[PATCH v2 0/5] avoid -Wuse-after-free [BZ #26779]

[Martin Sebor]

On 1/25/22 10:46, Carlos O'Donell wrote:
> On 1/24/22 19:52, Martin Sebor via Libc-alpha wrote:
>> This is a repost of the original patch but broken down by source
>> file and with some suppression done by #pragma GCC diagnostic
>> instead of conversion to intptr_t.  It also adds fixes for
>> the same problem in the test suite that I overlooked before.
> 
> Thanks for the repost! We really want gcc 12 and glibc 2.35 to work together.
> 
> For future posts please review the contribution checklist, we have some
> specific instructions to help reviewers and CI/CD that interacts with your
> patch.
> 
> (1) Allow the reviewer to review all of what you will push.
> 
> Your current posts do not use git format-patch and so do not provide me
> with the commit message for review.
> 
> The intent is that I as a reviewer can review your commit message as
> expected to be pushed. I want to be able to see all of the work you
> will push (like a PR/MR) and approve it all.
> 
> It should be possible for you to have pushed all 5 patches as distinct
> commits with commit messages, use git format-patch --cover-letter HEAD~5
> to generate 6 files to mail out, and then you fill in patch 0 and send.

Thanks for the review!  I've pushed the changes as distinct commits
with the adjusted descriptions (including Florian suggestion) after
rerunning the tests.  Some of the tests failed so I fixed those up
and posted an update.

> 
> (2) CI/CD
> 
> Your use of "Re:" in patches 2-5 has broken CI, and it sees these as
> follow-ups to your original messages.
> 
> The contribution checklist has some notes about this:
> ~~~
> In order for an in-reply-to with a new version of the patch to be
> treated as a new patch you must remove the "Re:" from the subject.
> If you leave the "Re:" then patchwork considers your reply a comment
> to the original patch. This is important to support reviewers using
> patchwork for pulling patches and for CI/CD systems testing your patches.
> ~~~
> 
> Is "Re:" common in other communities you are a part of?
> 

In GCC it doesn't matter.  Other things that don't matter here
are enforced there (e.g., like the minute details of ChangeLog
entries, the exact form of the PR reference and where it goes).

Martin