corner-cases.sh
> index b871dd4f86..1e89494ef6 100755
> --- a/t/t8003-blame-corner-cases.sh
> +++ b/t/t8003-blame-corner-cases.sh
> @@ -273,10 +273,6 @@ test_expect_success 'blame file with CRLF core.autocrlf=true' '
> grep "A U Thor" actual
> '
>
> -# Tests the splitting and merging of blame entries in blame_coalesce().
> -# The output of blame is the same, regardless of whether blame_coalesce() runs
> -# or not, so we'd likely only notice a problem if blame crashes or assigned
> -# blame to the "splitting" commit ('SPLIT' below).
> test_expect_success 'blame coalesce' '
> cat >giraffe <<-\EOF &&
> ABC
> @@ -302,10 +298,11 @@ test_expect_success 'blame coalesce' '
> git commit -m "same contents as original" &&
>
> cat >expect <<-EOF &&
> - $oid 1) ABC
> - $oid 2) DEF
> + $oid 1 1 2
> + $oid 2 2
> EOF
It has become a bit harder to grok, but for the purpose of the later
steps to see where things exactly came from (including their line
numbers), it is easier to see what is going on with the new format.
> - git -c core.abbrev=$(test_oid hexsz) blame -s giraffe >actual &&
> + git blame --porcelain giraffe >actual.raw &&
The original forced the abbrev length; by switching to the format
for porcelain-writers, we know we will get the full object name. OK.
> + grep "^$oid" actual.raw >actual &&
> test_cmp expect actual
> '
