TEST=try
BUG=none

Review URL: http://codereview.chromium.org/4949005

The issue was a builder failed:

http://build.chromium.org/p/chromium.memory/builders/Webkit%20Mac%20(valgrind)/builds/1052

Error was:

CompileC ../../../xcodebuild/build_angle.build/Release/translator_common.build/Objects-normal/i386/glslang_lex.o /b/build/slave/webkit-rel-mac-valgrind/build/src/third_party/angle/src/compiler/glslang_lex.cpp normal i386 c++ com.apple.compilers.gcc.4_2
    cd /b/build/slave/webkit-rel-mac-valgrind/build/src/third_party/angle/src
    /Developer/usr/bin/gcc-4.2 -x c++ -arch i386 -fmessage-length=0 -pipe -Wno-trigraphs -fno-exceptions -fno-rtti -O1 -Wnewline-eof -DTRACE_OUTPUT_FILE="angle-debug.txt" -DCHROMIUM_BUILD -DENABLE_REMOTING=1 -DENABLE_GPU=1 -DNDEBUG -DDYNAMIC_ANNOTATIONS_ENABLED=1 -isysroot /Developer/SDKs/MacOSX10.5.sdk -fvisibility=hidden -fvisibility-inlines-hidden -fno-threadsafe-statics -mmacosx-version-min=10.5 -gdwarf-2 -Wendif-labels -Wno-unused-parameter -Wno-missing-field-initializers -F/b/build/slave/webkit-rel-mac-valgrind/build/src/third_party/angle/src/../../../xcodebuild/Release -I/b/build/slave/webkit-rel-mac-valgrind/build/src/third_party/angle/src/../../../xcodebuild/Release/include -I. -I../include -I/b/build/slave/webkit-rel-mac-valgrind/build/src/third_party/angle/src/../../../xcodebuild/build_angle.build/Release/translator_common.build/DerivedSources -g -fno-inline -fno-omit-frame-pointer -fno-builtin -c /b/build/slave/webkit-rel-mac-valgrind/build/src/third_party/angle/src/compiler/glslang_lex.cpp -o /b/build/slave/webkit-rel-mac-valgrind/build/src/third_party/angle/src/../../../xcodebuild/build_angle.build/Release/translator_common.build/Objects-normal/i386/glslang_lex.o
i686-apple-darwin9-gcc-4.2.1: /b/build/slave/webkit-rel-mac-valgrind/build/src/third_party/angle/src/compiler/glslang_lex.cpp: No such file or directory
i686-apple-darwin9-gcc-4.2.1: warning: '-x c++' after last input file has no effect
i686-apple-darwin9-gcc-4.2.1: no input files


TBR=alokp@chromium.org

Review URL: http://codereview.chromium.org/5116001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66351 0039d316-1c4b-4281-b951-d872f2087c98

Add named testing interface. This allows you to connect to a pre-existing Chrome process and run tests on it. This is an addition to the low level interface underlying testing frameworks like PyAuto and WebDriver.

Normally, test frameworks communicate with Chrome over an unnamed socket pair on POSIX. The test creates the socket pair and then launches the browser as a child process, passing an open file descriptor for one end of the socket to the browser. This change adds a command line switch that, when passed to the browser, causes it to listen on a named socket instead, eliminating this parent/child process requirement. Therefore, you can potentially connect any number of tests to a preexisting browser process.

For ChromeOS, this allows you to run tests on the instance of Chrome that is launched on startup, which controls things like the login and lock screens, the battery meter, the wireless UI, etc. Currently there is no way to run tests on a pre-existing Chrome instance. Eventually this will also allow you to connect both PyAuto and WebDriver to the same Chrome instance and run both in the same test.

If you pass the browser the following command line switch:
./chrome --testing-channel=NamedTestingInterface:/path/to/file
This causes the browser to listen for incoming connections. An AutomationProxy can connect to the browser by connecting a Unix domain socket to the specified path and control the browser over the socket.

This is currently only for POSIX. Windows support will come in a future change. Also, this initial change only allows one connection; multiple connection support will come in a future change.

