[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Multiple build warnings in Fedora during mockbuild
From: |
Bruno Haible |
Subject: |
Re: Multiple build warnings in Fedora during mockbuild |
Date: |
Mon, 08 Apr 2024 17:26:13 +0200 |
Hello,
Manish Tiwari wrote:
> gettext_mockbuild_output.txt contains full mockbuild output details.
In there, I read:
Package "xz-5.6.1-1.fc41.x86_64" is already installed.
Update your system *IMMEDIATELY*. Haven't you heard about the most serious
attack on the internet ever, that is being discussed in all channels, for
a week already?
See https://packages.fedoraproject.org/pkgs/xz/xz/fedora-rawhide.html
> Getting multiple mockbuild warnings while building gettext on fedora, these
> warnings were occurring in previous versions of gettext also.
> repro steps: --> fedpkg prep
> --> fedpkg mockbuild
> I have attached two files below
> gettext_build_warnings.txt contains filtered build warnings from mockbuild
> output
Thanks for the report. But you could have saved me some time by eliminating
pointless warnings, before reporting. (You get paid by Red Hat, I don't :).)
* Wstack-protector warnings are pointless to me, the developer. The compiler
is telling us that it has a limitation. OK, then it has a limitation.
* The 3 Wformat warnings are pointless, because referencing an 'int' as an
'unsigned int' (or vice versa) is not dangerous. It may help the developer
to catch a mistake, but I have dealt with this warning already after the
gettext 0.22 release [1].
* The warning
careadlinkat.c:180:10: warning: function may return address of local variable
[-Wreturn-local-addr]
is preceded, in the code, by a pointer to
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93644
which is a GCC bug that is still not fixed.
* The Wdiscarded-qualifiers warnings are due to a newer version of libxml2
on your side. Fixed in git [2].
* The Wstringop-overflow warning is a GCC bug:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109995
that is still not fixed.
* The Walloc-size-larger-than warning is a new GCC bug. Since you are using a
GCC 14 prerelease, while I am using the GCC releases, would you please report
a GCC bug for it?
* The Wanalyzer-possible-null-argument warning is on my TODO list. Hasn't been
handled yet. But it is harmless, since xasprintf can only fail for results
larger than 2 GB.
* The Wanalyzer-infinite-loop warning is a new GCC bug. Since you are using a
GCC 14 prerelease, while I am using the GCC releases, would you please report
a GCC bug for it?
* The Wuse-after-free warnings are a GCC bug:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109990
that is still not fixed.
* The Wanalyzer-out-of-bounds warnings are a new GCC bug, since I generated the
file cldr-plural.c using GNU Bison 3.8.2. Again, would you please report
a GCC bug for it?
* The 2 Wsuggest-attribute=format warnings, I have already dealt with after the
gettext 0.22 release [1].
* The 3 Wswitch warnings are preceded, in the log, by this banner:
## ------------------- Gnulib tests ------------------- ##
## You can ignore compiler warnings in this directory. ##
## ---------------------------------------------------- ##
Therefore, nothing to do for them.
Bruno
[1]
https://git.savannah.gnu.org/gitweb/?p=gettext.git;a=commitdiff;h=635e16ad8120ad6e99fcc62f2ebde0cce8068c44
[2]
https://git.savannah.gnu.org/gitweb/?p=gettext.git;a=commitdiff;h=6456980949f3766b366d3a7afba05d7fefe27cd0