GitLab

The CL fixed a memory leak when requests are cancelled, but
caused crashes under certain circumstances.  The leaks will
still need to be fixed after this lands.

BUG=chrome-os-partner:27911,328092


> Fix a long standing leak on cancellation of renderer-
> initiated URLRequests.  ResourceDispatcher was not calling
> peer->OnCompletedRequest on cancellation of async requests,
> despite it being expected by WebURLLoaderImpl::Context
> (The only ResourceLoaderBridge::Peer).
>
> This shouldn't impact externally visible behavior at all,
> other than fixing the leak, because once
> WebURLLoaderImpl::Context::Cancel() is called, it will never
> call back into WebURLLoaderImpl.
>
> BUG=328092
>
> Review URL: https://codereview.chromium.org/113893007

Review URL: https://codereview.chromium.org/301753002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273149 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273148 0039d316-1c4b-4281-b951-d872f2087c98

r272637: Update libjpeg_turbo to use clz for bitcounting on ARM.
Review: https://codereview.appspot.com/97690043

BUG=None

Review URL: https://codereview.chromium.org/300553007

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273147 0039d316-1c4b-4281-b951-d872f2087c98

Changes StreamFramePacketOverhead to use offset in determining overhead.
Stream overhead changes as the offset changes.

Merge internal change: 67881361
https://codereview.chromium.org/304513006/


Fixes bugs in packet size computation in the creator and in the framer
mostly related to truncated ack size computation and verification. Fixes
a bug where sequence_number_length was incorrectly being allowed to
change mid-packet, and actively updates lengths in packets at
packet/FEC-group boundary. Adds tests to verify correct detection of ack
truncation.

Much yak hair was harvested, but no yaks were hurt in the making of
this CL.

Fixes bugs in packet-size computation in QuicPacketCreator and adds
tests.

Merge internal change: 67880898
https://codereview.chromium.org/300623009/


clang_tidy: remove redundant .get() calls on smart ptrs

Merge internal change: 67852120
https://codereview.chromium.org/304623002/


Update naming of members per style guideline with clang_tidy.
+ slowStart_  was changed to slow_start_.

Merge internal change: 67850333
https://codereview.chromium.org/300263003/


Added a helper function (SetIpInfoInCmsg) for WritePacket which fills in
the cmsg with the supplied self address.

This is part of "setting QoS on each quic sendmsg." internal CL.

Merge internal change: 67783235
https://codereview.chromium.org/300693006/


QUIC: rename GetWriteblockedStreams to GetWriteBlockedStreams.

Merge internal change: 67771815
https://codereview.chromium.org/302603007/


Change SendStreamId to take a stream ID argument in quic_session_test.cc

I think this makes the OnCanWriteBundlesStreams correct - it now writes
data with multiple stream IDs.

Also cleanup: use Invoke instead of InvokeWithoutArgs for calls which
have no args anyway...

Merge internal change: 67763247
https://codereview.chromium.org/307453006/


Rename QUIC's TransmissionInfo pending to in_flight and the associated
uses of pending in QuicUnackedPacketMap.

Merge internal change: 67762946
https://codereview.chromium.org/302783003/


Add a packet bundler in the internal server so that headers and body go
out in the same packet.

Tested by checking a small request is 1 packet; negative test fails.

Bundle headers with body in outgoing QUIC packets.

Merge internal change: 67729594
https://codereview.chromium.org/301563006/


Add an IncreaseLargestObserved to QuicUnackedPacketMap to remove the rtt
property from packets.

Merge internal change: 67725797
https://codereview.chromium.org/306573002/


Open a new port before closing the original in QUIC's port migration
EndToEndTest to ensure a new port is selected every time.

Merge internal change: 67703074
https://codereview.chromium.org/297213009/


Simplify QuicSentPacketManager MarkPacketHandled to take an iterator
instead of a sequence number and remove the case when it's called in
RetransmitUnackedPackets.

