GitLab

BUG=none
TEST=compiles and ProcessSingletonTest passes.


Review URL: http://codereview.chromium.org/8574030

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110395 0039d316-1c4b-4281-b951-d872f2087c98

The last try broke at least 2 things which I've fixed in this CL:
- Attempts to access geolocation from an extension would hang.
- We'd sometimes crash when shutting down the browser with an ExtensionPopup
open.

BUG=84146
TEST=extensions still work

Review URL: http://codereview.chromium.org/8587001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110394 0039d316-1c4b-4281-b951-d872f2087c98

BUG=88782
TEST=X509UtilNSSTest


Review URL: http://codereview.chromium.org/8537025

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110393 0039d316-1c4b-4281-b951-d872f2087c98

Data that's only pertinent to initially loading the document is stored in LoadTimes. These are carried over when the NavigationState changes due to in-document navigation.

I'd be interested in feedback on the position of LoadTimes. I'm not happy with it. I'd prefer to have it on WebDataSource, but I expect NavigationState was hidden in ExtraData for a reason.

This also impacts PLT histograms. In-document navigation is the cause of some of the missing start types. So, by fixing this, we should get more complete page load data. There are probably whole classes of sites where we have no PLT data, such as Google Docs.

BUG=79078
TEST=ui_tests


Review URL: http://codereview.chromium.org/8404018

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110392 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=none
TBR=OWNERS

Review URL: http://codereview.chromium.org/8539001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110391 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=none

R=groby@chromium.org

Review URL: http://codereview.chromium.org/8590010

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110390 0039d316-1c4b-4281-b951-d872f2087c98

This failure path has not received a lot of testing until now. Here are
the issues addressed by this patch:

- We usually check the return value of step() calls. However, we do not
check the return value of prepare() calls, which are more likely to fail.
If they do fail, we will DCHECK() or go on to dereference an invalid
pointer in step(). This patch checks the return value of one particular
prepare statement, the one in CheckIntegrity().

- Disable DCHECKs on sqlite errors, DirectoryManager open failure,
and SyncManager initialization failure. This will allow us to test these
error paths.

- Be careful in ShutdownOnSyncThread(). The directory will not be fully
intialized during shutdown if the database load failed.

- Add a ProfileSyncService unit test that simulates a load from an
unreadable database. The harness had to be modified slightly to make
this possible.

- Remove a setup_for_test_mode_ flag in SyncManager::SyncInternal::Init.
I don't know what the original intent of this flag was. However, I do
know that it prevents me from properly simulating a database load failure
and removing it seems to have no ill effects.

- Do not delete the database from DirectoryBackingStore. If this code
were to get executed it would put us into an inconsistent state. See
issue 103824. However, it's unlikely this code would get executed. If
the database were actually corrupt, we would DCHECK or de-reference an
invalid pointer on our way to this code because we don't check the
return value of the attempt to prepare an SQL statement in
DirectoryBackingStore::CheckIntegrity().

- Modify the DirectoryBackingStoreTest.Corruption unit test to expect the
new behaviour.

- Disable sync when backend initialize fails. Such a failure could
be due to bad local state. We don't know the actual cause because the
information is not available from the ProfileSyncService callback. The
safe course of action is to clear our local sync state and try again
later. It's the easiest way to get back to the most well travelled sync
initialization path.

- Fix error handling logic in OpenAndConfigureHandleHelper.  It used
to rely on a specially-crafted scoped_ptr to close the database if we
had to leave the function unsuccessfully.  This was wrong in two ways.
First, it did not reset the handle to NULL, which meant that the
DirectoryBackingStore would attempt to free the handle again when it
was destructed.  Second, it failed to clean up the handle when the
return value was not SQLITE_OK.  (Though I suppose it would have been
cleaned up by the DirectoryBackingStore destructor, thanks to the
previous issue).

BUG=103307, 103824
TEST=DirectoryBackingStoreTest.Corruption, ProfileSyncServiceTest.CorruptDatabase


Review URL: http://codereview.chromium.org/8568028

TBR=rlarocque@chromium.org
Review URL: http://codereview.chromium.org/8586004

TBR=cpu@chromium.org
Review URL: http://codereview.chromium.org/8574045

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110388 0039d316-1c4b-4281-b951-d872f2087c98

Revert 110355 - Use shared D3D9 texture to transport the compositor's backing buffer to the browser process for presentation.

Implemented ImageTransportSurface for Linux (without texture sharing), XP, Vista and 7. XP. The non-texture sharing Linux and XP paths just present directly to the compositing child window owned by the browser process as before.

