[cairo] regression in text rendering when not using Render xlib

[Behdad Esfahbod]

On Tue, 2007-03-06 at 11:04 -0500, Frederic Crozat wrote:
> Le mardi 06 mars 2007 à 10:28 -0500, Behdad Esfahbod a écrit :
> > On Tue, 2007-03-06 at 03:47 -0500, Frederic Crozat wrote:
> > > Le lundi 05 mars 2007 à 16:05 -0500, Behdad Esfahbod a écrit :
> > > > On Mon, 2007-03-05 at 08:53 -0500, Frederic Crozat wrote:
> > > > > 
> > > > > You need to apply first fix and second one in sequence (I've undo some
> > > > > of the change I did in my first patch, which aren't really needed). 
> > > > 
> > > > In that case, would you mind giving us a final combined patch, without
> > > > any formatting changes please?
> > > 
> > > I wanted git to give me a combined patch with changelog but I'm still a
> > > newbie for that ;)
> > > 
> > > Anyway, new patch attached.
> > 
> > Thanks.  BTW, cairo's configure tries to use -fno-strict-aliasing.  Did
> > you force it off or something to hit this bug?
> 
> Our default packages CFLAGS are :
> on x86 : -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions
> -fomit-frame-pointer -march=i586 -mtune=generic
> -fasynchronous-unwind-tables
> 
> and on x86-64 : -O2 -g -pipe -Wp,-D_FORTIFY_SOURCE=2 -fexceptions
> 
> and they overwrite default CFLAGS in an autoconf/automake setup.
> 
> Which is why we were the only one to see this problem.

I still don't see how that makes cairo not use -fno-strict-aliasing.

> So, I think we can say -fno-strict-aliasing can be removed from cairo
> with my patch :)

Well, it's less harmful now.  But theoretically we still need it,
because we cast different surface types (abstract and specific) all the
time.

-- 
behdad
http://behdad.org/

"Those who would give up Essential Liberty to purchase a little
 Temporary Safety, deserve neither Liberty nor Safety."
        -- Benjamin Franklin, 1759