GitLab

The gin::Runner should be destroyed when the JS context is so that its
weak pointers are invalidated and any outstanding callbacks that
interact with the JS context can detect its destruction. Currently,
ScriptContext implements gin::Runner but is not immediately destroyed
when the JS context is. This CL moves this gin::Runner implementation
into a separate class which is destroyed when the ScriptContext is
invalidated.

Review URL: https://codereview.chromium.org/831423005

Cr-Commit-Position: refs/heads/master@{#311358}

A bug in onc_translator_onc_to_shill.cc is triggering a DCHECK when
a property is set in a previously non-existent dictionary.

This is a trivial fix to code that is subtle but clearly incorrect
by inspection.

BUG=none

Review URL: https://codereview.chromium.org/838043003

Cr-Commit-Position: refs/heads/master@{#311357}

ForkWithFlags is a wrapper around the clone syscall that uses the libc clone
wrapper, and thus updates the libc's pid cache if it has one (using sys_clone
directly does not update the pid cache, so getpid may return an incorrect
result in the child). This exposes the ability to set clone flags, which is
needed to use Linux namespaces.

BUG=312380

Review URL: https://codereview.chromium.org/840893003

Cr-Commit-Position: refs/heads/master@{#311356}

Fixing incorrect handling of out of range descriptors passed as newfd to dup2.

Fixing incorrect free descriptor management when using dup2 on a previously
allocated descriptor value.

Replacing a heap with a std::set, to allow descriptor deletion from the free
set in O(lg n).

Dropping impossible case where dup2ing a null handle.

TEST=Added two nacl_io_tests.
BUG=None
R=binji@chromium.org

Review URL: https://codereview.chromium.org/849743002

Cr-Commit-Position: refs/heads/master@{#311355}

BUG=417463

Review URL: https://codereview.chromium.org/848963002

Cr-Commit-Position: refs/heads/master@{#311354}

On ChromeOS the accelerometer needs to be read by both ash/ and content/

This change removes the ownership of AccelerometerReader from Shell, changing it into a Singleton.

TEST=Confirmed that a device with an accelerometer enters maximize mode.
BUG=342908

Review URL: https://codereview.chromium.org/822503003

Cr-Commit-Position: refs/heads/master@{#311353}

If a ui::TouchEvent is missing a radius on one axis, use the value from the other axis.

BUG=406004

Review URL: https://codereview.chromium.org/822803005

Cr-Commit-Position: refs/heads/master@{#311352}

BUG=

Review URL: https://codereview.chromium.org/852623002

Cr-Commit-Position: refs/heads/master@{#311351}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/199ba8e..bb928a0

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

Review URL: https://codereview.chromium.org/808423003

Cr-Commit-Position: refs/heads/master@{#311350}

This forces WeakPtrs to be invalidated before other members' destructors run.
See the bug or the documentation of WeakPtrFactory for more information. I'm
in the process of trying to turn on a clang check for this (see the bug).

BUG=303818
R=morrita@chromium.org

Review URL: https://codereview.chromium.org/848533003

Cr-Commit-Position: refs/heads/master@{#311349}

Focus was not maintained because the elements in the row did not have focus when the menu was displayed. Change makes it so that the delete happens when the menu is no longer displayed. This allows the normal code path to run and maintain focus.

Updated unit test so that the appropriate event is tested.

BUG=431029

Review URL: https://codereview.chromium.org/751763003

Cr-Commit-Position: refs/heads/master@{#311348}

If the tab is instructed to print while the native side is not properly
initialized, it will now fail gracefully instead of crashing.

BUG=445467

Review URL: https://codereview.chromium.org/837773008

Cr-Commit-Position: refs/heads/master@{#311347}

R=qsr

Review URL: https://codereview.chromium.org/835223004

Cr-Commit-Position: refs/heads/master@{#311346}

"pushed_it" created early can point to expired entry.
DeleteExpiredPushedStreams will remove actual entry invalidating "pushed_it"
iterator. std::map::insert will crash when invalid iterator passed as hint.

Swap inserting new element and removing expired ones so iterator will stay
valid until late.

BUG=443490
R=bnc@chromium.org, rch@chromium.org

Review URL: https://codereview.chromium.org/813993002

Cr-Commit-Position: refs/heads/master@{#311345}

BUG=448521
TBR=enne@chromium.org

Review URL: https://codereview.chromium.org/851803002

Cr-Commit-Position: refs/heads/master@{#311344}

Multiple implementations of this functionality exist with slight modifications. This CL consolidates them.

