Removal of -pedantic and -pedantic-errors from acinclude.m4.in

[Stephan Kulow]

On Monday 18 August 2003 09:40, Alex Zepeda wrote:
> On Sun, Aug 17, 2003 at 04:40:07PM +0200, Stephan Kulow wrote:
> 
> > About removing -pedantic we could argue, but it's only triggering warnings.
> > That's why -pedantic-errors is named -errors - and that's why we use it
> > in configure as it makes detecting missing prototypes reliable. If FreeBSD
> > has problems in stdlib.h with gcc 3.3, then perhaps gcc 3.3 should add that
> > to it's fixincludes script.
> 
> Yes, I've read the man page.  Yes, I agree -pedantic-errors should create 
> errors, and -pedantic should not.  But this is NOT THE CASE.  This is what 
> I've been saying over and over.
Dirk found out meanwhile that's this is a gcc 3.3 regression and it's most likely
to be fixed for 3.3.2 (btw: the 3.3 prerelease on SuSE 8.2 doesn't show the
problem either). It just doesn't help you now I guess ;(

> 
> It's not just stdlib.h, quite a few libstdc++ headers trigger these 
> 'warnings' as well.
I thought gcc supresses warnings from system headers. Not any more?

> 
> Again, look at the kde-freebsd lists.  I believe Lauri pointed to the big
> thread about this.  This is indeed a problem with gcc, but by using
> --disable-warnings I should be able to run configure without -pedantic.
> 
> That said, these prototypes are not missing on FreeBSD, but since autoconf 
> can't determine the difference between a missing <whatever> and some other 
> incidental error...
If you have an idea how to check for these in another way, I'm glad to hear.

Greetings, Stephan

-- 
There may be no I in team, but a M and an E.