BUG=chromium-os:8512
TEST=Run Chrome with --testing-interface=/var/tmp/NamedTestingInterface, then run NamedInterfaceTest.BasicNamedInterface under ui_tests.

Review URL: http://codereview.chromium.org/4202004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66350 0039d316-1c4b-4281-b951-d872f2087c98

BUG=66386
TEST=None
TBR=kmadhusu

Review URL: http://codereview.chromium.org/5115001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66349 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=none
Review URL: http://codereview.chromium.org/5024004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66348 0039d316-1c4b-4281-b951-d872f2087c98

Moved it to GarbageCollectExtensions().  This avoids some problems with
shutdown ordering causing crashes.

Re-enable themes sync integration tests on OS X (which were crashing).

BUG=63285,62869
TEST=themes sync integration tests

Review URL: http://codereview.chromium.org/4957005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66347 0039d316-1c4b-4281-b951-d872f2087c98

- removed button that was used to fake payment process. 
- fixed race condition on connect/disconnect. 
- added UMA histograms to trace results of activation process.

BUG=chromium-os:9202, chromium-os:9220, chromium-os:9169
TEST=none


Review URL: http://codereview.chromium.org/5072001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66346 0039d316-1c4b-4281-b951-d872f2087c98

TBR=derat@chromium.org
BUG=none
TEST=chromium os build should cycle green

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66345 0039d316-1c4b-4281-b951-d872f2087c98

- Check for a NULL PluginObject at all NPAPI entry points. This fixes a crash in Chrome where the browser foolishly calls NPP_SetWindow with instance->pdata == NULL.
- Factor out obscene amounts of redundancy from main_<platform>.(cc|mm) into main.cc. This makes no change to the functionality, except that NP_Initialize on Linux now has HANDLE_CRASHES whereas before it did not.

TEST=built on Linux, Mac, and Windows; installed and loaded O3D on each platform and made sure it worked; repeatedly loaded on Mac & Win in Chrome 8.0.552.200 and verified no crashes; inspected "svn diff" output manually and verified that the code and order of execution of that code is unchanded for each platform
BUG=none

Review URL: http://codereview.chromium.org/4957002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66344 0039d316-1c4b-4281-b951-d872f2087c98

Change the font size to medium.

BUG=chromium-os:9194
TEST=login and lock the screen using short power keypress. Confirm the uesrname has the same font/color as in login screen.

Review URL: http://codereview.chromium.org/5074001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66343 0039d316-1c4b-4281-b951-d872f2087c98

BUG=63038
TEST=Install an extension with a bg page and a popup. Right-click the browser action and inspect the popup. In the task manager, kill the extension process. The browser should not crash.

Review URL: http://codereview.chromium.org/5038004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66342 0039d316-1c4b-4281-b951-d872f2087c98

These appear to have been following the pattern of:

  ScreenLocker::Show();
  ...
  ui_test_utils::WaitForNotification(
      NotificationType::SCREEN_LOCK_STATE_CHANGED);

I believe that this hangs until the test times out if the
notification is sent before we return from Show() (since we
won't yet be listening for the notification at the time).

I'm changing them to:

  ScreenLocker::Show();
  ...
  if (!chromeos::ScreenLocker::GetTester()->IsLocked())
    ui_test_utils::WaitForNotification(
        NotificationType::SCREEN_LOCK_STATE_CHANGED);

The old version looks like it started failing with my
r66287, which I don't understand -- I thought that we were
already generating the notification immediately as of
r66131, which fixed a bug where the initial attempt to grab
the pointer and keyboard always failed (thus ensuring that
the notifications wouldn't be sent until the next time that
the event loop was run).

BUG=none
TEST=all screen locker tests passed on my local machine

Review URL: http://codereview.chromium.org/5054004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66340 0039d316-1c4b-4281-b951-d872f2087c98

TEST=try
BUG=none

Review URL: http://codereview.chromium.org/4949005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66339 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=none

Review URL: http://codereview.chromium.org/5089001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66338 0039d316-1c4b-4281-b951-d872f2087c98

BUG=63026
TEST=none

Review URL: http://codereview.chromium.org/4850001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66337 0039d316-1c4b-4281-b951-d872f2087c98

When chrome.exe is launched with --chrome-frame switch, we need
to gather any crashes in this mode under 'ChromeFrame' product
id under go/crash. 

