This is the mail archive of the
mailing list for the GCC project.
Re: [PATCH] accept all C integer types in function parameters referenced by alloc_align (PR 88363)
On Tue, Dec 11, 2018 at 2:46 PM Martin Sebor <email@example.com> wrote:
> On 12/11/18 11:15 AM, Joseph Myers wrote:
> > On Tue, 11 Dec 2018, Martin Sebor wrote:
> >> I recently brought up the question of the write w/o approval
> >> policy on the gcc list:
> >> https://gcc.gnu.org/ml/gcc/2018-11/msg00165.html
> >> looking for clarification. Except for Jeff's comment (which
> >> I'm afraid didn't really clarify things), didn't get any.
> > I think "will the person who objects to my work the most be able to find a
> > fault with my fix?" in the policy on obviousness is clear enough. A
> > policy decision on what is or is not part of a language extension can't be
> > obvious, and nor can determining subtle questions of exactly what the
> > definition of some internal interface is or should be.
> Anything that someone might find fault with is so broad as to
> remove the ability to make the judgment in any case. Reviewers
> have been known to find fault with the slightest things, from
> trivial formatting nits, to punctuation in comments, to names
> of variables, to the location of new tests, to ChangeLogs.
> If the policy's intent is not to let people make judgment calls
> then it ought to be updated. I have no proposal for changes to
> it at the moment, but I don't see how anyone can reasonably
> object to someone posting a patch and saying "if there are no
> objections I will go ahead and commit this change because I think
> it's obviously correct." If even that is against the policy then
> change it to make that clear (though I sincerely hope that isn't
I don't think anyone objected to your mail, they were just disagreeing
that the patch was obvious. That is also a judgment call. IMO
there's no need to have an ironclad policy here, since the
consequences of a particular change being "wrongly" consider either
obvious or non-obvious are small.