OSDN Git Service

android-x86/external-webkit.git
12 years agoFix SVG text rendering
Steve Block [Thu, 24 Nov 2011 14:02:29 +0000 (14:02 +0000)]
Fix SVG text rendering

Android's implementation of GraphicsContext::setCTM() is broken, because Skia's
Picture mode does not support setMatrix(). This broken code was added in
https://android-git.corp.google.com/g/#/c/112636 as part of the merge to WebKit
r80534.

This change modifies SVGInlineTextBox to avoid calling
GraphicsContext::setCTM() and makes clear that the method should not be used.

Bug: 5590123
Change-Id: I45a3c8c356b7f068bb943a45b9f10cbc041331e2

12 years agoForce layers above overscroll/fixed elements to composite
Chris Craik [Wed, 23 Nov 2011 00:06:08 +0000 (16:06 -0800)]
Force layers above overscroll/fixed elements to composite

bug:5480098

This fixes the case where they were drawn on the base layer, beneath overscroll
elements.

Change-Id: If9dda8c3a25cf38e777847b834ae02ec530a8000

12 years agoMerge "Use isHighEndGfx to double textures" into ics-mr1
John Reck [Mon, 21 Nov 2011 22:11:48 +0000 (14:11 -0800)]
Merge "Use isHighEndGfx to double textures" into ics-mr1

12 years agoUse isHighEndGfx to double textures
John Reck [Mon, 21 Nov 2011 17:17:33 +0000 (09:17 -0800)]
Use isHighEndGfx to double textures

 Bug: 5639612

Change-Id: I308f6fbb8208a0d2e83f2b6d285d4c471137a0e7

12 years agoMerge "Set the number of current textures depending on the layers' count, to avoid...
Nicolas Roard [Mon, 21 Nov 2011 18:07:22 +0000 (10:07 -0800)]
Merge "Set the number of current textures depending on the layers' count, to avoid going to single drawing surface if we can. Also fix a crash when logging layers." into ics-mr1

12 years agoSet the number of current textures depending on the layers' count,
Nicolas Roard [Thu, 17 Nov 2011 02:25:59 +0000 (18:25 -0800)]
Set the number of current textures depending on the layers' count,
to avoid going to single drawing surface if we can.
Also fix a crash when logging layers.

bug:5279231
Change-Id: I1c3f2ce4bcedac1c172e87c7ec3c6692d8e35e14

12 years agoMerge "Content and viewport rects set in one step." into ics-mr1
Bart Sears [Fri, 18 Nov 2011 19:46:11 +0000 (11:46 -0800)]
Merge "Content and viewport rects set in one step." into ics-mr1

12 years agoMerge "Only allow stacking contexts to skip compositing" into ics-mr1
Bart Sears [Fri, 18 Nov 2011 03:25:21 +0000 (19:25 -0800)]
Merge "Only allow stacking contexts to skip compositing" into ics-mr1

12 years agoOnly allow stacking contexts to skip compositing
Chris Craik [Fri, 18 Nov 2011 01:33:50 +0000 (17:33 -0800)]
Only allow stacking contexts to skip compositing

bug:5635942
Change-Id: Ibda67bb638d9fca5f32c85ee39c1ae1429986eda

12 years agoMerge "Force tiles to swap whenever drawing tree is ready" into ics-mr1
Chris Craik [Thu, 17 Nov 2011 20:50:39 +0000 (12:50 -0800)]
Merge "Force tiles to swap whenever drawing tree is ready" into ics-mr1

12 years agoMerge "Add logging to help debug an ANR" into ics-mr1
Teng-Hui Zhu [Thu, 17 Nov 2011 18:07:14 +0000 (10:07 -0800)]
Merge "Add logging to help debug an ANR" into ics-mr1

12 years agoForce tiles to swap whenever drawing tree is ready
Chris Craik [Thu, 17 Nov 2011 02:24:33 +0000 (18:24 -0800)]
Force tiles to swap whenever drawing tree is ready

bug:5632237

Sometimes the drawing tree wouldn't swap it's tiles after a scroll or a view
resize, even though the tiles were painted ready. Now, force all tiles to swap
every time the entire tree is ready.

