GitLab

Add flackr to ash/wm/overview/OWNERS

BUG=None
NOTRY=true

Review URL: https://codereview.chromium.org/145973008

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247795 0039d316-1c4b-4281-b951-d872f2087c98

This eliminates two descriptor leaks -- the downloaded nmf file, and
the nexe itself.  One socket (on Linux) still leak.  Because Unixes
always use the lowest available number and the leaked descriptor is
the highest numbered one while the F@H nexe is running, this implies
that this might be a PPAPI proxy related leak, since proxy hookup is
the last thing that happens before the application itself runs.

See /home/bsy/fd-leak-log for time-lapsed view of what happens when a
F@H NaCl module finishes and another starts up.

TEST= manually running F@H and checking /proc/<renderer>/fd/*
BUG= 335186
R=bbudge@chromium.org, dmichael@chromium.org

Review URL: https://codereview.chromium.org/147083014

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247794 0039d316-1c4b-4281-b951-d872f2087c98

Currently tests initialize GL by calling into methods that should be
internal to the gl bindings code. Instead, everyone should go through
GLSurface::InitializeOneOff.

Also GLSurface::InitializeOneOff early outs if it was already called,
leading to a pattern of initializing GL all over the place just in case
and not having a clear idea of where it should be set up. Instead, DCHECK
that it is not called more than once, and move calls to this method to
be during process startup for unit test suites instead of mid-test.

This adds two test variants of InitializeOneOff for tests to call, that
set up OSMesa or Mock GL bindings, via GLSurface::InitializeOneOff.

R=piman, sievers
BUG=270918

Review URL: https://codereview.chromium.org/135213003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247793 0039d316-1c4b-4281-b951-d872f2087c98

This is part of a long-running background task to remove the remaining
uses of the unchecked IPC_ENUM_TRAITS() macro.

BUG=246708

Review URL: https://codereview.chromium.org/149193002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247791 0039d316-1c4b-4281-b951-d872f2087c98

Right now the default mode is to make requests against the server unless
the experiment group says not to, and that may be contributing to some
unexpected load we're seeing. So this CL changes it to not do that.

It also adds some histograms to help measure the success rate and
sources of requests to the server.

BUG=335379
R=finnur@chromium.org

Review URL: https://codereview.chromium.org/149353002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247790 0039d316-1c4b-4281-b951-d872f2087c98

Chrome for Android does not use chrome://help so we should not include
the strings for it.

BUG=326023
R=newt@chromium.org

Review URL: https://codereview.chromium.org/142573005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247789 0039d316-1c4b-4281-b951-d872f2087c98

This moves ExtensionSystem to extensions/browser.
It also moves setting the channel for Features to ChromeExtensionsBrowserClient.
It also removes the duplicate ExtensionSystem::GetForBrowserContext.

BUG=337707
TBR=sky@chromium.org

Review URL: https://codereview.chromium.org/147923005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247788 0039d316-1c4b-4281-b951-d872f2087c98

BUG=323015
TBR=tonyg
NOTRY=True

Review URL: https://codereview.chromium.org/149773002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247787 0039d316-1c4b-4281-b951-d872f2087c98

Mainly replacing bool parameters with enum values

BUG=none
R=mkwst@chromium.org

Review URL: https://codereview.chromium.org/149073007

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247785 0039d316-1c4b-4281-b951-d872f2087c98

BUG=297132

Patch set 1 is the original CL:

https://codereview.chromium.org/137543002/

which broke tests on the memory bots. Path set 2 contains the fix for the test failure.

Review URL: https://codereview.chromium.org/148213010

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247784 0039d316-1c4b-4281-b951-d872f2087c98

This refactors all the windows code out of external_protocol_dialog.cc
into the cross-platform ShellIntegration interface, and makes the
external_protocol_dialog.cc implementation use the existing
external_protocol_dialog_delegate.cc interface.

BUG=337660
R=sky@chromium.org

Review URL: https://codereview.chromium.org/132903018

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247783 0039d316-1c4b-4281-b951-d872f2087c98

BUG=chromedriver:690

Review URL: https://codereview.chromium.org/147083010

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247782 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
NOTRY=True
TBR=bengr1

Review URL: https://codereview.chromium.org/133793006

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247781 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=none
R=yoz@chromium.org

Review URL: https://codereview.chromium.org/149883002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247778 0039d316-1c4b-4281-b951-d872f2087c98

Telemetry had two classes that were tightly linked: TemporaryHTTPServer and
MemoryCacheHTTPServer. Before this patch. TemporaryHTTPServer's job was to
spawn a child process and then launch in that child process the
MemoryCacheHTTPServer, scrape its serving port, and then establish port
forwarding between the browser and that port.

