GitLab

Bug: 779687
Change-Id: Iac2ee692d25aa811353749ad5aa77eb01f5e9d3e
Reviewed-on: https://chromium-review.googlesource.com/c/1437199
Commit-Queue: Weidong Guo <weidongg@chromium.org>
Reviewed-by: Xiyuan Xia <xiyuan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626295}

Most code doesn't care to distinguish between an error and cancelation,
as both abort the transfer. The few places that do care (e.g., the code
that decides whether to show an error to the user) can explicitly check
the type. This also allows cancellation to be easily propagated through
the error path from the source, instead of requiring a secondary
cancellation path.

Bug: 679313
Change-Id: I3c90ff55e7a58043535244a9d92cfc9e6814a4e6
Reviewed-on: https://chromium-review.googlesource.com/c/1437657
Commit-Queue: Erik Jensen <rkjnsn@chromium.org>
Reviewed-by: Joe Downing <joedow@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626294}

https://skia.googlesource.com/skia.git/+log/f6a2348535c9..fad5c773f7b6


git log f6a2348535c9..fad5c773f7b6 --date=short --no-merges --format='%ad %ae %s'
2019-01-25 brianosman@google.com Convert DM's -r JSON parsing to SkJSON


Created with:
  gclient setdep -r src/third_party/skia@fad5c773f7b6

The AutoRoll server is located here: https://autoroll.skia.org/r/skia-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.

CQ_INCLUDE_TRYBOTS=luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux-blink-rel;luci.chromium.try:linux-chromeos-compile-dbg;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win_optional_gpu_tests_rel
TBR=djsollen@chromium.org

Change-Id: I61843c42714cedbd88396fe16876e9091eac8fb3
Reviewed-on: https://chromium-review.googlesource.com/c/1437855Reviewed-by: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#626293}

https://chromium.googlesource.com/chromium/tools/depot_tools.git/+log/db34d87aff44..eb2767b2eb24


git log db34d87aff44..eb2767b2eb24 --date=short --no-merges --format='%ad %ae %s'
2019-01-25 recipe-roller@chromium.org Roll recipe dependencies (trivial).


Created with:
  gclient setdep -r src/third_party/depot_tools@eb2767b2eb24

The AutoRoll server is located here: https://autoroll.skia.org/r/depot-tools-chromium-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


TBR=agable@chromium.org

Change-Id: I22e5432f3ebf087a62cf679177df7b5f2b62e5c2
Reviewed-on: https://chromium-review.googlesource.com/c/1437722Reviewed-by: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#626292}

This is part of the work to finish up the One Build Bot Step migration to make all
of the performance tests use testing/scripts/run_performance_tests.py,
both gtest perf tests and Telemetry tests.

Isolates for these tests are configured in gn_isolate_map.pyl.

Those tests can be run in two contexts:
1. Performance waterfall testing. This is configured by
src/tools/perf/core/perf_data_generator.py

2. CQ/smoke testing. This is configured by
test_suites.pyl

This changelist places flags always needed by the isolate into gn_isolate_map.pyl,
and it places flags that can vary when an isolate is used in different
contexts into perf_data_generator.py and test_suites.pyl.

run_performance_tests.py requires the --non-telemetry=true flag for
non-telemetry tests so it knows how to
run them. Therefore, that flag needs to be in gn_isolate_map.pyl because
if it isn't passed in then the isolate cannot run. Furthermore, it is
non-optional and doesn't change in different contexts.

--migrated-test=true flag is always needed since all the tests are migrated
now. I will remove it in the next changelist and make this behavior the
default.

--gtest-benchmark-name can vary by context. For example the
command_buffer_perftests isolate should be named
passthrough_command_buffer_perftests when you pass it in
'--use-cmd-decoder=passthrough' but it should be called
validating_command_buffer_perftests when you pass it in
'--use-cmd-decoder=validating'.