Change-Id: Id204ed9082b7ee4d41fa712c37febc3dacf6984e

12 years agoMerge "synchronous layer updates, and animation deferral during paint" into ics-mr1
Chris Craik [Thu, 17 Nov 2011 02:21:35 +0000 (18:21 -0800)]
Merge "synchronous layer updates, and animation deferral during paint" into ics-mr1

12 years agosynchronous layer updates, and animation deferral during paint
Chris Craik [Tue, 15 Nov 2011 19:28:34 +0000 (11:28 -0800)]
synchronous layer updates, and animation deferral during paint

bug:5522081
bug:5239801
bug:5297563

Change-Id: I600f66999e093f720a8ea97ef3e15d3d1d297a8f

12 years agoAdd logging to help debug an ANR
Teng-Hui Zhu [Wed, 16 Nov 2011 23:03:40 +0000 (15:03 -0800)]
Add logging to help debug an ANR

bug:5627977
Change-Id: Ie68a4ae44599063fa93938370a5dcff7e63d7b9b

12 years agoFix clipping issue causing the layers to disappear when IME is up
Nicolas Roard [Wed, 16 Nov 2011 22:40:14 +0000 (14:40 -0800)]
Fix clipping issue causing the layers to disappear when IME is up

bug:5622291
Change-Id: I0134b784092b8e3910fcabf3e647034eb31096c7

12 years agoMerge "Promote -ve z-index fixed/scrollable elements to layers" into ics-mr1
Ben Murdoch [Wed, 16 Nov 2011 19:20:29 +0000 (11:20 -0800)]
Merge "Promote -ve z-index fixed/scrollable elements to layers" into ics-mr1

12 years agoMerge "Position WebTextView correctly on layers" into ics-mr1
Michael Kolb [Wed, 16 Nov 2011 18:34:38 +0000 (10:34 -0800)]
Merge "Position WebTextView correctly on layers" into ics-mr1

12 years agoContent and viewport rects set in one step.
George Mount [Tue, 15 Nov 2011 19:21:25 +0000 (11:21 -0800)]
Content and viewport rects set in one step.

 Bug 5342918
 The content rect of the WebView was being retrieved during
 the draw while the viewport rect was being set when the
 draw functor was setup. During animations, the content
 rect was changing between the time the draw functor was
 retrieved and it was executed. The content rect is now
 being set with the viewport rect.

 Framework change: I1b0978ee

Change-Id: I05d68dcc73ede1e919ddc1478f48b372b7b64042

12 years agoPromote -ve z-index fixed/scrollable elements to layers
Ben Murdoch [Wed, 16 Nov 2011 11:47:08 +0000 (11:47 +0000)]
Promote -ve z-index fixed/scrollable elements to layers

We normally do not necessarily promote fixed or scrollable
elements to compositing layers to try and reduce the number
of layers that get created. However in the case that such
an element appears in the negative z-order, we must do
the promotion to ensure that the layer is painted at the
correct z-index in the hierarchy (i.e not too high - sites
like to use a fixed image with z-index:-1 to implement
a fixed background -- without this patch we end up painting
the background over the main page content).

Bug: 5152567

Change-Id: I36bfe14660a33d45ad54552ae41b9d7157f8dabf

12 years agoMerge "Fix CacheManager to correctly write CacheFiles to disk." into ics-mr1
Ben Murdoch [Wed, 16 Nov 2011 10:26:41 +0000 (02:26 -0800)]
Merge "Fix CacheManager to correctly write CacheFiles to disk." into ics-mr1

12 years agoMerge "Use the transformed clipping rect for media layers (flash plugin)" into ics-mr1
Bart Sears [Wed, 16 Nov 2011 06:44:58 +0000 (22:44 -0800)]
Merge "Use the transformed clipping rect for media layers (flash plugin)" into ics-mr1

12 years agoUse the transformed clipping rect for media layers (flash plugin)
Nicolas Roard [Wed, 16 Nov 2011 03:11:51 +0000 (19:11 -0800)]
Use the transformed clipping rect for media layers (flash plugin)

bug:5621574
Change-Id: I4e2925edc314872a628b81daffe4ab1be7fe366f

12 years agoPosition WebTextView correctly on layers
Michael Kolb [Tue, 15 Nov 2011 23:39:50 +0000 (15:39 -0800)]
Position WebTextView correctly on layers

    Bug: 5550648

Change-Id: If3d59e0d197e9383b64f6a9e17d94e022044b5c3

12 years agoFix text selection handle positioning
John Reck [Tue, 15 Nov 2011 19:53:29 +0000 (11:53 -0800)]
Fix text selection handle positioning

 Bug: 5620362

Change-Id: Ic75ba855b5b86f8b0c3a21ee5dcbcc92b0dc982a

12 years agoFix CacheManager to correctly write CacheFiles to disk.
Ben Murdoch [Tue, 15 Nov 2011 15:26:37 +0000 (15:26 +0000)]
Fix CacheManager to correctly write CacheFiles to disk.

