GitLab

This is a partial revert of:
https://codereview.chromium.org/1449953002

Most of that CL is inactive code, but it did change the way the
compositor thread was created. Instead of a base::Thread, I'd used a
WebThreadImplForWorkerScheduler and this seems to have had some
performance impact. While I try to figure out why, this CL should get
us back to our prior performance.

BUG=565789

Review URL: https://codereview.chromium.org/1550093002

Cr-Commit-Position: refs/heads/master@{#367051}

(╯^□^)╯︵ ❄☃❄

BUG=557422
R=avi@chromium.org
TBR=jam@chromium.org

Review URL: https://codereview.chromium.org/1553493002

Cr-Commit-Position: refs/heads/master@{#367050}

This reverts commit 83557154.

BUG=553569

Review URL: https://codereview.chromium.org/1538473002

Cr-Commit-Position: refs/heads/master@{#367049}

Cr-Commit-Position: refs/heads/master@{#367048}

Review URL: https://codereview.chromium.org/1547283002

Cr-Commit-Position: refs/heads/master@{#367047}

Previously, if a linearized PDF was loaded in a single callback (i.e. if
IsDocumentComplete was true in the call to LoadDocument), only the first
page of the document was visible because FPDFAvail_IsPageAvail was never
called for all other pages. This has been corrected.

BUG=572655

Review URL: https://codereview.chromium.org/1552603002

Cr-Commit-Position: refs/heads/master@{#367046}

After John's work in crbug.com/560559, telemetry project no longer depends
on pylib, hence tools/perf which is a user of telemetry must specify this
deps explictly.

BUG=572700, 560559
TBR=jbudorick@chromium.org

Review URL: https://codereview.chromium.org/1556463002

Cr-Commit-Position: refs/heads/master@{#367045}

(╯^□^)╯︵ ❄☃❄

BUG=557422
R=avi@chromium.org
TBR=lcwu@chromium.org

Review URL: https://codereview.chromium.org/1553503002

Cr-Commit-Position: refs/heads/master@{#367044}

std::move is not needed when returning local variables. This wasn't
caught in the initial commit because only Windows builds //third_party
with -Wall.

BUG=82385
TBR=thakis@chromium.org

Review URL: https://codereview.chromium.org/1549373004

Cr-Commit-Position: refs/heads/master@{#367043}

BUG=561190

Review URL: https://codereview.chromium.org/1550803002

Cr-Commit-Position: refs/heads/master@{#367042}

We currently resolve nearby URLs twice, once before showing a "nearby web
pages found" notification and once just before displaying the list of
nearby URLs. If a URL was not resolvable for the notification we avoid
re-resolving it for display. This prevents costly delays at display time
as we are only requesting data expected to already be cached by the
resolution service.

During onboarding (ie, before the user has explicitly opted into the
physical web feature) we still scan for nearby URLs but do not resolve
them for privacy reasons. This causes all URLs to be considered
"unresolvable" immediately after opting in as none have been sent to the
resolution service yet.

To fix, at display time allow ListUrlsActivity to receive the full
list of nearby URLs (including unresolved) only if the resolved list is
empty.

BUG=529962

Review URL: https://codereview.chromium.org/1556453002

Cr-Commit-Position: refs/heads/master@{#367041}

Exactly the same as issue 1528243002 at patchset 20001 (http://crrev.com/1528243002#ps20001)

BUG=534894

Review URL: https://codereview.chromium.org/1551683002

Cr-Commit-Position: refs/heads/master@{#367040}

BUG=571302,510410,532249

Review URL: https://codereview.chromium.org/1544443004

Cr-Commit-Position: refs/heads/master@{#367039}

The current blacklist entry will exclude any Optimus configuration when
GPU driver that isn't "Mesa >= 10.1".  This is (unintentionally) blocking
GPU on devices where the NVIDIA driver is used.

This CL changes the intent so that we blacklist "Mesa < 10.1" rather than
"anything except Mesa >= 10.1".

