GitLab

This rolls to a version of our mesa library that does not have the
CL, GLES2, EGL or VG headers.  We should never use CL/VG headers, we
should always use GLES2 and EGL headers from khronos.

Review URL: https://codereview.chromium.org/612523002

Cr-Commit-Position: refs/heads/master@{#297201}

Unset LDFLAGS before building. Otherwise we override the default LDFLAGS in the
makefile, breaking the build. We want to append our flags to the default flags
instead.

BUG=418175
TBR=glider@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/611083003

Cr-Commit-Position: refs/heads/master@{#297200}

This fixes a DCHECK failure on chrome://print, where no handler was
registered to handle chrome.send("dialogClose").

BUG=416783

Review URL: https://codereview.chromium.org/614683002

Cr-Commit-Position: refs/heads/master@{#297199}

If the fling update is close to the final state, then stop the fling. The
threshold for this used to be 1e-5f, but change it to 1e-1f, since there
won't be any visual changes for subsequent fling-update steps.

BUG=none
R=jdduke@chromium.org

Review URL: https://codereview.chromium.org/606093002

Cr-Commit-Position: refs/heads/master@{#297198}

It is a proprietary extension of the Web Manifest.

BUG=414873

Review URL: https://codereview.chromium.org/580513002

Cr-Commit-Position: refs/heads/master@{#297197}

Also, remove pepper_31 (the current oldest SDK) such
that the total number of SDKs we publish stays constant.

R=binji@chromium.org

Review URL: https://codereview.chromium.org/615573002

Cr-Commit-Position: refs/heads/master@{#297196}

Issue like 121085 or 320427 could be caused by writing request on busy sockets.

BUG=121085

Review URL: https://codereview.chromium.org/603783002

Cr-Commit-Position: refs/heads/master@{#297195}

TBR=maruel@chromium.org

Review URL: https://codereview.chromium.org/612113002

Cr-Commit-Position: refs/heads/master@{#297194}

While writing the unit tests I have found problems and fixed them. Since I am not sure when the unit tests will land however (they come with the new Athena Browser test framework) I created this patch to get a workable version for Sadrul's Overview mode work.

BUG=408837
TEST=tests are in different CL which introduces Browser Tests for Athena

Review URL: https://codereview.chromium.org/611773002

Cr-Commit-Position: refs/heads/master@{#297193}

Revert of Revert of Mojo: Convert some scoped_ptr<...>(new ...) to make_scoped_ptr(new ...) in mojo/{embedder,system}. (patchset #1 id:1 of https://codereview.chromium.org/609033002/)

Reason for revert:
scoped_ptr nullptr support was un-reverted.

Original issue's description:
> Revert of Mojo: Convert some scoped_ptr<...>(new ...) to make_scoped_ptr(new ...) in mojo/{embedder,system}. (patchset #2 id:20001 of https://codereview.chromium.org/610083002/)
>
> Reason for revert:
> scoped_ptr nullptr support was reverted.
>
> Original issue's description:
> > Mojo: Convert some scoped_ptr<...>(new ...) to make_scoped_ptr(new ...) in mojo/{embedder,system}.
> >
> > R=jamesr@chromium.org
> >
> > Committed: https://crrev.com/a60566701028a7344d17826076780eeb84b63052
> > Cr-Commit-Position: refs/heads/master@{#297095}
>
> TBR=jamesr@chromium.org,viettrungluu@chromium.org
> NOTREECHECKS=true
> NOTRY=true
>
> Committed: https://crrev.com/3f8f6ea2347da4437dfd56b567c7daddfb9d8940
> Cr-Commit-Position: refs/heads/master@{#297108}

TBR=jamesr@chromium.org,dcheng@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/613673004

Cr-Commit-Position: refs/heads/master@{#297192}

Revert of Revert of Mojo: Convert scoped_ptr<...>() to nullptr in mojo/embedder. (patchset #1 id:1 of https://codereview.chromium.org/593313006/)

Reason for revert:
scoped_ptr nullptr support was un-reverted.

Original issue's description:
> Revert of Mojo: Convert scoped_ptr<...>() to nullptr in mojo/embedder. (patchset #1 id:1 of https://codereview.chromium.org/606143003/)
>
> Reason for revert:
> scoped_ptr nullptr support was reverted.
>
> Original issue's description:
> > Mojo: Convert scoped_ptr<...>() to nullptr in mojo/embedder.
> >
> > R=jamesr@chromium.org
> >
> > Committed: https://crrev.com/1b198fa11edbe68701e12c27d5f4981398f5ca71
> > Cr-Commit-Position: refs/heads/master@{#297093}
>
> TBR=jamesr@chromium.org,viettrungluu@chromium.org
> NOTREECHECKS=true
> NOTRY=true
>
> Committed: https://crrev.com/76577a68922585f5621db8c63ba2ab6d5aa03f69
> Cr-Commit-Position: refs/heads/master@{#297109}