PassThroughImageTransportSurface still needs a proper name. I will move it into its own file once that is decided.

I moved AcceleratedSurfaceBuffersSwapped outside of the platform specific ifdefs and made the signature the same on all platforms for greater consistency.

I removed the code related to sharing surfaces between processes and synchronizing resize and swapping out of GpuCommandBufferStub. It is all now in ImageTransportSurface implementations.
Review URL: http://codereview.chromium.org/8060045

TBR=apatrick@chromium.org
Review URL: http://codereview.chromium.org/8587012

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110387 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=none
Review URL: http://codereview.chromium.org/8587003

TBR=gene@chromium.org
Review URL: http://codereview.chromium.org/8586005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110385 0039d316-1c4b-4281-b951-d872f2087c98

This failure path has not received a lot of testing until now. Here are
the issues addressed by this patch:

- We usually check the return value of step() calls. However, we do not
check the return value of prepare() calls, which are more likely to fail.
If they do fail, we will DCHECK() or go on to dereference an invalid
pointer in step(). This patch checks the return value of one particular
prepare statement, the one in CheckIntegrity().

- Disable DCHECKs on sqlite errors, DirectoryManager open failure,
and SyncManager initialization failure. This will allow us to test these
error paths.

- Be careful in ShutdownOnSyncThread(). The directory will not be fully
intialized during shutdown if the database load failed.

- Add a ProfileSyncService unit test that simulates a load from an
unreadable database. The harness had to be modified slightly to make
this possible.

- Remove a setup_for_test_mode_ flag in SyncManager::SyncInternal::Init.
I don't know what the original intent of this flag was. However, I do
know that it prevents me from properly simulating a database load failure
and removing it seems to have no ill effects.

- Do not delete the database from DirectoryBackingStore. If this code
were to get executed it would put us into an inconsistent state. See
issue 103824. However, it's unlikely this code would get executed. If
the database were actually corrupt, we would DCHECK or de-reference an
invalid pointer on our way to this code because we don't check the
return value of the attempt to prepare an SQL statement in
DirectoryBackingStore::CheckIntegrity().

- Modify the DirectoryBackingStoreTest.Corruption unit test to expect the
new behaviour.

- Disable sync when backend initialize fails. Such a failure could
be due to bad local state. We don't know the actual cause because the
information is not available from the ProfileSyncService callback. The
safe course of action is to clear our local sync state and try again
later. It's the easiest way to get back to the most well travelled sync
initialization path.

- Fix error handling logic in OpenAndConfigureHandleHelper.  It used
to rely on a specially-crafted scoped_ptr to close the database if we
had to leave the function unsuccessfully.  This was wrong in two ways.
First, it did not reset the handle to NULL, which meant that the
DirectoryBackingStore would attempt to free the handle again when it
was destructed.  Second, it failed to clean up the handle when the
return value was not SQLITE_OK.  (Though I suppose it would have been
cleaned up by the DirectoryBackingStore destructor, thanks to the
previous issue).

BUG=103307, 103824
TEST=DirectoryBackingStoreTest.Corruption, ProfileSyncServiceTest.CorruptDatabase


Review URL: http://codereview.chromium.org/8568028

TBR=rlarocque@chromium.org
Review URL: http://codereview.chromium.org/8586004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110384 0039d316-1c4b-4281-b951-d872f2087c98

Review URL: http://codereview.chromium.org/8589002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110383 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110380 0039d316-1c4b-4281-b951-d872f2087c98

BUG=104535
TEST=none

R=viettrungluu@chromium.org

Review URL: http://codereview.chromium.org/8574041

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110379 0039d316-1c4b-4281-b951-d872f2087c98

BUG=104534
TEST=none

R=thestig@chromium.org

Review URL: http://codereview.chromium.org/8586003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110378 0039d316-1c4b-4281-b951-d872f2087c98

Bug=None
Test=None


Review URL: http://codereview.chromium.org/8372034

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110377 0039d316-1c4b-4281-b951-d872f2087c98

The general approach is to require user gesture and put up the confirmation UI if an extension wants to re-enable a disabled-due-to-permissions-escalation extension.

Also added some new infrastructure for making this easier to test.

(This is re-landing http://codereview.chromium.org/8423069/ which had a compile
 error on the cros_tegra2 bot)

BUG=102579
TEST=See bug

TBR=aa@chromium.org

Review URL: http://codereview.chromium.org/8488012

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110376 0039d316-1c4b-4281-b951-d872f2087c98

This CL fixes a crash that occurs if --enable-bluetooth is set via the
chrome://flags page. The UI attempts to interact with the
BluetoothManager, but neither the BluetoothManager nor the underlying
bluetooth dbus clients have been created because the parsed command
line was tested for --enable-bluetooth prior to the about:flags flags
being added to it.

