GitLab

With this change attachment store ownership is moved to datatype. If
datatype wants to use attachments it needs to instantiate attachment
store and return it to sync via GetAttachmentStore.
GenericChangeProcessorFactory calls SyncableService::GetAttachmentStore
and passes attachment_store through GenericChangeProcessor::ctor to
AttachmentService::ctor.
GetAttachmentStore returning NULL means datatype doesn\u2019t use attachments,
therefore GenericChangeProcessor doesn\u2019t create AttachmentService.

R=maniscalco@chromium.org
BUG=412070

Review URL: https://codereview.chromium.org/548373003

Cr-Commit-Position: refs/heads/master@{#294070}

BUG=

Review URL: https://codereview.chromium.org/555843003

Cr-Commit-Position: refs/heads/master@{#294069}

needs_establish_peer_ has been used to skip actual estalishing a surface
texture peer based on many conditions. Not all places that modify
actually checks all these conditions so there are bound to be some bugs.

This CL aims to tease out all these different conditions and put them in
a single function EstablishSurfaceTexturePeerIfNeeded. This function
should be called any time these conditions are changed. Makes for much
saner reading.

BUG=412578

Review URL: https://codereview.chromium.org/557593002

Cr-Commit-Position: refs/heads/master@{#294068}

Remove storage-related redirect headers in webkit/common/, update #includes and DEPS files to use storage/ directly

BUG=338338
TBR=darin@chromium.org

Review URL: https://codereview.chromium.org/553983013

Cr-Commit-Position: refs/heads/master@{#294067}

Review URL: https://codereview.chromium.org/552373003

Cr-Commit-Position: refs/heads/master@{#294066}

Revert of Update metrics verification for dev-proxy. (patchset #3 id:80001 of https://codereview.chromium.org/553543004/)

Reason for revert:
This appears to have made ChromeProxyMetricsTest.testChromeProxyMetricForHeaderValidation fail in Android Tests: https://build.chromium.org/p/chromium.linux/builders/Android%20Tests/builds/15658/steps/chrome_proxy/logs/stdio

Traceback (most recent call last):
  run at /usr/lib/python2.7/unittest/case.py:327
    testMethod()
  testChromeProxyMetricForHeaderValidation at tools/chrome_proxy/integration_tests/chrome_proxy_metrics_unittest.py:143
    metric.AddResultsForHeaderValidation(None, results)
  AddResultsForHeaderValidation at tools/chrome_proxy/integration_tests/chrome_proxy_metrics.py:146
    bypassed, _ = self.IsProxyBypassed(tab)
  IsProxyBypassed at tools/chrome_proxy/integration_tests/chrome_proxy_metrics.py:167
    bad_proxies = [str(p['proxy']) for p in info['badProxies']]
KeyError: 'badProxies'

Locals:
  info : {'proxies': ['compress.googlezip.net:80', 'direct://'], 'enabled': True}
  tab  : None

Original issue's description:
> Update metrics verification for dev-proxy.
> Remove http://aws1.mdw.la/piatek/bypass-demo, which is not bypassed by the server now.
>
> BUG=412497
>
> Committed: https://chromium.googlesource.com/chromium/src/+/0bae38f3f8525a2b49d6df9f7152e1cc130851b1

TBR=bengr@chromium.org,bolian@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=412497

Review URL: https://codereview.chromium.org/555373002

Cr-Commit-Position: refs/heads/master@{#294065}

They don't need to run in Chrome's unit_tests suite.

Moved a test for a one-off filesystem permission hack into Chrome's
PermissionsTest because the filesystem API lives in src/chrome.

BUG=397165
TEST=extensions_unittests

Review URL: https://codereview.chromium.org/553153005

Cr-Commit-Position: refs/heads/master@{#294064}

Restored disabling preview.
Restored printing throttling.

Restoring printing without print preview removed in https://codereview.chromium.org/488853002

Restored code is not used in Chrome or Chromium browsers.

BUG=374321
TBR=gene@chromium.org

Review URL: https://codereview.chromium.org/550033002

Cr-Commit-Position: refs/heads/master@{#294063}

This partially reverts commit 74d15398 (r292900).

