GitLab

This pulls in the following Native Client changes:

r13932: (jvoung) Keep the perf data of the trusted version of performance_tests separate.
r13933: (jvoung) Enable arm callingconv test for varargs and va_list.
r13934: (jvoung) Allow 'opt' bots to use GN with is_debug=false ('dbg' bots get debug).
r13935: (nfullagar) Initial benchmark suite for toolchain side.
r13936: (nfullagar) Add third_party/computer_language_benchmarks_game to DEPS
r13937: (petarj) [MIPS] Add missing -mtls-use-call when building irt libraries
r13942: (yyanagisawa) Avoid using mismatch compilers from buildbot.
r13947: (shyamsundarr) allow avx/xop/fma instructions but disable all them via def file attributes.

BUG=none
TEST=browser_tests and nacl_integration
CQ_EXTRA_TRYBOTS=tryserver.chromium.linux:linux_rel_precise32,linux_nacl_sdk,linux_chromium_dbg

Review URL: https://codereview.chromium.org/667783002

Cr-Commit-Position: refs/heads/master@{#300355}

mojo_examlpes.gypi was removed in the following CL:
https://codereview.chromium.org/647263005

BUG=423889
TEST=NONE
R=jamesr@chromium.org

Review URL: https://codereview.chromium.org/642243005

Cr-Commit-Position: refs/heads/master@{#300354}

Release notes for Checkstyle 5.9
http://checkstyle.sourceforge.net/releasenotes.html

The most important part is all the fixes for Indentation check that allows
us to finally enable it for presubmit.

Also adding OWNERS file to checkstyle directory.

BUG=318404
TBR=yfriedman@chromium.org

Review URL: https://codereview.chromium.org/661153003

Cr-Commit-Position: refs/heads/master@{#300353}

When navigation away from a page fails, the displayed error page uses
the zoom level of the current page since the underlying RenderViewImpl
is re-used without resetting its zoom level before displaying the error
page.

This CL forces the RenderViewImpl's zoom level to be reset to 0 whenever
the url being navigated to is unkown to the RenderViewImpl.

BUG=422464

Review URL: https://codereview.chromium.org/663203002

Cr-Commit-Position: refs/heads/master@{#300352}

BUG=422142
R=asvitkine

Review URL: https://codereview.chromium.org/637953010

Cr-Commit-Position: refs/heads/master@{#300351}

This deadlock involves SyncBackendRegistrar running on the sync thread and at least two
workers - one of them (worker A) running on UI thread and
another (worker B) running on some other background thread.
In this particular case it was a password worker running on Password model thread.

The deadlock occurs when the SyncBackendRegistrar shutdown
code running in the sync thread blocks on a waitable event
set by worker B, at the same time the worker B's thread is
blocked for whatever reason waiting to make a synchronous
call into the UI thread, and at the same time the worker A
is blocked on a lock owned by the sync thread while it attempts
to unregister itself with SyncBackendRegistrar.

I think the main issue here is that the sync thread might
block on a worker. This is possible in only one situation -
when that worker hasn't have a chance to run any tasks because
its own thread was blocked doing something else.

The fix removes this dependency. The only reason this
blocking occurs is because ModelSafeWorker has a mechanism
for subscribing / unsubscribing to the MessageLoop descruction
which has to be done in the right thread context.
In order to unsubscribe the worker needs to remember the message loop it runs on which is done
in SetWorkingLoopToCurrent callback, which is called from a task
posted on the worker's thread. There is also a waitable
event which is signalled at that time. The even is used to
block ModelSafeWorker::UnregisterForLoopDestruction from
delegating the call on an uninitialized message loop which
is possible only when SetWorkingLoopToCurrent has never been called.

There is no need to block in UnregisterForLoopDestruction.
If the registration for even loop destruction hasn't been
been done yet (because the thread is blocked) then there
really no reason to wait for that and then immediately
unsubscribe. We can skip both subscription to the even loop
notification altogether and let the registrar know that
it is OK to remove the worker. That's essentially what the
fix does.

I added a test that verifies that SyncBackendRegistrar::Shutdown
doesn't block on a blocked model type thread anymore.

BUG=423078

Review URL: https://codereview.chromium.org/637413003

Cr-Commit-Position: refs/heads/master@{#300350}

Added a mobile and desktop testing in case the browser being tested has a
user agent that defaults to the tablet user agent.

