[llvm-dev] [RFC] Require PRs for XFAILing tests

[Renato Golin via llvm-dev]

On 28 September 2016 at 18:21, Chris Bieneman via llvm-dev
<llvm-dev at lists.llvm.org> wrote:
> This may be an unpopular opinion (and I don’t have the full context on those specific issues), but I believe that these are an abuse of XFAIL, and should probably be written in terms of REQUIRES instead of XFAIL.

Agreed.

We already have an unwritten rule to create PRs for XFAILs, and we
normally don't XFAIL lightly (I don't, at least). But creating one PR
for every existing XFAIL may end up as a long list of never looked
PRs. :)

This could be one of the commit hooks that were proposed a while ago,
to demand PR numbers for commits that have XFAIL in their change logs
(only if the line is changed). But I'm always sceptical as to the
value of such hooks...

cheers,
--renato