Since ChromeFrame exists only on windows, this change is limited
to breakpad_win.cc for now.

BUG=none
TEST=none


Review URL: http://codereview.chromium.org/4847001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66336 0039d316-1c4b-4281-b951-d872f2087c98

The GPU watchdog thread that seemed to cause this error is now disabled when running with valgrind. Leaks are expected when the watchdog aborts the GPU process.

TEST=mac_valgrind
BUG=58245

Review URL: http://codereview.chromium.org/5027003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66335 0039d316-1c4b-4281-b951-d872f2087c98

Collecting GPU info was taking enough time to make the GPU watchdog abort the GPU process.

Moved GTK initialization into GpuProcess (but still before gfx::GLContext::InitializeOneOff).

TEST=try
BUG=none

Review URL: http://codereview.chromium.org/5017004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66333 0039d316-1c4b-4281-b951-d872f2087c98

and extend RenderViewHost with a concept of reserved contents rect, a place to show extra stuff,
such as Sidebar's mini tab UI.
sidebar UI implementation warranted this change (mini tabs UI and resize corner area for sidebar
contents).

TabContentsDelegate::GetRootWindowResizerRect() is no more, reserved contents area is now cached
in RenderWidgetHostView and updated upon view resize.

Views: BrowserView is responsible for the actual layout and reserved contents area update.

Mac: TabContentsController now manages all TabContents views in the main browser window to solve
two problems, first to prevent contents flickering during tab change not only for the page,
but for the sidebar and devtools contents too and, second, to monitor contents view frame changes
and update reserved contents area accordingly.


BUG=51084
TEST=All tests should pass, this is a refactoring CL.

Review URL: http://codereview.chromium.org/3547008

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66332 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=none
Review URL: http://codereview.chromium.org/5095001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66331 0039d316-1c4b-4281-b951-d872f2087c98

This is pretty broad, since there are leaks in the test body itself, as well as in calls made by the test body.

BUG=63386
TEST=None
TBR=kmadhusu

Review URL: http://codereview.chromium.org/4985007

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66330 0039d316-1c4b-4281-b951-d872f2087c98

TBR=asargent
BUG=63078
TEST=no

Review URL: http://codereview.chromium.org/5020002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66329 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=better error messages.
Review URL: http://codereview.chromium.org/4999003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66328 0039d316-1c4b-4281-b951-d872f2087c98

- Overlapping styles are now bitwise-ORed together (instead of replacing
previous ones).
- The "length" parameter for style ranges is optional and can be 0.

BUG=63077,63078
TEST=covered by tests

Review URL: http://codereview.chromium.org/5064001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66327 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=none
Review URL: http://codereview.chromium.org/5034004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66326 0039d316-1c4b-4281-b951-d872f2087c98

It looks like all the Passwords integration tests are failing on the mac
builder. Marking them as FAILS until the associated bug is fixed.

For logs, see
http://build.chromium.org/p/chromium.fyi/builders/Chromium%20Mac%20Sync/builds/1238

TBR=akalin@chromium.org
BUG=59867
TEST=sync_integration_tests on mac builder

Review URL: http://codereview.chromium.org/4985005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66325 0039d316-1c4b-4281-b951-d872f2087c98

BUG=chromium-os:9035
TEST=Go to chrome://settings and observe that there's no Labs section.

Review URL: http://codereview.chromium.org/5025001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66324 0039d316-1c4b-4281-b951-d872f2087c98

Platform abstraction for a shared lock between processes. The process
that owns the lock will release it on exit even if exit is due
to a crash.

For cloud-print and remoting we want to be able to have a singleton service-process that can run independently of the browser process. This service process will communicate with the browser process via the standard IAC channels, but we want to have a way of signaling to other processes that a) there is a service-process running and b) that it is in a state where it is ready to be communicated with. The multi_process_lock class is intended to work as a simple flag that can be queried from multiple processes. If the service-process should crash, we would like the flag to be cleared automatically so that there is never confusion about the state of the service-process.