BUG=421229

Review URL: https://codereview.chromium.org/640453002

Cr-Commit-Position: refs/heads/master@{#300349}

Enable RenderTextHarfBuzz on Chrome OS.

BUG=321868,376077,423791

Review URL: https://codereview.chromium.org/634983002

Cr-Commit-Position: refs/heads/master@{#300348}

On some Android devices, the sysfs-based C-State info contains extra blank
lines which will cause an exception in C-State stat parser in Android platform
backend if the power metric and a Monsoon Power Monitor is in use. More
precisely, /sys/devices/system/cpu/cpu*/cpuidle/state*/name contains two
trailing newlines on at least Samsung Galaxy Note 3 and S5.

This patch filters out these blank lines from the sysfs output and fixes nearby
style.

BUG=421008
TEST=SysfsPowerMonitorMonitorTest.testGetCpuStateForAndroidDevices

Review URL: https://codereview.chromium.org/635313002

Cr-Commit-Position: refs/heads/master@{#300347}

This change causes sites using SSLv3 to show a "mixed content" level warning on
the lock icon and for the section in the connection status to go red.

There's no message added to the connection status because, if we plan on
removing SSLv3 by default in Chrome 40, then we want to merge this to Chrome 39
and thus there's no time to get a translation.

BUG=419870

Review URL: https://codereview.chromium.org/662883002

Cr-Commit-Position: refs/heads/master@{#300346}

Disable direct write if font count in registry is greater than or equal to threshold (currently 1750).

R=scottmg,ananta,cpu,asvitkine
BUG=421305

Review URL: https://codereview.chromium.org/667013003

Cr-Commit-Position: refs/heads/master@{#300345}

This will pave the way for a future CL that completely removes all //content
dependencies from the class.

TBR=ben@google.com
BUG=371476
TEST=No visible impact.

Review URL: https://codereview.chromium.org/658073004

Cr-Commit-Position: refs/heads/master@{#300344}

Also move a class implementation out of feature_channel.h.

BUG=424746

Review URL: https://codereview.chromium.org/661593005

Cr-Commit-Position: refs/heads/master@{#300343}

BUG=425248
TBR=nasko@chromium.org

Review URL: https://codereview.chromium.org/657993004

Cr-Commit-Position: refs/heads/master@{#300342}

Define a new macro S_MODEBITS to represent the
mode bits settable by chmod/fchmod.

Add test to verify that file type cannot be changed
with chmod.

Review URL: https://codereview.chromium.org/671513002

Cr-Commit-Position: refs/heads/master@{#300341}

This fixes keypresses on Logitech K400. There's no key repeat yet.

BUG=412996

Review URL: https://codereview.chromium.org/654073002

Cr-Commit-Position: refs/heads/master@{#300340}

Don't spam the logs on Windows, Mac, or Android when
GetFontRenderParams() is called with a non-empty query;
RenderTextHarfBuzz sends non-empty queries and is being
enabled for more platforms now.

Continue to log an error if a family name is requested on
these platforms, though.

BUG=424971

Review URL: https://codereview.chromium.org/645073004

Cr-Commit-Position: refs/heads/master@{#300339}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/83506e0..c6bd766

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

Review URL: https://codereview.chromium.org/665913003

Cr-Commit-Position: refs/heads/master@{#300338}

Handle STAGE_INITIALIZATION_ERROR with a logging message

BUG=420111
TEST=manual

Review URL: https://codereview.chromium.org/656503005

Cr-Commit-Position: refs/heads/master@{#300337}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/33e6466..43b8b36

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

Review URL: https://codereview.chromium.org/664203002

Cr-Commit-Position: refs/heads/master@{#300336}

Extensions: Split the constants from the extensions_common target into its own extensions_common_constants target.

The extensions constants are used in too many places and very hard to
get rid of. Making it into its own target so it can be included in all
builds, even when extensions are not enabled.

BUG=424746

Review URL: https://codereview.chromium.org/652623003

Cr-Commit-Position: refs/heads/master@{#300335}

Some of the updated files got accidentally stashed and I needed to pop
and recheck that the content was up-to-date.

Review URL: https://codereview.chromium.org/647763004

Cr-Commit-Position: refs/heads/master@{#300334}

Revert of Revert of Drop CreateChildFrame messages when swapping out. (patchset #1 id:1 of https://codereview.chromium.org/663183003/)

Reason for revert:
This revert was supposed to be on a branch but happened on master.

