GitLab

Created an ETC1 encoder that uses SSE2 to improve compression speed.
The SSE encoder extends TextureCompressor and uses the same algorithm as TextureCompressorETC1.

Added unittest for TextureCompressorETC1.

Moved shared code into a etc1 header.

Added new performance test scenarios.

Performance difference on Ubuntu x64, Haswell Processor:
Without SSE:
*RESULT Compress256x256BlackAndWhiteGradientImage: ETC1 Low= 1.966321587562561 us
*RESULT Compress256x256SolidBlackImage: ETC1 Low= .0956009104847908 us
*RESULT Compress256x256SolidColorImage: ETC1 Low= .4367307722568512 us
*RESULT Compress256x256RandomColorImage: ETC1 Low= 5.948055744171143 us

With SSE:
*RESULT Compress256x256BlackAndWhiteGradientImage: ETC1 Low= 1.0316201448440552 us
*RESULT Compress256x256SolidBlackImage: ETC1 Low= .25716209411621094 us
*RESULT Compress256x256SolidColorImage: ETC1 Low= .2768038809299469 us
*RESULT Compress256x256RandomColorImage: ETC1 Low= 1.834145426750183 us

BUG=434699
TEST=newly added unittest TextureCompressorETC1Test::Compress256x256CreateETC1, TextureCompressorETC1Test::Compress256x256RatioETC1

Review URL: https://codereview.chromium.org/1096703002

Cr-Commit-Position: refs/heads/master@{#329840}

WebRTC 9187:9189
Changes: https://chromium.googlesource.com/external/webrtc/trunk/webrtc.git/+log/65d8695..872cf25

TBR=sergeyu@chromium.org

Review URL: https://codereview.chromium.org/1136483003

Cr-Commit-Position: refs/heads/master@{#329839}

Calculates per-frame encode latency, which is basically the time delay from
input of each VideoFrame (VEA::Encode()) to output of the corresponding
BitstreamBuffer (VEA::Client::BitstreamBufferReady()).

It calculates encode latency values at the 50th(median), 75th, and 95th percentiles.

Also fix a wrong comment.

BUG=345181
TEST=Check "Encode latency for the 50%/75%/95%" in output logs.

Committed: https://crrev.com/b21ffeee2aff4da8edb9808b2227165c37e5ce86
Cr-Commit-Position: refs/heads/master@{#329819}

Review URL: https://codereview.chromium.org/1117853002

Cr-Commit-Position: refs/heads/master@{#329838}

The current version does not handle messages larger than a single characteristic size (around 150 bytes). Message receiving functionality is not implemented yet.

BUG=485123

Review URL: https://codereview.chromium.org/1116963002

Cr-Commit-Position: refs/heads/master@{#329837}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/dabd83a..33861c9

TBR=jsbell@chromium.org

Review URL: https://codereview.chromium.org/1142483002

Cr-Commit-Position: refs/heads/master@{#329836}

BUG=487989

Review URL: https://codereview.chromium.org/1135073003

Cr-Commit-Position: refs/heads/master@{#329835}

Revert of Revert of Use |IsOriginSecure| for mark-HTTP-nonsecure field trial (patchset #1 id:1 of https://codereview.chromium.org/1130843005/)

Reason for revert:
The initial patch was relanded: https://codereview.chromium.org/1124383007/

Original issue's description:
> Revert of Use |IsOriginSecure| for mark-HTTP-nonsecure field trial (patchset #1 id:1 of https://codereview.chromium.org/1143503003/)
>
> Reason for revert:
> Android Tests dbg bot is failing:
> http://build.chromium.org/p/chromium.linux/builders/Android%20Tests%20(dbg)
>
> The reason of this failure is following patch: https://codereview.chromium.org/1123943002/
>
> I had to revert this patch because it is a follow up.
>
> Original issue's description:
> > Use |IsOriginSecure| for mark-HTTP-nonsecure field trial
> >
> > When the flag or field trial is set such that non-secure origins will be
> > marked as dubious/non-secure, use |IsOriginSecure| to determine whether
> > an origin is non-secure, instead of simply treating http:// and ftp://
> > as non-secure. This means that http://nytimes.com will be marked as
> > dubious/non-secure, but not http://localhost (for example).
> >
> > BUG=487499
> > TEST=Enable the "Mark non-secure as non-secure" flag in
> > chrome://flags. Visit http://nytimes.com and observe that the lock icon
> > is red and crossed out. Visit an http:// URL on localhost and observe
> > that the lock icon is neutral.
> >
> > Committed: https://crrev.com/5f99677ac4f9b8180ee7fd2f0d5c85e91828a951
> > Cr-Commit-Position: refs/heads/master@{#329774}
>
> TBR=meacer@chromium.org,estark@chromium.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=487499
>
> Committed: https://crrev.com/236d015aa9bf3f61ece3e32c8af40fc1b9153727
> Cr-Commit-Position: refs/heads/master@{#329826}

TBR=meacer@chromium.org,estark@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=487499

Review URL: https://codereview.chromium.org/1144513003

Cr-Commit-Position: refs/heads/master@{#329834}

- Skip button aligned to Next button.
- Added Optional to step message.

