GitLab

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/660cc59..c7a9281

TBR=dstockwell@chromium.org,aboxhall@chromium.org

Review URL: https://codereview.chromium.org/665163004

Cr-Commit-Position: refs/heads/master@{#300404}

Speculatively re-enabling ServiceWorkerBrowserTest.Reload and ServiceWorkerBlackBoxBrowserTest.Registration

This doesn't reproduce neither locally nor on trybots after
running all night, and old test logs are inaccessible.
I'm going to re-enable these and see if they're still flaking.

BUG=387045,393486

Review URL: https://codereview.chromium.org/652743012

Cr-Commit-Position: refs/heads/master@{#300403}

Demoting a warning message to DLOG(INFO) due to high frequency.  Will
follow-up with a different CL to log the actual "unusual, we don't think
this should really happen" case at the WARNING level.

Changed the level of a warning message to dlog(info)

Merge internal change: 78113178
https://codereview.chromium.org/665883003/

Fixes two div by zero bugs in QUIC's BBR pacing code that caused server
crashes.

Fixes are in the internal source code.

Merge internal change: 78017204
https://codereview.chromium.org/670533004/

Simplify QuicUnackedPacketMap to expose AddSentPacket, instead of
AddPacket, OnRetransmittedPacket, and SetSent.

Merge internal change: 77986449
https://codereview.chromium.org/640853005/

Making a whole host of QUIC framing errors pre-decryption non-fatal.  If
someone is sending random junk prefixed with a real user's CID it should
not disrupt the connection.

making the QUIC connection resilliant to injected traffic

Merge internal change: 77961885
https://codereview.chromium.org/667573006/

Added helper methods to get/set previous_cached_network_params_.

Merging the following internal change:

Add varz to track accuracy of QUIC bandwidth predictions.

Plan is to use this data to calculate possible impact of using bandwidth
predictions, get an idea of how accurate our predictions are. Intended
varz output is something like:

.../spdy4_over_quic/stk_predicted_bw_accuracy_within_5_mins
.../spdy4_over_quic/stk_predicted_bw_accuracy_within_50_mins
.../spdy4_over_quic/stk_predicted_bw_accuracy_within_500_mins
.../spdy4_over_quic/stk_predicted_bw_accuracy_any_time
.../spdy4_over_quic/max_stk_predicted_bw_accuracy_within_5_mins
.../spdy4_over_quic/max_stk_predicted_bw_accuracy_within_50_mins
.../spdy4_over_quic/max_stk_predicted_bw_accuracy_within_500_mins
.../spdy4_over_quic/max_stk_predicted_bw_accuracy_any_time

Each of these is a map of 20 buckets, covering 10% chunks of the
[0%, 200%] range. For example, a bandwidth estimate from 25 minutes ago,
that turns out to be 125% of the observed bandwidth, will result in the
stk_predicted_bw_accuracy_within_50_mins_histogram{13} bucket being
incremented.

Merge internal change: 77877600
https://codereview.chromium.org/666003002/

Fix test which times out when FLAGS_quic_allow_more_open_streams is
true. Needed to update test to open additional streams to trigger server
close. Internal version of the test does something similar already.

Merge internal change: 77868811
https://codereview.chromium.org/661423005/

Restoring tests incorrectly removed in cr/77591875

Merge internal change: 77791119

Commented out headers_stream->HasBufferedData() call to fix flaky
HandshakeUnblocksFlowControlBlockedHeadersStream unit test.

BUG=423586
https://codereview.chromium.org/662233003/

Replace calls to set.count() with ContainsKey in QUIC.

Merge internal change: 77762867
https://codereview.chromium.org/665823004/

Allow QUIC's BBR TCP sender to pace based on min rtt and add an explicit
PacingRate() method to SendAlgorithmInterface for BBR.

Merge internal change: 77668996
https://codereview.chromium.org/664923003/

Removing QUIC_VERSION_18.
This also lets us remove the code for
 -hard-coded session limits, since those apply to v19 or greater
 -stats tracking % sessions with no flow control
 -the bulk of AdjustErrorForVersion since there's no outstanding
  versions with new errors

 Removing QUIC_VERSION_18.  Not flag protected.

Merge internal change: 77603158
https://codereview.chromium.org/663013003/

Change QUIC's pacing granularity from 5ms to 1ms to match the kernel's
default granularity.

Merge internal change: 77587738
https://codereview.chromium.org/665083004/

R=rch@chromium.org

Review URL: https://codereview.chromium.org/667763003

Cr-Commit-Position: refs/heads/master@{#300402}

BUG=422730

Review URL: https://codereview.chromium.org/665563002

Cr-Commit-Position: refs/heads/master@{#300401}

BUG=423645

Review URL: https://codereview.chromium.org/659883004

Cr-Commit-Position: refs/heads/master@{#300400}

In chrome OS, the file chooser dialog box may return paths that does not point to a native file. The CL lets FileSelectorHelper handle for the non-native paths and generate FileChooserFileInfo that refers non-native files.