This change is needed before
https://chromium-review.googlesource.com/c/chromium/src/+/1423840

TEST=running
python tools/mb/mb.py run -m chromium.perf -b "linux-builder-perf" -j 40 //out/Default tracing_perftests -- --isolated-script-test-output=$CTEMP/tracing/output.json --isolated-script-test-perf-output=$CTEMP/tracing/perf.json


BUG=757933, 920002, 739863

Change-Id: Idd63445951c9d63a41fe0537c351d95ab03da8b9
Reviewed-on: https://chromium-review.googlesource.com/c/1436559Reviewed-by: Brian Sheedy <bsheedy@chromium.org>
Reviewed-by: Stephen Martinis <martiniss@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626291}

If the primary profile is not manged, the kDevToolsAvailability pref
will default to 0 (allow, except for in force-installed extensions).
This matters *only* if there's a secondary profile with a more liberal
kDevToolsAvailabilityPref, because the primary profile's value would be
more retrictive then. Only evaluate the primary profile's value if it is
really managed.

      secondary profile with force-installed extensions. Set
      DeveloperToolsAvailability policy in the secondary profile's
      policy to 1 (allowed) and ensure that developer tools are usable
      in force-installed extensions.

Bug: 924973
Test: manual -- sign-in with an unmanaged primary profile and a managed
Change-Id: I1328a70664af759231ebc19bdb51a7ff1c193783
Reviewed-on: https://chromium-review.googlesource.com/c/1434316Reviewed-by: Dmitry Gozman <dgozman@chromium.org>
Reviewed-by: Lutz Justen <ljusten@chromium.org>
Commit-Queue: Pavol Marko <pmarko@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626290}

General testing idea:
For a network request, we store a token in stash. And after a fixed
period of time sufficient for the download to finish, if there are
any, we validate the token in the stash to see if the network request
was issued and finished successfully.

In the case of <a download> where the decision of download can be made
before resource fetching, the server sets the token immediately.

In the case of navigation to a download, the server will stream a
response over 1 seconds and set the token at the end only when the
socket has been connected. So it is able to detect any download
cancellation while fetching the resource.

Bug: 539938
Change-Id: I734c8cc18d1f761f16646c6b859a6b731ab40ff5
Reviewed-on: https://chromium-review.googlesource.com/c/1422667
Commit-Queue: Yao Xiao <yaoxia@chromium.org>
Reviewed-by: Mike West <mkwst@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626289}

BUG=873026

Change-Id: Ie0a4503a08d8f58a98bfaac619c4500debdf5d38
Reviewed-on: https://chromium-review.googlesource.com/c/1437523Reviewed-by: Demetrios Papadopoulos <dpapad@chromium.org>
Commit-Queue: Robert Liao <robliao@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626288}

Added HostResolverSource::LOCAL_ONLY setting to only allow local (eg
cache and HOSTS) resolution through the newer API. This will replace the
old ResolveFromCache() method (which was poorly named as it could do
other non-cache local sources).

Also turned |allow_cached_response| into |cache_usage| that adds a third
|STALE_ALLOWED| option. Setting this will replace the old
ResolveStaleFromCache() method, except that it doesn't do cache-only
unless also paired with LOCAL_ONLY.

EntryStaleness works a bit differently in the new API. Before it was
returned iff ResolveStaleFromCache() was used (and set to a default
non-stale value if the result doesn't come from the cache). Now it gets
set on the ResolveHostRequest object iff the result came from cache
(even if |STALE_ALLOWED| is not set). More consistent and more
reasonable for the newer pattern of results being on-demand through
Get*() methods rather that an out parameter.