BUG=530981

Review URL: https://codereview.chromium.org/1522683002

Cr-Commit-Position: refs/heads/master@{#367038}

BUG=532905

Review URL: https://codereview.chromium.org/1545763002

Cr-Commit-Position: refs/heads/master@{#367037}

Instead of allocating and mapping a new chunk of shared memory segment
in each draw synchronously, re-use the same segment if the sizes match
exactly.

Break software draws into 3 IPCs:
SetSharedMemory: synchronous, pass the shared memory handle to be used
                 for subsequent software draws to renderer
DemandDrawSw: synchronous, the draw itself
ZeroSharedMemory: asynchronous, zeros shared memory on renderer side,
                  sent after each DemandDrawSw after browser is done
                  reading from the shared memory.

ZeroSharedMemory is asynchronous so zero-ing the memory can be done in
parallel. Note messages on renderer are ordered, so ZeroSharedMemory
is guaranteed to be handled before the next DemandDrawSw.

BUG=545611

Review URL: https://codereview.chromium.org/1541203003

Cr-Commit-Position: refs/heads/master@{#367036}

BUG=532905

Review URL: https://codereview.chromium.org/1547683003

Cr-Commit-Position: refs/heads/master@{#367035}

The main GraphicsLayer of a composited scroller should be the size of the
scroller's frame.  Report this size in PaintLayer::boundingBoxForCompositing
and update ViewPainter to use the same rect when painting the background fill.

This fixes placement of scrollbar layers in RTL documents, which were confused
by a bogus m_offsetFromLayoutObject because the rect returned by
LayoutView::documentRect is relative to the scroll origin.

This patch makes compositing/rtl/rtl-overflow-invalidation.html pass (discovered
through flag-specific expectations!)

Sadly we regress several repaint-tracking tests due to the new invalidation in
FrameView::viewportSizeChanged.  These will need flag-specific baselines (not
yet implemented) at least until http://crbug.com/568847 is addressed.

BUG=542432

Review URL: https://codereview.chromium.org/1472083006

Cr-Commit-Position: refs/heads/master@{#367034}

Diff since last roll:
https://github.com/chromium/dom-distiller/compare/6e47a6727c...349b932fc2

Picked up changes:
349b932 Remove DOM Distiller dependency on scoped_ptr<T>::Pass().
8fa737b Add <pre> to nesting tags
7991208 Clean up attributes of image elements

BUG=557422,567955
TBR=mdjones@chromium.org

Review URL: https://codereview.chromium.org/1550133004

Cr-Commit-Position: refs/heads/master@{#367033}

Review URL: https://codereview.chromium.org/1550693003 .

Cr-Commit-Position: refs/heads/master@{#367032}

If the dialog is initially in route-details view, and the route is
closed before the 3s startup timer fires, then the dialog will flash
between the route-details and sink-list view before settling on
sink-list. See bug for details.

The fix is to only allow maybeShowRouteDetailsOnOpen to be invoked
once, on startup. It is removed from |rebuildRouteMaps_| and is now
called from setInitialData. Also removed |localRouteCount_| field from
media-router-container since it's no longer needed.

BUG=571463

Review URL: https://codereview.chromium.org/1538363003

Cr-Commit-Position: refs/heads/master@{#367031}

BUG=568353

Review URL: https://codereview.chromium.org/1542593003

Cr-Commit-Position: refs/heads/master@{#367030}

Apparently cmake doesn't like it if the compiler changes like it did in
https://codereview.chromium.org/1555443002/ without a clobber build.

Since this might help with issue 569337 too (who knows), do it everywhere
for a short while instead of just on linux.

BUG=509739,569337
TBR=hans

Review URL: https://codereview.chromium.org/1546323004 .

Cr-Commit-Position: refs/heads/master@{#367029}

If we do not already have the location permission, we need to request
it when the user enables the Physical Web feature in the privacy
settings.

BUG=529962

Review URL: https://codereview.chromium.org/1535883002

Cr-Commit-Position: refs/heads/master@{#367028}

Display a dialog describing the Physical Web and inviting the user to
enable or decline the feature.