There are two problems with the CacheManager -
1. Need to null terminate the file path used for the cache file on
disk.
2. A race condition between starting the background task to write
the file and updating internal state variables.

These were causing a CTS test to fail.

Bug: 5619303
Change-Id: I61f06d50b7ef560ede2f1141fc51d92255d4efbd

12 years agoOnly mark tiles as dirty if the inval region actually intersects
Nicolas Roard [Tue, 15 Nov 2011 01:41:26 +0000 (17:41 -0800)]
Only mark tiles as dirty if the inval region actually intersects

bug:5553084
Change-Id: Ic5cf86ff8f53e86fca9ae98745d9556eadc7f8d2

12 years agoUse a single shared bitmap instead of one per BaseTile.
Nicolas Roard [Mon, 14 Nov 2011 17:48:20 +0000 (09:48 -0800)]
Use a single shared bitmap instead of one per BaseTile.

CL https://android-git.corp.google.com/g/#/c/148708/ uncovered an
existing memory leak on BaseTile. As we now keep the bitmap around,
this triggers an OOM.
We still need to track down the BaseTile leak, but at least with
this CL we won't OOM (and it's also using less memory...)

bug:5613031
Change-Id: Iac56d3b5c408b5e3e87c0f621e227bd3f4e37c3d

12 years agoMerge "Enable SPDY protocol." into ics-mr1
Selim Gurun [Fri, 11 Nov 2011 22:29:33 +0000 (14:29 -0800)]
Merge "Enable SPDY protocol." into ics-mr1

12 years agoRemove inset for WebTextView
Michael Kolb [Fri, 11 Nov 2011 18:28:34 +0000 (10:28 -0800)]
Remove inset for WebTextView

    Bug: 5383911
    By removing the inset we cover webkit's text field completely

Change-Id: Ib3edf21beb4e26b38b2195a9e336523336f26eb7

12 years agoMerge "Fix combobox metrics and asset scaling" into ics-mr1
Michael Kolb [Fri, 11 Nov 2011 17:13:46 +0000 (09:13 -0800)]
Merge "Fix combobox metrics and asset scaling" into ics-mr1

12 years agoMerge "Fallback mode when running out of textures for layers" into ics-mr1
Nicolas Roard [Fri, 11 Nov 2011 05:18:09 +0000 (21:18 -0800)]
Merge "Fallback mode when running out of textures for layers" into ics-mr1

12 years agoEnable SPDY protocol.
Selim Gurun [Fri, 11 Nov 2011 02:46:47 +0000 (18:46 -0800)]
Enable SPDY protocol.

Bug: 5524112
Reenable SPDY protocol.

Change-Id: I0144791020b64e20a12b7cd45ac2427100bef8a7
Signed-off-by: Selim Gurun <sgurun@google.com>
12 years agoFallback mode when running out of textures for layers
Nicolas Roard [Fri, 4 Nov 2011 20:56:14 +0000 (13:56 -0700)]
Fallback mode when running out of textures for layers

- Implements a degraded rendering mode when running out of textures
- Speed up rendering by reusing the same SkBitmap in RasterRenderer
- Some refactoring

bug:5279231

Change-Id: I52943a87bed56c49b01b47fea4fa1a5c49e09e93

12 years agoMerge "Use anti-aliasing in prefetch tiles" into ics-mr1
Chris Craik [Fri, 11 Nov 2011 01:54:30 +0000 (17:54 -0800)]
Merge "Use anti-aliasing in prefetch tiles" into ics-mr1

12 years agoUse anti-aliasing in prefetch tiles
Chris Craik [Thu, 10 Nov 2011 23:05:05 +0000 (15:05 -0800)]
Use anti-aliasing in prefetch tiles

bug:5534278
Change-Id: I5929752a9edcf3cd1c51360e0dc4d8df6ae6bb1e

12 years agoFix combobox metrics and asset scaling
Michael Kolb [Thu, 10 Nov 2011 19:31:26 +0000 (11:31 -0800)]
Fix combobox metrics and asset scaling

    Bug: 5521553

Change-Id: I630eda2ffd90d0c986f00c5a2a4b523e93537192

12 years agoMerge "Fix CTS regressions from merge to r74534." into ics-mr1
Ben Murdoch [Thu, 10 Nov 2011 19:54:29 +0000 (11:54 -0800)]
Merge "Fix CTS regressions from merge to r74534." into ics-mr1

12 years agoMerge "Support pausing drawing" into ics-mr1
John Reck [Thu, 10 Nov 2011 17:20:02 +0000 (09:20 -0800)]
Merge "Support pausing drawing" into ics-mr1

12 years agoMerge "DO NOT MERGE Offset scroll area by the scroll origin on Android" into ics-mr1
George Mount [Thu, 10 Nov 2011 16:32:10 +0000 (08:32 -0800)]
Merge "DO NOT MERGE Offset scroll area by the scroll origin on Android" into ics-mr1

12 years agoFix CTS regressions from merge to r74534.
Ben Murdoch [Thu, 10 Nov 2011 11:52:39 +0000 (11:52 +0000)]
Fix CTS regressions from merge to r74534.

