glibc 2.35 - Status update: ABI freeze, and release blockers.

glibc 2.35 - Status update: ABI freeze, and release blockers.

[Carlos O'Donell]

Community,

The glibc 2.35 release machine testing should start shortly. I'll
send a distinct email about that.

Please consider the ABI frozen. Any ABI changes should be OK'd with
the release manager.

(a) Blockers.

(a.1) LD_AUDIT fixes.

I'm an reviewing the LD_AUDIT fixes (series 6823). I would like to see
these fixes go in.

(a.2) glibc 2.34 ABI omission patch.

Florian,

What is the current status of the glibc 2.34 ABI omission patch?

I did a review here:
https://sourceware.org/pipermail/libc-alpha/2022-January/135271.html

I think we should commit what you have, and look at HJ's suggestions
for the future as you noted here:
https://sourceware.org/pipermail/libc-alpha/2022-January/135635.html

(a.3) ancillary 64-bit time timestamp conversion.

Adhemerval,

What is the current status of the ancillary 64-bit time timestamp
conversion fixes?

(b) Desirable.

(b.1) Predictable ELF destructor ordering.

Florian,

We haven't had much time to review this, and I'd like to spend more
time reviewing the changes and testing in Fedora. Would it be OK if
we keep reviewing this and commit early in February? This would give
us lots of time to include this in Fedora Rawhide and other rolling
releases to look for issues.

(b.2) pthread condvar stealing bug.

I'm going to switch to reviewing this right after (a.1), and if we
get a fix after the relase we can backport to the release branch,
but I won't hold the release for this bug.


-- 
Cheers,
Carlos.

Re: glibc 2.35 - Status update: ABI freeze, and release blockers.

[Florian Weimer]

* Carlos O'Donell:

> (a.2) glibc 2.34 ABI omission patch.
>
> Florian,
>
> What is the current status of the glibc 2.34 ABI omission patch?

I'm going to post a new version with or1k fixes (which have so far been
obscured by other build-many-glibcs.py failures).

> (b.1) Predictable ELF destructor ordering.
>
> Florian,
>
> We haven't had much time to review this, and I'd like to spend more
> time reviewing the changes and testing in Fedora. Would it be OK if
> we keep reviewing this and commit early in February? This would give
> us lots of time to include this in Fedora Rawhide and other rolling
> releases to look for issues.

That's okay.

We have two more issues that need investigation:

* The new terminal API for posix_spawn looks wrong to me (post
  forthcoming).

* The C.UTF-8 inconsistency reported by Michael Hudson-Doyle.

Thanks,
Florian

Re: glibc 2.35 - Status update: ABI freeze, and release blockers.

[Stefan Liebler]

On 27/01/2022 06:11, Carlos O'Donell via Libc-alpha wrote:
> Community,
> 
> The glibc 2.35 release machine testing should start shortly. I'll
> send a distinct email about that.
I've already added my test results for s390/s390x to the wiki as I'll be
out of office and will be back on 2022-02-01.
I've used commit 501246c5e2dfcc278f0ebbdb72345cdd239521c7 for testing.
I'll retest as soon as I'm back and will update the wiki if something
changes.

Bye,
Stefan