diff -r 000000000000 -r 6474c204b198 gfx/cairo/missing-cairo-clip-init.diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/gfx/cairo/missing-cairo-clip-init.diff	Wed Dec 31 06:09:35 2014 +0100
@@ -0,0 +1,21 @@
+diff --git a/gfx/cairo/cairo/src/cairo-gstate.c b/gfx/cairo/cairo/src/cairo-gstate.c
+--- a/gfx/cairo/cairo/src/cairo-gstate.c
++++ b/gfx/cairo/cairo/src/cairo-gstate.c
+@@ -1841,16 +1841,17 @@ _cairo_gstate_show_text_glyphs (cairo_gs
+ 							transformed_glyphs,
+ 							&num_glyphs,
+ 							transformed_clusters);
+ 
+     if (status || num_glyphs == 0)
+ 	goto CLEANUP_GLYPHS;
+ 
+     _cairo_gstate_copy_transformed_source (gstate, &source_pattern.base);
++    _cairo_clip_init(&clip);
+ 
+     /* For really huge font sizes, we can just do path;fill instead of
+      * show_glyphs, as show_glyphs would put excess pressure on the cache,
+      * not all components below us correctly handle huge font sizes, and
+      * path filling can be cheaper since parts of glyphs are likely to be
+      * clipped out.  256 seems like a good limit.  But alas, seems like cairo's
+      * rasterizer is something like ten times slower than freetype's for huge
+      * sizes.  So, no win just yet when we're using cairo's rasterizer.