TBR=jamesr@chromium.org,dcheng@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/612923002

Cr-Commit-Position: refs/heads/master@{#297191}

* Making sure GCMAccountTracker is only depenedent on the GCMDriver
* Replacing GCMDesktopDriver.SetAccountsForCheckin with
  virtual GCMDesktop.SetAccountTokens
* Producing a vector of AccountTokenInfo in GCMAccountTracker

BUG=374969
R=jianli@chromium.org

Review URL: https://codereview.chromium.org/594383003

Cr-Commit-Position: refs/heads/master@{#297190}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/c06d9a7..4ceb039

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

Review URL: https://codereview.chromium.org/611003003

Cr-Commit-Position: refs/heads/master@{#297189}

When a volume is unmounted, while it was selected, then the list model is
updated, and since there is no selection anymore, the current entry from the
directory model is selected. However, the corresponding VolumeInfo is gone
so we can't call resolveRootEntry().

This patch adds a check. If the volume is gone, then the selection is aborted.

TEST=Tested manually with FSP.
BUG=418513

Review URL: https://codereview.chromium.org/595903003

Cr-Commit-Position: refs/heads/master@{#297188}

And replaces with a callback. Also renames SetBounds to SetSize.

BUG=none
TEST=none
R=ben@chromium.org

Review URL: https://codereview.chromium.org/607233002

Cr-Commit-Position: refs/heads/master@{#297187}

NOTRY=true
BUG=418600

Review URL: https://codereview.chromium.org/613693003

Cr-Commit-Position: refs/heads/master@{#297186}

According to specs hr tag should have xml role as separator.

BUG=384723

Review URL: https://codereview.chromium.org/610973002

Cr-Commit-Position: refs/heads/master@{#297185}

In order for CrOS Core services to speak Mojo, the mojo_shell will need to be able to accept incoming connections from externally-managed processes.

POR is to have the shell listen on a unix domain socket and convert incoming connections to a proper Mojo MessagePipe and then wire it to an InterfaceImpl<ExternalApplicationRegistrar>, which will make the appropriate calls on ApplicationManager.

BUG=418289
TEST=mojo_external_application_tests
STATUS=Fixed

Review URL: https://codereview.chromium.org/522443003

Cr-Commit-Position: refs/heads/master@{#297184}

On Debug ChromeOS bot, this takes 5 minutes.

BUG=414808
TBR=sky

Review URL: https://codereview.chromium.org/612673002

Cr-Commit-Position: refs/heads/master@{#297183}

Adding this handler will make authoring tests which exercise cross-site navigations much easier. This is harder in the current codebase, because testserver runs on a random port each time it is ran which requires boilerplate code in each test that needs it. Instead of computing the real URL in C++ code, just navigate using "/cross-site/hostname/path/and/more" and the server will redirect to the new host.
This requires browser tests to map all hosts to localhost, which is done by adding 'host_resolver()->AddRule("*", "127.0.0.1");' to the beginning of the test.

BUG=418236

Review URL: https://codereview.chromium.org/612533002

Cr-Commit-Position: refs/heads/master@{#297182}

This patch removes low quality mode for low resolution tiles. It also
cleansup some logic with tile versions, since there's only one now.

As a follow-up we can see if we should move the tile version stuff
onto tile directly, although I kind of prefer having a separate class
just for managing what mode we're drawing.

BUG=417876
R=enne, reveman, danakj

Review URL: https://codereview.chromium.org/603683006

Cr-Commit-Position: refs/heads/master@{#297181}

BUG=392145
TBR=benm

Review URL: https://codereview.chromium.org/459953002

Cr-Commit-Position: refs/heads/master@{#297180}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/c13bc57..c06d9a7

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

Review URL: https://codereview.chromium.org/612883002

Cr-Commit-Position: refs/heads/master@{#297179}

Quick Report reports they have fixed the issue and the dll now works
correctly with chromium.

BUG=382561

Review URL: https://codereview.chromium.org/617453002

Cr-Commit-Position: refs/heads/master@{#297178}

One (1!) line fix to remove Valgrind issue. Original review https://codereview.chromium.org/543753002/

Keeps the fetcher functionality as ImageFetcherImpl in chrome/browser/search/suggestions.

BUG=387751
TBR=blundell
TEST=ImageManagerTest,ImageFetcherBrowserTest

Review URL: https://codereview.chromium.org/607183002

Cr-Commit-Position: refs/heads/master@{#297177}

Fix issue with method with empty response.
Fix issue with local variable shadowing user parameters.

R=ppi@chromium.org

Review URL: https://codereview.chromium.org/607933002

Cr-Commit-Position: refs/heads/master@{#297176}

Mojom targets containing files referencing imported entities need to
depend on targets containing the imported .mojom files. This patch fixes
existing omissions.