BUG=487353

Review URL: https://codereview.chromium.org/1131073004

Cr-Commit-Position: refs/heads/master@{#329833}

Revert of Revert of Move SecurityLevel into a class of its own (patchset #1 id:1 of https://codereview.chromium.org/1135373003/)

Reason for revert:
Reland initial patch. There is a fix for SecurityLevelTest:

https://codereview.chromium.org/1135073003

Original issue's description:
> Revert of Move SecurityLevel into a class of its own (patchset #19 id:420001 of https://codereview.chromium.org/1123943002/)
>
> Reason for revert:
> Android Tests dbg bot is failing:
> http://build.chromium.org/p/chromium.linux/builders/Android%20Tests%20(dbg)
>
> Failure example: http://build.chromium.org/p/chromium.linux/builders/Android%20Tests%20%28dbg%29/builds/27901
>
> Original issue's description:
> > Move SecurityLevel into a class of its own
> >
> > This CL moves |GetSecurityLevelForWebContents| and the |SecurityLevel|
> > enum into a class of their own: |ConnectionSecurityHelper| in
> > //chrome/browser/ssl (since it is applying Chrome-specific security
> > policy to determine the security level of a page).
> >
> > Later, we will use the new |ConnectionSecurityHelper| class to compute a
> > security level for a page to send to the DevTools Security
> > panel. Specifically, we need to be able to calculate the security level
> > for a given page, not just the WebContents for the active tab.
> >
> > BUG=484392
> >
> > Committed: https://crrev.com/ca7c0a3374c21ab10c821f4f3a890afc79182acf
> > Cr-Commit-Position: refs/heads/master@{#329591}
> >
> > Committed: https://crrev.com/f50c9b4834e151aa6b5538545a503614cb74f5f9
> > Cr-Commit-Position: refs/heads/master@{#329724}
>
> TBR=felt@chromium.org,pkasting@chromium.org,rsleevi@chromium.org,bauerb@chromium.org,sky@chromium.org,tedchoc@chromium.org,estark@chromium.org
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=484392
>
> Committed: https://crrev.com/39722bb693112be3d91c729246c0ade351e7e900
> Cr-Commit-Position: refs/heads/master@{#329827}

TBR=felt@chromium.org,pkasting@chromium.org,rsleevi@chromium.org,bauerb@chromium.org,sky@chromium.org,tedchoc@chromium.org,estark@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=484392

Review URL: https://codereview.chromium.org/1124383007

Cr-Commit-Position: refs/heads/master@{#329832}

Favicon is resolved as data URL to avoid discrepancies if favicon depends on environment (e.g. user-agent).

BUG=461581

Review URL: https://codereview.chromium.org/1124023004

Cr-Commit-Position: refs/heads/master@{#329831}

In preparation for reusing touch transformations for transforming on
screen stylus events into display co-ordinates this CL does a simple
refactor of TouchTransformerController.

This change will be followed up by a subsequent CL that enables support
for having multiple touchscreens per display with the per-device looping
happening in the new methods: UpdateTouchRadius() and
UpdateTouchTransform().

TEST=ash_unittests passes, interactive testing on link_freon in mirrored
and non-mirrored mode.
BUG=448467

Review URL: https://codereview.chromium.org/1130983005

Cr-Commit-Position: refs/heads/master@{#329830}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/5f528ff..dabd83a

TBR=jsbell@chromium.org

Review URL: https://codereview.chromium.org/1141453004

Cr-Commit-Position: refs/heads/master@{#329829}

BUG=484690
R=nkostylev@chromium.org

Review URL: https://codereview.chromium.org/1138883003

Cr-Commit-Position: refs/heads/master@{#329828}

Revert of Move SecurityLevel into a class of its own (patchset #19 id:420001 of https://codereview.chromium.org/1123943002/)

Reason for revert:
Android Tests dbg bot is failing:
http://build.chromium.org/p/chromium.linux/builders/Android%20Tests%20(dbg)

Failure example: http://build.chromium.org/p/chromium.linux/builders/Android%20Tests%20%28dbg%29/builds/27901

