GitLab

> [ui/surface] Separate image processing logic from presentation logic.
> 
> The D3D image processing code from AcceleratedPresenter moves to its own class, "ui/surface/accelerated_surface_transformer_win.h". This split allows the image processing code to be tested independently of the present-scheduling system, and so we add unit tests doing exactly that. Utility functions (loading d3d, creating temp surfaces) doing things commonly required by test, transform, and present code -- these functions are moved to a third location, "ui/surface/d3d9_utils_win.h"
> 
> The new unit tests -- which live in the ui_unittests binary -- make extensive use of pseudorandom image content, which I think is kind of neat. The tests use D3D HAL devices; I tried to use REF, but it didn't work, as StretchRect+LINEAR is not supported by refrast. So as a result we may have a GPU vendor dependency in these results.
> 
> BUG=161537
> TEST=new ui_unittests
> 
> Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=174028
> 
> Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=174338
> 
> Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=174943
> 
> Review URL: https://codereview.chromium.org/11464017

TBR=nick@chromium.org
Review URL: https://codereview.chromium.org/11753031

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175161 0039d316-1c4b-4281-b951-d872f2087c98

Follow up post-review comments on change
https://src.chromium.org/viewvc/chrome?view=rev&revision=174967

See https://codereview.chromium.org/11748015/ for more information.

NOTRY=true

Review URL: https://chromiumcodereview.appspot.com/11748032

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175159 0039d316-1c4b-4281-b951-d872f2087c98

This removes the extra commands from the command buffer. I've left the
Jump struct and handler in place as it's used to advance the buffer
if there isn't enough room at the end to place the next command.

TESTING=Ran the gpu_unittests and everything worked correctly.


Review URL: https://chromiumcodereview.appspot.com/11613021

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175158 0039d316-1c4b-4281-b951-d872f2087c98

R=mmenke@chromium.org
BUG=None


Review URL: https://chromiumcodereview.appspot.com/11746030

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175157 0039d316-1c4b-4281-b951-d872f2087c98

Remove the "Java" and rename to ExecuteScript*.  This better matches
conventions elsewhere in the codebase.

Introduce ExecuteScriptInFrame* variants for the less common case
where script execution in a subframe is desired.

Make it possible to pass the ExecuteScript* family of functions either
a WebContents pointer or a RenderViewHost pointer to further simplify
callsites.

R=jam@chromium.org
BUG=none


Review URL: https://chromiumcodereview.appspot.com/11753009

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175156 0039d316-1c4b-4281-b951-d872f2087c98

R=asargent@chromium.org
BUG=156311

Review URL: https://codereview.chromium.org/11744034

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175155 0039d316-1c4b-4281-b951-d872f2087c98

BUG=152731


Review URL: https://chromiumcodereview.appspot.com/11747016

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175154 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175153 0039d316-1c4b-4281-b951-d872f2087c98

The D3D image processing code from AcceleratedPresenter moves to its own class, "ui/surface/accelerated_surface_transformer_win.h". This split allows the image processing code to be tested independently of the present-scheduling system, and so we add unit tests doing exactly that. Utility functions (loading d3d, creating temp surfaces) doing things commonly required by test, transform, and present code -- these functions are moved to a third location, "ui/surface/d3d9_utils_win.h"

The new unit tests -- which live in the ui_unittests binary -- make extensive use of pseudorandom image content, which I think is kind of neat. The tests use D3D HAL devices; I tried to use REF, but it didn't work, as StretchRect+LINEAR is not supported by refrast. So as a result we may have a GPU vendor dependency in these results.

BUG=161537
TEST=new ui_unittests

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=174028

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=174338

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=174943

Review URL: https://codereview.chromium.org/11464017

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175152 0039d316-1c4b-4281-b951-d872f2087c98

This pulls in the following Native Client changes:

r10504: (shcherbina) Validator_ragel: remove some dead code.
r10505: (shcherbina) Validator_ragel: move .def files to separate dir.
r10506: (shcherbina) Validator_ragel: move docs to separate dir.
r10507: (shcherbina) Validator_ragel: get rid of generated .._consts.h files.
r10508: (shcherbina) Validator_ragel: simplify libraries and dependencies structure.
r10509: (jvoung) Re-enable archived llc.pexe test for x86-32 and arm, disable on x86-64.
r10510: (shcherbina) Validator_ragel: document missing operand types and sizes in .def format.
r10511: (jvoung) Only create ARM untrusted env for validator testing when needed.
r10512: (jvoung) Update PNaCl rev to 10509 in TOOLS_REVISION.
r10513: (shcherbina) Validator_ragel: get rid of dfa_validate_x86_xx libraries in scons build.
r10514: (shcherbina) Validator_ragel: fix minor irregularities in .def files.
r10515: (shcherbina) Build rdfa validator and ncval_new on arm as well.
r10516: (shcherbina) New ncval: Don't print "Validating <filename>..." message.
r10517: (shcherbina) New ncval: move segment size and offset checks into x86-specific function.
r10518: (shcherbina) Validator_ragel: rudimentary new gen_dfa and test.
r10519: (shcherbina) Validator_ragel: don't run gen_dfa_test under qemu.
r10520: (shcherbina) Validator_ragel: minor rename.
r10521: (halyavin) Debug stub: fix test_interrupt test.
r10522: (halyavin) Debug stub: Remove busy-waiting in debug stub loop on Linux.
r10523: (petarj) [MIPS] Do not use pnacl-irt-link for MIPS target.
r10527: (jfb) See below comments for rationale.
r10529: (mcgrathr) Use assembly rather than _control87 function to get x87 CW on Windows
r10530: (mseaborn) Provide more informative errors if we fail to load the IRT inside Chromium
r10531: (jvoung) Revert r10512: "Update PNaCl rev to 10509 in TOOLS_REVISION."
r10532: (mseaborn) Clean up declarations of DoFnstcw(), DoFxsave() and DoFxrstor()
r10533: (jvoung) Partly change IRT load-skipping to be explicit.

