Offline Previews metrics will not be reported currently because the
MIME-type sniffing only allows text/html or application/xhtml+xml.
MHTML, entires pages saved as one file are mime-type multipart/related
(https://en.wikipedia.org/wiki/MIME#Multipart_subtypes). Offline
previews are treated as HTTP/HTTPS MHTML pages from the perspective of
the chrome page load metrics harness, and treated as HTML documents in
blink. This should mean that offline pages are the only class of MHTML
pages that will be tracked in the harness.

BUG=693776

Review-Url: https://codereview.chromium.org/2699213002
Cr-Commit-Position: refs/heads/master@{#451906}
(cherry picked from commit 07205c9)

Review-Url: https://codereview.chromium.org/2708213005 .
Cr-Commit-Position: refs/branch-heads/2987@{#642}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

…focus advancing (https://codereview.chromium.org/2295843006)

Revert reason:
Revert to get rid of the crash before having a proper fix.

TBR=xiyuan@chromium.org, stevenjb@chromium.org, achuith@chromium.org
BUG=692784
BUG=468969
TEST=crash doesn't happen now, leaving the bug in crbug.com/468969 comment 7 reopen.

Review-Url: https://codereview.chromium.org/2709593002
Cr-Commit-Position: refs/heads/master@{#451983}
(cherry picked from commit 26f0153)

Review-Url: https://codereview.chromium.org/2711793002 .
Cr-Commit-Position: refs/branch-heads/2987@{#643}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Also expose first_run::IsOnWelcomePage() so it can be reused.

BUG=693292
NOTRY=true
NOPRESUBMIT=true

Review-Url: https://codereview.chromium.org/2692203009
Cr-Commit-Position: refs/heads/master@{#451794}
(cherry picked from commit 31b2815)

Review-Url: https://codereview.chromium.org/2710073002
Cr-Commit-Position: refs/branch-heads/2987@{#644}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Fixed an issue where persistent pages would be deleted by storage manager
when clearing expired pages.

BUG=691023

Review-Url: https://codereview.chromium.org/2700563005
Cr-Commit-Position: refs/heads/master@{#451405}
(cherry picked from commit 0386b22)

Review-Url: https://codereview.chromium.org/2707363002 .
Cr-Commit-Position: refs/branch-heads/2987@{#645}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

…ong place

The previous code I introduced in DocumentMarkerController::setComposition()
was passing an offset measured in DOM children to addCompositionUnderlines()
(and it wasn't even measured correctly) when it should have been passing an
offset measured in plain text characters. Oops

I'd like to add a test case for this bug, unfortunately I'm having trouble
reproducing it in a test.

BUG=692699

Review-Url: https://codereview.chromium.org/2689233006
Cr-Commit-Position: refs/heads/master@{#451903}
(cherry picked from commit 6923513)

Review-Url: https://codereview.chromium.org/2709343002 .
Cr-Commit-Position: refs/branch-heads/2987@{#646}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

To discourage development patterns leading to bad user experience, we
block wasm compilation/instantiation through the sync APIs
for modules larger than a certain size.

Promise-based APIs still succeed; also, sync APIs succeed in worker
contexts.

BUG=693945

R=bradnelson@chromium.org,esprehn@chromium.org,jochen@chromium.org

Review-Url: https://codereview.chromium.org/2702953002
Cr-Commit-Position: refs/heads/master@{#451811}
(cherry picked from commit 50e18a2)

Review-Url: https://codereview.chromium.org/2713733002 .
Cr-Commit-Position: refs/branch-heads/2987@{#647}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

This is in preparation of v8 supporting that version only. It currently
supports both 0x1 and 0xD.

BUG=chromium:575167
BUG=chromium:694657

Review-Url: https://codereview.chromium.org/2710033002
Cr-Commit-Position: refs/heads/master@{#452165}
(cherry picked from commit 85df088)

Review-Url: https://codereview.chromium.org/2707263007 .
Cr-Commit-Position: refs/branch-heads/2987@{#648}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Currently, one preconnect job is allowed in-flight for one proxy server.
This CL changes it to allow one in-flight preconnect job for one
(proxy server, privacy mode) tuple. This is needed since requests that
use different privacy modes cannot share the same connection.

