Re: attribute warn_unused_result

[Eli Zaretskii]

> Date: Thu, 03 Feb 2011 13:40:21 -0800
> From: Paul Eggert <address@hidden>
> CC: Emacs-Devel devel <address@hidden>, Eli Zaretskii <address@hidden>
> 
> On 02/03/11 12:42, Chad Brown wrote:
> 
> > ..and it broke the W32 and nextstep builds (at least) again.
> 
> Sorry, I forgot to add the new file lib/ignore-value.h.
> That should be fixed now.
> 
> On 02/03/11 13:16, Eli Zaretskii wrote:
> 
> > And is butt-ugly, IMO.
> 
> I don't like ignore-value either.  I'd rather that we simply
> tell GCC never to generate those bogus warnings anywhere.
> However, as I understand it, that's not so easily done.

The -Wno-unused-result compiler options looks easy to me, and is IMO
cleaner than using ignore-value.h.  Heck, even the corresponding
pragma in the affected source files looks cleaner.

The only advantage of ignore-value.h is that it will work with
compilers other than GCC (if some of them also insist on this
warning).