WebKit change http://trac.webkit.org/changeset/73385 removed
overflow dimensions from the calculation of layer dimensions.
This means that the optimisation inside the nav cache to skip
building a frame cache when one of the layer dimensions is 0
is no longer valid (as the layer may contain only overflow
content).

Bug: 5572431

Change-Id: I3c973ad1ed57b07217d19f5d4fba7bfe522a2165

12 years agoSupport pausing drawing
John Reck [Thu, 10 Nov 2011 00:22:25 +0000 (16:22 -0800)]
Support pausing drawing

 Bug: 5594608

Change-Id: I09590f8c04904957b079f0458022b960599eb47a

12 years agoDO NOT MERGE Offset scroll area by the scroll origin on Android
George Mount [Mon, 24 Oct 2011 22:45:31 +0000 (15:45 -0700)]
DO NOT MERGE Offset scroll area by the scroll origin on Android

 Bug 5268793
 The scrolled area for WebView is 0,0 - maxX,maxY while
 ScrollView can have any origin. On RTL pages, the origin
 is negative when it can scroll. This scrolls the page
 to the right when it detects an RTL page.

 Cherry-picked from master CL: I6a91b97f

 Framework CL: I2537b84a

Change-Id: I6a91b97f90f5ba4c323c56dda9d125108d9502c4

12 years agoMerge "Fix font size handling for harfbuzz hack" into ics-mr1
Russell Brenner [Wed, 9 Nov 2011 23:23:16 +0000 (15:23 -0800)]
Merge "Fix font size handling for harfbuzz hack" into ics-mr1

12 years agoMerge "Add support for Bengali and Tamil gpos/gsub tables" into ics-mr1
Russell Brenner [Wed, 9 Nov 2011 18:11:22 +0000 (10:11 -0800)]
Merge "Add support for Bengali and Tamil gpos/gsub tables" into ics-mr1

12 years agoAdd support for Bengali and Tamil gpos/gsub tables
Russell Brenner [Tue, 8 Nov 2011 00:10:32 +0000 (16:10 -0800)]
Add support for Bengali and Tamil gpos/gsub tables

HB_Script_Bengali and HB_Script_Tamil are now recognized in
FontAndroid.cpp and map to Lohit-Bengali.ttf and Lohit-Tamil.ttf,
allowing Harfbuzz to properly load the gpos/gsub tables for those
languages.

Bug: 5578901
Change-Id: I20ed542dd5a616535222dcb5c808f3716a062e80

12 years agoMerge "Revert crash on bad scale - do not merge" into ics-mr1
Bart Sears [Wed, 9 Nov 2011 02:30:04 +0000 (18:30 -0800)]
Merge "Revert crash on bad scale - do not merge" into ics-mr1

12 years agoRevert crash on bad scale - do not merge
Bart Sears [Wed, 9 Nov 2011 01:11:44 +0000 (17:11 -0800)]
Revert crash on bad scale - do not merge

bug:5367327

Cherry pick of: dc193f393a845e88f55ecf0055386fc4e4204537

Change-Id: Iea6e4a564fc003d8e2ad8e1d6d9d9b64d016e43d

12 years agoMerge "Add necessary null pointer check to avoid crash on youtube video" into ics-mr1
Teng-Hui Zhu [Wed, 9 Nov 2011 00:20:56 +0000 (16:20 -0800)]
Merge "Add necessary null pointer check to avoid crash on youtube video" into ics-mr1

12 years agoAdd necessary null pointer check to avoid crash on youtube video
Teng-Hui Zhu [Tue, 8 Nov 2011 21:34:17 +0000 (13:34 -0800)]
Add necessary null pointer check to avoid crash on youtube video

bug:5562347
Change-Id: I9b53a58137d70955bde5334c7c8b581343a8c2ce

12 years agoFix font size handling for harfbuzz hack
Russell Brenner [Thu, 3 Nov 2011 23:22:00 +0000 (16:22 -0700)]
Fix font size handling for harfbuzz hack

Complex languages (Arabic, Thai, etc.) currently need to override the
base font (e.g. Roboto)  with an explicitly loaded copy of the right
fallback font (e.g. DroidNaskh). To keep from running out of file
descriptors, a caching mechanism was put in place, but the cached
data contained the font size, so the first size seen on the page was
applied to all subsequent renderings of that font, regardless of the
size requested. Size is now also considered in the cache.

