[llvm-dev] [RFC] Documentation clarification: Phabricator, not the lists is the main entry point for new patches
James Henderson via llvm-dev
llvm-dev at lists.llvm.org
Wed Jun 19 07:50:42 PDT 2019
I don't think I've ever looked at llvm-commits for patches that need
reviewing except for when I was looking to actively get involved in
reviewing an area I wasn't already subscribed to by default. Certainly, if
somebody posted a patch on llvm-commits to one of the binutils like
llvm-readelf or llvm-objcopy, I'd be interested in it, but probably
wouldn't see it nowadays, as I have too great a workload to filter through
the mailing lists, whereas I can setup Herald rules to get myself
auto-subscribed/added as a reviewer to relevant patches on Phabricator.
Registering isn't a complicated process, compared to developing a patch in
most cases anyway.
That's a long way of saying +1.
On Wed, 19 Jun 2019 at 15:30, Roman Lebedev via llvm-dev <
llvm-dev at lists.llvm.org> wrote:
> The current documentation talks about both the Phabricator review, and
> as mail replies on -commits lists. It also talks about submitting
> patches to lists,
> with the subtext that it may be friendlier for outsiders.
> It is true that Phabricator has some entry threshold, larger than
> github, or maillists,
> so the attempt is not unwarranted. But from what i can tell, 99.9% patches
> via Phabricator. There is a large chance that such a mail-only patch
> will simply be
> overlooked, ignored, or the very first reply will be "Please post the
> patch to
> Both of these cases i would call counter-welcoming.
> I don't think that is what we want?
> I propose to fix the docs to specify that all new patches should go
> via Phabricator, not lists:
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the llvm-dev