Note: This is a reland of https://codereview.chromium.org/823143004/, which was reverted due to memory leaks, which are now fixed.

BUG=329911
TBR=phajdan.jr@chromium.org,thestig@chromium.org,stevenjb@chromium.org  # Reland.

Review URL: https://codereview.chromium.org/851703002

Cr-Commit-Position: refs/heads/master@{#311343}

Get rid of flakiness in android_dbg_tests_recipe bots.

BUG=448518
TEST=
TBR=dmazzoni@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/847263002

Cr-Commit-Position: refs/heads/master@{#311342}

This is the first CL to move the test template to //testing/test.gni

This CL is just creating an empty test.gni file. It is needed because
the test template is used in Blink, so we need to makes this change in
stages.

R=cjhopman@chromium.org

Review URL: https://codereview.chromium.org/850743003

Cr-Commit-Position: refs/heads/master@{#311341}

https://chromium.googlesource.com/angle/angle/+log/04184fb046..9e0478f62

BUG=447140

Review URL: https://codereview.chromium.org/847253002

Cr-Commit-Position: refs/heads/master@{#311340}

The active user is always Guest. This is dead code from way back when
the User Manager was displayed under a specific, regular profile.

BUG=NONE

Review URL: https://codereview.chromium.org/788993006

Cr-Commit-Position: refs/heads/master@{#311339}

TBR=miu@chromium.org
BUG=448511

Review URL: https://codereview.chromium.org/841873005

Cr-Commit-Position: refs/heads/master@{#311338}

GBM surface mode requires synchronous page flips at the moment. We need
to set DriWrapper to synchronously page flip.

BUG=none
NOTRY=true

Review URL: https://codereview.chromium.org/847213002

Cr-Commit-Position: refs/heads/master@{#311337}

Instead of storing information in TouchFactory (which is X11 specific), store
the maximum number of touch-points supported by a device in platform agnostic
TouchscreenDevice. Use this TouchscreenDevice to compute the maximum number of
simultaneous touch-points supported at runtime.

BUG=none
R=dnicoara@chromium.org, oshima@chromium.org, spang@chromium.org

Review URL: https://codereview.chromium.org/845413002

Cr-Commit-Position: refs/heads/master@{#311336}

Refactor code converting located (e.g. VKEY_LSHIFT) and non-located
(e.g. VKEY_SHIFT) versions of ui::KeyboardCode.

BUG=444048

Review URL: https://codereview.chromium.org/796003005

Cr-Commit-Position: refs/heads/master@{#311335}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/9877b72..2a7b861

TBR=hayato@chromium.org,zmo@chromium.org

Review URL: https://codereview.chromium.org/852613002

Cr-Commit-Position: refs/heads/master@{#311334}

Make the Chrome OS implementation of
CheckIdleStateIsLocked() (used by the chrome.idle API) get
the lock state from chromeos::SessionManagerClient instead
of depending on ash::SessionStateDelegate (which reaches
into chrome/browser/chromeos/login/). This makes it possible
for the idle code to be later moved under ui/base/ instead
of living in chrome/browser/, permitting the chrome.idle API
to be moved from chrome/ to extensions/.

As part of this, make SessionManagerClient cache the
most-recently-received-from-session_manager screen-lock
state. Previously it just passed state updates to its
observers, but the idle functions lack a class that can
register itself as an observer.

BUG=446320

Review URL: https://codereview.chromium.org/852453005

Cr-Commit-Position: refs/heads/master@{#311333}

BUG=None

Review URL: https://codereview.chromium.org/835683004

Cr-Commit-Position: refs/heads/master@{#311332}

R=rch@chromium.org

Review URL: https://codereview.chromium.org/852513002

Cr-Commit-Position: refs/heads/master@{#311331}

This CL is to fix crash in cronet builder,
http://build.chromium.org/p/chromium.fyi/builders/Android%20Cronet%20Builder/builds/6248

Review URL: https://codereview.chromium.org/848933002

Cr-Commit-Position: refs/heads/master@{#311330}

The newlib version of getcwd in libnacl now supports
passing NULL of buffer, and recent version of gtest
will gracefully handle getcwd failure in static
constructors so we no longer need our own getcwd
implementation.