This refactors this slightly:
- TemporaryServer takes on the job of staritng a subprocess inside of which
  a server will be launched, plus the logic for matchmaking to that child
  process' selected serving port.
- The portions specific to what the child process actually does on that port
  are abstracted behind TemporaryServerBackend class.
- MemoryCacheHTTPServer is rephrased in terms of a TemporaryServer and
  TemporaryServerBackend.

R=tonyg

Review URL: https://codereview.chromium.org/148583014

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247777 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247776 0039d316-1c4b-4281-b951-d872f2087c98

http://build.chromium.org/f/chromium/perf/dashboard/ui/changelog_blink.html?url=/trunk&range=165996:166035&mode=html
TBR=
BUG=

Review URL: https://codereview.chromium.org/147593009

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247775 0039d316-1c4b-4281-b951-d872f2087c98

This patch changes all metrics in RenderingStats to be lists of lists
(one per timeline range). Right now, we only need this feature for frame_timestamps,
but it is cleaner to handle all metrics the same way and it preserves more
information.
The discrepancy calculation was modified to handle list-of-list inputs
correctly.

R=nduca@chromium.org,dominikg@chromium.org
BUG=338796

Review URL: https://codereview.chromium.org/149093003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247773 0039d316-1c4b-4281-b951-d872f2087c98

R=zork@chromium.org
BUG=249853

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=246992

Review URL: https://codereview.chromium.org/141433011

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247772 0039d316-1c4b-4281-b951-d872f2087c98

This CL creates the PasswordManager component and moves a leaf file needed as
part of unit test support into it.

BUG=334616,334637
TBR=zea

Review URL: https://codereview.chromium.org/148383003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247771 0039d316-1c4b-4281-b951-d872f2087c98

This action does a sanity check of the JavaScript files to make sure that the types are used consistently.

The jscompile build target is off by default. Use run_jscompile=1 to enable.

BUG=
R=jamiewalch@chromium.org

Review URL: https://codereview.chromium.org/149043003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247770 0039d316-1c4b-4281-b951-d872f2087c98

Any struct of size > 4 bytes should be passed by const ref. 
Passing by ref for these structs is faster than passing by value,
especially when invoking function has multiple parameters and some
other scenarios mentioned in the bug.

BUG=159273

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=247426

Review URL: https://codereview.chromium.org/145313006

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247769 0039d316-1c4b-4281-b951-d872f2087c98

As part of issue 330908, I've scrutinized JSONWriter on all angles. Despite failing to optimize it much; here are some micro optimizations and style cleanups I came up with, might as well put them in (although their impact appears to be minimal from local profiling).

BUG=330908

Review URL: https://codereview.chromium.org/139903013

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247765 0039d316-1c4b-4281-b951-d872f2087c98

That way, event converters can be organized in whichever way is most useful
for the platform. For evdev, organize by device path. This will let us find
out which which devices have already been opened and is needed for hotplug.

Review URL: https://codereview.chromium.org/133233008

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247764 0039d316-1c4b-4281-b951-d872f2087c98

BUG=
R=sergeyu@chromium.org

Review URL: https://codereview.chromium.org/139283010

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247763 0039d316-1c4b-4281-b951-d872f2087c98

It's no longer possible to revoke individual tokens, which means that signing
out of the web-app will cause all of user's hosts to drop off-line. This CL
"fixes" this by not revoking tokens at all. Note that users can still hit the
OAuth token limit, in which case hosts will start dropping off-line one at a
time; this will be addressed by migrating hosts to use robot accounts.

BUG=222527
R=rmsousa@chromium.org

Review URL: https://codereview.chromium.org/149863002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247762 0039d316-1c4b-4281-b951-d872f2087c98

This prevents a use-after-free bug where it is possible WebContentsImpl attempts
to access a newly created window after BrowserPluginGuest has freed it because the
permission was instantly denied. This can happen if BrowserPluginGuest has no
delegate: it's not a <webview> or <adview>.

