GitLab

R=michaelpg@chromium.org
BUG=none
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation
NOPRESUBMIT=true  # crisper.js

Review-Url: https://codereview.chromium.org/2167383003
Cr-Commit-Position: refs/heads/master@{#407701}

Reland of HttpChunkedDecoder: Support chunks longer than 2^31-1 bytes. (patchset #1 id:1 of https://codereview.chromium.org/2183433003/ )

Reason for revert:
The CL was blamed for compile failures it didn't cause, re-landing.

Original issue's description:
> Revert of HttpChunkedDecoder: Support chunks longer than 2^31-1 bytes. (patchset #4 id:60001 of https://codereview.chromium.org/2170133004/ )
>
> Reason for revert:
> Speculative revert for bug 631246.  Have no better ideas.
>
> Apologies if this wasn't the cause.
>
> Original issue's description:
> > HttpChunkedDecoder: Support chunks longer than 2^31-1 bytes.
> >
> > We were using HexStringToInt to parse chunk size, which returns a
> > 32-bit int.  This CL switches to using HexStringToInt64, which uses
> > 64-bit ints, so we can now support chunks up to 2^63-1 bytes.
> >
> > That should be enough for anybody. [Cue dramatic music]
> >
> > BUG=630680
> >
> > Committed: https://crrev.com/8ea249f8b70f154f9995ed538fc853fe9cf46503
> > Cr-Commit-Position: refs/heads/master@{#407549}
>
> TBR=eroman@chromium.org,mmenke@chromium.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=630680
>
> Committed: https://crrev.com/2321e8a45603ca595296b67e4d064c070f2b3591
> Cr-Commit-Position: refs/heads/master@{#407651}

TBR=eroman@chromium.org,mpearson@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=630680

Review-Url: https://codereview.chromium.org/2180063002
Cr-Commit-Position: refs/heads/master@{#407700}

This patch updates the code to prevent auto deducing to a raw pointer.

R=dalecurtis@chromium.org
BUG=554600

Review-Url: https://codereview.chromium.org/2181163002
Cr-Commit-Position: refs/heads/master@{#407699}

WebUI URLRequest source data is (on desktop) backed by a memory-mapped
file. Copying that data in CompleteRead can trigger jankiness due to
potential file IO caused by the read.

This CL moves the memcpy on the file thread.

R=dbeam@chromium.org
BUG=455423

Review-Url: https://codereview.chromium.org/2158123003
Cr-Commit-Position: refs/heads/master@{#407698}

Cr-Commit-Position: refs/heads/master@{#407697}

Per-request from rkaplow

BUG=

Review-Url: https://codereview.chromium.org/2168233002
Cr-Commit-Position: refs/heads/master@{#407696}

1/5: Introduce NavigationHintSender.
     https://codereview.chromium.org/2043863003/
2/5: Pipe NavigationHints from NavigationHintSender to ChromeRenderMessageFilter
     https://codereview.chromium.org/2043083002/
3/5: Call StartServiceWorkerForNavigationHint() from ChromeRenderMessageFilter
     This CL.
4/5: Measure the precision of the speculative launch of Service Workers for
NavigationHints
     https://codereview.chromium.org/2045153003/
5/5: Add flags to enable SupeculativeLaunchServiceWorker
     https://codereview.chromium.org/2053573002/

When ChromeRenderMessageFilter receives NavigationHint IPC message, it calls
ServiceWorkerContextWrapper::StartServiceWorkerForNavigationHint().
ServiceWorkerContextWrapper will try to start a Service Worker for the document
which is linked from the anchor element.

This CL depends on https://codereview.chromium.org/2181553003/

BUG=616502

Review-Url: https://codereview.chromium.org/2052613003
Cr-Commit-Position: refs/heads/master@{#407695}

This patch converts "editing/insert-text-with-newlines.html" to use w3c test
harness to simplify test files.

This patch is a preparation of getting rid of grammar checking feature.

