[cairo] 3rd release candidate for cairo 1.4.12 available

[Baz]

On Nov 22, 2007 2:27 AM, Baz <brian.ewins at gmail.com> wrote:
> On Nov 21, 2007 7:03 PM, Carl Worth <cworth at cworth.org> wrote:
> > After a couple of bugs were reported (missing cairo-malloc-private.h
> > and missing definition of INT32_MAX on win32) we now have a third
> > release-candidate snapshot for cairo 1.4.12:
> >
> > http://cairographics.org/snapshots/cairo-1.4.11-20071121-4df0d2.tar.gz
> >
> > And, as before, there's a corresponding 1.4 branch in the central git
> > repository, (which has been rebuilt again with the latest fix).
> >
> > Thanks again for all the reports so far.
>
> I'm afraid it's still bad :(
> 63e0bc28584c7d661a3a6f7e91630b39792b6de0
> This fails to build when you enable atsui. The cherry uses 'return
> _cairo_error(...)'  but the change that made that work
> (d90d4bb6b99e0a912650234e28d097ea76c1cecc) isn't in. Looks like the
> other quartz fixes after this have the same problem.
>
> I cherry-picked that patch, it fixed things up enough to go to the next problem.
>
> 63e0bc also contains this:
> -       return CAIRO_STATUS_NO_MEMORY;
> +       status = CAIRO_STATUS_NO_MEMORY);
> +       goto BAIL1;
>
> Looks like a partial fix has been applied for the _cairo_error
> problems, since this wasn't in the original commit. Reinserting the
> '_cairo_error(' fixes this, and I can build.
>
> Tests look much the same as 1.4.10 after that (I'll pic the REF_DIR
> patches and check that properly tomorrow)

Tested ok except I'd also revert this one:
aea3eb85b04fb61a05ed1bf4fca8401af4f29eef "implement clipping with
surface masks."

While the quartz surface has a bunch of visual glitches, that change
doesn't make things 100% correct, it just changes which set of
glitches you get; the with-fallback glitches or the without-fallback
glitches. On the 1.4 branch I'd leave the fallbacks in.

Cheers,
Baz