Original issue's description:
> Revert of Drop CreateChildFrame messages when swapping out. (patchset #1 id:1 of https://codereview.chromium.org/649683007/)
>
> Reason for revert:
> It broke compile since this CL relies on https://crrev.com/cce56cd951f6685a0120db63418aa7e6d3df28f2 and it isn't in this branch.
>
> Original issue's description:
> > Drop CreateChildFrame messages when swapping out.
> >
> > There is a race condition in the current state of the code where in cross-process navigation we swap the existing RenderFrameHost with a new RenderFrameHost. If the existing host sends an IPC message to create a new child frame, it arrives on the IO thread, allocates a routing id based of the existing process (p1) and does a PostTask to the UI thread. If there is a CommitPending event either executing on the UI thread or in the task queue before the task posted from the IO thread, it will end up putting the existing RenderFrameHost in swapped out state (or waiting for swapped out). When the task to create a child frame is executed after that, it creates a new RenderFrameHost, but it uses the "current" process (p2), which is different than the process that sent the message (p1). This manifests sometimes as adding duplicate routing ids to RenderProcessHost and is in general really bad bug.
> >
> > BUG=415059, 423691, 381990
> >
> > Review URL: https://codereview.chromium.org/642813007
> >
> > Cr-Commit-Position: refs/heads/master@{#299939}
> > (cherry picked from commit dcdb02fa)
> >
> > Conflicts:
> > 	content/browser/frame_host/navigator_impl_unittest.cc
> >
> > R=creis@chromium.org
> >
> > Committed: https://chromium.googlesource.com/chromium/src/+/a6a21983b1a29ad11d30782c66d95facca65ab55
>
> TBR=creis@chromium.org
> NOTREECHECKS=true
> NOTRY=true
> BUG=415059, 423691, 381990
>
> Committed: https://crrev.com/d2c7272b80fe61a11cd2f04f17fb6caf52631010
> Cr-Commit-Position: refs/heads/master@{#300332}

TBR=creis@chromium.org,nasko@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=415059, 423691, 381990

Review URL: https://codereview.chromium.org/667813002

Cr-Commit-Position: refs/heads/master@{#300333}

Revert of Drop CreateChildFrame messages when swapping out. (patchset #1 id:1 of https://codereview.chromium.org/649683007/)

Reason for revert:
It broke compile since this CL relies on https://crrev.com/cce56cd951f6685a0120db63418aa7e6d3df28f2 and it isn't in this branch.

Original issue's description:
> Drop CreateChildFrame messages when swapping out.
>
> There is a race condition in the current state of the code where in cross-process navigation we swap the existing RenderFrameHost with a new RenderFrameHost. If the existing host sends an IPC message to create a new child frame, it arrives on the IO thread, allocates a routing id based of the existing process (p1) and does a PostTask to the UI thread. If there is a CommitPending event either executing on the UI thread or in the task queue before the task posted from the IO thread, it will end up putting the existing RenderFrameHost in swapped out state (or waiting for swapped out). When the task to create a child frame is executed after that, it creates a new RenderFrameHost, but it uses the "current" process (p2), which is different than the process that sent the message (p1). This manifests sometimes as adding duplicate routing ids to RenderProcessHost and is in general really bad bug.
>
> BUG=415059, 423691, 381990
>
> Review URL: https://codereview.chromium.org/642813007
>
> Cr-Commit-Position: refs/heads/master@{#299939}
> (cherry picked from commit dcdb02fa)
>
> Conflicts:
> 	content/browser/frame_host/navigator_impl_unittest.cc
>
> R=creis@chromium.org
>
> Committed: https://chromium.googlesource.com/chromium/src/+/a6a21983b1a29ad11d30782c66d95facca65ab55

TBR=creis@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=415059, 423691, 381990

Review URL: https://codereview.chromium.org/663183003

Cr-Commit-Position: refs/heads/master@{#300332}

We don't actually use the embedded test server and we don't need most
of what the file path conversion utilities provide. We only need to
support very simple URLs for loading off of file paths and we do not
need to support windows at all for mojo_shell.

Review URL: https://codereview.chromium.org/640403002

Cr-Commit-Position: refs/heads/master@{#300331}

BUG=424755
R=oshima@chromium.org
TEST=manually && athena_unittests

Review URL: https://codereview.chromium.org/668443002

Cr-Commit-Position: refs/heads/master@{#300330}

The default URLFileRequestJob::IsRedirectResponse implementation is incorrect
for URLRequestExtensionJob (chrome-extension:-scheme) and ExtensionResourcesJob (chrome-extension-resource:-scheme).