BUG=none
TEST=nacl_integration

Review URL: https://codereview.chromium.org/11747023

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175151 0039d316-1c4b-4281-b951-d872f2087c98

BUG=156929
TEST=none

Original CL: https://chromiumcodereview.appspot.com/11744014


Review URL: https://chromiumcodereview.appspot.com/11744030

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175150 0039d316-1c4b-4281-b951-d872f2087c98

This is another clean-up patch for https://codereview.chromium.org/11348217/#msg10.
These globals have been complicating things, so it should be cleaner to have them
explicitly passed in when needed.

BUG=157459

TBR=kkania

Review URL: https://chromiumcodereview.appspot.com/11740020

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175149 0039d316-1c4b-4281-b951-d872f2087c98

BUG=None
TEST=None
R=bartfab@chromium.org
TBR=nkostylev@chromium.org

Review URL: https://chromiumcodereview.appspot.com/11753026

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175148 0039d316-1c4b-4281-b951-d872f2087c98

BUG=None
NOTRY=true


Review URL: https://chromiumcodereview.appspot.com/11744031

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175146 0039d316-1c4b-4281-b951-d872f2087c98

TBR=kaiwang@chromium.org
BUG=166977


Review URL: https://chromiumcodereview.appspot.com/11730005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175145 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175141 0039d316-1c4b-4281-b951-d872f2087c98

control: https://codereview.chromium.org/11743030/
Review URL: https://codereview.chromium.org/11747018

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175140 0039d316-1c4b-4281-b951-d872f2087c98

On Android JavascriptAppModalDialog is implemented with a Java class and its
native counterpart.
The Java class holds a pointer to the native instance. When UI events (e.g.
button click) are processed, the Java side calls a native method (e.g.
JavascriptAppModalDialog::DidAcceptAppModalDialog()). When this native method
completes the instance deletes itself.
This is only correct in case it is guaranteed that no further native method
call is performed by the Java side since the pointer was freed.
The problem is that this can happen in some rare circumstances. For instance
the user could manage to click on two buttons before the dialog is closed which
might happen if the first click event is not processed immediately or takes a
long time to be processed.

This CL fixes the crash by invalidating the native pointer on the Java side
when it is deleted so that the Java side can perform a native method call only
if the native instance pointer is still valid.

BUG=167585

Review URL: https://codereview.chromium.org/11728004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175139 0039d316-1c4b-4281-b951-d872f2087c98

The Android NetworkChangeNotifier is initialized with connection_type =
CONNECTION_UNKNOWN.
This means that the result returned by NetworkChangeNotifier::IsOffline() is
inaccurate until the first network change happens.
This can happen for the whole initialization path if the user starts Chrome
after he explicitly disabled WiFi and 3G.

This CL updates NetworkChangeNotifier.java to support multiple native
observers. This can happen in case NetworkChangeNotifierDelegateAndroid is
instantiated multiple times (e.g. when multiple factories are instantiated).

Still on the Java side, NetworkChangeNotifier.java was also changed to fetch
the actual connection type as soon as the auto-detector (interacting with the
Android platform) is enabled.

On the native side, NetworkChangeNotifierDelegateAndroid now fetches at
construction time the current connection type from the Java side singleton
through a direct JNI function call (possible since the delegate is constructed
on the JNI thread).
NetworkChangeNotifierAndroid::GetCurrentConnectionType() is now a simple
wrapper around the delegate's GetCurrentConnectionType() method (which is
thread-safe).

BUG=166883

Review URL: https://codereview.chromium.org/11628008

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175138 0039d316-1c4b-4281-b951-d872f2087c98

In README.chromium, BSD license should be written as "BSD N-Clause License".
Also, update Android WebView's license checker to accept BSD 2-Clause.

TBR=torne@chromium.org,bcwhite@chromium.org
NOTRY=true


Review URL: https://chromiumcodereview.appspot.com/11742039

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175137 0039d316-1c4b-4281-b951-d872f2087c98

TBR=torne@chromium.org
NOTRY=true

Review URL: https://chromiumcodereview.appspot.com/11722029

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175136 0039d316-1c4b-4281-b951-d872f2087c98

Introduce the required code to get SW rendering working once RenderView is able
to provide the required picture piles from the compositor.