BUG=126902
TEST=manually test to upload drive/local files

Review URL: https://codereview.chromium.org/663473002

Cr-Commit-Position: refs/heads/master@{#300399}

This is a reland of https://codereview.chromium.org/651983002/ which
was reverted as part of bug 424923.  It turns out this CL was not
related and should reland.

BUG=420159

Review URL: https://codereview.chromium.org/653753004

Cr-Commit-Position: refs/heads/master@{#300398}

A change in http://crrev.com/612573006 to ProfileSyncServiceAndroid
caused isSyncInitialized() to check whether sync was active instead of
whether the backend was initialized. This broke something in the disable
sync flow; probably an assertion.

BUG=423450

Review URL: https://codereview.chromium.org/666033002

Cr-Commit-Position: refs/heads/master@{#300397}

when passive auth fails

tl;dr: when passive auth fails, go to accounts.google.com/ServiceLogin rather than accounts.google.com/AddAccount

BUG=323327

Review URL: https://codereview.chromium.org/663643002

Cr-Commit-Position: refs/heads/master@{#300396}

Stacks grow slowly (via page faults) and never shrink.

TBR=rvargas
BUG=none

Review URL: https://codereview.chromium.org/640373006

Cr-Commit-Position: refs/heads/master@{#300395}

ChannelMojo::CanBeUsed() let the code enable the feature for specific
platform regardless of the command line switch.

TEST=none
R=viettrungluu@chromium.org
BUG=377980

Review URL: https://codereview.chromium.org/665993002

Cr-Commit-Position: refs/heads/master@{#300394}

BUG=267773

Review URL: https://codereview.chromium.org/666803003

Cr-Commit-Position: refs/heads/master@{#300393}

Cleanup to make future migration to std::move() easier.

BUG=423621

Review URL: https://codereview.chromium.org/665663002

Cr-Commit-Position: refs/heads/master@{#300392}

BUG=414920
R=rockot@chromium.org

Review URL: https://codereview.chromium.org/667053002

Cr-Commit-Position: refs/heads/master@{#300391}

BUG=chromedriver:944
TBR=stgao@chromium.org

Review URL: https://codereview.chromium.org/668783002

Cr-Commit-Position: refs/heads/master@{#300390}

On Android, each object can only have a single accessible text string
associated with it - so a text field can't simultaneously have a name
and a value. Have it always return the value if nonempty, and the
name/label otherwise.

BUG=425310

Review URL: https://codereview.chromium.org/661393005

Cr-Commit-Position: refs/heads/master@{#300389}

Re-enable the disabled AddBakedInConfigs test after poking  bake_in_configs.py with a stick.

Also, tweak error handling in bake_in_configs.py so we print all failures (even JSON parse errors), not just the first one.

BUG=

Review URL: https://codereview.chromium.org/656033008

Cr-Commit-Position: refs/heads/master@{#300388}

Specially being both locks.

BUG=none

Review URL: https://codereview.chromium.org/661213002

Cr-Commit-Position: refs/heads/master@{#300387}

there is no Athena port of AppListService (crbug.com/417571)

In addition this CL:
- Fixes crashes in webstorePrivate.completeInstall() when the Javascript
requests the "App installed" bubble to be shown. In practice this is only
requested for extensions which should not be enabled (but are enabled) in
Athena.
- Moves ExtensionInstallUI and CrxInstallerError to
extensions/browser/install so that they can be used by code in Athena
- Refactors ExtensionInstallUI to remove the static methods

BUG=414341,417571
TEST=Manual, see bug

Review URL: https://codereview.chromium.org/634313004

Cr-Commit-Position: refs/heads/master@{#300386}

The related CL:
https://codereview.chromium.org/657293002/

BUG=141116

Review URL: https://codereview.chromium.org/665583002

Cr-Commit-Position: refs/heads/master@{#300385}

BUG=

Review URL: https://codereview.chromium.org/658803004

Cr-Commit-Position: refs/heads/master@{#300384}

The DisplayVirtualKeyboard function on Windows 8 and beyond should not be displaying the OSK if a physical keyboard is attached to the machine.

Added a function IsKeyboardPresent locally to win_util.cc which uses the setup API's to enumerate
keyboard devices and check for the common ones like ACPI\PNP and HID\VID. This function returns
true if the number of keyboards is greater than 1. This is a hacky approach and is being done
for lack of a good way to determine the presence of a keyboard. The documented methods don't work
well.

BUG=335735

Review URL: https://codereview.chromium.org/644783005

Cr-Commit-Position: refs/heads/master@{#300383}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/43b8b36..b0e89dc

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

Review URL: https://codereview.chromium.org/642293006

Cr-Commit-Position: refs/heads/master@{#300382}

By using base::ScopedFD, the ownership of the returned FD becomes
clearer.

BUG=415294
R=brettw@chromium.org, jam@chromium.org, yzshen@chromium.org, sehr@chromium.org
TBR=gene@chromium.org

Review URL: https://codereview.chromium.org/621613002

Cr-Commit-Position: refs/heads/master@{#300381}

app_shell needs the client ID and oauth2 scopes for its identity API.

