summaryrefslogtreecommitdiffstats
path: root/player/core.h
diff options
context:
space:
mode:
authorwm4 <wm4@nowhere>2015-07-06 21:55:37 +0200
committerwm4 <wm4@nowhere>2015-07-06 21:55:37 +0200
commit385febe27632075160c903a22502fcb78b160ae4 (patch)
tree8b9019f75492b42c22000558a9507294095eb1e5 /player/core.h
parentdcd167ca37fcb14719f97ca637ec5f3efc3952d8 (diff)
downloadmpv-385febe27632075160c903a22502fcb78b160ae4.tar.bz2
mpv-385febe27632075160c903a22502fcb78b160ae4.tar.xz
sub: protect ASS_Renderer state
Each subtitle track gets its own decoder instance (sd_ass). But they use a shared ASS_Renderer. This is done mainly because of fontconfig. Initializing fontconfig is very slow when using it with memory fonts, so there's a practical need to cache this memory font state, which is done by not creating separate ASS_Renderers. This is very dirty and very evil, but we probably can't get rid of it any time soon. The shared ASS_Renderer was not properly synchronized. While the program logic guarantees that only one sd_ass instance is visible at a time, there are other interactions that require synchronization. In particular, I suspect concurrent execution of mp_ass_configure_fonts() and sd_ass.get_bitmaps cause issues in a newer libass development branch. So here's a shitty hack that hopefully fixes things, hopefully only until libass becomes less dependent on fontconfig.
Diffstat (limited to 'player/core.h')
-rw-r--r--player/core.h1
1 files changed, 1 insertions, 0 deletions
diff --git a/player/core.h b/player/core.h
index 907c997b96..7b3e47d8d2 100644
--- a/player/core.h
+++ b/player/core.h
@@ -317,6 +317,7 @@ typedef struct MPContext {
* loaded across ordered chapters, instead of reloading and rescanning
* them on each transition. (Both of these objects contain this state.)
*/
+ pthread_mutex_t ass_lock;
struct ass_renderer *ass_renderer;
struct ass_library *ass_library;
struct mp_log *ass_log;