GitLab

This is more consistent overall, GetGLImplementation() doesn't tell exactly
whether we're using the core profile, or which version of ES etc.
It also helps testing & fuzzing, because mocks/stubs may implement ES2 but don't
report kGLImplementationEGLGLES2, causing inconsistencies.

BUG=None
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2170293002
Cr-Commit-Position: refs/heads/master@{#407276}

This CL fixes both (1) the renderer crash originally fixed by
https://codereview.chromium.org/2165043002, and (2) reverts the part
of that CL which prevents ChromeOS from having zero-velocity
GestureFlingStarts, which caused a regression in two-finger swipe
navigation for touchpad. It also moves the DCHECK from the original CL
to a new location which would have prevented the renderer crashes from
being introduced in the first place.

This CL doesn't include a test, though the test proposed by
khmel@ in https://codereview.chromium.org/2168693003 presumably will
land shortly after this CL.

BUG=627227
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_site_isolation

Review-Url: https://codereview.chromium.org/2173603002
Cr-Commit-Position: refs/heads/master@{#407275}

This should allow some unittests (specifically offline pages related
unittests) to create the UINQEService despite not being a browser
process with no IOThread.

BUG=

Review-Url: https://codereview.chromium.org/2172393002
Cr-Commit-Position: refs/heads/master@{#407274}

In the near future, SurfaceIdAllocator will live in clients
and so ui::ContextFactory should not create a
SurfaceIdAllocator directly. Instead, ui::ContextFactory
allocates a new surface client ID, which can be used to
create a SurfaceIdAllocator in clients.

This CL also moves calls to RegisterSurfaceClientId and
InvalidateSurfaceClientId close to
RegisterSurfaceClientFactory and
UnregisterSurfaceClientFactory respectively. In a
subsequent CL, these two will be merged together into
SurfaceFactory.

BUG=627283, 629940
TBR=reveman@chromium.org
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_site_isolation

Review-Url: https://codereview.chromium.org/2166423002
Cr-Commit-Position: refs/heads/master@{#407273}

Context: https://codereview.chromium.org/2128793003/#msg29

BUG=

Review-Url: https://codereview.chromium.org/2176843002
Cr-Commit-Position: refs/heads/master@{#407272}

Unit tests for https://codereview.chromium.org/2149523002/.
BUG=628496, 624506
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2166923002
Cr-Commit-Position: refs/heads/master@{#407271}

On most platforms the monotonic clock is system wide which means it is
consistent across processes making conversion unnecessary. In fact not
only it is unnecessary but harmfull because the
InterProcessTimeTickConverter (IPTTC) conversion [1] is an estimation
which introduces some error.

The patch limits the usage of IPTTC to platforms that require it which
is any Windows where we cannot use |QueryPerformanceCounter|.

After this change we will only collect UMA metrics[2] for IPTTC on
platforms where it is used.

[1] https://docs.google.com/document/d/1RcaOZ1G8ttAez9YuilvbDJCkCaYuLKqNkVpQvIBqsvM/edit
[2] Affects InterProcessTimeTicks.*

BUG=616574
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation

Review-Url: https://codereview.chromium.org/2090783002
Cr-Commit-Position: refs/heads/master@{#407270}

This avoids problems with the timer being delayed by nested message
loops and ScopedPageLoadDeferrers, where code in the nested message
loop might access the document and not generate a notification.

The timer was originally added to prevent re-entrant calls to V8,
but that can now be prevented with ScriptForbiddenScope.

BUG=629542
TEST=See bug for repro steps.

Review-Url: https://codereview.chromium.org/2171063002
Cr-Commit-Position: refs/heads/master@{#407269}

The current implementation relies on SkPath::getBounds(), which
computes a conservative result (includes control points).

Use SkPathOps::TightBounds() instead, for an exact answer.

BUG=230599
R=fs@opera.com,pdr@chromium.org,schenney@chromium.org,caryclark@google.com

Review-Url: https://codereview.chromium.org/2168133002
Cr-Commit-Position: refs/heads/master@{#407268}

This provides autorelease support for Mac tasks.

BUG=553459

Review-Url: https://codereview.chromium.org/2172123003
Cr-Commit-Position: refs/heads/master@{#407267}

Review-Url: https://codereview.chromium.org/2179543002
Cr-Commit-Position: refs/heads/master@{#407266}

No-op cleanup extracted from https://codereview.chromium.org/2135413003/#ps40001

In this CL:

 - Explicitly document the threading requirements of the API after grokking its use of locks
   in some places and lack of in some others)
   - (It's mostly thread unsafe; except for a few calls which are okay from the owner
      as well as the underlying thread; and one method which is actually thread-safe)
   - Add assertions as such (and catch issues on try bots :-O -- http://crbug.com/629139)
   - Remove |thread_lock_| which is unnecessary under the documented conditions
     (delayed until http://crbug.com/629139 is fixed..).