Bug: 5509443
Change-Id: I616954980122e89da14ce86efe99d5e4343a40c2

12 years agoMerge "Fix extras clipping" into ics-mr1
John Reck [Tue, 8 Nov 2011 01:29:07 +0000 (17:29 -0800)]
Merge "Fix extras clipping" into ics-mr1

12 years agoFix extras clipping
John Reck [Mon, 7 Nov 2011 19:21:27 +0000 (11:21 -0800)]
Fix extras clipping

 Bug: 5556494

Change-Id: I82e8282271dfc623afa9b1ef0d0061ca9c3df5f4

12 years agoMerge "Change the default webView behavior to minimize the memory consumption" into...
Teng-Hui Zhu [Mon, 7 Nov 2011 23:53:51 +0000 (15:53 -0800)]
Merge "Change the default webView behavior to minimize the memory consumption" into ics-mr1

12 years agoRemove the global button lock
John Reck [Mon, 7 Nov 2011 21:53:18 +0000 (13:53 -0800)]
Remove the global button lock

 Bug: 5558699
 Bug: 5572238

Change-Id: Ic199ffda5bc4aa09d39ee54221c6fcce6a91ca05

12 years agoChange the default webView behavior to minimize the memory consumption
Teng-Hui Zhu [Mon, 7 Nov 2011 20:47:43 +0000 (12:47 -0800)]
Change the default webView behavior to minimize the memory consumption

For now, we just limit the expansion on x and y direction.
In the future, we need to dynamically allocate and deallocate the memory.

browser change is : c/148242

bug:5522175
Change-Id: I8562c39c55913ab89848afe8a36e8a2784c12727

12 years agoMerge "Doesn't lock the entire drawing..." into ics-mr1
Nicolas Roard [Mon, 7 Nov 2011 18:21:50 +0000 (10:21 -0800)]
Merge "Doesn't lock the entire drawing..." into ics-mr1

12 years agoReconcile with ics-mr0-release
The Android Open Source Project [Mon, 7 Nov 2011 16:22:50 +0000 (08:22 -0800)]
Reconcile with ics-mr0-release

Change-Id: If02ce708d74c5d627e4e6f83cfc5ee003fb42b58

12 years agoDoesn't lock the entire drawing...
Chris Craik [Tue, 25 Oct 2011 01:39:29 +0000 (18:39 -0700)]
Doesn't lock the entire drawing...

This caused us to potentially waiting on the completion of a tile
painting. On sites where skia struggle with, this had a large impact on
scrolling performances.

This is only part one of the solution -- we need to also get rid of the
global button lock.

We also lower the priority of the texture gneeration thread.

bug:5558699

Change-Id: I251354d6cabb2fd7ebfd665c30eff7fe90c3d316

12 years agomerge in ics-mr0-release history after reset to ics-mr0
The Android Automerger [Fri, 4 Nov 2011 22:39:42 +0000 (15:39 -0700)]
merge in ics-mr0-release history after reset to ics-mr0

12 years agoam 55b5da34: (-s ours) Merge "Don\'t force composited layers to be stacking contexts...
Bart Sears [Fri, 4 Nov 2011 11:59:22 +0000 (11:59 +0000)]
am 55b5da34: (-s ours) Merge "Don\'t force composited layers to be stacking contexts - DO NOT MERGE" into ics-mr0

* commit '55b5da341a376006b2e0ab9cc6dfba5a76c65c7c':
  Don't force composited layers to be stacking contexts - DO NOT MERGE

12 years agoam 209db4d4: (-s ours) Fix plugin rendering bug when video surface is not ready....
Derek Sollenberger [Fri, 4 Nov 2011 11:59:21 +0000 (11:59 +0000)]
am 209db4d4: (-s ours) Fix plugin rendering bug when video surface is not ready. do not merge

* commit '209db4d462bfff0c9f13896f72611a82c2c6639c':
  Fix plugin rendering bug when video surface is not ready. do not merge

12 years agoMerge "Don't force composited layers to be stacking contexts - DO NOT MERGE" into...
Bart Sears [Fri, 4 Nov 2011 03:57:38 +0000 (20:57 -0700)]
Merge "Don't force composited layers to be stacking contexts - DO NOT MERGE" into ics-mr0

12 years agoDon't force composited layers to be stacking contexts - DO NOT MERGE
Chris Craik [Fri, 21 Oct 2011 22:28:20 +0000 (15:28 -0700)]
Don't force composited layers to be stacking contexts - DO NOT MERGE

Cherry-pick of c0fb399b4644c319c15e0a6a7e0f6c29917d9099

bug:5264240

Was causing crashes, and layer ordering problems.