We resolve this issue by unconditionally creating the bluetooth dbus
clients and deferring conditional creation of the BluetoothManager until
after the about:flags flags are added to the parsed command line.

BUG=chromium-os:22877
TEST=Visit chrome://flags. Enable bluetooth. Restart. Visit
chrome://settings/system and observe that Chrome doesn't crash.

Change-Id: Ib5deed32200afa8c1e9455abc72cefb7e5183da2


Review URL: http://codereview.chromium.org/8575011

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110375 0039d316-1c4b-4281-b951-d872f2087c98

Relanding with fix for component build: use NET_EXPORT_PRIVATE
in network_change_notifier_linux.h.

BUG=53473
TEST=see attachment on bug 7469
R=willchan@chromium.org

Review URL: http://codereview.chromium.org/8575013

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110374 0039d316-1c4b-4281-b951-d872f2087c98

aura (doesn't support NPAPI).

BUG=103496
TEST=none
R=mmenke@chromium.org


Review URL: http://codereview.chromium.org/8488016

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110373 0039d316-1c4b-4281-b951-d872f2087c98

BUG=104403
TEST=none
Review URL: http://codereview.chromium.org/8520034

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110372 0039d316-1c4b-4281-b951-d872f2087c98

Extension Settings API: make it so that when changes received from sync fail to apply to an extension's settings, all sync for those settings is disabled.  Likewise if local changes fail to be sent to sync.  This makes sure that no invalid outgoing changes are sent to sync, or any local consistency violated.

If sync is restarted, re-enable sync for that extension.

BUG=103514
TEST=*ExtensionSettings* (unit_tests, browser_tests)


Review URL: http://codereview.chromium.org/8539037

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110371 0039d316-1c4b-4281-b951-d872f2087c98

Review URL: http://codereview.chromium.org/8574038

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110370 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=none
Review URL: http://codereview.chromium.org/8587003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110369 0039d316-1c4b-4281-b951-d872f2087c98

Will be re-enabled tomorrow.

R=dennisjeffrey@chromium.org
BUG=
TEST=

Review URL: http://codereview.chromium.org/8513022

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110368 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=none

Review URL: http://codereview.chromium.org/8566028

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110367 0039d316-1c4b-4281-b951-d872f2087c98

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=641497
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=647524
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=647526
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=647527

And also a patch by Adam Langley (agl):

http://anonscm.debian.org/gitweb/?p=devscripts/devscripts.git;a=commitdiff;h=917d833db3495bdc1252d3e739e8ef68f5710105

BUG=28291


Review URL: http://codereview.chromium.org/8486004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110366 0039d316-1c4b-4281-b951-d872f2087c98

fix a bug caused by redirects (such as in the GWS case).
We must copy all alias URLs from the original prerender
into the dummy.
R=cbentzel, dominich
Review URL: http://codereview.chromium.org/8583001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110365 0039d316-1c4b-4281-b951-d872f2087c98

the aura build, and delete the stubs (which were ifdef'd out in Windows)


BUG=102946
TEST=none


Review URL: http://codereview.chromium.org/8543001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110364 0039d316-1c4b-4281-b951-d872f2087c98

Needed to delete an argument

BUG=none
TEST=Run CloudPrintProxyPolicyTest unit tests under valgrind, no leaks.

Review URL: http://codereview.chromium.org/8586001

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110363 0039d316-1c4b-4281-b951-d872f2087c98

BUG=NONE
TEST=NONE
Review URL: http://codereview.chromium.org/8571010

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110362 0039d316-1c4b-4281-b951-d872f2087c98

Preallocating the shared memory object saves an IPC round trip.
It also makes the code path more similar for Linux and Mac.
It also lays the groundwork for changing the allocation method
later without having to replumb the proxying layer.

BUG= http://code.google.com/p/chromium/issues/detail?id=103377
TEST= nacl still works on Linux

R=mseaborn@chromium.org,jam@chromium.org

Review URL: http://codereview.chromium.org/8513018

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110361 0039d316-1c4b-4281-b951-d872f2087c98

This version fixes a problem with the windows incremental linking build. We can't incrementally link chrome when content is being built as a DLL because chrome links in webkit_glue and webkit_glue depends on symbols defined in content. We can remove this when we fix glue.

R=jam@chromium.org
BUG=90442, 98755.