BUG=619452
TEST=n/a; this patch changes a test file

Review-Url: https://codereview.chromium.org/2180863002
Cr-Commit-Position: refs/heads/master@{#407694}

TBR=michaelpg@chromium.org
BUG=none
NOPRESUBMIT=true  # crisper.js
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2171073002
Cr-Commit-Position: refs/heads/master@{#407693}

R=michaelpg@chromium.org
BUG=none
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation
NOPRESUBMIT=true  # crisper.js

Review-Url: https://codereview.chromium.org/2172833002
Cr-Commit-Position: refs/heads/master@{#407692}

https://chromium.googlesource.com/skia.git/+log/0f7748ab388c..b8113780c3cf

$ git log 0f7748ab3..b8113780c --date=short --no-merges --format='%ad %ae %s'
2016-07-25 bungeman Update typeface serialization style.

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

Review-Url: https://codereview.chromium.org/2177413002
Cr-Commit-Position: refs/heads/master@{#407691}

This change implements a tweak to the Web Animations spec:
https://github.com/w3c/web-animations/commit/a9ba51338ed09170d16c47317f8e4e2eef122a82

Negative end delay now affects the progress used in the after phase.

Review-Url: https://codereview.chromium.org/2030843002
Cr-Commit-Position: refs/heads/master@{#407690}

R=michaelpg@chromium.org
TBR=michaelpg@chromium.org
BUG=629710
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:closure_compilation
NOPRESUBMIT=true  # crisper.js

Review URL: https://codereview.chromium.org/2158913007 .

Cr-Commit-Position: refs/heads/master@{#407689}

https://chromium.googlesource.com/external/github.com/catapult-project/catapult.git/+log/139402a9056e..47a907219954

$ git log 139402a90..47a907219 --date=short --no-merges --format='%ad %ae %s'

TBR=catapult-sheriff@chromium.org

Review-Url: https://codereview.chromium.org/2179013003
Cr-Commit-Position: refs/heads/master@{#407688}

The metric is recorded by Chrome OS session manager:
https://chromium-review.googlesource.com/#/c/361042/

BUG=b/28815532

Review-Url: https://codereview.chromium.org/2175423002
Cr-Commit-Position: refs/heads/master@{#407687}

If user navigates forward and then immidiately back the following
sequence of events happens:

1.) ForwardURL is registered for load
2.) BackURL is registered for load
3.) Loading is stopped and current web view URL is ForwardURL
4.) Did Finish Loading is called and current web view URL is BackURL

On step 3 web should not signal didFinishLoad if
current web view URL != last registered URL, because telling embedder
that load has finished for ForwardURL is false information.

BUG=629565

Review-Url: https://codereview.chromium.org/2169703002
Cr-Commit-Position: refs/heads/master@{#407686}

BUG=598180

Review-Url: https://codereview.chromium.org/2182813002
Cr-Commit-Position: refs/heads/master@{#407685}

Review-Url: https://codereview.chromium.org/2183683002
Cr-Commit-Position: refs/heads/master@{#407684}

he problem of an iterator is that it will be invalid when the vector
reallocates buffer. An index won't be invalid as long as we don't
add/remove any item before the index.

An examples that we had used indices before this CL is
DisplayItemIndicesByClientMap which maps DisplayItemClient to indices
of display items. Iterators don't work because the map is built when we
are adding items into the list. Once the list reallocates the underlying
buffer, all existing iterators might become invalid.

Another example is
m_underInvalidationCheckingBegin/m_underInvalidationCheckingEnd.
Before this CL we had to reset them in appendDebugDrawingAfterCommit()
when the list changes. (I had a hard time debugging a flaky crash caused
by an invalid iterator before adding the reset code.)

Because we have to use indices, if we use iterators we have to convert
between them.

The main purpose of this CL is for https://codereview.chromium.org/2176573004/
which requires us to find a cached chunk by index of a display item.
Otherwise we would add another place converting iterator to index.

This CL doesn't expose indices across public api boundary.

Run 1093 of ct (https://ct.skia.org/chromium_perf_runs/) showed no performance
impact.