Bug: 922699
Change-Id: Ia5c55beaf3c3c931a9503bb80e0b95f3334c5266
Reviewed-on: https://chromium-review.googlesource.com/c/1423577
Commit-Queue: Eric Orth <ericorth@chromium.org>
Reviewed-by: Tom Sepez <tsepez@chromium.org>
Reviewed-by: Asanka Herath <asanka@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626287}

https://chromium.googlesource.com/chromiumos/chromite.git/+log/19ec8ef577ff..1b0dd7244802


git log 19ec8ef577ff..1b0dd7244802 --date=short --no-merges --format='%ad %ae %s'
2019-01-25 jflat@chromium.org Add samus-kernelnext to internal release boards


Created with:
  gclient setdep -r src/third_party/chromite@1b0dd7244802

The AutoRoll server is located here: https://autoroll.skia.org/r/chromite-chromium-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


TBR=chrome-os-gardeners@chromium.org

Change-Id: Ic412f1df6d57728930e1b7b15884ad3453d80e24
Reviewed-on: https://chromium-review.googlesource.com/c/1436765Reviewed-by: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#626286}

Makes clang package ~7% smaller and lets the tot and upload bots
cycle ~7% faster. Since we don't use either of these features,
let's disable them.

Bug: 925145
Change-Id: I2570531caca6ca38173da35b8d4a76a925b31044
Reviewed-on: https://chromium-review.googlesource.com/c/1437255Reviewed-by: Hans Wennborg <hans@chromium.org>
Commit-Queue: Nico Weber <thakis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626285}

Bug: 916299
Change-Id: I533f0f9847892cdf2cdd5c49f9c7ba957c391298
Reviewed-on: https://chromium-review.googlesource.com/c/1437524
Commit-Queue: Brandon Jones <bajones@chromium.org>
Commit-Queue: Klaus Weidner <klausw@chromium.org>
Auto-Submit: Brandon Jones <bajones@chromium.org>
Reviewed-by: Klaus Weidner <klausw@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626284}

The settings ui should only show when feature is enabled.
Also hide the "delete" button for now per ux request.

Bug: b:123254533
Test: locally build and run
Change-Id: Ice44a9888b2d02f901206518967aea2d7bdb43a4
Reviewed-on: https://chromium-review.googlesource.com/c/1434755Reviewed-by: Steven Bennetts <stevenjb@chromium.org>
Commit-Queue: Xiaohui Chen <xiaohuic@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626283}

Removes some redundant code that appears to have been as
accidentally introduced by
https://chromium-review.googlesource.com/c/chromium/src/+/1409551

Change-Id: I784b6cb91f788dfc2e2c8bd3b1e888e8b54699b2
Reviewed-on: https://chromium-review.googlesource.com/c/1433144Reviewed-by: Camille Lamy <clamy@chromium.org>
Commit-Queue: Aaron Colwell <acolwell@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626282}

This turns on both features by default. They are slated to launch fully
via Finch in M72.

Also removes the obsolete flag for testing the Windows WebAuthn API
integration with an incompatible API version.

Bug: 897741,898718
Change-Id: I90b6d68331e74befb807aa7bc18e72ed253dc779
Reviewed-on: https://chromium-review.googlesource.com/c/1435760
Commit-Queue: Martin Kreichgauer <martinkr@google.com>
Reviewed-by: Kim Paulhamus <kpaulhamus@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626281}

Add //chrome/browser/safe_browsing to the allow_circular_includes_from
list for //chrome/browser/ui, and //chrome/browser/extensions. A real
solution would involve rather substantial refactoring of the Safe
Browsing code base. So do this to prevent build instability in the
meantime.

Bug: 919683, 925153
Change-Id: Iff0c9d0c67098c6255217721b55a390999e65737
Reviewed-on: https://chromium-review.googlesource.com/c/1436035
Commit-Queue: Daniel Rubery <drubery@chromium.org>
Reviewed-by: Peter Kasting <pkasting@chromium.org>
Reviewed-by: Istiaque Ahmed <lazyboy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626280}

https://pdfium.googlesource.com/pdfium.git/+log/578e24839231..a8c86c2b674e


git log 578e24839231..a8c86c2b674e --date=short --no-merges --format='%ad %ae %s'
2019-01-25 tsepez@chromium.org Delete "NULL terminator" CFXA_Node property/attribute records.