Merge internal change: 67652345
https://codereview.chromium.org/293383008/

R=rch@chromium.org

Review URL: https://codereview.chromium.org/300683008

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273146 0039d316-1c4b-4281-b951-d872f2087c98

This patch renames the cryptic fsp name to easier to understand provided "file
systems" in the C++ code. Also, the url has changed accordingly.

TEST=Tested manually by invoking chrome://provided-file-systems with a running
providing extension.
BUG=376095

Review URL: https://codereview.chromium.org/300023008

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273145 0039d316-1c4b-4281-b951-d872f2087c98

Revert of Content side of languagechange event fired on accept languages changes. (https://codereview.chromium.org/280953002/)

Reason for revert:
Broke two layout tests: fast/forms/validationMessage.html and fast/forms/email-idn-conversion.html


Original issue's description:
> Content side of languagechange event fired on accept languages changes.
> 
> This call the WebView API to notify the pages that the accept languages
> has changed and setup the test runner to allow Blink to change the
> accept languages for test purposes.
> 
> Depends on https://codereview.chromium.org/284793003/
> 
> BUG=365123
> 
> Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=273076

BUG=378129
NOTRY=true
TBR=mlamouri@chromium.org,jochen@chromium.org

Review URL: https://codereview.chromium.org/305633005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273144 0039d316-1c4b-4281-b951-d872f2087c98

BUG=356716
TEST=existing tests
TBR=pavely@chromium.org,kalman@chromium.org,arv@chromium.org

Review URL: https://codereview.chromium.org/290013011

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273140 0039d316-1c4b-4281-b951-d872f2087c98

There patch fixes the following two bugs.

1. The main cause: we assign null to metadata if we failed to fetch the data.
   But old code check the existance of data by using the 'in' operator and null
   value passes the check unexpectedlly.

2. Specifying OVER_FILL instead of FILL for an image in a tile of the mosaic
   view in case that the thumbnail image is samller than the tile.

BUG=375676
TEST=manually

Review URL: https://codereview.chromium.org/297323003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273139 0039d316-1c4b-4281-b951-d872f2087c98

If the requests are received before the bubble is shown, and there are main
frame requests pending, they will not be shown with the main frame requests.

Once the main frame requests have been handled, the iframe requests will be
shown in a new bubble.

All requests are flushed from the queues when a page is reloaded.

BUG=369685

Review URL: https://codereview.chromium.org/292453009

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273137 0039d316-1c4b-4281-b951-d872f2087c98

These is no platform implementation, with no available APIs on at least
Chrome OS and OS X to implement these, and Android does not support
Out of Band Pairing.

Thus remove the APIs, they can always be put back later if we find an
implementation.

BUG=376492

Review URL: https://codereview.chromium.org/293063015

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273136 0039d316-1c4b-4281-b951-d872f2087c98

It may fail on database errors. We should not crash Chrome for such errors.
The CHECK is replaced by a log message. We already handles  database error in general in OnDatabaseError, so no additional error handling is needed.

BUG=376884

Review URL: https://codereview.chromium.org/289343005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273135 0039d316-1c4b-4281-b951-d872f2087c98

This CL adds TileItemViews to the experimental app list which currently
show the first 5 apps in the app list. This CL also changes the width
of the placeholder bar on the start page to align with the tiles.

BUG=349727

Review URL: https://codereview.chromium.org/297643002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273134 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273128 0039d316-1c4b-4281-b951-d872f2087c98

Now we introduced create_nonsfi_test_nmf.py. So, by this CL
it starts to be used to generate nmf file for non-SFI ppapi_tests.
Also, this CL removes ppapi_nacl_tests_pnacl_nonsfi.nmf.

BUG=368949
TEST=Ran browser_tests --gtest_filter=*NonSfi* and trybots.

Review URL: https://codereview.chromium.org/302533007

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273127 0039d316-1c4b-4281-b951-d872f2087c98

This uses the RNDIS forwarder on mac since the custom forwarder is pretty far
from being able to compile on mac.

The main work of the patch was just making the RNDIS forwarder work with
ifconfig output instead of ip addr output.