Original issue's description:
> Move SecurityLevel into a class of its own
>
> This CL moves |GetSecurityLevelForWebContents| and the |SecurityLevel|
> enum into a class of their own: |ConnectionSecurityHelper| in
> //chrome/browser/ssl (since it is applying Chrome-specific security
> policy to determine the security level of a page).
>
> Later, we will use the new |ConnectionSecurityHelper| class to compute a
> security level for a page to send to the DevTools Security
> panel. Specifically, we need to be able to calculate the security level
> for a given page, not just the WebContents for the active tab.
>
> BUG=484392
>
> Committed: https://crrev.com/ca7c0a3374c21ab10c821f4f3a890afc79182acf
> Cr-Commit-Position: refs/heads/master@{#329591}
>
> Committed: https://crrev.com/f50c9b4834e151aa6b5538545a503614cb74f5f9
> Cr-Commit-Position: refs/heads/master@{#329724}

TBR=felt@chromium.org,pkasting@chromium.org,rsleevi@chromium.org,bauerb@chromium.org,sky@chromium.org,tedchoc@chromium.org,estark@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=484392

Review URL: https://codereview.chromium.org/1135373003

Cr-Commit-Position: refs/heads/master@{#329827}

Revert of Use |IsOriginSecure| for mark-HTTP-nonsecure field trial (patchset #1 id:1 of https://codereview.chromium.org/1143503003/)

Reason for revert:
Android Tests dbg bot is failing:
http://build.chromium.org/p/chromium.linux/builders/Android%20Tests%20(dbg)

The reason of this failure is following patch: https://codereview.chromium.org/1123943002/

I had to revert this patch because it is a follow up.

Original issue's description:
> Use |IsOriginSecure| for mark-HTTP-nonsecure field trial
>
> When the flag or field trial is set such that non-secure origins will be
> marked as dubious/non-secure, use |IsOriginSecure| to determine whether
> an origin is non-secure, instead of simply treating http:// and ftp://
> as non-secure. This means that http://nytimes.com will be marked as
> dubious/non-secure, but not http://localhost (for example).
>
> BUG=487499
> TEST=Enable the "Mark non-secure as non-secure" flag in
> chrome://flags. Visit http://nytimes.com and observe that the lock icon
> is red and crossed out. Visit an http:// URL on localhost and observe
> that the lock icon is neutral.
>
> Committed: https://crrev.com/5f99677ac4f9b8180ee7fd2f0d5c85e91828a951
> Cr-Commit-Position: refs/heads/master@{#329774}

TBR=meacer@chromium.org,estark@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=487499

Review URL: https://codereview.chromium.org/1130843005

Cr-Commit-Position: refs/heads/master@{#329826}

Revert of vea_unittest: Calculate per-frame encode latency (patchset #15 id:280001 of https://codereview.chromium.org/1117853002/)

Reason for revert:
Fail to compile on x86 Chrome OS.

Original issue's description:
> vea_unittest: Calculate per-frame encode latency
>
> Calculates per-frame encode latency, which is basically the time delay from
> input of each VideoFrame (VEA::Encode()) to output of the corresponding
> BitstreamBuffer (VEA::Client::BitstreamBufferReady()).
>
> It calculates encode latency values at the 50th(median), 75th, and 95th percentiles.
>
> Also fix a wrong comment.
>
> BUG=345181
> TEST=Check "Encode latency for the 50%/75%/95%" in output logs.
>
> Committed: https://crrev.com/b21ffeee2aff4da8edb9808b2227165c37e5ce86
> Cr-Commit-Position: refs/heads/master@{#329819}

TBR=owenlin@chromium.org,posciak@chromium.org,wuchengli@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=345181

Review URL: https://codereview.chromium.org/1137643003

Cr-Commit-Position: refs/heads/master@{#329825}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/62b675e..5f528ff

TBR=jsbell@chromium.org

Review URL: https://codereview.chromium.org/1138503004

Cr-Commit-Position: refs/heads/master@{#329824}

BUG=478250
R=mkwst@chromium.org

Review URL: https://codereview.chromium.org/1145463002

Cr-Commit-Position: refs/heads/master@{#329823}

This patch is just a cleanup. No functionality is changed.

BUG=none
TEST=none

Review URL: https://codereview.chromium.org/1129023005

Cr-Commit-Position: refs/heads/master@{#329822}

This patch adds the code to reshow the notification after user clicks "Keep" button on dangerous download notification.

This patch also adds the tests of dangerous downloads.

BUG=480489
TEST=browsertest run

Review URL: https://codereview.chromium.org/1130353004

Cr-Commit-Position: refs/heads/master@{#329821}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/49baa47..62b675e

TBR=jsbell@chromium.org

Review URL: https://codereview.chromium.org/1141043002

Cr-Commit-Position: refs/heads/master@{#329820}

Calculates per-frame encode latency, which is basically the time delay from
input of each VideoFrame (VEA::Encode()) to output of the corresponding
BitstreamBuffer (VEA::Client::BitstreamBufferReady()).

It calculates encode latency values at the 50th(median), 75th, and 95th percentiles.

Also fix a wrong comment.