Change-Id: Ieecff6d64405f13b459ea013da2a309b00077d47

12 years agoFix plugin rendering bug when video surface is not ready. do not merge
Derek Sollenberger [Fri, 28 Oct 2011 17:11:02 +0000 (13:11 -0400)]
Fix plugin rendering bug when video surface is not ready. do not merge

cherry-pick of bf6900fe796a697400203ab777001716f3d83a0a

The browser incorrectly jumps out of the compositing step for a
plugin early if it detects a video surface is not ready. The
correct behavior is to skip rendering the video and proceed to
rendering the other plugin contents.

bug: 5525272
Change-Id: I12acb477890303d03c99ffa13729fd7e7eefc800

12 years agoFix crash in Autofill.
Ben Murdoch [Thu, 3 Nov 2011 12:30:30 +0000 (12:30 +0000)]
Fix crash in Autofill.

Fix a crash in autofill that can be caused by toggling the
setting on and off.

Bug: 5452927
Change-Id: I80ba0b823fd16cb4e2df3359be95f3d063a9b091

12 years agoReport a load error when network loads blocked. Do not merge
Ben Murdoch [Mon, 24 Oct 2011 11:42:19 +0000 (12:42 +0100)]
Report a load error when network loads blocked. Do not merge

Cherry pick from master.

When network loads are blocked, we should fail those
loads rather than leaving them dangling. In particular,
this fixes an issue where WebCore will wait to do the
first layout while CSS loads are pending. If network
loads are blocked, we need to tell WebCore they are never
going to complete.

To implement this, we make use of the Chromium stack's
load flags and move the code to a more suitable location.

Bug: 5416543
Change-Id: I4bd9987611364b29d3ccaa50437d418a5eb9a5cd

12 years agoReconcile with ics-mr0-release
The Android Open Source Project [Fri, 28 Oct 2011 20:05:22 +0000 (13:05 -0700)]
Reconcile with ics-mr0-release

Change-Id: I50da917e834f4ba4e0ba3e43e5af0ce523a706a4

12 years agoMerge "Fix plugin rendering bug when video surface is not ready." into ics-mr1
Derek Sollenberger [Fri, 28 Oct 2011 19:39:49 +0000 (12:39 -0700)]
Merge "Fix plugin rendering bug when video surface is not ready." into ics-mr1

12 years agoMerge "Don't discardQueue if the upload type didn't change" into ics-mr1
John Reck [Fri, 28 Oct 2011 19:16:34 +0000 (12:16 -0700)]
Merge "Don't discardQueue if the upload type didn't change" into ics-mr1

12 years agoFix plugin rendering bug when video surface is not ready.
Derek Sollenberger [Fri, 28 Oct 2011 17:11:02 +0000 (13:11 -0400)]
Fix plugin rendering bug when video surface is not ready.

The browser incorrectly jumps out of the compositing step for a
plugin early if it detects a video surface is not ready. The
correct behavior is to skip rendering the video and proceed to
rendering the other plugin contents.

bug: 5525272
Change-Id: Id838f6cef4913e817f58bf2b7e17d871f9f0250e

12 years agoDon't discardQueue if the upload type didn't change
John Reck [Fri, 28 Oct 2011 00:17:20 +0000 (17:17 -0700)]
Don't discardQueue if the upload type didn't change

 Bug: 5529012

Change-Id: I0b57c4cf9d295524f675f2cb7504d5a9c0363871

12 years agoMerge "Keep text handles in sync with native touch targets" into ics-mr1
John Reck [Thu, 27 Oct 2011 20:40:23 +0000 (13:40 -0700)]
Merge "Keep text handles in sync with native touch targets" into ics-mr1

12 years agoKeep text handles in sync with native touch targets
John Reck [Thu, 27 Oct 2011 16:07:40 +0000 (09:07 -0700)]
Keep text handles in sync with native touch targets

 Bug: 5522153
 Also add some slop to the handle targets, cleaned
 up the JNI and improved debugging.

