On Feb 9, 12:15pm, u...@stderr.spb.ru (Valery Ushakov) wrote: -- Subject: Re: CVS commit: src/include
| On Fri, Feb 09, 2018 at 02:44:05 +0000, Christos Zoulas wrote: | | > In article <20180208234417.ga23...@britannica.bec.de>, | > Joerg Sonnenberger <jo...@bec.de> wrote: | > >On Thu, Feb 08, 2018 at 10:56:22PM +0000, Christos Zoulas wrote: | > >> In article <20180207130259.ga21...@britannica.bec.de>, | > >> Joerg Sonnenberger <jo...@bec.de> wrote: | > >> >On Tue, Feb 06, 2018 at 03:21:21PM -0500, Christos Zoulas wrote: | > >> >> Module Name: src | > >> >> Committed By: christos | > >> >> Date: Tue Feb 6 20:21:21 UTC 2018 | > >> >> | > >> >> Modified Files: | > >> >> src/include: unistd.h | > >> >> | > >> >> Log Message: | > >> >> detect duplicate declaration of pthread_atfork() in pthread.h | > >> > | > >> >Is this for some new broken GCC warning? | > >> > | > >> | > >> This is what the compiler said, and we have prior art for this (macro | > >> protect to avoid dup declarations -- grep for _DECLARED). | > > | > >Normally only for typedefs, since pre-C11 (?) it was invalid to typedef | > >the same thing twice. That's not true for prototypes though. | > | > We have been doing this for functions for a while; check unistd.h | | But *why* have we been doing that? I added the redundand decls warning in bsd.sys.mk in 2001, but it is still commented out. Is that now part of -Wall? christos