BUG=529962

Review URL: https://codereview.chromium.org/1526323003

Cr-Commit-Position: refs/heads/master@{#367027}

That way, the libstdc++ that's used by default should work on the bots.
Before this change, one had to pass --gcc-toolchain pointing to a precise-built
gcc when building linux clang binaries.

BUG=509739
TBR=hans@chromium.org

Review URL: https://codereview.chromium.org/1555443002 .

Cr-Commit-Position: refs/heads/master@{#367026}

BUG=529962

Review URL: https://codereview.chromium.org/1512113002

Cr-Commit-Position: refs/heads/master@{#367025}

As the OCSP Init code (for NSS) is a non-leaky Singleton, but
the other NSS code is, we end up calling into it during shutdown
while other requests may still be pending. Since the code is,
unfortunately, not exactly the most thread safe, the right answer
is to just not attempt to cleanup on shutdown, same as all the
other NSS code.

BUG=572727

Review URL: https://codereview.chromium.org/1550853002

Cr-Commit-Position: refs/heads/master@{#367024}

BUG=571488

Review URL: https://codereview.chromium.org/1554443002

Cr-Commit-Position: refs/heads/master@{#367023}

BUG=569691

Review URL: https://codereview.chromium.org/1522233002

Cr-Commit-Position: refs/heads/master@{#367022}

BUG=456525
TBR=pfeldman@chromium.org

Review URL: https://codereview.chromium.org/1551713003

Cr-Commit-Position: refs/heads/master@{#367021}

This is a follow-up to https://codereview.chromium.org/1525033002 --
this is a small change that was left out of that CL which makes it so
that --required_initial_confidence is added to the command line parameters
passed to bisect_perf_regression.py.

BUG=568041

Review URL: https://codereview.chromium.org/1543003002

Cr-Commit-Position: refs/heads/master@{#367020}

It was missing at least one source file, so until we make the GYP->GN
switch, it's better to just use the glob.

BUG=532905

Review URL: https://codereview.chromium.org/1538393003

Cr-Commit-Position: refs/heads/master@{#367019}

This includes a minor change to settings_page_browsertest.js to use $$ instead of shadowRoot.querySelector (thanks to dschuyler@ for the reminder).

BUG=543299
For simple change to BUILD.gn:
TBR=sky@chromium.org

Review URL: https://codereview.chromium.org/1466433002

Cr-Commit-Position: refs/heads/master@{#367018}

runTest function has parameter that enable watchdog for test. Watchdog terminates test after 20 seconds. When runTest function is set as load event listener then it get event as first parameter and enables watchdog.

BUG=498782
TBR=pfeldman@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/1552593002

Cr-Commit-Position: refs/heads/master@{#367017}

Otherwise NSS may cache verification results even after the test DB is gone.

BUG=569657
TEST=net_unittests --gtest_filter='CertDatabaseNSSTest.ImportCaAndServerCert:TestRootCertsTest.OverrideTrust' --gtest_shuffle --gtest_repeat=25

Review URL: https://codereview.chromium.org/1532223005

Cr-Commit-Position: refs/heads/master@{#367016}

BUG=

Review URL: https://codereview.chromium.org/1545343002

Cr-Commit-Position: refs/heads/master@{#367015}

Not required in GYP since it doesn't assert outputs.
Build fails without this change.

BUG=532905

Review URL: https://codereview.chromium.org/1542153002

Cr-Commit-Position: refs/heads/master@{#367014}

https://chromium.googlesource.com/skia.git/+log/93eb4b97d4b8..636270245fb4

$ git log 93eb4b97d..636270245 --date=short --no-merges --format='%ad %ae %s'
2015-12-28 mtklein initialize offscreen in StrokeZeroGM

CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel
TBR=mtklein@google.com

Review URL: https://codereview.chromium.org/1546373002

Cr-Commit-Position: refs/heads/master@{#367013}

BUG=572726
TBR=pfeldman@chromium.org
NOTRY=true

Review URL: https://codereview.chromium.org/1547333002

Cr-Commit-Position: refs/heads/master@{#367012}