Re: make test Unexpected passes

From: Elijah Newren <newren@gmail.com>
Date: Apr 22, 2016, 22:19 UTC
Message-ID: <CABPp-BFdzu4stEfbGAiqDwRAXt7EcvYBEVz1kChJaR4udC5SXA@mail.gmail.com>
In-Reply-To: <571A8404.5030200@ramsayjones.plus.com>

On Fri, Apr 22, 2016 at 1:05 PM, Ramsay Jones <ramsay@ramsayjones.plus.com> wrote:

> Hi Ben, Junio,
>
> In the second case, t6036-*.sh, git bisect fingered commit b61f9d6e
> ("ll-merge: use a longer conflict marker for internal merge", 14-04-2016).

Yeah, the t6036 testcase 'git detects conflict w/ criss-cross+contrived resolution' could be made to pass by tweaking the conflict markers. In fact, any tweak would make it appear to pass, but the test could be updated to still fail by updating the contrived resolution of a previous merge to use the newer conflict marker style as well.

The test is kind of fragile in this way, and is really there just to document this slightly weird and never seen in practice corner case. I don't think we'll ever fix the underlying "problem"; not even sure if it's possible without fundamentally re-thinking our merge strategy altogether. I just thought that when I was writing all those tests that documenting this particular behavior as a testcase had some value, but if the conflict markers are going to be updated periodically, then the cost of the test may outweigh its value and we should just toss this one test from that file.

Elijah