 - C++11 member initializers (makes it cleaner to add new POD members and
                              makes constructors simpler -- e.g. |run_loop_| was uninitialized)
 - Assertions and tests for current API which allows Start/Stop/Start cycles
   but wasn't tested... (actually more than that is poorly/not tested in this class
   but I'm merely adding tests for parts I'm stressing in my actual follow-up CL).

 - Comment nits

 - Make event-based tests use WaitableEvent instead of sleeping + toggling a bool.

BUG=629716, 629139

Review-Url: https://codereview.chromium.org/2145463002
Cr-Commit-Position: refs/heads/master@{#407265}

Changes the associated bindings implementation for ChannelMojo
such that remote interfaces can be acquired immediately upon
ChannelMojo construction rather than having to wait for connection
on the IO thread.

Simplifies the Channel bootstrapping process, removing a round-trip
Init message (and in fact the entire IPC::mojom::Boostrap interface)
since there's no need to actually exchange associated interface handles
over the pipe. Instead both sides can assume the other will use a fixed,
reserved endpoint ID for their IPC::mojom::Channel interface.

This also removes the restriction that associated interfaces must be
added to a Channel after Init. Instead the same constraints apply as
with AddFilter: an associated interface, like a filter, may be added
at any time as long as either Init hasn't been called OR the remote
process hasn't been launched.

The result of this CL is that any place it's safe to AddFilter,
it's also safe to AddAssociatedInterface; and any place it's safe to
Send, it's also safe to GetRemoteAssociatedInterface and begin using
any such remote interface immediately.

Remote interface requests as well as all messages to remote interfaces
retain FIFO with respect to any Send calls on the same thread. Local
interface request dispatch as well as all messages on locally bound
associated interfaces retain FIFO with respect to any OnMessageReceived
calls on the same thread.

BUG=612500,619202

Committed: https://crrev.com/e1037f997da9e1d44ca3b09d4ff32f0465673091
Committed: https://crrev.com/508da24622f957a01b076ccd058bfdccc79068a4
Review-Url: https://codereview.chromium.org/2163633003
Cr-Original-Original-Commit-Position: refs/heads/master@{#406720}
Cr-Original-Commit-Position: refs/heads/master@{#407050}
Cr-Commit-Position: refs/heads/master@{#407264}

This CL also let chrome compile against 24.1.1 support library.

BUG=609085

Review-Url: https://codereview.chromium.org/2175513002
Cr-Commit-Position: refs/heads/master@{#407263}

The current plan is to remove the code for M54 but
keep the Panels enabled for M53 to provide more time for transition.

BUG=571511

Review-Url: https://codereview.chromium.org/2176823002
Cr-Commit-Position: refs/heads/master@{#407262}

This DCHECK is clearly not intentional and can't possibly be correct.
It's not being hit in practice today due to the latency of ChannelMojo
setup, but reducing that latency easily exposes it.

TBRing because it's trivial.

BUG=None
TBR=yzshen@chromium.org

Review-Url: https://codereview.chromium.org/2178633002
Cr-Commit-Position: refs/heads/master@{#407261}

This reverts commit 58ac1ade, which
disabled a few flaky tests, and fixes those tests so they are no longer
flaky.

In a follow up change, we'll roll back
https://codereview.chromium.org/2172183002
since this change addresses the flakes caused by that patch.

BUG=630549

Review-Url: https://codereview.chromium.org/2176803002
Cr-Commit-Position: refs/heads/master@{#407260}

platform/ParsingUtilities.h caters to the same needs, so transition
users of skipString(...) to skipToken(...) and remove
platform/text/ParserUtilities.h.

Review-Url: https://codereview.chromium.org/2176623003
Cr-Commit-Position: refs/heads/master@{#407259}

BUG=621070

Review-Url: https://codereview.chromium.org/2167843004
Cr-Commit-Position: refs/heads/master@{#407258}

Revert of [system-health] Add full Google Search story to the System Health story sets (patchset #5 id:80001 of https://codereview.chromium.org/2148843002/ )

Reason for revert:

I see a failure for
benchmarks.system_health_smoke_test.SystemHealthBenchmarkSmokeTest.system_health.memory_desktop.search:portal:google
on
https://build.chromium.org/p/chromium.win/builders/Win7%20Tests%20%28dbg%29%281%29/builds/50880

This is obviously related.

