This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
RE: [PATCH] Work around in-tree gmp configure problems
- From: Bernd Edlinger <bernd dot edlinger at hotmail dot de>
- To: Jan-Benedict Glaw <jbglaw at lug-owl dot de>, Richard Biener <richard dot guenther at gmail dot com>
- Cc: "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>, Jakub Jelinek <jakub at redhat dot com>
- Date: Mon, 24 Nov 2014 14:58:15 +0100
- Subject: RE: [PATCH] Work around in-tree gmp configure problems
- Authentication-results: sourceware.org; auth=none
- References: <DUB118-W4008C5DFFAB255FA19320CE4740 at phx dot gbl>,<20141122153438 dot GQ8445 at lug-owl dot de>,<CAFiYyc3YSE8LQvyr6VCFW45kkSu-tKkGQ3V--=vM6n=Ljydm8Q at mail dot gmail dot com>,<20141124131119 dot GC8445 at lug-owl dot de>
Hi Richard,
On Mon, 24 Nov 2014 14:11:19 +0100, Jan-Benedict Glaw wrote:
>
> Hi Richard,
>
> On Mon, 2014-11-24 12:53:11 +0100, Richard Biener <richard.guenther@gmail.com> wrote:
>> On Sat, Nov 22, 2014 at 4:34 PM, Jan-Benedict Glaw <jbglaw@lug-owl.de> wrote:
>>> On Sat, 2014-11-22 14:33:29 +0100, Bernd Edlinger <bernd.edlinger@hotmail.de> wrote:
>>>> since r217627 we use an updated AutoMake "missing" script. However
>>>> that revealed a hidden bug in gmp-4.3.2's (up to gmp-6.0.0a)
>>>> configure script. That is: an in-tree gmp/configure fails now if
>>>> flex is missing. The gmp configure uses our missing flex script,
>>>> and previously that emitted an error message and created a dummy
>>>> lex.yy.c, The new version of that script does no longer create any
>>>> lex.yy.c.
>>> [...]
>>>> 2014-11-22 Bernd Edlinger <bernd.edlinger@hotmail.de>
>>>>
>>>> * Makefile.def (module=gmp): Work around in-tree gmp configure
>>>> bug with missing flex.
>>>> * Makefile.in: Regenerated.
>>>
>>> Bernd and I discussed that patch; I'm all for it! Looks like that's
>>> the least-ugly thing we'd possibly come up with.
>>
>> I suppose it is fixed upstream now?
>
> *cough* No. Not at all.
>
> I subscribed to gmp-devel and sent a longish email with all the
> details, just to get it rejected: gmp-devel seems to be used for
> internal discussion between core developers. Re-sent it to
> gmp-discuss, but no response there at all. So up to now, there's no
> upstream change on the GMP side at all.
>
> In the mean time, I received other emails about specific problematic
> cases; the outcome there is that they're asking the GCC guys to not
> recommend the old 4.3.2 release, but to use some modern version and
> build it with --disable-assembly. Of course, that doesn't fix the
> observed configure problem. (However, using a modern version sounds
> like a fair request.)
>
Yes.
We disable the asm-parts by passing host=target="none-${host_vendor}-${host_os}".
Hope that still works.
Maybe I should note, that gmp-6.0.0a turned out to be incomatible to
mpfr-2.4.2 and mpc-0.8.1: gmp-6.0.0a does no longer define __gmp_const,
so these need to be updated at the same time.
Bernd.