Re: Git Test Coverage Report (v2.20.0-rc0)

From: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Date: Nov 19, 2018, 18:33 UTC
Message-ID: <871s7g29zy.fsf@evledraar.gmail.com>
In-Reply-To: <9e293b1b-1845-1772-409b-031c0bf4d17b@gmail.com>

On Mon, Nov 19 2018, Derrick Stolee wrote:

> Here is a test coverage report for the uncovered lines introduced in
> v2.20.0-rc0 compared to v2.19.1.

Thanks a lot for this.

Show 6 quoted lines

> [...]
> builtin/rebase.c
> 62c23938fa 55) return env;
> [...]
> Ævar Arnfjörð Bjarmason 62c23938f: tests: add a special setup
> where rebase.useBuiltin is off

This one would be covered with GIT_TEST_REBASE_USE_BUILTIN=false. Obviously trivial, but I wonder if the rest of the coverage would look different when passed through the various GIT_TEST_* options.

That would make it take at least 12x as long if you did them one at a time. Probably just 2-3x as long in practice since most can be combined in some way, and somewhat computationally infeasible if you're going to try all possible combinations.

Show 6 quoted lines

> [...]
> fsck.c
> fb8952077d 214) die_errno("Could not read '%s'", path);
>
> René Scharfe fb8952077: fsck: use strbuf_getline() to read
> skiplist file

The fault of a patch I submitted. Couldn't find a sane & portable way to emulate cases where ferror() would trigger.