This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa] Enable warnings for libbanshee
- From: Daniel Berlin <dberlin at dberlin dot org>
- To: Andreas Jaeger <aj at suse dot de>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Sun, 29 Feb 2004 09:22:08 -0500
- Subject: Re: [tree-ssa] Enable warnings for libbanshee
- References: <m3ptbyutdd.fsf@gromit.moeb> <m3llmmut45.fsf@gromit.moeb>
On Feb 29, 2004, at 4:34 AM, Andreas Jaeger wrote:
Looking closer at the warnings, we should use IMO (As we do in GCC and
not as we do in libiberty):
ac_libbanshee_warn_cflags='-W -Wall -pedantic -Wwrite-strings
-Wstrict-prototypes -Wmissing-prototypes
With that change we get the appended list of warnings currently.
80% of these are from macros or are worthless.
Only about 10% of them are valid. And of those, all but about 1-2% of
them occur in optional functionality we don't use (and won't ever use).
Fixing all of them will require significant local changes to
libbanshee, making merges incredibly hard.
Banshee itself isn't compiled with these flags, so why should we add
them?
If you want, I can happily fix the integer signedness warnings and
pointer->int casts.
Do you really want prototypes for every single function?
Do we really care if it uses variadic macros?
I'm much more inclined to just require turning off libbanshee if the
host compiler doesn't support these things (There has been 1 person who
has tried to compile libbanshee with a compiler that isn't gcc, AFAIK)
--Dan