Just a refactor, no functional changes.

BUG=424653
TEST=extensions_unittests
TBR=xiyuan@chromium.org for header file rename touching chrome/browser/signin/easy_unlock_toggle_flow.cc

Review URL: https://codereview.chromium.org/671553002

Cr-Commit-Position: refs/heads/master@{#300380}

BUG=424997

Review URL: https://codereview.chromium.org/664023002

Cr-Commit-Position: refs/heads/master@{#300379}

BUG=423621

Review URL: https://codereview.chromium.org/657313003

Cr-Commit-Position: refs/heads/master@{#300378}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/c6bd766..660cc59

TBR=dstockwell@chromium.org,aboxhall@chromium.org

Review URL: https://codereview.chromium.org/647553007

Cr-Commit-Position: refs/heads/master@{#300377}

-Wtype-limits will throw warnings if an enum is used interchangably
with an integer. Replace some such usages with proper integer
constants.

BUG=424334
TEST=build, unit test

Review URL: https://codereview.chromium.org/662893002

Cr-Commit-Position: refs/heads/master@{#300376}

* Refactors the code to allow adding metrics as part of metrics provider
  interface
* Implements a function in android_metrics_provider to record a new
  histogram which records the current document mode. This is needed for
  better analysis in Hera. I am adding a new histogram because the
  semantics are a bit different from DocumentActivity.RunningMode. The
  previous histogram is used only to record eligible devices that have
  manually opted in or out. The new histogram will record for all devices
  regardless of whether they could possibly be in document mode.

BUG=418642

Review URL: https://codereview.chromium.org/658903002

Cr-Commit-Position: refs/heads/master@{#300375}

The LEA instruction within the inline assembly statement was throwing
off glibc's backtrace() function, because it lost track of where the
stack was.  The easy fix for this is to convert SyscallAsm() to simply
use the standard C calling convention on X86-64, and make it into a
normal C function call so the compiler can ensure CFI information is
correct for us.

While here, there's no need to use the "call/pop/addq" trick to
compute a PC-relative address because we have %rip-based addressing.
So simply use "lea 2f(%rip), %rax" to compute the return address (and
avoid branch mispredictions from desync'ing the call stack).

BUG=424973

Review URL: https://codereview.chromium.org/661393004

Cr-Commit-Position: refs/heads/master@{#300374}

Mean frame time is in other benchmarks, but it's a good
sanity check for understanding performance changes. So is
tasks-per-frame on each thread, so that is also added in this
patch.

BUG=None

Review URL: https://codereview.chromium.org/643343002

Cr-Commit-Position: refs/heads/master@{#300373}

BUG=425164

Review URL: https://codereview.chromium.org/646753003

Cr-Commit-Position: refs/heads/master@{#300372}

Mojo: mojob.sh: Add support for gn, and remove support for gyp and shared library (component) builds.

R=jamesr@chromium.org

Review URL: https://codereview.chromium.org/668603003

Cr-Commit-Position: refs/heads/master@{#300371}

This CL allows web pages to intercept the Ctrl+Shift+0 accelerator when no UI scaling is applied (which is the common case)

BUG=404473
TEST=None

Review URL: https://codereview.chromium.org/668573002

Cr-Commit-Position: refs/heads/master@{#300370}

Data reduction proxy statistics are accumulated in local
state preferences. To support desktop and other platforms
with multiple profiles, these are being ported to profile
preferences.

BUG=411373

Review URL: https://codereview.chromium.org/652743004

Cr-Commit-Position: refs/heads/master@{#300369}

There were several places in this file where an error code value
was checked to see if it was less than QUIC_NO_ERROR or
QUIC_STREAM_NO_ERROR, which are both 0. On some compilers this
causes a warning, because such checks are trivially true.

BUG=424334
TEST=net_unittests
R=rch@chromium.org

Review URL: https://codereview.chromium.org/662513007

Cr-Commit-Position: refs/heads/master@{#300368}

Remove references to cancel-button CSS class, which is not defined.

BUG=none
R=dbeam@chromium.org

Review URL: https://codereview.chromium.org/645223004

Cr-Commit-Position: refs/heads/master@{#300367}

bridged_native_widget_unittest.mm currently fails to compile on 10.7+
SDKs because there is no such thing as
NSWindowDidFailToEnterFullScreenNotification (whoops).

The failures don't go through NSNotificationCenter, but to the
NSWindowDelegate protocol directly. So, add a dummy protocol for 10.6
and, in the simulate fullscreen failure test, send
windowDidFailTo{Enter,Exit}FullScreen to the window delegate directly.

Call via `id` so the test compiles on 10.6 SDKs, but return early so the
tests pass when actually running on 10.6.

BUG=378134
TEST=By tracing the simulation test, and manually to generate an actual
failure.

Review URL: https://codereview.chromium.org/665823002

Cr-Commit-Position: refs/heads/master@{#300366}

R=lcwu@chromium.org,damienv@chromium.org
BUG=336640

Review URL: https://codereview.chromium.org/652353003

Cr-Commit-Position: refs/heads/master@{#300365}