Created with:
  gclient setdep -r src/third_party/pdfium@a8c86c2b674e

The AutoRoll server is located here: https://autoroll.skia.org/r/pdfium-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


TBR=dsinclair@chromium.org

Change-Id: I815a46fd300522dec7aa1f7c40db9f57519ce059
Reviewed-on: https://chromium-review.googlesource.com/c/1437814Reviewed-by: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#626279}

They are never displayed, and it ends up requiring a bunch of null
checks elsewhere in the layout code.

Bug: 578349
Change-Id: I318c86c27200e5ea270c077d9e458dcdaa7e4b3a
Reviewed-on: https://chromium-review.googlesource.com/c/1431872
Commit-Queue: Daniel Cheng <dcheng@chromium.org>
Auto-Submit: Daniel Cheng <dcheng@chromium.org>
Reviewed-by: Chris Harrelson <chrishtr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626278}

This reverts commit c6df4d27.

Reason for revert: Broke Fuchsia browsertests.
Bug: 925574

Original change's description:
> [Fuchsia] Implement extra headers support in NavigationController.
> 
> This implements setting and using the |headers| data in
> LoadUrlParams. A new test fixture is added to ensure the extra headers
> are properly received by the test server.
> 
> Bug: 923598
> Change-Id: I59a44c886f1853bbc631ae016b0655dd1d4d539b
> Reviewed-on: https://chromium-review.googlesource.com/c/1427565
> Commit-Queue: Fabrice de Gans-Riberi <fdegans@chromium.org>
> Reviewed-by: Kevin Marshall <kmarshall@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#625912}

TBR=wez@chromium.org,kmarshall@chromium.org,fdegans@chromium.org

Change-Id: I82c9a5dead06edd58553a0999b708c6cc8b8dfd7
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: 923598
Reviewed-on: https://chromium-review.googlesource.com/c/1437914Reviewed-by: Fabrice de Gans-Riberi <fdegans@chromium.org>
Commit-Queue: Fabrice de Gans-Riberi <fdegans@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626277}

Bug: 925035
Change-Id: I2966cb89faa5a2c3db593f522b6cc8f50f064ab3
Reviewed-on: https://chromium-review.googlesource.com/c/1434754Reviewed-by: Ben Kelly <wanderview@chromium.org>
Reviewed-by: Nasko Oskov <nasko@chromium.org>
Commit-Queue: Łukasz Anforowicz <lukasza@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626276}

This change removes transaction error-tracking for IndexedDBCursor error
cases in which the cursor has been closed. Closing a cursor involves
nulling out the transaction associated with it, so attempting to track
the number of errors on the transaction after cursor close was causing a
null dereference.

Bug: https://crbug.com/925235
Change-Id: Ia0b854bca327f875d0f31f606cd636799f880137
Reviewed-on: https://chromium-review.googlesource.com/c/1437261
Commit-Queue: Andreas Butler <andreasbutler@google.com>
Commit-Queue: Daniel Murphy <dmurph@chromium.org>
Reviewed-by: Daniel Murphy <dmurph@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626275}

R=hcarmona@chromium.org
BUG=925278

Change-Id: Ia5c5f4770609a3794737732188312b5da4ebc43a
Reviewed-on: https://chromium-review.googlesource.com/c/1437272
Auto-Submit: Dan Beam (slow until 1/29) <dbeam@chromium.org>
Commit-Queue: Hector Carmona <hcarmona@chromium.org>
Reviewed-by: Hector Carmona <hcarmona@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626274}

Once an HTML validation alert appears, do not remove the errormessage
relation until the form control becomes valid. The disappearance of
the validation alert is a timeout for the convenience of sighted users,
but does not indicate whether the error message still applies.

Bug: 924996
Change-Id: Icb43dba7bb29f64fb7f4814dc86cb56c24ed33c9
Reviewed-on: https://chromium-review.googlesource.com/c/1434796Reviewed-by: Dominic Mazzoni <dmazzoni@chromium.org>
Commit-Queue: Aaron Leventhal <aleventhal@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626273}

Previously, it was possible to switch active sessions via hotkey and
keep Assistant open from the primary profile. In this case, LibAssistant
would not be running so interactions with Assistant would fail, but the
desired UX is to close Assistant completely in this case.