(But we keep a small refactor to make it easier to reorder these buttons
in the future.)

BUG=406222

Review URL: https://codereview.chromium.org/555763002

Cr-Commit-Position: refs/heads/master@{#294062}

There seems to be a lot of them.

BUG=None
Test=None, code cleanup, no visible change.

Review URL: https://codereview.chromium.org/554333003

Cr-Commit-Position: refs/heads/master@{#294061}

Review URL: https://codereview.chromium.org/558663004

Cr-Commit-Position: refs/heads/master@{#294060}

 * Actually running the tasks in the ordered you asked!
 * Allow running only pending tasks, all tasks until idle, to a given time or
   for a given period.
 * Allow stopping of running tasks on *any* arbitrary condition. No longer will
   your tasks stop working when someone adds a new task or changes the task
   order!
 * Task runner intimately connected to time and controls Now().
   Supports both automatic management and manual control.

This change makes it possible for the scheduler_unit tests to be 100%
deterministic. It also allows them to be more flexible and less brittle.

BUG=380889

Review URL: https://codereview.chromium.org/387493002

Cr-Commit-Position: refs/heads/master@{#294059}

Natural (aka australian) scrolling occurs outside Chrome in the X11
touchpad driver. Chrome dictates via an inputcontrol script what the
value (true/false) should be. The script notes (in the user's homedir)
what the last set value is so that added devices get it applied.

Unfortunately, what was happening was Chrome would update the pref
before the user logged in, so the inputcontrol script couldn't make a
note. Later, when a user was logged in and a new device was attached,
the script wouldn't know if australian scrolling should be on or off.

To fix, we always set australian scrolling whenever we consider
sending touchpad value changes to the inputcontrol script.

BUG=chromium:406480
TEST=Tested that when user logs in, chrome always sends the current
value of natural scrolling to the inputcontrol script. Also tested
that changing non-mouse/tp settings doesn't trigger calls to
inputcontrol script.

Review URL: https://codereview.chromium.org/553153004

Cr-Commit-Position: refs/heads/master@{#294058}

This is implemented generically in terms of GetStat().

Review URL: https://codereview.chromium.org/547053002

Cr-Commit-Position: refs/heads/master@{#294057}

The dev marker is removed, making this a stable pepper API.

The context around this decision can be found here:
https://groups.google.com/a/chromium.org/forum/#!topic/pepper-dev/Dw-UGrTuYuI

Review URL: https://codereview.chromium.org/547523002

Cr-Commit-Position: refs/heads/master@{#294056}

Per msan report, the line in question reads:
  if (timer_id == fade_out_timer_id_) {
so initialize fade_out_timer_id_ in the constructor.

BUG=411177

Review URL: https://codereview.chromium.org/557493002

Cr-Commit-Position: refs/heads/master@{#294055}

BufferingFrameProvider is a CodedFrameProvider which buffers frames
up to a given level.

BUG=408189

Review URL: https://codereview.chromium.org/554893003

Cr-Commit-Position: refs/heads/master@{#294054}

This CL also adds a test that exercises the code path and makes
sure destroying an opener that has unattached window doesn't
cause any undesired side effects.

If we destroy an opener that had a pending newwindow, the
newwindow's GC would try to setPermission to deny the
newwindow later. But the opener is gone and its guestInstanceId
isn't valid anymore, this would throw a JavaScript exception b/c
we call webViewInternal.setPermission with undefined instance id:
"Uncaught Error: Invocation of form ..."
The fix is to ignore calling webViewInternal.setPermission in this
case.

Note that the added test doesn't necessarily check for the above
exception because GC can be delayed arbitrarily. The test checks
the regular code path.

BUG=406616
Test=See the test added in newwindow/embedder.js for details.
Open a chrome app window that has a <webview>.
Add a newwindow event listener to the <webview>.
Now trigger a new window from the <webview> so the listener
fires.
In the listener, destroy the <webview>.
Every thing should work fine and should not exhibit any bad
behavior. Without this change it would throw a JavaScript
exception.