BUG=167913

Review URL: https://codereview.chromium.org/11727005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175135 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175134 0039d316-1c4b-4281-b951-d872f2087c98

BUG=


Review URL: https://chromiumcodereview.appspot.com/11722028

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175133 0039d316-1c4b-4281-b951-d872f2087c98

This limit is defined in the notification product requirement document,
section "Inbox Format for Multiple Messages". The limit only affects
the number of items displayed (to avoid notifications growing too tall).
The back end still handles any number of items. The maximum number is
declared in the public message_center_constants.h file so future code
can use it, for example to output a warning or return an error if
callers to the chrome.experimental.notification API ask for a
notification with more than the maximum number of displayable items.

BUG=161101


Review URL: https://chromiumcodereview.appspot.com/11638035

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175132 0039d316-1c4b-4281-b951-d872f2087c98

(This is a reland of r173539, adding early bails to some implementations of OnBeforeWidgetInit so that we don't clobber an already set Widget::InitParams::native_widget.)

There are actual cases where we have to show dialogs with no parent or context,
when there are no browser windows. All of the cases that I can find go through
the CreateWindow() interface.

CreateWindow() now explicitly sets InitParams::top_level to true. The
ViewsDelegate now can do arbitrary manipulations on the InitParams so that when
top_level is true and we have no context, we either create a native_widget or
set the context based on desktop type.

BUG=165759
TBR=ben@chromium.org (who did the original review)

First Review URL: https://chromiumcodereview.appspot.com/11578014

Review URL: https://chromiumcodereview.appspot.com/11758002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175131 0039d316-1c4b-4281-b951-d872f2087c98

In theory the #include could be replaced with a forward declaration.  However, forward declaring string16 is more bother than it's worth, since at least on some platforms it's a typedef rather than a proper class.

BUG=none


Review URL: https://chromiumcodereview.appspot.com/11761026

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175130 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175125 0039d316-1c4b-4281-b951-d872f2087c98

BUG=168084

Review URL: https://chromiumcodereview.appspot.com/11647050

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175123 0039d316-1c4b-4281-b951-d872f2087c98

Notification mock-ups show ellipses at the end of text when there is not
enough room for that text. Previously title text in multi-item
notifications had a trailing fade instead of trailing ellipses. This
changes that to ellipses.

BUG=161101


Review URL: https://chromiumcodereview.appspot.com/11636043

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175122 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175121 0039d316-1c4b-4281-b951-d872f2087c98

TBR=pkasting@chromium.org
BUG=165765

Review URL: https://codereview.chromium.org/11747021

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175118 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175117 0039d316-1c4b-4281-b951-d872f2087c98

http://trac.webkit.org/log/?rev=138767&stop_rev=138712&verbose=on

TBR=macpherson@chromium.org

BUG=


Review URL: https://chromiumcodereview.appspot.com/11740039

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175114 0039d316-1c4b-4281-b951-d872f2087c98

BUG=168252


Review URL: https://chromiumcodereview.appspot.com/11753017

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175113 0039d316-1c4b-4281-b951-d872f2087c98

Vista Tests (1) is timing out a lot.

TBR=sail@chromium.org
BUG=140882

Review URL: https://codereview.chromium.org/11741031

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175112 0039d316-1c4b-4281-b951-d872f2087c98

Test was flaky due to event timing differences between single and
multi-process WebKit ports. The test was made more resilient in
http://wkrev.com/138402 which has now rolled, so re-enabling.

R=dgrogan@chromium.org,alecflett@chromium.org
BUG=163694


Review URL: https://chromiumcodereview.appspot.com/11746026

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175111 0039d316-1c4b-4281-b951-d872f2087c98

NOTRY=true
BUG=167798,168084
TEST=Install https://chrome.google.com/webstore/detail/slinky-elegant/bmanlajnpdncmhfkiccmbgeocgbncfln?hl=en
Look at new tab button in active and inactive windows. Also check the new tab
button with the default theme. Look at startup perf graphs, should improve by
~1.8ms.


Review URL: https://chromiumcodereview.appspot.com/11748026

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175108 0039d316-1c4b-4281-b951-d872f2087c98

There's a race condition between the audio shutdown process and calls to OnMoreData() (in FillBufferQueue()).  During shutdown, or after an error occurs within AudioOutputController, OnMoreData() can return 0 to indicate "no data."  In this case, FillBufferQueue() should not attempt to call simple_buffer_queue_->Enqueue() with a "zero bytes" argument, which is an invalid argument value.

Also changed a DLOG(FATAL) to DLOG(ERROR), per Chromium style guidelines, since there is code that explicitly handles the event after the log statement.

BUG=168260
TEST=build/android/run_tests.py -s media_unittests -f *PlayStateChangeClose*
TBR=tommi@chromium.org


Review URL: https://chromiumcodereview.appspot.com/11744026

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175107 0039d316-1c4b-4281-b951-d872f2087c98

BUG=153048

Review URL: https://chromiumcodereview.appspot.com/11770002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@175106 0039d316-1c4b-4281-b951-d872f2087c98