BUG=667471
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.android:android_cronet_tester

Review-Url: https://codereview.chromium.org/2648593002
Cr-Commit-Position: refs/heads/master@{#445392}
(cherry picked from commit 37e33d5)

Review-Url: https://codereview.chromium.org/2707053006 .
Cr-Commit-Position: refs/branch-heads/2987@{#649}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Remove GMS from upstream WebView (it's only needed downstream) and
create a google_play_services_package variable with the upstream GMS
location, so it can be overridden downstream.

BUG=662166

Review-Url: https://codereview.chromium.org/2697543004
Cr-Commit-Position: refs/heads/master@{#450780}
(cherry picked from commit 635a452)

Review-Url: https://codereview.chromium.org/2712863002 .
Cr-Commit-Position: refs/branch-heads/2987@{#650}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

We've seen enough breakages to justify reverting for now.
We'll try again in a later release.

BUG=624061

Review-Url: https://codereview.chromium.org/2693423008
Cr-Commit-Position: refs/heads/master@{#451933}
(cherry picked from commit 1cb8abd)

Review-Url: https://codereview.chromium.org/2710953005 .
Cr-Commit-Position: refs/branch-heads/2987@{#651}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

For the chrome os recovery tool, it used to use "/var/tmp" to store
the downloaded image and intermediate unzipped files. The folder is
on a disk partition with a max size of ~3GB. Since arc++, the image
size grows bigger over the size limitation. We decide to use the
current user's Download folder instead.

Cherry pick change list 2662103005 into M57.

BUG=687422
TEST=manually tested on a chromebox.
NOTRY=true
NOPRESUBMIT=true

Review-Url: https://codereview.chromium.org/2662103005
Cr-Commit-Position: refs/heads/master@{#449198}
(cherry picked from commit 1275040)

Review-Url: https://codereview.chromium.org/2711473009
Cr-Commit-Position: refs/branch-heads/2987@{#652}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Cr-Commit-Position: refs/branch-heads/2987@{#653}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

If text insertion causes focus to be lost, then FocusController calls
InputMethodController::willChangeFocus(), and eventually calls
finishComposingText(DoNotKeepSelection).

But DoNotKeepSelection actually moves the caret to the end of the current
composition, so the caret will end up at the start of the inserted text.
Simply switching to KeepSelection fixes this issue.

BUG=691666

Review-Url: https://codereview.chromium.org/2705183002
Cr-Commit-Position: refs/heads/master@{#452002}
(cherry picked from commit c203f3a)

Review-Url: https://codereview.chromium.org/2706403005 .
Cr-Commit-Position: refs/branch-heads/2987@{#654}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Clears out the external protocol data stored on Profile, when user
clears browsing history, and checks 'cookies and other site and plugin
data'.

BUG=457254
NOTRY=true
NOPRESUBMIT=true

Review-Url: https://codereview.chromium.org/2664253006
Cr-Commit-Position: refs/heads/master@{#451517}
(cherry picked from commit 561a9cd)

Review-Url: https://codereview.chromium.org/2710603007
Cr-Commit-Position: refs/branch-heads/2987@{#655}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

This updates from revision
48ff55eab9b82ebc88762eed45277448d1f4f79 (2016-11-29)

to revision
1b9ceb762bda9cba75066b8be3c73aed634fd82f (2017-02-23)

TBR=pkasting@chromium.org
BUG=686800

Review-Url: https://codereview.chromium.org/2712883002
Cr-Commit-Position: refs/heads/master@{#452364}
(cherry picked from commit 5c1bebe)

Review-Url: https://codereview.chromium.org/2713823002 .
Cr-Commit-Position: refs/branch-heads/2987@{#656}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

This bug occurs because during an animation we may change the amount of layout
overflow (via FrameView::recalcOverflowAfterStyleChange). Previously, this
would cause a change in scrollbar existance until r432965 made it so that the
FrameView would prevent scrollbar creation earlier in cases where they're known
to be unneeded. Changing scrollbar existence causes a layout so we'd implicitly
invalidate and repaint. Since that change, we (correctly) don't bother doing a
layout but because of that we no longer invalidate so the additionally exposed
background region doesn't get painted.