BUG=338345

Review URL: https://codereview.chromium.org/149643002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247761 0039d316-1c4b-4281-b951-d872f2087c98

Reported metrics:
- number of extensions in cache
- size of cache on disk

BUG=316371
TEST=manual

Review URL: https://codereview.chromium.org/148013004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247760 0039d316-1c4b-4281-b951-d872f2087c98

Core types don't respect off by default state yet.

TBR=petewil@chromium.org
BUG=339094, 280266

Review URL: https://codereview.chromium.org/146913004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247759 0039d316-1c4b-4281-b951-d872f2087c98

. get rid of most of the separate objects, which were making my head explode. There's now just ViewManagerImpl and ViewImpl. ViewManagerImpl implements NativeViewportClient, LauncherClient, ShellClient, etc.

R=sky@chromium.org
BUG=

Review URL: https://codereview.chromium.org/149363006

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247758 0039d316-1c4b-4281-b951-d872f2087c98

Review URL: https://codereview.chromium.org/144033004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247757 0039d316-1c4b-4281-b951-d872f2087c98

This is necessary in order to fix the presubmit error that happened in
https://codereview.chromium.org/147803002/#msg7

BUG=None
TEST=None
R=scottmg@chromium.org
TBR=ben@chromium.org

Review URL: https://codereview.chromium.org/143263021

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247755 0039d316-1c4b-4281-b951-d872f2087c98

In sending handles across processes, the ProxyMessagePipeEndpoint will
need to go from a generic Dispatcher to specific dispatcher types. (The
alternative is to have layers of complex abstraction....)

Also, at some point, I'll probably add the ability to query the type of
a handle (along with other information).

R=darin@chromium.org

Review URL: https://codereview.chromium.org/136343004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247754 0039d316-1c4b-4281-b951-d872f2087c98

Move handling of dispatchers when enqueueing messages into the endpoint
code. That handling will need to be very specific to the type of
endpoint.

R=darin@chromium.org

Review URL: https://codereview.chromium.org/147983009

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247753 0039d316-1c4b-4281-b951-d872f2087c98

Make sure DeviceRequestMessageFilter call MediaStreamManager::CancelRequest for all pending requests to MediaStreamManager::EnumerateDevices when the IPC channel is teared down.

Before this change DeviceRequestMessageFilter call CancelRequest with the render process pid. But that pid was not been used as an identifier when calling EnumerateDevices. 

BUG=338561

Review URL: https://codereview.chromium.org/144363005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247744 0039d316-1c4b-4281-b951-d872f2087c98

Fixed regression caused by 179554 which changed the meaning of ToggleFullscreenMode(). Updated BrowserWindowCocoa::Show() accordingly.

BUG=222605
TEST=(1) Start Chrome and enter fullscreen mode.
(2) Quit and restart.
(3) Verify that Chrome started in fullscreen mode, not presentation mode.

Review URL: https://codereview.chromium.org/149093004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247743 0039d316-1c4b-4281-b951-d872f2087c98

This is a quick and dirty fix for a problem that occurs to me when
trying to run instrumentation tests in a simulator.
For some reason 'pm path org.chromium.content_shell_apk.tests' outputs
a warning message in my environment that confuses the python scripts
and they end up calling md5sum for the root directory which is a bit
slow... The real fix would be to be able drop the stderr of the command
but it needs a lot of plumbing and probably there is not that much chance
that something like that will happen again for someone.

BUG=334439

Review URL: https://codereview.chromium.org/138143019

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247742 0039d316-1c4b-4281-b951-d872f2087c98

BUG=299507

TBR=lazyboy@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/149783002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247741 0039d316-1c4b-4281-b951-d872f2087c98

Currently, the title was one behind, because the title was being updated before the current volume was remembered to a member variable. This patch resolves this issue by storing the current volume to a member variable ASAP.

TEST=Tested manually with two windows.
BUG=339015

Review URL: https://codereview.chromium.org/145123005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247740 0039d316-1c4b-4281-b951-d872f2087c98

This is the first of three patches converting WebTouchPoint to use
WebFloatPoint instead of WebPoint. This introduces some temporary
casts to WebPoint in chromium, which will be removed as soon as
blink has WebTouchPoint storing it's location in a WebFloatPoint.

The second patch (https://codereview.chromium.org/149053002/) 
will be a blink modification to store WebFloatPoint,
and the third patch (https://codereview.chromium.org/148453012/) 
will clean up the chromium side.

BUG=336807

Review URL: https://codereview.chromium.org/140973005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@247739 0039d316-1c4b-4281-b951-d872f2087c98