BUG=417289

Review URL: https://codereview.chromium.org/611023002

Cr-Commit-Position: refs/heads/master@{#297175}

BUG=418258
TBR=pfeldman

Review URL: https://codereview.chromium.org/615533002

Cr-Commit-Position: refs/heads/master@{#297174}

This gets Put() closer to the cache spec.

Downstream of: https://codereview.chromium.org/606843002/

BUG=392621

Review URL: https://codereview.chromium.org/608593003

Cr-Commit-Position: refs/heads/master@{#297173}

BUG=418369
TBR=zelidrag@chromium.org

Review URL: https://codereview.chromium.org/607413004

Cr-Commit-Position: refs/heads/master@{#297172}

BUG=418572
TBR=dmazzoni@chromium.org

Review URL: https://codereview.chromium.org/607333003

Cr-Commit-Position: refs/heads/master@{#297171}

Revert of Rebaseline for findbugs for Android. (patchset #1 id:1 of https://codereview.chromium.org/611013002/)

Reason for revert:
This has created issues in other bots:
https://chromegw.corp.google.com/i/chromium.linux/builders/Android%20Builder%20(dbg)/builds/66548/steps/findbugs/logs/stdio

Looks like a different configuration is used on different bots, yielding a different output. Reverting for now.

Original issue's description:
> Rebaseline for findbugs for Android.
>
> BUG=418549
> NOTRY=true
>
> Committed: https://crrev.com/26fb9711a189ed3c49ae90873981e074c38534b0
> Cr-Commit-Position: refs/heads/master@{#297165}

TBR=yfriedman@chromium.org,ppi@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=418549

Review URL: https://codereview.chromium.org/613653005

Cr-Commit-Position: refs/heads/master@{#297170}

BUG=None
TBR=tommi

Review URL: https://codereview.chromium.org/613613003

Cr-Commit-Position: refs/heads/master@{#297169}

Summary of changes available at:
https://chromium.googlesource.com/skia/+log/b5fae93..c13bc57

This roll includes SkDevice::drawPosText API changes for GatherPixelRefDevice and VectorPlatformDeviceEmf, needed after https://codereview.chromium.org/605533002.

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

Review URL: https://codereview.chromium.org/607853003

Cr-Commit-Position: refs/heads/master@{#297168}

On the Blink side:
The clamping method used by computeScaleAndScrollForBlockRect now
uses a new method PinchViewport::clampDocumentOffsetAtScale to clamp
the requested viewport offset taking into account the inner and outer
viewports.

Also added an overload of applyViewportDeltas, called in virtual
viewport mode only, that applies the inner and outer deltas along
page scale in one shot. This makes the viewport layers a special case
again in that their scroll offsets are set through this call, rather
than the standard 'did_scroll' callback of most layers. This is needed
since the scroll offsets may be invalid until the page scale is
applied so this is handled in one method to prevent unintended
clamping.

On the Compositor side:
Calls the appropriate applyViewportDeltas based on which pinch mode
we're running in. The page scale animation now prefers to scroll the
inner viewport first, scrolling the outer only when the inner has
reached its scroll extent.

Blink-side: https://codereview.chromium.org/584833003

BUG=364106

Review URL: https://codereview.chromium.org/585063002

Cr-Commit-Position: refs/heads/master@{#297167}

Note: will land on Monday; need to watch bots so that they execute the
test right.

TBR=vrk@chromium.org
BUG=418034

Review URL: https://codereview.chromium.org/599103006

Cr-Commit-Position: refs/heads/master@{#297166}

BUG=418549
NOTRY=true

Review URL: https://codereview.chromium.org/611013002

Cr-Commit-Position: refs/heads/master@{#297165}

The Put callback functions took an absurd number of paramters.  This
CL simplifies by putting them in a context struct.  This is needed by
a downstream CL that needs to add another member to the struct.

Upstream of: https://codereview.chromium.org/608593003

BUG=392621

Review URL: https://codereview.chromium.org/606843002

Cr-Commit-Position: refs/heads/master@{#297164}

Allow component IME extensions use app.window and add 'ime' boolean in CreateWindowOptions for app window

To create an IME window, an IME must be whitelisted, must have app.window.ime permission and must use set 'ime' and 'frame' this way:
chrome.app.window.create(url, {
  'ime': true,
  'frame': 'none
}, function() {})
Note only whitelisted component IME extensions can use this boolean parameter. See doc here: https://docs.google.com/a/google.com/document/d/1JQHpLu_RjN8C1Yi0i_eApS5SZtPoMDqvlq06YOwXCFU/edit

BUG=401984

Review URL: https://codereview.chromium.org/503383002

Cr-Commit-Position: refs/heads/master@{#297163}

Cr-Commit-Position: refs/heads/master@{#297162}