Only one run with this patch has been completed thus far.  Either this change caused the failure or, even if not, this change has caused this test to become flaky (I don't see any failures before this).

Original issue's description:
> [system-health] Add Google Search browsing story to the System Health story sets
>
> This patch adds a single 'search:portal' story (both desktop and
> mobile) which does the following:
>
>   1. Open google.co.uk.
>   2. Type 'what is science' and hit return.
>   3. Scroll down to the Wikipedia result.
>   4. Click on the Wikipedia result and wait until it loads.
>
> This is a prototype of the SH stories with same origin and cross-origin
> navigtation (see
> https://docs.google.com/spreadsheets/d/1t15Ya5ssYBeXAZhHm3RJqfwBRpgWsxoib8_kwQEHMwI/#gid=712129109).
> Equivalent stories for other search pages (Baidu, Yahoo, Amazon, ...)
> will follow shortly (in subsequent CLs).
>
> BUG=589726
> CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.perf:android_s5_perf_cq;master.tryserver.chromium.perf:mac_retina_perf_cq
>
> Committed: https://crrev.com/d6d935821c3afc9c0ce21e2a74a5411b59cb3c66
> Cr-Commit-Position: refs/heads/master@{#407225}

TBR=nednguyen@google.com,petrcermak@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=589726

Review-Url: https://codereview.chromium.org/2173943003
Cr-Commit-Position: refs/heads/master@{#407257}

Use the page transition, which seems more reliable, rather than the
HasUserGesture method, to detect whether a given navigation event was
spawned by the user clicking on a link.

BUG=630072

Review-Url: https://codereview.chromium.org/2165193002
Cr-Commit-Position: refs/heads/master@{#407256}

It turns out that
roundedIntPoint(LayoutUnit::max()) != LayoutUNit::max()
so disable the assert in BackgroundImageGeometry that is really just
checking that we have pixel snapped the phase. We don't check any other
values for pixel snapping, so there's no reason to keep checking phase.

And there's no danger to having an unrounded value here. Nothing breaks.

R=fmalita@chromium.org
BUG=626745

Review-Url: https://codereview.chromium.org/2180433002
Cr-Commit-Position: refs/heads/master@{#407255}

https://chromium.googlesource.com/chromium/src/+/81de34207

BUG=526594
TBR=schenney@chromium.org

Review URL: https://codereview.chromium.org/2173053003 .

Cr-Commit-Position: refs/heads/master@{#407254}

BUG=572726
TBR=caseq@chromium.org

Review-Url: https://codereview.chromium.org/2174893002
Cr-Commit-Position: refs/heads/master@{#407253}

https://chromium.googlesource.com/skia.git/+log/4c6e47a8a827..f1491693527a

$ git log 4c6e47a8a..f14916935 --date=short --no-merges --format='%ad %ae %s'
2016-07-22 bungeman Correct advances for 'monospace' fonts in PDF.
2016-07-22 mtklein Add SkRasterPipeline blitter.
2016-07-22 brianosman Bundle SkShader::asFragmentProcessor arguments in a struct
2016-07-22 caryclark limit number of searched roots

BUG=630649

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_blink_rel
TBR=jcgregorio@google.com

Review-Url: https://codereview.chromium.org/2180483002
Cr-Commit-Position: refs/heads/master@{#407252}

Goma builds necessarily don't use PDBs. That means that symbols have to be put in .obj files. This means that symbols get stored redundantly, which makes the job of the linker much harder - greatly increased linker working sets and greatly increased link times.

gyp deals with this by disabling symbol generation in the compile stage for goma builds. gn has not been doing this, which has made gn goma builds *significantly* slower, unless symbol_level is explicitly set to 1.

gyp has an override switch (win_z7) but it appears that nobody uses it so for now I am leaving it unimplemented in gn.

BUG=630074

Review-Url: https://codereview.chromium.org/2174873002
Cr-Commit-Position: refs/heads/master@{#407251}

https://chromium.googlesource.com/external/github.com/catapult-project/catapult.git/+log/20358d6bf8c9..07aa51b63e8c

$ git log 20358d6bf..07aa51b63 --date=short --no-merges --format='%ad %ae %s'

TBR=catapult-sheriff@chromium.org

Review-Url: https://codereview.chromium.org/2176783002
Cr-Commit-Position: refs/heads/master@{#407250}

BUG=None

Review-Url: https://codereview.chromium.org/2173543002
Cr-Commit-Position: refs/heads/master@{#407249}

This allows us to restore to maximized mode when fullscreen
is toggled by the user pressing F4.

