summaryrefslogtreecommitdiffstats
path: root/config_host
diff options
context:
space:
mode:
authorLuboš Luňák <l.lunak@collabora.com>2020-01-26 21:41:01 +0100
committerLuboš Luňák <l.lunak@collabora.com>2020-01-27 16:35:57 +0100
commit3f362beeda762e2b122ca9c3978852ef162cc046 (patch)
treee5e2fb113527d833af047093bf257b0120aafc21 /config_host
parenttdf#124176: Use pragma once instead of include guards (diff)
downloadcore-3f362beeda762e2b122ca9c3978852ef162cc046.tar.gz
core-3f362beeda762e2b122ca9c3978852ef162cc046.zip
disable 32bpp bitmap usage for Skia again
The code is just not ready for 32bpp bitmaps that actually include alpha directly in the data instead of the stupid VCL way of using a separate bitmap for that. And it seems the performance difference is not that large, so revert and maybe somewhen later. Change-Id: Icb61663665c843fb426206967d0fe7667ed88477 Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87466 Tested-by: Jenkins Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
Diffstat (limited to 'config_host')
-rw-r--r--config_host/config_skia.h.in4
1 files changed, 2 insertions, 2 deletions
diff --git a/config_host/config_skia.h.in b/config_host/config_skia.h.in
index 1a8b64a1dfd6..e31d23ad443e 100644
--- a/config_host/config_skia.h.in
+++ b/config_host/config_skia.h.in
@@ -21,8 +21,8 @@ are the same.
// So ultimately the 24+8 split should be dumped (preferably in all of LO,
// not just the Skia-related code), but until all of LO works correctly
// with 32bpp disabling this will avoid such breakages.
-#define SKIA_USE_BITMAP32 1
-//#define SKIA_USE_BITMAP32 0
+//#define SKIA_USE_BITMAP32 1
+#define SKIA_USE_BITMAP32 0
/* TODO SKIA check all these */