BUG=683814

Review-Url: https://codereview.chromium.org/2685883004
Cr-Commit-Position: refs/heads/master@{#452170}
(cherry picked from commit 717789e)

Review-Url: https://codereview.chromium.org/2711033004 .
Cr-Commit-Position: refs/branch-heads/2987@{#657}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

BUG=682067

Review-Url: https://codereview.chromium.org/2690613003
Cr-Commit-Position: refs/heads/master@{#449998}
(cherry picked from commit 34bda99)

Review-Url: https://codereview.chromium.org/2716603002 .
Cr-Commit-Position: refs/branch-heads/2987@{#658}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

This patch fixes to choose the correct hyphenation point when the word
has multiple hyphenation points on Android.

The line breaker computes the number of characters that can fit and
computes that last hyphenation point within the number. The number was
not used correctly in the Android hyphenation engine.

The fix is about 5 lines in |HyphenationMinikin::lastHyphenLocation|,
but the test isn't small. The hyphenation engine loads dictionaries
using a mojo service, but since the mojo service is not available in
unit tests, the test creates the engine from the dictionary file it
loaded.

BUG=692094

Review-Url: https://codereview.chromium.org/2692233005
Cr-Commit-Position: refs/heads/master@{#451519}
(cherry picked from commit dd1e643)

Review-Url: https://codereview.chromium.org/2713553007 .
Cr-Commit-Position: refs/branch-heads/2987@{#659}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

…bute usage

We need to analysis which sites are using autoplay muted by attribute.
This CL adds rappor metrics to record the eTLD+1 of the child frame and
top-level frame.

BUG=693889

Review-Url: https://codereview.chromium.org/2706593002
Cr-Commit-Position: refs/heads/master@{#452010}
(cherry picked from commit d0698b2)

Review-Url: https://codereview.chromium.org/2709013006 .
Cr-Commit-Position: refs/branch-heads/2987@{#660}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

In the cl https://codereview.chromium.org/2618383003/,
"set_separator_visibility to false" is called for all trays
that the separator is missing on the right side. I forget to
set it for the overview_tray.

TEST=MANUAL

BUG=692374

Review-Url: https://codereview.chromium.org/2702583002
Cr-Commit-Position: refs/heads/master@{#451514}
(cherry picked from commit c8cc142)

Review-Url: https://codereview.chromium.org/2703393011 .
Cr-Commit-Position: refs/branch-heads/2987@{#661}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Cr-Commit-Position: refs/branch-heads/2987@{#662}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

…for MSE."

The counts can be inferred by the sample volume for UnderflowDuration,
so there's no need to track both.

BUG=641633
TEST=manually checked histograms in chrome://histograms.

Review-Url: https://codereview.chromium.org/2707973005
Cr-Commit-Position: refs/heads/master@{#452113}
(cherry picked from commit 4f6d14d)

Review-Url: https://codereview.chromium.org/2713883003 .
Cr-Commit-Position: refs/branch-heads/2987@{#663}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Conceptually similar to the code removed in https://codereview.chromium.org/2594543004
but now takes padding into account so that the arrow isn't offset.

BUG=687079

Review-Url: https://codereview.chromium.org/2706273004
Cr-Commit-Position: refs/heads/master@{#452160}
(cherry picked from commit ac719a4)

Review-Url: https://codereview.chromium.org/2716503005 .
Cr-Commit-Position: refs/branch-heads/2987@{#664}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Allow inline install for user initiated fullscreen mode since this
mode requires that the user initiate it and due to a nuance with mac
windows where the green maximize button enters fullscreen.

BUG=690523

Review-Url: https://codereview.chromium.org/2693183002
Cr-Commit-Position: refs/heads/master@{#450753}
(cherry picked from commit e6d00f1)

Review-Url: https://codereview.chromium.org/2716693002 .
Cr-Commit-Position: refs/branch-heads/2987@{#665}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

…rd navigation.'

Cherry-picked patchset nwjs#1 and nwjs#2 from cl/2691693002

BUG=688047, 677356
Review-Url: https://codereview.chromium.org/2685803002
NOTRY=true
NOPRESUBMIT=true

Review-Url: https://codereview.chromium.org/2711123004
Cr-Commit-Position: refs/branch-heads/2987@{#666}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

I.e., how many players completed preload but never saw playback. This
will be handy in terms of guiding efforts to reduce the number of these
types of "useless" players. I.e. maybe deferred offscreen loading, etc.