BUG=630696,b/29817350
TEST=exo_unittests --gtest_filter=ShellSurfaceTest.ToggleFullscreen

Review-Url: https://codereview.chromium.org/2177623002
Cr-Commit-Position: refs/heads/master@{#407248}

BUG=630719
TEST=N/A

Review-Url: https://codereview.chromium.org/2169173002
Cr-Commit-Position: refs/heads/master@{#407247}

BUG=630001
R=dtapuska@chromium.org
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2167883002
Cr-Commit-Position: refs/heads/master@{#407246}

The old v8::Isolate::GetCpuProfiler API is going to be deprecated soon.

Review-Url: https://codereview.chromium.org/2127423003
Cr-Commit-Position: refs/heads/master@{#407245}

This CL makes sure that MHTMLGenerationManager::JobFinished is
idempotent and doesn't do anything when called for a second time.
This avoids the crash from the linked bug.

BUG=612098

Review-Url: https://codereview.chromium.org/2163073002
Cr-Commit-Position: refs/heads/master@{#407244}

This is required by subsequence caching for spv2. When we copy a cached
subsequence, we won't do actual paint and will use the existing
PaintChunkProperties which points to the paint property nodes created
during the previous paint. We are sure that the values of these property
nodes didn't change since the previous paint because we can use the
cached subsequence.

To meet this requirement, when building the paint property tree, instead
of rebuilding the whole tree by creating all new nodes, we should reuse
the existing nodes if their values don't change. The reused nodes should
be placed in the new tree.

In the future we may also reuse a whole subtree of property nodes.

BUG=596983

Review-Url: https://codereview.chromium.org/2144823006
Cr-Commit-Position: refs/heads/master@{#407243}

Pull out the extensions code drag-and-drop handler in order to reuse it
in the MD extensions page. In the process, closure compiler DragWrapper
and create a real interface for the DragWrapperHandler, which previously
was only specified in comments. Update drag wrapper handler implementors
with @override.

BUG=529395
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2164693002
Cr-Commit-Position: refs/heads/master@{#407242}

Add a constructFlexItem function that does the computations
for creating a FlexItem, and returns it. Also add a couple more
cached pieces of data to FlexItem which simplifies code in
a couple of places.

constructFlexItem will also be useful for implementing the
new intrinsic size algorithm.

R=eae@chromium.org,dgrogan@chromium.org
TEST=n/a, no change in behavior

Review-Url: https://codereview.chromium.org/2170533002
Cr-Commit-Position: refs/heads/master@{#407241}

TouchPad pinch to zoom on Windows typically works via a WM_MOUSEWHEEL
message with the CONTROL flag in wParam and varying deltas. On touchscreen
Windows devices like Surface, the WM_MOUSEWHEEL event is flagged as a
synthesized message occurring due to a touch.

I think it is safe to allow all WM_MOUSEWHEEL events to be processed by
ignoring the touch flag as they don't appear to get synthesized via
standard touch scrolls.

Changes in this patch are as below:

1. Ignore the touch flag on WM_MOUSEWHEEL and WM_MOUSEHWHEEL messages.
2. Removed dead code in hwnd_message_handler.cc/.h which tried to set
   scroll styles on the window.

BUG=620182

Review-Url: https://codereview.chromium.org/2172843002
Cr-Commit-Position: refs/heads/master@{#407240}

Clearing them at the end of NavigateInternal causes problems when a
ScopedPageLoadDeferrer is in use.

Credit to thestig@ for the test framework changes.

BUG=626838
TEST=See bug for repro steps.

Review-Url: https://codereview.chromium.org/2140393002
Cr-Commit-Position: refs/heads/master@{#407239}

Cr-Commit-Position: refs/heads/master@{#407238}

Current Dockerfiles were written with a misunderstanding of how wildcards and directory sources interact with the ADD command. This was discovered in testing using the generated tarballs.

The change fixes an incorrect comment in the engine Dockerfile.

Additionally, create_bundle.py was updated to tar up the filesystem state from a context other than the build directory.

These changes mean that for the env_tests bundle, running the Dockerfile in the context of the tarball results in a system file structure such that /blimp/ contains the set of dpendencies for the testing binaries in the same directory structure as // (the chromium src directory).

This change does not remove the duplicate ./third_party/blimp_fonts (./fonts is used instead) for the engine environment.

BUG=616945, 608487, 630438

Review-Url: https://codereview.chromium.org/2154873002
Cr-Commit-Position: refs/heads/master@{#407237}