Posting to that thread a bit late (only just jumped on the list) - Ryan Schmidt, you said:
Note that cairo 1.4.0 has been released and that the cairo portfile will soon be updated -- just as soon as we can figure out how to enable the quartz backend without having to disable the xlib backend, which is unfortunately the current situation with 1.4.0 [1].
Theres a patch here: https://bugs.freedesktop.org/show_bug.cgi?id=10407 I'm fairly certain that some version of this fix will make it into the next developer snapshot, but if you want to use the patch in the portfile for the last full release, be my guest. One thing to watch out for is that if you use this and enable both quartz and atsui by default, then atsui will be used by the toy api in xlib as well - with ugly results. Toy font selection is much worse in atsui. but shouldn't affect anything using the real api, eg via Pango. Nothing should be relying on bling from that api anyway - its there for testing. Cheers, Baz
On Mar 30, 2007, at 05:07, Baz wrote:
Posting to that thread a bit late (only just jumped on the list) - Ryan Schmidt, you said:
Note that cairo 1.4.0 has been released and that the cairo portfile will soon be updated -- just as soon as we can figure out how to enable the quartz backend without having to disable the xlib backend, which is unfortunately the current situation with 1.4.0 [1].
Theres a patch here: https://bugs.freedesktop.org/show_bug.cgi?id=10407
I'm fairly certain that some version of this fix will make it into the next developer snapshot, but if you want to use the patch in the portfile for the last full release, be my guest.
One thing to watch out for is that if you use this and enable both quartz and atsui by default, then atsui will be used by the toy api in xlib as well - with ugly results. Toy font selection is much worse in atsui. but shouldn't affect anything using the real api, eg via Pango. Nothing should be relying on bling from that api anyway - its there for testing.
The cairo port was already updated to 1.4.2, because it did not seem to exhibit the problem. You're saying the problem still exists in (the development version) 1.4.3? What happens if you try "sudo port install cairo +quartz +atsui"? It compiles without complaining for me.
On 30/03/07, Ryan Schmidt <ryandesign@macports.org> wrote:
The cairo port was already updated to 1.4.2, because it did not seem to exhibit the problem. You're saying the problem still exists in (the development version) 1.4.3? What happens if you try "sudo port install cairo +quartz +atsui"? It compiles without complaining for me.
Whoops yes, my bad - behdad put in a fix so that now only 'make test' (or perf, etc) hits the bug.
Ryan Schmidt <ryandesign@macports.org> on Friday, March 30, 2007 at 1:31 PM -0800 wrote:
The cairo port was already updated to 1.4.2, because it did not seem to exhibit the problem. You're saying the problem still exists in (the development version) 1.4.3? What happens if you try "sudo port install cairo +quartz +atsui"? It compiles without complaining for me.
Someone has posted patches for gtk2 2.10.6. Is this applicable or are they necessary with the current gtks, 2.10.11? http://trac.macosforge.org/projects/macports/ticket/11138 Mark
Am 31.03.2007 um 05:41 schrieb Mark Duling:
Ryan Schmidt <ryandesign@macports.org> on Friday, March 30, 2007 at 1:31 PM -0800 wrote:
The cairo port was already updated to 1.4.2, because it did not seem to exhibit the problem. You're saying the problem still exists in (the development version) 1.4.3? What happens if you try "sudo port install cairo +quartz +atsui"? It compiles without complaining for me.
Someone has posted patches for gtk2 2.10.6. Is this applicable or are they necessary with the current gtks, 2.10.11?
Any updates on that? Yesterday I did a selfupdate of my ports installation and then tried sudo port install gtk2 +quartz but that still failed: gdkdrawable-quartz.c: In function 'gdk_quartz_ref_cairo_surface': gdkdrawable-quartz.c:64: error: incompatible type for argument 1 of 'cairo_quartz_surface_create' gdkdrawable-quartz.c:64: error: too many arguments to function 'cairo_quartz_surface_create' make[4]: *** [gdkdrawable-quartz.lo] Error 1 make[3]: *** [all-recursive] Error 1 make[2]: *** [all] Error 2 make[1]: *** [all-recursive] Error 1 make: *** [all] Error 2 Error: Status 1 encountered during processing.
Mark
_______________________________________________ macports-users mailing list macports-users@lists.macosforge.org http://lists.macosforge.org/mailman/listinfo/macports-users
Thanks, Lars
On Jun 5, 2007, at 08:15, Lars Sonchocky-Helldorf wrote:
Am 31.03.2007 um 05:41 schrieb Mark Duling:
Ryan Schmidt on Friday, March 30, 2007 at 1:31 PM -0800 wrote:
The cairo port was already updated to 1.4.2, because it did not seem to exhibit the problem. You're saying the problem still exists in (the development version) 1.4.3? What happens if you try "sudo port install cairo +quartz +atsui"? It compiles without complaining for me.
Someone has posted patches for gtk2 2.10.6. Is this applicable or are they necessary with the current gtks, 2.10.11?
Any updates on that? Yesterday I did a selfupdate of my ports installation and then tried
sudo port install gtk2 +quartz
but that still failed:
gdkdrawable-quartz.c: In function 'gdk_quartz_ref_cairo_surface': gdkdrawable-quartz.c:64: error: incompatible type for argument 1 of 'cairo_quartz_surface_create' gdkdrawable-quartz.c:64: error: too many arguments to function 'cairo_quartz_surface_create' make[4]: *** [gdkdrawable-quartz.lo] Error 1 make[3]: *** [all-recursive] Error 1 make[2]: *** [all] Error 2 make[1]: *** [all-recursive] Error 1 make: *** [all] Error 2
Error: Status 1 encountered during processing.
I updated cairo to 1.4.8. I understand some things about Quartz and ATSUI have changed again. I don't know if this improves or worsens what happens with gtk2. You ask if the posted patches for gtk2 2.10.6 are still applicable to 2.10.11. I don't know; I don't use gtk2. If you use gtk2, perhaps you could try to build gtk2 with and without the patches and report which works better.
On 11/06/07, Ryan Schmidt <ryandesign@macports.org> wrote:
I updated cairo to 1.4.8. I understand some things about Quartz and ATSUI have changed again. I don't know if this improves or worsens what happens with gtk2.
I did several of those changes... nothing changed in the public API, so it shouldn't have changed anything wrt the gtk2 patch situation. Mostly the fixes were perf/correctness fixes driven by mozilla bug reports.
You ask if the posted patches for gtk2 2.10.6 are still applicable to 2.10.11. I don't know; I don't use gtk2. If you use gtk2, perhaps you could try to build gtk2 with and without the patches and report which works better.
The 1.6 release of cairo is planned for some time in July and is meant to target the needs of the release schedules of firefox and gnome. Its also going to be the first release where quartz/atsui are publicly documented and not experimental (yay!). We get a fair amount of feedback from mac firefox folk, but not nearly enough from mac gnome users - if you file rendering bugs on bugzilla.gnome.org, please feel free to add a cc to me (brian.ewins@gmail.com). Cheers, Baz
participants (4)
-
Baz
-
Lars Sonchocky-Helldorf
-
Mark Duling
-
Ryan Schmidt