> Can you help add the related commits here? That will help us to understand
> the current issues and patch.
I think the commits are:
commit 0075c53724f78c78aa1692cc8e3bf1433eeb0b9f
Author: Nick Clifton <nickc@redhat.com>
Date: Thu Nov 24 12:30:22 2022 +0000
Impport libiberty commit: 885b6660c17f from gcc mainline. Fix gas's acinclude.m4 to stop a potwntial configure time warning message.
commit b4f26d541aa7224b70d363932e816e6e1a857633
Author: Tom Tromey <tom@tromey.com>
Date: Tue Mar 2 13:42:37 2021 -0700
Import GNU Readline 8.1
commit 9c9d63b15ad548f65f8bfd41cb7a4ef9af5ccb28
Author: Paul E. Murphy <murphyp@linux.ibm.com>
Date: Fri Jan 22 13:55:45 2021 -0600
gnulib: update to 776af40e0
This fixes PR27184, a failure to compile gdb due to
cdefs.h being out of sync with glibc on ppc64le targets
which are compiled with -mabi=ieeelongdouble and glibc
2.32.
commit 3f3328b816ee3486ac0621cfd6d808faa2c0a689
Author: Martin Liska <mliska@suse.cz>
Date: Thu Mar 18 15:16:54 2021 +0100
Use startswith more for strncmp function calls.
Thank you for the information, Florian. This really helps us.
> Does it have any specific gcc compile options(stricter c99 mode)? Or
> could you please explain more details?
It's possible to emulate the future behavior to some extent using
-Werror=implicit-function-declaration -Werror=implicit-int. In some
cases, autotools automatically remove -Werror=* options, I think, but I
don't know if this relevant here.
Sounds like a good idea, I will try the -Werror= options later.
The Fedora Change documentation has more details, and there are
instrumented compilers/build roots you can use for testing:
<https://fedoraproject.org/wiki/Changes/PortingToModernC>
<https://fedoraproject.org/wiki/Toolchain/PortingToModernC>
> While at it, also update the bundled copy of <sys/cdefs.h> in
> gnulib. This header file unfortunately shadows the glibc version,
> causing build failures on ppc64le if it is too old.
>
>
> Is this a specific Fedora related issue? Currently, it has a Fedora
> Only patch as below:
No, it shows up with all current GCC versions on ppc64le if they default
to the binary128 format for long double. Fedora was just very early in
switching, so it may appear as if it is a Fedora-specific problem, but
is really not.
Thanks for the explanation, Florian.
BTW: Could you please follow the current styles of gdb-10.2patch?
+diff --git gdb-10.2.orig/bfd/elf-bfd.h gdb-10.2/bfd/elf-bfd.h
+index eebdf9a..775d96c 100644
+--- gdb-10.2.orig/bfd/elf-bfd.h
++++ gdb-10.2/bfd/elf-bfd.h
Usually, the above styles are not recommended.
Thanks.
Lianbo
Thanks,
Florian