Review URL: https://codereview.chromium.org/558813002

Cr-Commit-Position: refs/heads/master@{#294053}

Mac creates a mini-framework in `ui_unittests Framework` so that ui
tests can get a resource bundle without depending on the Chrome
framework.

The way it generated locale information was causing build flakes since
two instances of repack_locales.py could be writing to
out/../gen/repack. Tests always use en-US, so instead just do what iOS
does and copy ui_test.pak to en.lproj/locale.pak and roll it into the
mini-framework bundle.

BUG=410468

Review URL: https://codereview.chromium.org/552963002

Cr-Commit-Position: refs/heads/master@{#294052}

https://chromium.googlesource.com/chromium/blink/+log/396df151b75a49950fcb0a1184f3520cf4627fbc..02fbb58eafa10f9ccc920c3f2683808342f83b5e
TBR=keishi@chromium.org,jianli@chromium.org

Review URL: https://codereview.chromium.org/556003004

Cr-Commit-Position: refs/heads/master@{#294051}

Delete the window immediately in WindowOverviewModeImpl::CloseDragWindow() and
use wm::ScopedHidingAnimationSettings to make then hide animation look right.
This fixes a potential crash if overview mode is exited while a window is
animating closed.

BUG=409724
TEST=None

Review URL: https://codereview.chromium.org/555463003

Cr-Commit-Position: refs/heads/master@{#294050}

I don't think we will ever roll back ANGLE to anything less than 126,
so it should be safe to do s0.

BUG=
TEST=tree
R=kbr@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/552013007

Cr-Commit-Position: refs/heads/master@{#294049}

BUG=405075, 386145

Review URL: https://codereview.chromium.org/549193002

Cr-Commit-Position: refs/heads/master@{#294048}

This mainly involves adjusting chrome_tests_unit.gypi to exclude
chrome's unit_tests that use views UI not enabled by default with
toolkit-views=1 on Mac.

Mostly the unit tests compile, but fail to link since their
corresponding UI parts are not in the browser.
(confirm_bubble_views_unittest.cc has a simple compile fix, but isn't
linked in yet).

An events_test_support dependency is needed in views_test_support for
ui/events/test/event_generator which some unit tests use, and otherwise
comes in through aura_test_support or views_unittests.

BUG=412234, 399191

Review URL: https://codereview.chromium.org/507373002

Cr-Commit-Position: refs/heads/master@{#294047}

When setting an exception we check that no other exceptions have been set.
However, sometimes it is preferable not to bail out of the
codepath immediately because the code becomes messier. In particular, if
the context is destroyed, an exception will be set which can happen upon
construction of the try/catch. Keeping this NOTREACHED means that we have to check
for an exception after the construction of every try/catch. I can't see a
problem with not checking the exception immediately - other operations will
fail and the result of the try-catch will eventually be set.

BUG=303491

Review URL: https://codereview.chromium.org/554173002

Cr-Commit-Position: refs/heads/master@{#294046}

Update BitmapCursorOzone type to store a vector of bitmaps, a hotspot,
and a frame delay. Plumb the vector through to the GPU process, where
we'll update the cursor on a timer. This is plumbing only; it does not
change the user-visible behavior of showing only the first cursor frame.

R=dnicoara, kenrb
BUG=343245
TEST=viewed css "wait" cursor on cros link_freon

Review URL: https://codereview.chromium.org/543643003

Cr-Commit-Position: refs/heads/master@{#294045}

There is no reason for sys/select.h to depend on
sdk_util headers.  The __BEGIN_DECLS/__END_DECLS
macros are available in all our toolchains.

This was causing various problems in the naclports
repo.

Review URL: https://codereview.chromium.org/539163004

Cr-Commit-Position: refs/heads/master@{#294044}

Underlying bugs have been fixed.

BUG=398337,401716

Review URL: https://codereview.chromium.org/557943002

Cr-Commit-Position: refs/heads/master@{#294043}

Otherwise, the old layout remains at the next time.

