> > Fix compilation errors with clang that masquerades as gcc 13.
>
> And here is a patch that fixes the clang warnings that appear to be caused
> by this version of __GNUC__.
And here is a patch that avoids a test failure. gcc >= 8 generates correct
floating point comparisons, gcc 4.2 and cl
I did:
> Fix compilation errors with clang that masquerades as gcc 13.
And here is a patch that fixes the clang warnings that appear to be caused
by this version of __GNUC__.
Previously I did not like the __GNUC_PREREQ (glibc) or _GL_GNUC_PREREQ (gnulib)
macro, because they seemed like unne
Bruno Haible writes:
> 1) The web site makes it hard to download the OS. See:
> - Go to https://www.openmandriva.org/
> - Click the "Download" button.
> - Now you are in a maze of twisty little hyperlinks that will not allow you
> to download anything:
The __GNUC__ and __GNUC_MINOR__ patch see
Hi Allison,
Based on the files named '3' and '4' in data-after.tar.gz, I believe the
attached change (not yet committed) has good chances of working. Can you
please give it a try?
Bruno
diff --git a/lib/readutmp.c b/lib/readutmp.c
index 10d79d1d81..3c4f97b146 100644
--- a/lib/readutmp.c
+++ b/li
hi,
Please see attached :)
Allison
On Mon, 26 Aug 2024 at 16:47, Bruno Haible wrote:
> Allison Karlitskaya wrote:
> > This impacts the output of the "who" command when someone is logged in
> via
> > Cockpit. The old version printed:
> >
> > adminweb console 2024-08-26 08:46 (:::1
> * build-aux/config.libpath: Accept a second argument. Use it on
> FreeBSD/powerpc64.
> (func_usage): Document it.
Oops, there was a mistake here. Fixed as follows:
2024-08-26 Bruno Haible
relocatable-prog: Fix config.libpath failure (regression 2024-08-24).
Sam James wrote:
> OpenMandriva Lx is what you want:
> *
> https://github.com/OpenMandrivaAssociation/llvm/blob/master/clang-default-newer-gnuc-version.patch
OK. Here's what I found regarding this distro:
1) The web site makes it hard to download the OS. See:
- Go to https://www.openmandriva.org
Hi Marc,
> > The problem with the glibc code is that they have specialized code for each
> > format:
> > - IEEE 754 single-precision,
> > - IEEE 754 double-precision,
> > - 'long double' with LDBL_MANT_DIG == 106 (a.k.a. "double double"),
> > - 'long double' with LDBL_MANT_DIG == 113 (a.k.
Hi Bruno,
Am Mo., 26. Aug. 2024 um 12:37 Uhr schrieb Bruno Haible :
> Hi Marc,
>
> > Gnulib
> > offers strtod replacements, but these refer to conversion routines of the
> > host systems and may lack the accuracy described in Clinger's well-known
> > paper [1] IIUC.
> >
> > Wouldn't it make sense
Allison Karlitskaya wrote:
> This impacts the output of the "who" command when someone is logged in via
> Cockpit. The old version printed:
>
> adminweb console 2024-08-26 08:46 (:::172.27.0.2)
>
> and the new version simply doesn't list the login.
The goal of the new readutmp impl
On 2024-08-26 01:51, Allison Karlitskaya wrote:
So I guess my wish here is open-ended, but something like "please allow for
a way to ensure that web logins appear in the output of who"
If we're simply trying to emulate the old readutmp it sounds like this
should be the default behavior.
On 2024-08-26 00:17, Sam James wrote:
It would happen on mandriva where they patch their Clang to identify as
newer GCC.
It can also happen in Fedora or Ubuntu if you configure with CC='clang
-fgnuc-version=14.2.1'. Not clear whether we need to support that sort
of thing.
Fedora recently added systemd-devel to the build dependencies for coreutils:
-
https://src.fedoraproject.org/rpms/coreutils/c/d3137f0cd04f6c716ef7c64ae35aa653e22aefea?branch=f41
This had the side-effect of the ./configure script defining
READUTMP_USE_SYSTEMD, enabling read_utmp_from_systemd() in
Hi Marc,
> Gnulib
> offers strtod replacements, but these refer to conversion routines of the
> host systems and may lack the accuracy described in Clinger's well-known
> paper [1] IIUC.
>
> Wouldn't it make sense to include an accurate conversion algorithm in
> Gnulib like [2] with the small imp
Gnulib contains modules like dtoastr that accurately convert floating point
numbers to strings. For the opposite direction of conversion, Gnulib
offers strtod replacements, but these refer to conversion routines of the
host systems and may lack the accuracy described in Clinger's well-known
paper
Bruno Haible writes:
> Sam James wrote:
>> It would happen on mandriva ...
>
> Mandriva is dead [1][2][3].
>
> Did you mean OpenMandriva Lx? Or Mageia? Or ROSA?
Sorry, bad old habits. OpenMandriva Lx is what you want:
*
https://github.com/OpenMandrivaAssociation/llvm/blob/master/clang-default-n
Sam James wrote:
> It would happen on mandriva ...
Mandriva is dead [1][2][3].
Did you mean OpenMandriva Lx? Or Mageia? Or ROSA?
Bruno
[1] https://en.wikipedia.org/wiki/Mandriva
[2] https://distrowatch.com/table.php?distribution=mandriva
[3] https://www.mandriva.com
Sam James wrote:
> It would happen on mandriva where they patch their Clang to identify as
> newer GCC.
That's an interesting info. Let me run a gnulib testdir on that platform...
Thanks!
Bruno
Paul Eggert writes:
> On 2024-08-24 09:05, Bruno Haible wrote:
>> So, when we have a conditional that tests for GCC >= 4.7 or 10 or 14,
>> we don't need to exclude clang, because it's already excluded.
>
> Weird. I can no longer reproduce the problem with clang. So I undid
> that change. If the p
19 matches
Mail list logo
