Hi Andy,
> > I was going to test the attached patch, but find that I cannot test it:
> > The commit
> > https://github.com/illumos/illumos-gate/commit/004345e48064ccd168d15f66eba2031c6090ccee
> > is from 2025-04-24.
> > - In OmniOS, https://downloads.omnios.org/media/, the latest 'bloody'
> > b
Hi Ludovic,
Ludovic Courtès writes:
>> I am only a relatively new user of guix, so I can't speak for which
>> approach would be preferable from their perspective. I CC'ed Ludovic
>> since he implemented the unprivileged guix daemon approach.
>
> I’m not sure how ‘modules/test-xfail’ is supposed
Thanks for jumping in, Collin.
> In the file 'modules/test-xfail' there is a few conditionals similar to
> this:
>
> gl_CONDITIONAL([OS_IS_NETBSD],
>[case "$host_os" in netbsd*) true ;; *) false ;; esac])
>
> Then in 'cat modules/sigprocmask-tests', for example, we have:
Bruno Haible via Gnulib discussion list writes:
> Finally some good use of the GCC analyzer. gcc reports this warning:
>
> -Wanalyzer-possible-null-argument
> ../../../gettext-tools/gnulib-lib/vc-mtime.c:374:7: warning: use of
> possibly-NULL 'dir2' where non-null expected [CWE-690]
I fi
Collin Funk wrote:
> > The cause is that xgetcwd() may return NULL, which I did not remember.
>
> Good catch. I assumed from the name it would always be successful or
> exit. But it only exits when OOM, not for other errors.
And I made this assumption because when I looked at xgetcwd.h, it has
no
Bruno Haible via Gnulib discussion list writes:
> Building a testdir with a GCC 15 snapshot, I see these warnings:
>
> ../../gllib/gc-gnulib.c:177:24: warning: 'calloc' sizes specified with
> 'sizeof' in the earlier argument and not in the later argument
> ../../gllib/gc-gnulib.c:566:24: warning
After running the following commands:
$ gnulib-tool --create-testdir --dir testdir1 --single-configure
`gnulib-tool --list | grep -E '^crypto/gc-?' | xargs`
$ cd testdir1 && ./configure --with-libgcrypt
CFLAGS='-Wcalloc-transposed-args'
make[4]: Entering directory
'/home/collin
Oops, the code uses one more array element than allocated.
Found with 'valgrind'.
Fixed like this:
2025-05-04 Bruno Haible
vc-mtime: Fix a buffer overflow.
* lib/vc-mtime.c (max_vc_mtime): Increase the size of argv by one
element.
diff --git a/lib/vc-mtime.c b/lib/vc-
Finally some good use of the GCC analyzer. gcc reports this warning:
-Wanalyzer-possible-null-argument
../../../gettext-tools/gnulib-lib/vc-mtime.c:374:7: warning: use of
possibly-NULL 'dir2' where non-null expected [CWE-690]
The cause is that xgetcwd() may return NULL, which I did not re
Hi Bruno and all,
keinflue writes:
>> Instead, how about
>> a) if you just ignore the test failures — since you know you are
>> running them in a very special environment, or
>> b) we can mark these 3 tests as expected test failures if you
>> provide an identifier to mark them with
Collin Funk writes:
> I think something like this would work:
>
> gl_CONDITIONAL([IS_OVERFLOW_GID],
>[test `cat /proc/sys/kernel/overflowgid` = `id -G | sed
> 's/ /\n/g' | awk '{ if (NR == 2) print }'`])
>
> Let me see if I can figure out how to set up a guix vm.
Unfortu
On Sun, 4 May 2025, Bruno Haible wrote:
> Hi,
>
> Andy Fiddaman wrote:
> > The attached patch adds support for using the POSIX getlocalename_l()
> > function
> > in illumos to avoid having to reach into undocumented/opaque structs.
>
> Thanks for notifying us about this. Yes, the purpose of ge
Hi,
Andy Fiddaman wrote:
> The attached patch adds support for using the POSIX getlocalename_l() function
> in illumos to avoid having to reach into undocumented/opaque structs.
Thanks for notifying us about this. Yes, the purpose of getlocalename_l() in
POSIX is to avoid peeking into undocumente
[sorry for the private email before. please reply to this one, CCing the list]
Hi Alan, Paul,
You contributed a lot of changes to the files lib/obstack.{h,in.h,c}
between 2007 and 2021.
We would like to have the ability to flow back these changes into
glibc. glibc is under LGPLv2+.
In this peri
14 matches
Mail list logo