To address this, we associate the Assistant window with the active
user. The window manager will appropriately close the window in
response to profile changes.

Bug: b:121223805
Change-Id: I9253cd58aa90e97645513e1b640ec3c7f8d77ce5
Reviewed-on: https://chromium-review.googlesource.com/c/1385850
Commit-Queue: David Black <dmblack@google.com>
Reviewed-by: Xiyuan Xia <xiyuan@chromium.org>
Reviewed-by: Xiaohui Chen <xiaohuic@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626272}

As a step to make to use IdentityManager, this CL replaces all uses of ATS
with IdentityManager's ones in chrome_password_protection_service_unittest.cc.

TEST: All tests of ChromePasswordProtectionServiceTest should work well.
Bug: 922768

Change-Id: I551403c7d7c4333b1246fe4b5b933c12783637e7
Reviewed-on: https://chromium-review.googlesource.com/c/1433259Reviewed-by: Colin Blundell <blundell@chromium.org>
Reviewed-by: Nathan Parker <nparker@chromium.org>
Commit-Queue: Gyuyoung Kim <gyuyoung@igalia.com>
Cr-Commit-Position: refs/heads/master@{#626271}

Testing four site both in the main profile and incognito is slow and causing test timeouts.
Separate the tests.

Bug: 923917
Change-Id: I2f93b0a200dd0c811fa66681afc4ef6da9f81aa5
Reviewed-on: https://chromium-review.googlesource.com/c/1427619Reviewed-by: Tommy Li <tommycli@chromium.org>
Commit-Queue: Mustafa Emre Acer <meacer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626270}

Have RenderWidget clear itself as the MainThreadEventQueue client when
it is closing, in OnClose(). The queue already deals with trying to
handle an event when the client is gone.

However, it does not deal with the fact that when not handling the event
it must still run the mojo reply callback or suffer the wrath of a
DCHECK.

RenderWidget does deal with this fact already. But instead of dealing
with this callback behaviour at multiple levels, have the handle
methods return a boolean back up to Dispatch(), where if it gets
false it can run HandledEvent() with the failure args from a single
place.

Remove the closing early outs for MainThreadEventQueue methods in
RenderWidget.

R=bokan@chromium.org

Change-Id: I0f836ed4df801ce470afb10894c9c3caa70c5711
Bug: 912193
Reviewed-on: https://chromium-review.googlesource.com/c/1435520Reviewed-by: Navid Zolghadr <nzolghadr@chromium.org>
Reviewed-by: Avi Drissman <avi@chromium.org>
Reviewed-by: David Bokan <bokan@chromium.org>
Commit-Queue: danakj <danakj@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626269}

This CL converts download notification to use new builder API which
support general notification metric recording, including:

Mobile.SystemNotification.Action.Click
Mobile.SystemNotification.Content.Click
Mobile.SystemNotification.Dismiss

TBR=dtrainor@chromium.org

Bug: 898269
Change-Id: I0a0634d9fc4bc52da2258e54dca0de7cdd3d4f09
Reviewed-on: https://chromium-review.googlesource.com/c/1433137
Commit-Queue: Xing Liu <xingliu@chromium.org>
Reviewed-by: Min Qin <qinmin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626268}

WebUINavigationThrottle blocks WebUI (chrome://...) pages from loading
non-WebUI pages in subframes.

When a renderer navigates to a WebUI page, a "WebUI bindings" bit is set
in that renderer's ChildProcessSecurityPolicyImpl::SecurityState, and it
remains for the lifetime of the renderer process.
WebUINavigationThrottles are created for renderers with this bit set.

In multi-process mode, navigating away from the WebUI page creates a
fresh renderer with a fresh SecurityState, so subsequent non-WebUI pages
are again able to load non-WebUI subframes.

In single-process mode, the WebUI bit remains. After visiting a WebUI
page, iframes on all other pages will be broken until Chrome is
restarted.