BUG=355646

Review URL: https://codereview.chromium.org/294663015

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273125 0039d316-1c4b-4281-b951-d872f2087c98

Calculate non client hittest values by converting the corresponding system metrics returned via the GetSystemMetrics API on Windows to DIP.

The API returns pixel values which are compared against points in DIP which gives wrong non client hittest
values.

The bug only happens on glass.

BUG=377983
R=sky

Review URL: https://codereview.chromium.org/305623004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273124 0039d316-1c4b-4281-b951-d872f2087c98

I notcied this bug while trying to fix the WebGL VAO related regressions
on Linux AMD.

Before the generated ID is not set, so by default it's 0, which
is wrong, because 0 is reserved to unbind instead of bind.

Also, turn on VAO support for majority of tests as the emulated code path
is more isolated and its tests are also more isolated.  I feel it's better
to test with VAO on by default.

BUG=373974
TEST=gpu_unittests
R=vmiura@chromium.org,kbr@chromium.org,piman@chromium.org,jbauman@chromium.org

Review URL: https://codereview.chromium.org/303543005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273123 0039d316-1c4b-4281-b951-d872f2087c98

Refactors the autofill integration tests so they no longer need to rely
on snapshots or self-notifications.

Deletes the integration tests around disabling the type or disabling
sync, since we decided in the review of r269198 that those tests are not
valuable.

Introduces KeysMatchStatusChangeChecker, a fairly typical
MultiClientStatusChangeChecker based on the existing KeysMatch()
function.

Introduces ProfilesMatchStatusChecker, which might be the first
StatusChangeChecker to not register as an observer of the
ProfileSyncService.  Autofill profiles are stored and synced on the DB
thread but cached on the UI thread.  There's no guarantee that the UI
thread cache of autofill profile data would be up to date by the time
the sync cycle notificaiton is emitted.  Instead of listening for sync
notifications, the ProfileMatchStatusChangeChecker registers with the
owner of the UI thread cache of autofill profile data, the
PersonalDataManager, so it can be notified directly when the cache is
updated.

Since it uses a new notification mechanism, there is some reason to
believe this could fix issue 152551.

BUG=97780,95742,152551

Review URL: https://codereview.chromium.org/303643003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273122 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273121 0039d316-1c4b-4281-b951-d872f2087c98

http://build.chromium.org/f/chromium/perf/dashboard/ui/changelog_blink.html?url=/trunk&range=174878:174895&mode=html
TBR=alph@chromium.org,esprehn@chromium.org,eseidel@chromium.org
BUG=

Review URL: https://codereview.chromium.org/303693003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273120 0039d316-1c4b-4281-b951-d872f2087c98

- import checkdeps

BUG=none
TBR=brettw@chromium.org

Review URL: https://codereview.chromium.org/303503006

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273119 0039d316-1c4b-4281-b951-d872f2087c98

"image" is specified as a MIME type but is invalid. We should just omit
the MIME type parameter.

BUG=377409

Review URL: https://codereview.chromium.org/303483006

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273118 0039d316-1c4b-4281-b951-d872f2087c98

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273117 0039d316-1c4b-4281-b951-d872f2087c98

. validate on the client that the node can only be deleted or have its bounds changed by the connection that created the node. note that it's still possible for the client to call the interface directly, but the service will validate also, and then the client's inconsistent state is the client's own fault.
. pass bounds through the INodes array

R=sky@chromium.org
http://crbug.com/365012

Review URL: https://codereview.chromium.org/296133012

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273116 0039d316-1c4b-4281-b951-d872f2087c98

Use ExtensionRegistry instead.

R=kalman@chromium.org
BUG=376293

Review URL: https://codereview.chromium.org/298253002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273115 0039d316-1c4b-4281-b951-d872f2087c98

I also renamed classes to NestedAcceleratorXxx. I felt this is a bit more clearer than NestedDispatcher, especially in ui/wm/core. Please let me know if you disagree or have better suggestion. I'm happy to rename them.