Change-Id: I3b2c3c8b1543d4a1aa599ae2938fa78fe95e01bf

12 years agoAdd missing save/restore calls to GraphicsContext::fillRect
Ben Murdoch [Wed, 26 Oct 2011 14:14:51 +0000 (15:14 +0100)]
Add missing save/restore calls to GraphicsContext::fillRect

Android is missing some calls to GraphicsContext::save/restore
inside our GraphicsContext::fillRect implementations. The calls
were added to the Skia GraphicsContext in http://trac.webkit.org/changeset/63102

Bug: 5361502
Change-Id: Ifcbc116d997473e3a184561102dc5cb74f5aa5ec

12 years agoMerge remote branch 'goog/ics-mr0' into ics-mr1
Ed Heyl [Wed, 26 Oct 2011 12:47:43 +0000 (12:47 +0000)]
Merge remote branch 'goog/ics-mr0' into ics-mr1

12 years agoMerge "DO NOT MERGE Adjust text selection position" into ics-mr1
John Reck [Wed, 26 Oct 2011 02:14:04 +0000 (19:14 -0700)]
Merge "DO NOT MERGE Adjust text selection position" into ics-mr1

12 years agoMerge "DO NOT MERGE Fix SW extras for layers" into ics-mr1
John Reck [Wed, 26 Oct 2011 02:13:24 +0000 (19:13 -0700)]
Merge "DO NOT MERGE Fix SW extras for layers" into ics-mr1

12 years agoDO NOT MERGE Adjust text selection position
John Reck [Mon, 24 Oct 2011 23:49:45 +0000 (16:49 -0700)]
DO NOT MERGE Adjust text selection position

 Bug: 5507800

Change-Id: Icfcfdb9ff019fac7e3e40df679723594fab15dc6

12 years agoDO NOT MERGE Fix SW extras for layers
John Reck [Mon, 24 Oct 2011 22:10:04 +0000 (15:10 -0700)]
DO NOT MERGE Fix SW extras for layers

 Bug: 5507239
 Also remove all the old, unused layer setExtras stuff

Change-Id: Id1670010536712b7a9f50e74821c9fb7f708e851

12 years agoMerge "Don't draw prefetch page if not necessary" into ics-mr1
Chris Craik [Tue, 25 Oct 2011 22:27:26 +0000 (15:27 -0700)]
Merge "Don't draw prefetch page if not necessary" into ics-mr1

12 years agoDon't draw prefetch page if not necessary
Chris Craik [Fri, 21 Oct 2011 02:10:41 +0000 (19:10 -0700)]
Don't draw prefetch page if not necessary

bug:5494086

Also only try to draw visible tiles.

Change-Id: Ic850afab94309d9d4d031600a40dc889298fef24

12 years agoTake scale into account when greedily prefetching an entire layer
Chris Craik [Tue, 25 Oct 2011 01:39:29 +0000 (18:39 -0700)]
Take scale into account when greedily prefetching an entire layer

bug:5509149

Scale wasn't being taken into account when estimating layer tile consumption, so
zooming in on layer content could attempt to fetch and draw hundreds of layers
per tile, causing lots of missing content.

Change-Id: Ib26e203827ad93be12d29561b9d8094cb712f014

12 years agoSwapLayersTextures on all layer changes
Chris Craik [Tue, 25 Oct 2011 00:55:31 +0000 (17:55 -0700)]
SwapLayersTextures on all layer changes

bug:5507262

This restores behavior, and allows a single setBaseLayer to set content (such as
is used in browser saved pages).

Change-Id: Ifeda068ce5c22275553d0e1ec99ee74647806a6c

12 years agoMark layers as ready to swap if invisible
Chris Craik [Mon, 24 Oct 2011 22:40:09 +0000 (15:40 -0700)]
Mark layers as ready to swap if invisible

bug:5505364

This avoids the problem of waiting infinitely during initialization (when a
layer doesn't have content) for the layer to be ready.

Change-Id: I364eff66616ea304a68976d5ad06ce5fdb1a6d4b

12 years agoRemove unnecessary log
Nicolas Roard [Sun, 23 Oct 2011 23:26:12 +0000 (16:26 -0700)]
Remove unnecessary log

CL https://android-git.corp.google.com/g/#/c/144158/ introduced a debug
log that we don't need, spamming logcat

Change-Id: Iea8a1784153dd9f6f53e67f7cb62504440344c7c

12 years agoDon't force composited layers to be stacking contexts - DO NOT MERGE
Chris Craik [Fri, 21 Oct 2011 22:28:20 +0000 (15:28 -0700)]
Don't force composited layers to be stacking contexts - DO NOT MERGE

bug:5264240

Was causing crashes, and layer ordering problems.