BUG=412505
R=sadrul@chromium.org
TEST=manually

Review URL: https://codereview.chromium.org/557083002

Cr-Commit-Position: refs/heads/master@{#294042}

It does not seem to run due to resource issues.

R=jamesr@chromium.org

Review URL: https://codereview.chromium.org/559763002

Cr-Commit-Position: refs/heads/master@{#294041}

BUG=none

Review URL: https://codereview.chromium.org/555223003

Cr-Commit-Position: refs/heads/master@{#294040}

This cl fixes a bug which was causing the final successful proxy
to be passed to NotifyProxyFallback rather than the proxy for which
fallback happened.

BUG=395769

Review URL: https://codereview.chromium.org/556933003

Cr-Commit-Position: refs/heads/master@{#294039}

Revert of [Android] Temporarily suppress RTL start&end check error. (patchset #1 id:1 of https://codereview.chromium.org/445753002/)

Reason for revert:
Android SDK fixed the issue, so we can convert back from warning to error.

Original issue's description:
> [Android] Temporarily suppress RTL start&end check error.
>
> The newest Android SDK had start&end attributes in non -v17 resources.
> Temporarily change the checking script from error to warning so that
> people can continue to work before it is fixed. This should be
> reverted once the issue is addressed
>
> BUG=None
>
> Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=287711

TBR=newt@chromium.org,tedchoc@chromium.org,dfalcantara@chromium.org
BUG=None

Review URL: https://codereview.chromium.org/555253004

Cr-Commit-Position: refs/heads/master@{#294038}

Known types occur very frequently. As it happens, we don't need to know
how prevalent unknown ones are in relation to known types.

BUG=412045
R=robertshield@chromium.org

Review URL: https://codereview.chromium.org/559633002

Cr-Commit-Position: refs/heads/master@{#294037}

This fixes most, but not all, errors identified by "gn check"

R=scottmg@chromium.org

Review URL: https://codereview.chromium.org/551103002

Cr-Commit-Position: refs/heads/master@{#294036}

Make the DataCompressionProxyDevRollout trial be able to override both the primary host and the fallback host of the data reduction proxy.

Related Chrome Android CL:
https://chrome-internal-review.googlesource.com/#/c/175697/

BUG=410961

Review URL: https://codereview.chromium.org/549153003

Cr-Commit-Position: refs/heads/master@{#294035}

BUG=412242
TEST=manual, see bug.

Review URL: https://codereview.chromium.org/549313003

Cr-Commit-Position: refs/heads/master@{#294034}

https://skia.googlesource.com/skia/+log/f5b6bf775c331784964bfcc9b9ac707dc6a7c62b..ef53f158de34ffaebd6537f1afd4adf38717987d

CQ_EXTRA_TRYBOTS=tryserver.blink:linux_blink_rel,linux_blink_dbg
TBR=fmalita@google.com

Review URL: https://codereview.chromium.org/555303002

Cr-Commit-Position: refs/heads/master@{#294033}

This changes the "test" template to create a shared_library (instead of
an executable) on Android. After this, we can actually run
base_unittests with
`build/android/test_runner.py gtest -s base_unnittests`
as normal (though may need to CHROMIUM_OUT_DIR and BUILDTYPE as
appropriate).

This requires adding the following targets:

//testing/android:native_test_native_code
//testing/android:native_test_util
//testing/android:native_test_jni_headers
//tools/android/md5sum:md5sum
//tools/android/md5sum:md5sum_bin
//tools/android/md5sum:md5sum_prepare_dist
//tools/android/md5sum:md5sum_copy_host

Also, makes it so that native executables are stripped (just like shared
libraries). Adds a simple create_native_execuatable_dist template that
sets up a dist directory for the executable (see
build/android/gyp/native_app_dependencies.gyp).

BUG=359249
TBR=rlarocque

Review URL: https://codereview.chromium.org/557463002

Cr-Commit-Position: refs/heads/master@{#294032}

BUG=388893

Review URL: https://codereview.chromium.org/555633003

Cr-Commit-Position: refs/heads/master@{#294031}