BUG=694855
TEST=manual

Review-Url: https://codereview.chromium.org/2709903002
Cr-Commit-Position: refs/heads/master@{#452135}
(cherry picked from commit 35848da5131b081ee4751be9669230223baf7c43)

Review-Url: https://codereview.chromium.org/2711803006 .
Cr-Commit-Position: refs/branch-heads/2987@{#667}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

This is important in order to avoid exposing detached WebFrames through the blink public API.

BUG=691012

Review-Url: https://codereview.chromium.org/2706963004
Cr-Commit-Position: refs/heads/master@{#452079}
(cherry picked from commit ba0384c)

NOTRY=true
NOPRESUBMIT=true
TBR=dcheng@chromium.org,haraken@chromium.org

Review-Url: https://codereview.chromium.org/2713963002
Cr-Commit-Position: refs/branch-heads/2987@{#668}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

In this change, we examine the error code returned when one of the P2P channels
is closed.  If no error is reported, then we pass OK to the error callback,
otherwise we convert to CHANNEL_CONNECTION_ERROR as before.

I am looking into the correct long-term fix for the session-terminate message
being dropped, however this change should be safer and can be merged into our
M57 release to allow us to correct the incorrect logging errors sooner.

BUG=692165

Review-Url: https://codereview.chromium.org/2708263002
Cr-Commit-Position: refs/heads/master@{#451912}
(cherry picked from commit 92f2a8c)

Review-Url: https://codereview.chromium.org/2715833002 .
Cr-Commit-Position: refs/branch-heads/2987@{#669}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Fix a crash due to reentrancy in the widget while processing a theme changed notification after coming out of fullscreen mode

The crash occurs as views are removed/added while we are propagating theme changed notifications to the view hierarchy.

Proposed fix is to add a recursion guard flag processing_theme_changed_ which is set in
Widget::OnNativeThemeUpdated before propagating the theme changed notification and reset after
the function returns. We check the flag in Widget::FrameTypeChanged() and bail. This function is called becaue
Widget is registered as an observer on the BrowserView.

BUG=681525
TBR=sky
NOTRY=true
NOPRESUBMIT=true

Review-Url: https://codereview.chromium.org/2703933002
Cr-Commit-Position: refs/heads/master@{#451505}
(cherry picked from commit f7c7217)

Review-Url: https://codereview.chromium.org/2711853004
Cr-Commit-Position: refs/branch-heads/2987@{#670}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Cr-Commit-Position: refs/branch-heads/2987@{#880}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

BUG=694382,698622

Review-Url: https://codereview.chromium.org/2742853003
Cr-Commit-Position: refs/heads/master@{#457363}
(cherry picked from commit 746da1c)

Review-Url: https://codereview.chromium.org/2775133002 .
Cr-Commit-Position: refs/branch-heads/2987@{#881}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Merge to M57.

BUG=704783

Review-Url: https://codereview.chromium.org/2775893002
Cr-Commit-Position: refs/heads/master@{#459554}
(cherry picked from commit e5d0b0e)

Review-Url: https://codereview.chromium.org/2780653002 .
Cr-Commit-Position: refs/branch-heads/2987@{#882}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Cr-Commit-Position: refs/branch-heads/2987@{#883}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Cr-Commit-Position: refs/branch-heads/2987@{#884}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

node-set is polymorphic; if called with a string, it returns a
node-set containing a text node with that content. The libxslt roll in
r425586 broke this behavior because Blink's fork of the node-set
extension returned a text node without an owner document. Fix that.

BUG=689977

Review-Url: https://codereview.chromium.org/2750943004
Cr-Commit-Position: refs/heads/master@{#457692}
(cherry picked from commit 3f6e482)

Review-Url: https://codereview.chromium.org/2784623002 .
Cr-Commit-Position: refs/branch-heads/2987@{#885}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Cr-Commit-Position: refs/branch-heads/2987@{#886}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

Cr-Commit-Position: refs/branch-heads/2987@{#887}
Cr-Branched-From: ad51088-refs/heads/master@{#444943}

The default background color of window is white, a flash of white
will be shown if the loaded page uses a dark background. So add a new
property named background_color to CreateWindowOptions that can be configured
the same as page dark background instead of white default color.