BUG=None

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=272740

R=ben@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=272995

Review URL: https://codereview.chromium.org/298703007

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273114 0039d316-1c4b-4281-b951-d872f2087c98

BUG=

Review URL: https://codereview.chromium.org/300173002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273113 0039d316-1c4b-4281-b951-d872f2087c98

Added a histogram "Network.Shill.Wifi.FallbackDNSTestResult" for tracking
the result of the fallback DNS test. The fallback DNS test is performed
when portal detection is failed due to DNS failure.

BUG=chromium:366301
TEST=Start chrome, and browse to chrome://histograms to make sure
this histogram exist after connect to a wifi network without internet
access.

Review URL: https://codereview.chromium.org/291823003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273110 0039d316-1c4b-4281-b951-d872f2087c98

BUG=373098
NOTRY=true
R=kbr@chromium.org

Review URL: https://codereview.chromium.org/303503009

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273109 0039d316-1c4b-4281-b951-d872f2087c98

Omaha configuration parameters now use Wow6432Node (32-bit registry) on 64-bit installs.  GCAPI configuration parameters also use these same accesses since they read/write the same data.

Also, force installation of Win64 into C:\Program Files (x86) until the code is added to support moving the files to the right place during upgrade and downgrade.  Uninstall information will remain stored in the 32-bit registry (even on 64-bit) until the installation location is changed.

This fixes system installs on Win64.

BUG=338706
R=grt@chromium.org

Review URL: https://codereview.chromium.org/300593002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273108 0039d316-1c4b-4281-b951-d872f2087c98

BUG=355735

Review URL: https://codereview.chromium.org/296873002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273107 0039d316-1c4b-4281-b951-d872f2087c98

glBufferSubData 's second argument type is GLintptr, not GLint. And the third argument type is GLsizeiptr, not GLsizei.
A mistake in glBufferData and glBufferSubData 's arguments.

BUG=

Review URL: https://codereview.chromium.org/291493005

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273106 0039d316-1c4b-4281-b951-d872f2087c98

BUG=364141, 330264

Review URL: https://codereview.chromium.org/297183010

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273105 0039d316-1c4b-4281-b951-d872f2087c98

protocol servers.
++ Net.AlternateProtocolUsage.200Truncated
++ Net.AlternateProtocolUsage.1000Truncated

R=jar@chromium.org

Review URL: https://codereview.chromium.org/301743004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273103 0039d316-1c4b-4281-b951-d872f2087c98

This CL adds some extra overloads for CalculateKMeanColorOfBitmap and
CalculateKMeanColorOfPNG to provide a more convenient API. More
specifically, it provides both parameterized and unparameterized
versions of both functions.

BUG=376613

Review URL: https://codereview.chromium.org/291653004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273102 0039d316-1c4b-4281-b951-d872f2087c98

This CL adds the utilities needed for using the current WebContents when
distilling web pages.

BUG=361939

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=272611

Review URL: https://codereview.chromium.org/266073003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273101 0039d316-1c4b-4281-b951-d872f2087c98

Change the default parameters for ETLD+1 Rappor metrics.
Allow number of cohorts to be set per metric.
Modify metric names to distinguish from reports with old parameters.

BUG=370599

Review URL: https://codereview.chromium.org/264123004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273100 0039d316-1c4b-4281-b951-d872f2087c98

R=xhwang@chromium.org

Review URL: https://codereview.chromium.org/286883002

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273099 0039d316-1c4b-4281-b951-d872f2087c98

When applying animations we squash the transforms and therefore
end up applying them in reverse order. When we have to calculate bounds
we end up applying them one at a time, so we also have to apply this
in reverse order.

Also modified how we calculate bounds for arcs to simplify the derivation.

R=ajuma@chromium.org

Review URL: https://codereview.chromium.org/297163010

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273098 0039d316-1c4b-4281-b951-d872f2087c98

BUG=377555

Review URL: https://codereview.chromium.org/302463004

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@273097 0039d316-1c4b-4281-b951-d872f2087c98