BUG=596983

Review-Url: https://codereview.chromium.org/2176683003
Cr-Commit-Position: refs/heads/master@{#407683}

Since HistoryBackend would remove username/password in an url
when write it into DB, Chrome Sync would ignore such url.
But ignore that url won't lost the url since HistoryBackend
will send username/password removed url to sync later.

BUG=629170

Review-Url: https://codereview.chromium.org/2176893002
Cr-Commit-Position: refs/heads/master@{#407682}

This will give content embedders (such as Chromecast) a way to override
supported video codecs logic.

BUG=477103

Review-Url: https://codereview.chromium.org/2153643002
Cr-Commit-Position: refs/heads/master@{#407681}

https://pdfium.googlesource.com/pdfium.git/+log/e283e47..d8cc503

BUG=629919
TBR=thestig@chromium.org

Review-Url: https://codereview.chromium.org/2177383002
Cr-Commit-Position: refs/heads/master@{#407680}

Revert of Remove NativeThemeWin::PaintIndirect (patchset #3 id:40001 of https://codereview.chromium.org/2090003003/ )

Reason for revert:
Directly causes crbug.com/628574, which is ship-blocking; we'll have to find another way to clean up the mess.

Original issue's description:
> Remove NativeThemeWin::PaintIndirect
>
> We don't seem to actually use this complicated codepath in practice.
>
> In particular:
>
> 1. // This block will only get hit with --enable-accelerated-drawing flag.
> That flag is obsolete, and its replacement also seems to no longer exist.
>
> 2. // Scrollbar components on Windows Classic theme (on all Windows versions)
>    // have particularly problematic alpha values, so always draw them
>    // indirectly.
> Turning on the Windows Classic theme on 64b Windows 7, I can't cause the block
> to be hit today.
>
> 3. // In addition, scrollbar thumbs and grippers for the Windows XP
>    // theme (available only on Windows XP) also need their alpha values
>    // fixed.
> We're no longer supporting XP and have been cleared to remove relevant code.
>
> Once it's gone, so are several copies of functions that paint the
> same UI element in different ways depending on whether they're drawing
> to HDC or SkCanvas: Gutter, MenuSeparator, MenuBackground, and
> MenuItemBackground.
>
> R=pkasting@chromium.org
> BUG=543755,579196,622692
>
> Committed: https://crrev.com/83a3c26417896b207efe22b116a03a9a162c0ca5
> Cr-Commit-Position: refs/heads/master@{#404155}

TBR=pkasting@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=543755,579196,622692

Review-Url: https://codereview.chromium.org/2181153002
Cr-Commit-Position: refs/heads/master@{#407679}

This refactoring decouples storage models from the ResourceTreeModel.

BUG=624494
R=dgozman, pfeldman

Review-Url: https://codereview.chromium.org/2172753002
Cr-Commit-Position: refs/heads/master@{#407678}

This reverts commit 010e9500.

I set this workaround flag based on misreading a crash report, and
there's actually no evidence that Adreno 5xx has any bugs with
framebuffer discarding.

BUG=612474
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/2177393003
Cr-Commit-Position: refs/heads/master@{#407677}

R=dgozman@chromium.org
BUG=none

Review-Url: https://codereview.chromium.org/2181453002
Cr-Commit-Position: refs/heads/master@{#407676}

As far as I can tell, SRGB gamma and 2.2 gamma dst color
spaces are very common, with all other possibilities
being extremely rare.

We'd like to limit our color correct implementations
to approximations of SRGB and 2Dot2 destinations:
(1) Allows software and gpu to stay in sync.
(2) Avoids code bloat caused by supporting any
    arbitrary representation that can be packed into an
    ICC profile.

To be confident in making this decision, we need some
metrics to help verify our assumptions about the dst color
spaces.

BUG=

Review-Url: https://codereview.chromium.org/2177933002
Cr-Commit-Position: refs/heads/master@{#407675}

AnimationCompositorAnimationsTest.ConvertTimingForCompositorIterationCount
contains an EXPECT_DEATH test which checks that
CompositorAnimations::convertTimingForCompositor hits an assertion failure if
passed a Timing object with a value of -1 for iterationCount. This test is
probably included because -1 is a valid value for the equivalent field in the
output CompositorTiming object.

