[Mesa-dev] [PATCH mesa] meson: merge C and C++ compiler arguments check

[Eric Engestrom]

On Thursday, 2018-03-22 11:17:55 -0700, Dylan Baker wrote:
> Quoting Eric Engestrom (2018-03-22 05:08:55)
> > On Monday, 2018-03-12 10:16:33 -0700, Dylan Baker wrote:
> > > Quoting Emil Velikov (2018-03-12 09:09:50)
> > > > On 12 March 2018 at 15:01, Eric Engestrom <eric.engestrom at imgtec.com> wrote:
> > > > > Signed-off-by: Eric Engestrom <eric.engestrom at imgtec.com>
> > > > > ---
> > > > > Dylan, was there any reason to have -Werror=missing-prototypes and
> > > > > -Werror=implicit-function-declaration in C but not C++?
> > > > > Both sound to me like something we always want.
> > > > 
> > > > Seems to be copied from the autotools setup.
> > > > Reason being, both are not valid for C++.
> > > > Although since we probe for them, everything should be fine. The C++
> > > > test will bail out and the flags won't be set during the actual build.
> > > 
> > > Which is why I left them separate, since it avoids having to compile for
> > > arguments we know that C++ doesn't support.
> > 
> > Pushed now (cb2ddcefa5196fdfeff7), but to explain, my point was: we want
> > those warnings when possible, so we might as well test for them and use
> > them if/when support is added in the compilers.
> > 
> > The configure-time cost is very low, and the build-time cost is
> > non-existent :)
> > 
> > > It probably doesn't matter either way,
> > > 
> > > Reviewed-by: Dylan Baker <dylan at pnwbakers.com>
> 
> I didn't notice before, but this breaks compiling basically all of our C++ code
> with clang since it adds -Werror=missing-prototypes, and clang gets very angry
> at us.

Ah crap :(
Are you pushing a revert, or should I?

Question though, why is the code full of missing prototypes? Is that
a compiler mistake (false-positives), or is the code really missing
a bunch of includes?

> 
> Dylan