[cairo] Quartz fallback fix

[Andrea Canciani]

Robert O'Callahan reported a bug in the quartz fallback handling and
proposed a fix (see the bugreport
https://bugzilla.mozilla.org/show_bug.cgi?id=507939).
I think that both current code and the proposed fix are incorrect, as
forcing the backend to always fallback makes tests in the testsuite
fail (notably caps, which passes without forcing the fallback).
I attach a patch to force falling back (so testing can easily be done)
and I propose a patch that makes the test suite behave as expected (no
new failures are introduced by the fallback, instead (as expected)
some more tests pass because the output is computed by the fallback).
Please review (in particular I've been suggested to refer to someone
more knowledgeable about this code) as I think the other patch already
made into mozilla-central and it should be reverted asap if incorrect.
Thank you
Andrea
-------------- next part --------------
A non-text attachment was scrubbed...
Name: test-fallback.diff
Type: application/octet-stream
Size: 644 bytes
Desc: not available
Url : http://lists.cairographics.org/archives/cairo/attachments/20090822/458e5c31/attachment.obj 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fix-fallback.diff
Type: application/octet-stream
Size: 1848 bytes
Desc: not available
Url : http://lists.cairographics.org/archives/cairo/attachments/20090822/458e5c31/attachment-0001.obj