Other approaches considered for some Unix/Mac:
- Standard unix domain sockets depend on the file system and don't clean up properly in the case of a crash.
- Shared memory on unix depend on the file system and don't clean up properly in the case of a crash.
- System V semaphores on unix again depend on the file system.
- named_mach_ports on Mac OS. Bootstrap_register_name has been deprecated on 10.6, so we are doing essentially the same thing using CFMessagePort.

On Windows it is implemented as an event.
On Mac it is implemented using a CFMessagePort name.
On Linux it is implement using an abstract name port socket.

TEST=none
BUG=none

Review URL: http://codereview.chromium.org/4721001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66323 0039d316-1c4b-4281-b951-d872f2087c98

Remove a DCHECK when failing to secure the shared memory used to hold the current version. This may well fail on older (read non-NTFS) systems since you can't secure file mappings without a securable file system underneath.

BUG=61609
TEST=No DCHECKing at startup when running on a non-NTFS system.

Review URL: http://codereview.chromium.org/5057007

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66322 0039d316-1c4b-4281-b951-d872f2087c98

copy of http://codereview.chromium.org/4991006/
had to do it again in 'commitable' version of chrome tree

Review URL: http://codereview.chromium.org/4979006

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66321 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=none
TBR=antonm@chromium.org

Review URL: http://codereview.chromium.org/5076002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66320 0039d316-1c4b-4281-b951-d872f2087c98

For print preview, a PP tab should be linked with the tab that initiated the printing operation. If the tab initiates a second printing operation while the first print preview tab is still open, that PP tab should be focused/activated. There may be more than one PP tab open.

Hook ctrl+p to show print preview tab.

BUG=59653, 57893
TEST=new unittest created.

Review URL: http://codereview.chromium.org/4338001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66319 0039d316-1c4b-4281-b951-d872f2087c98

TEST=try
BUG=none

Review URL: http://codereview.chromium.org/5087001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66318 0039d316-1c4b-4281-b951-d872f2087c98

Add scoped_allowio to InitSpeechInput until we can evaluate whether this IO access is ok on the main thread.

NOTE: The previous revision to this file was a simple line-ending change which I had to check in without Rietveld (uploading fails if there is a line-ending mismatch).

BUG=63335
TEST=None

Review URL: http://codereview.chromium.org/5046003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66317 0039d316-1c4b-4281-b951-d872f2087c98

Sometimes crashes on the Mac 10.5 tests bot.

BUG=63358
TEST=greener tree

Review URL: http://codereview.chromium.org/5072002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66316 0039d316-1c4b-4281-b951-d872f2087c98

Implement InputStub for the host.

BUG=None
TEST=None

Review URL: http://codereview.chromium.org/4726003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66314 0039d316-1c4b-4281-b951-d872f2087c98

M    browser_render_process_host.cc



git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66313 0039d316-1c4b-4281-b951-d872f2087c98

Modifies the Instant loader to listen for RENDER_VIEW_HOST_CHANGED notifications on Mac.  This notification is sent when a given TabContents changes its renderer, and on Mac we need to call SetTakesFocusOnlyOnMouseDown() on the new RWHV.

BUG=63330
TEST=Type in omnibox with Instant on.  Arrow down through results, then click on preview contents.  Preview contents should not disappear.
Review URL: http://codereview.chromium.org/5073002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66312 0039d316-1c4b-4281-b951-d872f2087c98

TBR=amit

Review URL: http://codereview.chromium.org/4980004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66311 0039d316-1c4b-4281-b951-d872f2087c98

Sends select field text values to browser when form is submitted.

BUG=49189
TEST=FormManagerTest.SelectOneAsText

Review URL: http://codereview.chromium.org/5080001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66310 0039d316-1c4b-4281-b951-d872f2087c98

BUG=57799
TEST=None
TBR=bulach

Review URL: http://codereview.chromium.org/5012002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@66309 0039d316-1c4b-4281-b951-d872f2087c98