However, EXPECT_DEATH is flaky and dangerous in a multi-threaded context. Since
we're not gaining much by having this test, this patch removes it.

BUG=631021

Review-Url: https://codereview.chromium.org/2180973003
Cr-Commit-Position: refs/heads/master@{#407674}

This CL
-- Migrates all ContentAutofillDriver<-->AutofillAgent
   IPC messages into mojo interfaces.
-- Revises related unit tests and browser tests.

TEST=components_unittests, browser_tests
BUG=582391
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_site_isolation

Review-Url: https://codereview.chromium.org/2007473004
Cr-Commit-Position: refs/heads/master@{#407673}

Follow up to crbug/609863.

Constrained dialogs should always be 1) fixed size or 2) autoresizing during their lifetime, not switching between one or the other.

Previously, autoresizing was done at DocumentOnLoadCompletedInMainFrame() rather than in the constructor, which caused a race condition in CrOS. This change will keep implementations consistent across desktop platforms.

BUG=630449

Review-Url: https://codereview.chromium.org/2174573002
Cr-Commit-Position: refs/heads/master@{#407672}

This patch converts "editing/inserting/insert-tab-00[1-4].html" to use w3c
test harness and consolidates into "insert_tab.html" to simplify test case
for ease of maintenance.

This patch is a preparation of getting rid of grammar checking feature.

Review-Url: https://codereview.chromium.org/2177173002
Cr-Commit-Position: refs/heads/master@{#407671}

This patch converts "editing/insert-div-22.html" to use w3c test harness to
simplify test files.

This patch is a preparation of getting rid of grammar checking feature.

BUG=619452
TEST=n/a; this patch changes a test file

Review-Url: https://codereview.chromium.org/2178863002
Cr-Commit-Position: refs/heads/master@{#407670}

BUG=628161

When set one of TextEdit, which are mEnterPassphrase and
mConfirmPassphrase, to have an error message, set another one
to empty.

Review-Url: https://codereview.chromium.org/2179093002
Cr-Commit-Position: refs/heads/master@{#407669}

Some context:
https://codereview.chromium.org/1653533002
added code to ItemChooserDialog to notify it that the discovery is complete
and no new devices will be added. It also adds a new string to show when
the discovery is complete.

However,
https://codereview.chromium.org/1739523002
accidentally removed this feature.

This CL added this feature back.

BUG=580346

Review-Url: https://codereview.chromium.org/2175613004
Cr-Commit-Position: refs/heads/master@{#407668}

BUG=630472

Review-Url: https://codereview.chromium.org/2172013002
Cr-Commit-Position: refs/heads/master@{#407667}

Implemented the same functionality on embedder side.
Also, sanitizing "settings" query param for remote frontends.

BUG=618037

Review-Url: https://codereview.chromium.org/2177983004
Cr-Commit-Position: refs/heads/master@{#407666}

BUG=627306

Review-Url: https://codereview.chromium.org/2174863003
Cr-Commit-Position: refs/heads/master@{#407665}

BUG=630732

Review-Url: https://codereview.chromium.org/2179593003
Cr-Commit-Position: refs/heads/master@{#407664}

Adding a ScrollView so we can handle cases when ARC has more that 3 app
candidates to handle a given URL. We are also marking the app selected
by the user by changing the background color.

Notice that we are displaying 3.5 rows when the app candidates exceed 3,
this is the intended behavior.

Removing EventMonitor and algorithm as these are not used anymore.

BUG=620129

Review-Url: https://codereview.chromium.org/2134293002
Cr-Commit-Position: refs/heads/master@{#407663}

Crashing thread might be in "running" state, i.e. after sys_sendmsg()
and before sys_read().

Wait for a short period in browser process, allow the crashing thread to
catch up and retry.

BUG=internal b/30003601
TEST=Manually crash the main thread in utility process and make sure
     thread ID is translated.

Review-Url: https://codereview.chromium.org/2169063002
Cr-Commit-Position: refs/heads/master@{#407662}