Cherry-pick of b8c237486bcf299d8a04bdae8d8e35fd37866172

Change-Id: I1f76516d4e313101ea9c90759ee63d4cef2d6ced

12 years agomerge in ics-mr0-release history after reset to ics-mr0
The Android Automerger [Tue, 25 Oct 2011 03:28:04 +0000 (20:28 -0700)]
merge in ics-mr0-release history after reset to ics-mr0

12 years agoMerge "Take scale into account when greedily prefetching an entire layer" into ics-mr0
Chris Craik [Tue, 25 Oct 2011 01:50:07 +0000 (18:50 -0700)]
Merge "Take scale into account when greedily prefetching an entire layer" into ics-mr0

12 years agoTake scale into account when greedily prefetching an entire layer
Chris Craik [Tue, 25 Oct 2011 01:39:29 +0000 (18:39 -0700)]
Take scale into account when greedily prefetching an entire layer

bug:5509149

Scale wasn't being taken into account when estimating layer tile consumption, so
zooming in on layer content could attempt to fetch and draw hundreds of layers
per tile, causing lots of missing content.

Change-Id: Ib26e203827ad93be12d29561b9d8094cb712f014

12 years agoSwapLayersTextures on all layer changes
Chris Craik [Tue, 25 Oct 2011 00:55:31 +0000 (17:55 -0700)]
SwapLayersTextures on all layer changes

bug:5507262

This restores behavior, and allows a single setBaseLayer to set content (such as
is used in browser saved pages).

Change-Id: Ifeda068ce5c22275553d0e1ec99ee74647806a6c

12 years agoMerge "Mark layers as ready to swap if invisible" into ics-mr0
Chris Craik [Mon, 24 Oct 2011 22:49:43 +0000 (15:49 -0700)]
Merge "Mark layers as ready to swap if invisible" into ics-mr0

12 years agoMark layers as ready to swap if invisible
Chris Craik [Mon, 24 Oct 2011 22:40:09 +0000 (15:40 -0700)]
Mark layers as ready to swap if invisible

bug:5505364

This avoids the problem of waiting infinitely during initialization (when a
layer doesn't have content) for the layer to be ready.

Change-Id: I364eff66616ea304a68976d5ad06ce5fdb1a6d4b

12 years agomerge in ics-mr0-release history after reset to ics-mr0
The Android Automerger [Mon, 24 Oct 2011 19:55:54 +0000 (12:55 -0700)]
merge in ics-mr0-release history after reset to ics-mr0

12 years agoMerge "Remove unnecessary log" into ics-mr0
Bart Sears [Mon, 24 Oct 2011 18:11:45 +0000 (11:11 -0700)]
Merge "Remove unnecessary log" into ics-mr0

12 years agoReconcile with ics-mr0-release
The Android Open Source Project [Mon, 24 Oct 2011 13:32:14 +0000 (06:32 -0700)]
Reconcile with ics-mr0-release

Change-Id: If0751cf1dcf857e8e01b65bf347690292b4d0645

12 years agoRemove unnecessary log
Nicolas Roard [Sun, 23 Oct 2011 23:26:12 +0000 (16:26 -0700)]
Remove unnecessary log

CL https://android-git.corp.google.com/g/#/c/144158/ introduced a debug
log that we don't need, spamming logcat

Change-Id: Iea8a1784153dd9f6f53e67f7cb62504440344c7c

12 years agomerge in ics-mr0-release history after reset to ics-mr0
The Android Automerger [Fri, 21 Oct 2011 22:55:04 +0000 (15:55 -0700)]
merge in ics-mr0-release history after reset to ics-mr0

12 years agoam 5dee9ed2: Merge "Cherry pick 97075" into ics-mr0
John Reck [Fri, 21 Oct 2011 20:59:22 +0000 (13:59 -0700)]
am 5dee9ed2: Merge "Cherry pick 97075" into ics-mr0

* commit '5dee9ed228c914254062b62ae459884e3060a2ae':
  Cherry pick 97075

12 years agoam 5236e6a2: Merge "Cherry pick 94857" into ics-mr0
John Reck [Fri, 21 Oct 2011 20:59:21 +0000 (13:59 -0700)]
am 5236e6a2: Merge "Cherry pick 94857" into ics-mr0

* commit '5236e6a2f70997cf9dec9e93261867dd79b21738':
  Cherry pick 94857