Review URL: http://codereview.chromium.org/8570024

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110360 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=none

R=mark@chromium.org

Review URL: http://codereview.chromium.org/8513019

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110359 0039d316-1c4b-4281-b951-d872f2087c98

BUG=104314
TEST=no change

Review URL: http://codereview.chromium.org/8539047

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110357 0039d316-1c4b-4281-b951-d872f2087c98

This failure path has not received a lot of testing until now. Here are
the issues addressed by this patch:

- We usually check the return value of step() calls. However, we do not
check the return value of prepare() calls, which are more likely to fail.
If they do fail, we will DCHECK() or go on to dereference an invalid
pointer in step(). This patch checks the return value of one particular
prepare statement, the one in CheckIntegrity().

- Disable DCHECKs on sqlite errors, DirectoryManager open failure,
and SyncManager initialization failure. This will allow us to test these
error paths.

- Be careful in ShutdownOnSyncThread(). The directory will not be fully
intialized during shutdown if the database load failed.

- Add a ProfileSyncService unit test that simulates a load from an
unreadable database. The harness had to be modified slightly to make
this possible.

- Remove a setup_for_test_mode_ flag in SyncManager::SyncInternal::Init.
I don't know what the original intent of this flag was. However, I do
know that it prevents me from properly simulating a database load failure
and removing it seems to have no ill effects.

- Do not delete the database from DirectoryBackingStore. If this code
were to get executed it would put us into an inconsistent state. See
issue 103824. However, it's unlikely this code would get executed. If
the database were actually corrupt, we would DCHECK or de-reference an
invalid pointer on our way to this code because we don't check the
return value of the attempt to prepare an SQL statement in
DirectoryBackingStore::CheckIntegrity().

- Modify the DirectoryBackingStoreTest.Corruption unit test to expect the
new behaviour.

- Disable sync when backend initialize fails. Such a failure could
be due to bad local state. We don't know the actual cause because the
information is not available from the ProfileSyncService callback. The
safe course of action is to clear our local sync state and try again
later. It's the easiest way to get back to the most well travelled sync
initialization path.

- Fix error handling logic in OpenAndConfigureHandleHelper.  It used
to rely on a specially-crafted scoped_ptr to close the database if we
had to leave the function unsuccessfully.  This was wrong in two ways.
First, it did not reset the handle to NULL, which meant that the
DirectoryBackingStore would attempt to free the handle again when it
was destructed.  Second, it failed to clean up the handle when the
return value was not SQLITE_OK.  (Though I suppose it would have been
cleaned up by the DirectoryBackingStore destructor, thanks to the
previous issue).

BUG=103307, 103824
TEST=DirectoryBackingStoreTest.Corruption, ProfileSyncServiceTest.CorruptDatabase


Review URL: http://codereview.chromium.org/8568028

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110356 0039d316-1c4b-4281-b951-d872f2087c98

Use shared D3D9 texture to transport the compositor's backing buffer to the browser process for presentation.

Implemented ImageTransportSurface for Linux (without texture sharing), XP, Vista and 7. XP. The non-texture sharing Linux and XP paths just present directly to the compositing child window owned by the browser process as before.

PassThroughImageTransportSurface still needs a proper name. I will move it into its own file once that is decided.

I moved AcceleratedSurfaceBuffersSwapped outside of the platform specific ifdefs and made the signature the same on all platforms for greater consistency.

I removed the code related to sharing surfaces between processes and synchronizing resize and swapping out of GpuCommandBufferStub. It is all now in ImageTransportSurface implementations.
Review URL: http://codereview.chromium.org/8060045

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110355 0039d316-1c4b-4281-b951-d872f2087c98

R=estade@chromium.org
BUG=103822
TEST=+1 action in SyncPromo.UserFlowAction and +2 new histograms.


Review URL: http://codereview.chromium.org/8528054

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110354 0039d316-1c4b-4281-b951-d872f2087c98

BUG=none
TEST=make compositor_model_bench


Review URL: http://codereview.chromium.org/8576003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110353 0039d316-1c4b-4281-b951-d872f2087c98

loading. This effectively reverts the change to LocationBar. The change to LocationBarView was problematic because the boolean passed to OmniboxView::SelectAll means something different than the boolean passed to LocationBarView::FocusLocation.

BUG=104201
TEST=see bug
R=alicet@chromium.org,ben@chromium.org


Review URL: http://codereview.chromium.org/8486022

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110352 0039d316-1c4b-4281-b951-d872f2087c98

BUG=90644
TEST=none

R=brettw@chromium.org

Review URL: http://codereview.chromium.org/8486018

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@110351 0039d316-1c4b-4281-b951-d872f2087c98