[-- Attachment #1: Type: text/plain, Size: 186 bytes --] Hi all, Please CC me on replies. It seems like musl is missing these confstr parameters which are necessary to portably get the build flags for building multithreaded programs. [-- Attachment #2: This is a digitally signed message part --] [-- Type: application/pgp-signature, Size: 252 bytes --]
On Tue, Mar 08, 2022 at 06:50:51AM +0000, John Scott wrote:
> Hi all,
>
> Please CC me on replies.
>
> It seems like musl is missing these confstr parameters which are
> necessary to portably get the build flags for building multithreaded
> programs.
Yes, we were waiting to add them to pick common numbers that glibc
would also use. Any idea if they've done that yet? If not I'll ping
them again, and if they still don't respond I guess we just pick our
own and let them potentially diverge...
Rich
On 08/03/2022 10:02, Rich Felker wrote:
> On Tue, Mar 08, 2022 at 06:50:51AM +0000, John Scott wrote:
>> Hi all,
>>
>> Please CC me on replies.
>>
>> It seems like musl is missing these confstr parameters which are
>> necessary to portably get the build flags for building multithreaded
>> programs.
>
> Yes, we were waiting to add them to pick common numbers that glibc
> would also use. Any idea if they've done that yet? If not I'll ping
> them again, and if they still don't respond I guess we just pick our
> own and let them potentially diverge...
Do you have the initial thread on libc-alpha in hand? I think I missed it.
In any case I would like to not diverge.
On Tue, Mar 08, 2022 at 10:52:48AM -0300, Adhemerval Zanella wrote:
>
>
> On 08/03/2022 10:02, Rich Felker wrote:
> > On Tue, Mar 08, 2022 at 06:50:51AM +0000, John Scott wrote:
> >> Hi all,
> >>
> >> Please CC me on replies.
> >>
> >> It seems like musl is missing these confstr parameters which are
> >> necessary to portably get the build flags for building multithreaded
> >> programs.
> >
> > Yes, we were waiting to add them to pick common numbers that glibc
> > would also use. Any idea if they've done that yet? If not I'll ping
> > them again, and if they still don't respond I guess we just pick our
> > own and let them potentially diverge...
>
> Do you have the initial thread on libc-alpha in hand? I think I missed it.
> In any case I would like to not diverge.
From 2020:
Subject: [RFC][PATCH] * bits/confname.h: Define _CS_POSIX_V7_THREADS_CFLAGS, _CS_POSIX_V7_THREADS_LDFLAGS
Message-Id: <20201026233303.16034-1-ericonr@disroot.org>
I thought there was a bugzilla item for it too but I can't find it.
Rich
On 08/03/2022 11:01, Rich Felker wrote: > On Tue, Mar 08, 2022 at 10:52:48AM -0300, Adhemerval Zanella wrote: >> >> >> On 08/03/2022 10:02, Rich Felker wrote: >>> On Tue, Mar 08, 2022 at 06:50:51AM +0000, John Scott wrote: >>>> Hi all, >>>> >>>> Please CC me on replies. >>>> >>>> It seems like musl is missing these confstr parameters which are >>>> necessary to portably get the build flags for building multithreaded >>>> programs. >>> >>> Yes, we were waiting to add them to pick common numbers that glibc >>> would also use. Any idea if they've done that yet? If not I'll ping >>> them again, and if they still don't respond I guess we just pick our >>> own and let them potentially diverge... >> >> Do you have the initial thread on libc-alpha in hand? I think I missed it. >> In any case I would like to not diverge. > > From 2020: > > Subject: [RFC][PATCH] * bits/confname.h: Define _CS_POSIX_V7_THREADS_CFLAGS, _CS_POSIX_V7_THREADS_LDFLAGS > Message-Id: <20201026233303.16034-1-ericonr@disroot.org> Thanks, the patch missed some bits but I think the rationale is ok to include. If Érico can update the patch I will apply it. > > I thought there was a bugzilla item for it too but I can't find it. It is BZ#25003 [1]. https://sourceware.org/bugzilla/show_bug.cgi?id=25003
On 3/8/22 11:32, Adhemerval Zanella wrote:
>
>
> On 08/03/2022 11:01, Rich Felker wrote:
>> On Tue, Mar 08, 2022 at 10:52:48AM -0300, Adhemerval Zanella wrote:
>>>
>>>
>>> On 08/03/2022 10:02, Rich Felker wrote:
>>>> On Tue, Mar 08, 2022 at 06:50:51AM +0000, John Scott wrote:
>>>>> Hi all,
>>>>>
>>>>> Please CC me on replies.
>>>>>
>>>>> It seems like musl is missing these confstr parameters which are
>>>>> necessary to portably get the build flags for building multithreaded
>>>>> programs.
>>>>
>>>> Yes, we were waiting to add them to pick common numbers that glibc
>>>> would also use. Any idea if they've done that yet? If not I'll ping
>>>> them again, and if they still don't respond I guess we just pick our
>>>> own and let them potentially diverge...
