[cairo] Cairo support for Mac OS 10.4

[Andrea Canciani]

On Fri, Jan 20, 2017 at 10:20 PM, Bryce Harrington <bryce at osg.samsung.com>
wrote:

> On Wed, Jan 18, 2017 at 03:29:55PM +0100, Andrea Canciani wrote:
> > The wip/restore-10.4 was updated to actually work on 10.4 with the help
> of
> > G 3.
> > I also manged to set up a virtual machine running OSX 10.4, so I will now
> > be able to check for regressions when updating new patches for
> cairo-quartz.
> > Given that there seems to still be some interest in keeping 10.4 working
> > and that the code is already available, I would go for it and merge
> > https://cgit.freedesktop.org/~ranma42/cairo/commit/?h=wip/
> restore-10.4&id=5a8a9c97ed268004cbac510d39739ff56c0fb43c
> >
> > Bryce, does this sound like a reasonable plan?
>
> Yes, that sounds fine.  I looked up the past discussions for the
> rationale we'd had for upping the version requirement to 10.5:
>
>     https://lists.cairographics.org/archives/cairo/2014-
> November/025792.html
>     https://lists.cairographics.org/archives/cairo/2015-
> February/025949.html
>
> The main issue seems to have been testing on 10.4, which sounds like you
> now have resolved to your satisfaction.  We were also uncertain whether
> maintaining support for such an old version was worth the effort, but
> there's been ample interest voiced by users, so bringing it back does
> seem like it would provide tangible value.
>

> My only question is, I recall you mentioned there were other code
> cleanups that would be made possible by dropping the 10.4 requirement?
> If those cleanups were performed, then they may need to be revisited in
> light of this change.
>

The cleanups were ready (in the
https://cgit.freedesktop.org/~ranma42/cairo/log/?h=wip/remove-10.4 branch)
but they have not been merged into master
The changes are sufficient to fix the compile-time errors and to bring
cairo-quartz on 10.4 basically on par with cairo on modern (10.11) releases
(I found that the older version of quartz is slightly more prone to some
numerical approximation and instability issues, but the test suite does not
show any major regression).

In fact the patch has already been merged in Tigerbrew (a fork of Homebrew
aimed at supporting older OSX releases):
https://github.com/mistydemeo/tigerbrew/commit/071470409272b955b10e6041c94be6539e11b24f

> If there are no concerns, I will proceed with this before working on some
> > test failures that were uncovered by the record-replay-extend.c test
> > (Quartz does not handle correctly bounded recording surfaces).
> >
> > Andrea
>
> Sounds good, thanks for following up on this.
>
> Bryce
>

Awesome, tomorrow I will do the merge :)

Andrea
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cairographics.org/archives/cairo/attachments/20170120/5875adec/attachment.html>