This breaks Android WebView as well, since WebView uses single-process
on old Android versions and low-memory devices. It breaks CTS because
testOnSafeBrowsingHit loads chrome://safe-browsing/match?type=malware,
which sets the bit, and then testShouldOverrideUrlLoadingOnCreateWindow
tries to use an iframe.

Fix this by not creating a WebUINavigationThrottle in single-process
mode.

BUG=924799

Change-Id: I5a61f45dc065681f0e6b97e0b205e804edbd7402
Reviewed-on: https://chromium-review.googlesource.com/c/1432953
Commit-Queue: Paul Miller <paulmiller@chromium.org>
Reviewed-by: Nasko Oskov <nasko@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626267}

TBR=rouslan@chromium.org
BUG=924131

Change-Id: I4662993b74824089e8828a42be75cda6f913a3a4
Reviewed-on: https://chromium-review.googlesource.com/c/1437617Reviewed-by: Ted Choc <tedchoc@chromium.org>
Commit-Queue: Ted Choc <tedchoc@chromium.org>
Auto-Submit: Ted Choc <tedchoc@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626266}

See https://github.com/w3c/ServiceWorker/issues/1385

Bug: 924959
Change-Id: I5e2850b743d0702b36f1f20a84c87591c3baab19
Reviewed-on: https://chromium-review.googlesource.com/c/1433657Reviewed-by: Ben Kelly <wanderview@chromium.org>
Commit-Queue: Matt Falkenhagen <falken@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626265}

Add test to ensure that handling the fetch doesn't trigger stale
while revalidate loading.

Add test to ensure that scripts and css loaded trigger a stale
while revalidate cache hit.

The PR for the spec changes is here:
https://github.com/whatwg/fetch/pull/853

This reland removes the resource timing and moves to css (instead of
images) which were both unreliable.

BUG=348877

Change-Id: Ibabd8d3fd0295bedc8259594fc926da6ab5cfc43
Reviewed-on: https://chromium-review.googlesource.com/c/1434776
Commit-Queue: Dave Tapuska <dtapuska@chromium.org>
Reviewed-by: Ben Kelly <wanderview@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626264}

https://skia.googlesource.com/skia.git/+log/d41c1876d834..f6a2348535c9


git log d41c1876d834..f6a2348535c9 --date=short --no-merges --format='%ad %ae %s'
2019-01-25 reed@google.com remove tricky asint
2019-01-25 recipe-roller@chromium.org Roll recipe dependencies (trivial).
2019-01-25 halcanary@google.com DEPS: use our mirror for Nima-Cpp
2019-01-25 bsalomon@google.com Avoid checking extension string in GrGLGpu::onResetContext.
2019-01-25 ethannicholas@google.com fixed Metal matrix constructors
2019-01-25 brianosman@google.com Revert "Remove --readPath/-r from DM"


Created with:
  gclient setdep -r src/third_party/skia@f6a2348535c9

The AutoRoll server is located here: https://autoroll.skia.org/r/skia-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.

CQ_INCLUDE_TRYBOTS=luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux-blink-rel;luci.chromium.try:linux-chromeos-compile-dbg;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win_optional_gpu_tests_rel
TBR=djsollen@chromium.org

Change-Id: I606168691534d565d9a1bd9fbe5969ead8e7713d
Reviewed-on: https://chromium-review.googlesource.com/c/1437402Reviewed-by: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#626263}

This patch adds support to clear stored touch calibration and
association data from display prefs. To achieve this, any and all touch
data is cleared from the stored pref before adding new pref values. The
system already does this indirectly when it instantiates a new
dictionary pref value for the touch data to be stored. However if the
touch data map that the system is about to store is empty to begin with,
this instantiation of a new dictionary never happens and the change to
delete any stale stored touch data fails to propogate.