BUG=388852,424961

Review URL: https://codereview.chromium.org/650453003

Cr-Commit-Position: refs/heads/master@{#300329}

These tests currently crash.

R=viettrungluu@chromium.org

Review URL: https://codereview.chromium.org/637953009

Cr-Commit-Position: refs/heads/master@{#300328}

BUG=399090

Review URL: https://codereview.chromium.org/661653002

Cr-Commit-Position: refs/heads/master@{#300327}

Note: This change is a prerequisite for an upcoming change that will add
extensive interactive UI testing around Flash fullscreen, including the
Fullscreen-Within-Tab UI.

For all interactive_ui_tests, the UI widget being referenced by
VIEW_ID_TAB_CONTAINER was wrong whenever Flash fullscreen mode was
engaged.  This change tags the correct NSView with the identifier, and
gives it some simple focus-passing behavior, similar to
ui::views::WebView, to ensure that tests checking focus and/or
simulating a click on the tab container get the desired behavior.

Because the code changes to TabContentsContainerView are pretty much the
same, this change also addresses bug 421728, which makes clicking in the
"gray area" during Fullscreen-Within-Tab mode a request to give the tab
contents the focus.

BUG=403641,421728

Review URL: https://codereview.chromium.org/642143004

Cr-Commit-Position: refs/heads/master@{#300326}

NOTRY=true

Review URL: https://codereview.chromium.org/663823004

Cr-Commit-Position: refs/heads/master@{#300325}

In r298580, I added a bug where jumps to the very beginning of the data (bit
offset zero) would be rejected. This was found by fuzzing.

BUG=none

Review URL: https://codereview.chromium.org/663903002

Cr-Commit-Position: refs/heads/master@{#300324}

When the number of font files exceeds 1750, we fallback to this list.
Include Verdana, Georgia, and Trebuchet MS in this list until font
loading is properly fixed.

R=shrikant@chromium.org,ananta@chromium.org
BUG=421305

Review URL: https://codereview.chromium.org/668743002

Cr-Commit-Position: refs/heads/master@{#300323}

This is a reland of https://codereview.chromium.org/659533002/

Previous CL broke telemetry unittests and perf unittests because it
didn't handle properly shell commands which may end with a newline
character when check_rc is not None.

This revised patch strips any whitespace at the end of shell commands
to prevent this.

BUG=267773

Review URL: https://codereview.chromium.org/669573003

Cr-Commit-Position: refs/heads/master@{#300322}

I stumbled on this while working on the dff change.  It looks
like the only place this function is used is in the test.

I removed the test as well, since its only purpose was to check
that calling leak would indeed leak.

Review URL: https://codereview.chromium.org/652583002

Cr-Commit-Position: refs/heads/master@{#300321}

At that occasion, make it owned by rvargas@.

BUG=none

Review URL: https://codereview.chromium.org/662273003

Cr-Commit-Position: refs/heads/master@{#300320}

We will be using that function.

BUG=424244

Review URL: https://codereview.chromium.org/669753002

Cr-Commit-Position: refs/heads/master@{#300319}

changes to disk every 500ms instead of 5secs. We are hoping if QUIC is
used as an alternate protocol, we will persist that information to disk
quicker. Hopefully, this would improve the QUIC winning the alternate
protocol race.

R=rch@chromium.org

Review URL: https://codereview.chromium.org/667473003

Cr-Commit-Position: refs/heads/master@{#300318}

The old way involved start-up code reading a value and setting a global
in template_url_prepopulate_data. Since the value can be computed
on-demand from device-information, just fetch it as needed instead.

BUG=169106
TBR=thakis

Review URL: https://codereview.chromium.org/628263004

Cr-Commit-Position: refs/heads/master@{#300317}

When it retried uploads using Channels, it would restart the upload with
data from the middle of the file, since Channels can't be rewound.  It
now just fails those requests instead.

Also, it was impossible to set the method for uploads - the default
value (POST) would overwrite any method set by the embedder.

Adds a new set of upload tests using the embedded test server, as
uploads were basically untested before.  Chunked uploads are still not
being tested, as the test server doesn't support them yet.

BUG=424712

Review URL: https://codereview.chromium.org/640593004

Cr-Commit-Position: refs/heads/master@{#300316}