Re: gl_MANYWARN_ALL_GCC() leads to many spurious warnings

[Samuel Bronson]

On Jul 6, 2012, at 5:41 PM, Paul Eggert wrote:

> On 07/06/2012 02:17 PM, Samuel Bronson wrote:
>
> > Why not simply use "-W -Wall -Wextra" and perhaps a few others, plus
> > a few overrides to turn off the warnings those enable that aren't
> > useful with Emacs?  That way, we'd only get warnings that had been
> > reasonably well tested.
>
> Something like that might well be useful, but it's a different option
> than what's intended here.  --enable-gcc-warnings is intended for
> developers, and it tries to take advantage of the latest and greatest
> GCC features.  If it enabled only flags that work well on all GCC
> versions, the static checking would be of considerably lower quality.

So, why does it use flags like -Wunreachable-code (which was so buggy  
that it eventually got yanked -- it's now silently ignored) and -Wsync- 
nand (which is totally irrelevant, and warns that this warning isn't  
allowed for Objective C code) and -Wunused-macros (which seems rather  
buggy in GCC 4.7.1 -- it doesn't seem to count appearing in an #ifdef  
as a use?).

> It might be useful to have another configure-time option that would
> enable just the warnings that work fairly well back to GCC version
> something-or-another, if someone wanted to maintain that.  It might be
> a maintenance hassle, though, as there are lots of old and buggy GCCs
> out there.

Might make sense for this version, though:
powerpc-apple-darwin9-gcc-4.2.1 (GCC) 4.2.1 (Apple Inc. build 5577)

(That seems to be the last version Apple released for use on PowerPC  
Macs.)  Old, yes.  Buggy, yes.  But it has its advantages...