Bug: 925551
Change-Id: I7230e05d0106e87b9223b2b75d2883f2269cf8c5
Component: Display prefs, touch calibration
Reviewed-on: https://chromium-review.googlesource.com/c/1437519Reviewed-by: Ahmed Fakhry <afakhry@chromium.org>
Commit-Queue: Malay Keshav <malaykeshav@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626262}

https://crrev.com/624035 made all javascript url navigations async.
Some sites appear to depend on "javascript:''" behaving like
"about:blank" and committing synchronously.

The spec doesn't say anything about this, but given the real-world
breakage, I'm inclined to consider this a very narrow partial revert
of https://crrev.com/624035 while we figure out whether this is a spec
oversight or a case where we just need to encourage libraries to update.

Bug: 923585
Change-Id: Ia030d0a06b5d8f20df319916a1a02297d916e4db
Reviewed-on: https://chromium-review.googlesource.com/c/1435821
Commit-Queue: Nate Chapin <japhet@chromium.org>
Reviewed-by: Dmitry Gozman <dgozman@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626261}

It is uninteresting to fuzz messages received from the trusted broker process
to non-broker client processes. On Windows it's also effectively impossible
to safeguard against invalid HANDLE transmissions in such cases, and this is
the mode the existing channel fuzzers have effectively been emulating. The
result is crashy fuzzers.

This changes the Channel and NodeChannel fuzzers to more closely emulate the
trusted broker environment, where junk HANDLEs from untrusted clients can be
effectively detected and discarded without risk of crashing.

Bug: 925213
Change-Id: I562df5e3b97b3f109a649b34300cbbe18d2151f0
Reviewed-on: https://chromium-review.googlesource.com/c/1436887Reviewed-by: Oksana Zhuravlova <oksamyt@chromium.org>
Commit-Queue: Ken Rockot <rockot@google.com>
Cr-Commit-Position: refs/heads/master@{#626260}

1. make updates so that the ChromeVox menus are up to date (e.g. remove the prefix key command, remove the ChromeVox activate/deactivate command from the menus).
2. apply the |disallowOOBE| bit to appropriate commands; remove it from the help command.
3. add support for |disallowOOBE| to CommandHandler

Bug: 921803
Change-Id: Ia3ff73c90b2c4aa94a455ccd9cc15ec20f4c0a41
Reviewed-on: https://chromium-review.googlesource.com/c/1436016
Commit-Queue: David Tseng <dtseng@chromium.org>
Reviewed-by: Dominic Mazzoni <dmazzoni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626259}

TBR=piman@chromium.org

Change-Id: I8b7e62b44c596e95f14a4644c27bc4e99d883a5e
Reviewed-on: https://chromium-review.googlesource.com/c/1432293
Commit-Queue: danakj <danakj@chromium.org>
Reviewed-by: Albert J. Wong <ajwong@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626258}

Use the IdentityManager and observe events from IdentityManager::Observer
to replace all direct usage of AccountTrackerService's APIs, leaving for
now just a reference to AccountTrackerService, required to initialize the
chromeos::AccountMapperUtil private member from the constructor.

Bug: 922785, 922786
Change-Id: Ieec3c3818a6993739fa204b112add0e5cc2f7f7b
Reviewed-on: https://chromium-review.googlesource.com/c/1437194
Commit-Queue: Mario Sanchez Prada <mario@igalia.com>
Reviewed-by: Xiyuan Xia <xiyuan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626257}

This code change results in a nearly 50% speedup of the
line_breaks test in blink_perf.accessibility by avoiding
the use of Positions.

Also fixes CSS first-letter accessibility, the inline text
boxes were missing the first letter and in fact the
Layout NG code seemed to have extra logic to duplicate
the broken behavior.

Bug: 921789
Change-Id: I58a21de6286e040ba0c10a9a591932de52908fa5
Reviewed-on: https://chromium-review.googlesource.com/c/1419222Reviewed-by: Koji Ishii <kojii@chromium.org>
Reviewed-by: Yoshifumi Inoue <yosin@chromium.org>
Commit-Queue: Dominic Mazzoni <dmazzoni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#626256}