This is the second attempt and landing this change.  The
first was reverted:
https://codereview.chromium.org/756333005

R=bradnelson

Review URL: https://codereview.chromium.org/851723004

Cr-Commit-Position: refs/heads/master@{#311329}

Upstreams parts of AutocompleteController dealing with voice
suggestions. Upstreams VoiceSuggestionProvider and
VoiceSuggestionProviderTest.

BUG=428863

Review URL: https://codereview.chromium.org/789103011

Cr-Commit-Position: refs/heads/master@{#311328}

Testing for this fix to be added by
https://codereview.chromium.org/845863002/

BUG=448269

Review URL: https://codereview.chromium.org/848823002

Cr-Commit-Position: refs/heads/master@{#311327}

Previously, if CHROMIUM_OUT_DIR was set but blank, it could (and did!)
delete your entire source directory, including .git folder.

BUG=446921

Review URL: https://codereview.chromium.org/832933004

Cr-Commit-Position: refs/heads/master@{#311326}

The OpenSSL and BoringSSL ports support one additional CHACHA20_POLY1305
variant. Add it to the list so it gets categorized correctly.

(diff is large because presubmit wanted a git cl format.)

BUG=none

Review URL: https://codereview.chromium.org/809743005

Cr-Commit-Position: refs/heads/master@{#311325}

BUG=447299

Review URL: https://codereview.chromium.org/851783002

Cr-Commit-Position: refs/heads/master@{#311324}

options. Deprecates quic_prober.cc in the process as all (used)
functionality now also in quic_client_bin.cc.

Uses std::cout/cerr instead of LOG() statements, and is has hopefully
more easily parseable output (especially for those off-team) than the
previous tool.

Removed the "print_response" argument to QuicClient, instead QuicClient
has accessor's to the latest response code/headers/body if a user wants
to print details (the only user of this feature is quic_client_bin).

Examples of output:

$ quic_client http://www.google.com/ipz --quiet --host=74.125.137.103
Connected to 74.125.137.103:80
Request succeeded (200).

$ quic_client http://www.google.com/invalid --quiet --host=74.125.137.103
Connected to 74.125.196.105:80
Request failed (404).

$ quic_client http://www.example.com --host=93.184.216.34
Failed to connect to 93.184.216.34:80. Error: QUIC_CONNECTION_TIMED_OUT

$ quic_client http://www.google.com/ipz --host=74.125.137.103
Connected to 173.194.37.84:80
Request:
headers:
 :authority: http://www.google.com
 :method: GET
 :path: /ipz
 :scheme: http
body:

Response:headers:
 HTTP/1.1 200 OK
 Content-Length: 13
 content-type: text/html; charset=UTF-8
 date: Tue, 30 Dec 2014 20:34:50 GMT
 server: GFE/2.0
 x-google-backends: acdd12:80
 x-google-gfe-request-trace: acdd12:80
 x-google-gfe-response-code-details-trace: handling_ipz
 x-google-gfe-service-trace:
 x-google-service: _internal

body: 172.26.18.85

Request succeeded (200).

TODO: make --host optional by getting IP Address of URL's host.

Merge internal change: 83518478

R=rch@chromium.org

Review URL: https://codereview.chromium.org/847813005

Cr-Commit-Position: refs/heads/master@{#311323}

* Style changes to make the content look more like said site.
* Change the path to the css file to match said site so that pages may be duplicated with changing them.

BUG=

Review URL: https://codereview.chromium.org/844073002

Cr-Commit-Position: refs/heads/master@{#311322}

TBR=nduca

Review URL: https://codereview.chromium.org/853523002

Cr-Commit-Position: refs/heads/master@{#311321}

Observe OnExtensionLoaded to trigger update of chrome.commands keybindings for component extensions.

ExtensionRegistryObserver::OnExtensionWillBeInstalled only triggers when a component extension loads for the first time on a new profile. As a result, component extensions using the commands api never get their commands manifest block re-parsed on an update.

The fix is to observe load to trigger update of keybindings which differs from what happens in ExtensionKyebindingRegistry's OnExtensionLoaded; namely, the kExtensionBindings pref gets set in the former only.

BUG=440459
TEST=manually edit a component extension's manifest command keybinding. Load the component extension and observe the keybinding changes and triggers on the new keybinding.

Review URL: https://codereview.chromium.org/788973002

Cr-Commit-Position: refs/heads/master@{#311320}

BUG=448482
TBR=jamescook@chromium.org

Review URL: https://codereview.chromium.org/848083002

Cr-Commit-Position: refs/heads/master@{#311319}