Chromium Code Reviews| Index: cc/layers/content_layer_client.h |
| diff --git a/cc/layers/content_layer_client.h b/cc/layers/content_layer_client.h |
| index 8271cd5e27f1711d2e2e7ba8ee38c2b39d82c7ea..14c276e850a749669986f11a0775f175d7805fcd 100644 |
| --- a/cc/layers/content_layer_client.h |
| +++ b/cc/layers/content_layer_client.h |
| @@ -25,12 +25,13 @@ class CC_EXPORT ContentLayerClient { |
| virtual void PaintContents(SkCanvas* canvas, |
| const gfx::Rect& clip, |
| + bool can_paint_lcd_text, |
| gfx::RectF* opaque, |
| GraphicsContextStatus gc_status) = 0; |
| - // Called by the content layer during the update phase. |
| - // If the client paints LCD text, it may want to invalidate the layer. |
| - virtual void DidChangeLayerCanUseLCDText() = 0; |
| + // If true the layer may need to invalidate itself when LCD text status |
| + // changes. |
| + virtual bool PaintsLCDText() const = 0; |
|
enne (OOO)
2014/06/06 16:59:32
Please don't change this. It should be up to the
alokp
2014/06/06 19:15:05
That is what this function is supposed to prevent.
alokp
2014/06/06 22:05:40
I realized that there is no test for this. I have
enne (OOO)
2014/06/06 22:20:59
Sure, but this also means that ui layers can no lo
alokp
2014/06/06 22:31:27
But if you actually paint LCD text and it is no lo
alokp
2014/06/10 17:18:59
piman: Do you see any issues with modifying this A
piman
2014/06/10 22:57:45
I can't tell for sure if that happens.
What is tru
alokp
2014/06/10 23:33:47
AFAICT in the current world ui::Layer is never inv
|
| // If true the layer may skip clearing the background before rasterizing, |
| // because it will cover any uncleared data with content. |