>>>
>>> Do you have the initial thread on libc-alpha in hand? I think I missed it.
>>> In any case I would like to not diverge.
>>
>> From 2020:
>>
>> Subject: [RFC][PATCH] * bits/confname.h: Define _CS_POSIX_V7_THREADS_CFLAGS, _CS_POSIX_V7_THREADS_LDFLAGS
>> Message-Id: <20201026233303.16034-1-ericonr@disroot.org>
>
> Thanks, the patch missed some bits but I think the rationale is ok to
> include. If Érico can update the patch I will apply it.
Agreed, there is no reason not to add them.
With libpthread merged into libc (as you noted) this is simpler too.
--
Cheers,
Carlos.
On Tue, Mar 08, 2022 at 03:23:08PM -0500, Carlos O'Donell wrote:
> On 3/8/22 11:32, Adhemerval Zanella wrote:
> >
> >
> > On 08/03/2022 11:01, Rich Felker wrote:
> >> On Tue, Mar 08, 2022 at 10:52:48AM -0300, Adhemerval Zanella wrote:
> >>>
> >>>
> >>> On 08/03/2022 10:02, Rich Felker wrote:
> >>>> On Tue, Mar 08, 2022 at 06:50:51AM +0000, John Scott wrote:
> >>>>> Hi all,
> >>>>>
> >>>>> Please CC me on replies.
> >>>>>
> >>>>> It seems like musl is missing these confstr parameters which are
> >>>>> necessary to portably get the build flags for building multithreaded
> >>>>> programs.
> >>>>
> >>>> Yes, we were waiting to add them to pick common numbers that glibc
> >>>> would also use. Any idea if they've done that yet? If not I'll ping
> >>>> them again, and if they still don't respond I guess we just pick our
> >>>> own and let them potentially diverge...
> >>>
> >>> Do you have the initial thread on libc-alpha in hand? I think I missed it.
> >>> In any case I would like to not diverge.
> >>
> >> From 2020:
> >>
> >> Subject: [RFC][PATCH] * bits/confname.h: Define _CS_POSIX_V7_THREADS_CFLAGS, _CS_POSIX_V7_THREADS_LDFLAGS
> >> Message-Id: <20201026233303.16034-1-ericonr@disroot.org>
> >
> > Thanks, the patch missed some bits but I think the rationale is ok to
> > include. If Érico can update the patch I will apply it.
>
> Agreed, there is no reason not to add them.
>
> With libpthread merged into libc (as you noted) this is simpler too.
OK. As long as glibc is on board with adding these now, can someone
confirm what the numbers will be so I can go ahead and add them on the
musl side?
Rich
On Tue, Mar 08, 2022 at 04:22:27PM -0500, Rich Felker wrote:
> On Tue, Mar 08, 2022 at 03:23:08PM -0500, Carlos O'Donell wrote:
> > On 3/8/22 11:32, Adhemerval Zanella wrote:
> > >
> > >
> > > On 08/03/2022 11:01, Rich Felker wrote:
> > >> On Tue, Mar 08, 2022 at 10:52:48AM -0300, Adhemerval Zanella wrote:
> > >>>
> > >>>
> > >>> On 08/03/2022 10:02, Rich Felker wrote:
> > >>>> On Tue, Mar 08, 2022 at 06:50:51AM +0000, John Scott wrote:
> > >>>>> Hi all,
> > >>>>>
> > >>>>> Please CC me on replies.
> > >>>>>
> > >>>>> It seems like musl is missing these confstr parameters which are
> > >>>>> necessary to portably get the build flags for building multithreaded
> > >>>>> programs.
> > >>>>
> > >>>> Yes, we were waiting to add them to pick common numbers that glibc
> > >>>> would also use. Any idea if they've done that yet? If not I'll ping
> > >>>> them again, and if they still don't respond I guess we just pick our
> > >>>> own and let them potentially diverge...
> > >>>
> > >>> Do you have the initial thread on libc-alpha in hand? I think I missed it.
> > >>> In any case I would like to not diverge.
> > >>
> > >> From 2020:
> > >>
> > >> Subject: [RFC][PATCH] * bits/confname.h: Define _CS_POSIX_V7_THREADS_CFLAGS, _CS_POSIX_V7_THREADS_LDFLAGS
> > >> Message-Id: <20201026233303.16034-1-ericonr@disroot.org>
> > >
> > > Thanks, the patch missed some bits but I think the rationale is ok to
> > > include. If Érico can update the patch I will apply it.
> >
> > Agreed, there is no reason not to add them.
> >
> > With libpthread merged into libc (as you noted) this is simpler too.
>
> OK. As long as glibc is on board with adding these now, can someone
> confirm what the numbers will be so I can go ahead and add them on the
> musl side?
Ok, to answer myself, it looks like they would be 1150 and 1151.
Rich