BUG=345181
TEST=Check "Encode latency for the 50%/75%/95%" in output logs.

Review URL: https://codereview.chromium.org/1117853002

Cr-Commit-Position: refs/heads/master@{#329819}

Every call to base::Bind results in a new class declaration. Since
base::internal::BindState is a virtual class, this also results in a
bespoke vtable being defined. And that's a lot of vtables. Like 1MB
worth of vtables, or 1% of Chrome's binary size.

However, this vtable's sole purpose is for RefCountedThreadSafe::
Release(). Release() _does_ need to know precisely how to delete the
BindState, but it doesn't need a vtable to do it. All it needs is a
function pointer.

This CL de-virtualizes base::BindState and instead moves the vtable
pointer (which is merely an array holding the destructor) to a function
pointer that contains the destructor instead. Since BindState previously
contained a pointer-to-vtable the net memory effect of having the
function pointer instead should be offset.

And of course a 1MB binary size drop also means reduced memory
requirements for the shared text-segment pages.

Locally produced numbers for ChromiumFramework on Mac:
pre-strip:
 - before: 192,031kB
 - after: 185,800kB
post-strip:
 - before: 101,353kB
 - after: 100,353kB

BUG=486594

Review URL: https://codereview.chromium.org/1129353003

Cr-Commit-Position: refs/heads/master@{#329818}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/1b19b6f..49baa47

TBR=alancutter@chromium.org,jsbell@chromium.org

Review URL: https://codereview.chromium.org/1135823006

Cr-Commit-Position: refs/heads/master@{#329817}

Cr-Commit-Position: refs/heads/master@{#329816}

The all slow tests have a possibility to be failed on slow bots (Debug build, ASAN, MSAN, TSAN bots). So this patch disabled them as a whole by defining DISABLE_SLOW_FILESPAPP_TEST macro.

BUG=none
TEST=run Debug, ASAN, MSAN, TSAN bots

Review URL: https://codereview.chromium.org/1139823003

Cr-Commit-Position: refs/heads/master@{#329815}

BUG=486571
R=thakis@chromium.org

Review URL: https://codereview.chromium.org/1137373002

Cr-Commit-Position: refs/heads/master@{#329814}

An N pattern cadence is simply figuring out if N times the
frame duration has an integer cadence (M) relative to the
render interval, then figuring out how to distribute N
values such that sum(x1 + x2 + x3 ... xN) == M.

I haven't dug through our discussion on how to compute this
for N > 2, but for N == 2, it's simple to just use
x1 == ceil(M/2), x2 == floor(M/2)

This change refactors the main function in VideoCadenceEstimator
into smaller pieces (an original review request :) in order to
reuse them for detecting 2-pattern cadence.

BUG=439548
TEST=all unit tests still pass, 3:2 cadence tests pass.

Review URL: https://codereview.chromium.org/1125893002

Cr-Commit-Position: refs/heads/master@{#329813}

BUG=487474
TEST=manual

Review URL: https://codereview.chromium.org/1132983005

Cr-Commit-Position: refs/heads/master@{#329812}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/0f5ff62..1b19b6f

TBR=alancutter@chromium.org,jsbell@chromium.org

Review URL: https://codereview.chromium.org/1137963003

Cr-Commit-Position: refs/heads/master@{#329811}

TBR=suzhe@chromium.org
BUG=407456
TEST=Verified on linux_chromeos & clapper device.

Review URL: https://codereview.chromium.org/1132333002

Cr-Commit-Position: refs/heads/master@{#329810}

Previously BatchableRequestBase can be used both as a normal request and as a
child request in BatchUploadRequest. Even when BatchableRequestBase is used as
child request, its lifetime must be managed by RequestSender, and it makes hard
to track the lifetime in BatchUploadRequest.

To simplify lifetime management in BatchUploadRequest, the CL extracts the core
part of BatchableRequestBase into BatchableDelegate, and lets BatchUploadRequest
use the delegate as a child request. Because the delegate is nolonger owned by
RequestSender, BatchUploadRequest can own its child requests without caring
RequestSender.

BUG=451917
TEST=googleapis_unittests

Review URL: https://codereview.chromium.org/1132693006

Cr-Commit-Position: refs/heads/master@{#329809}

And avoid dispatching onActivate & onDeactivated events to IME
extensions if the profile is not the active profile.

BUG=485835
TEST=Verified on linux_chromeos & clapper device.

Review URL: https://codereview.chromium.org/1136463005

Cr-Commit-Position: refs/heads/master@{#329808}

MessageLoopProxy is deprecated.
This basically does a search and replace of:

MessageLoopProxy::current() -> ThreadTaskRunnerHandle::Get().

MessageLoopProxy -> SingleThreadTaskRunner

BUG=391045

Review URL: https://codereview.chromium.org/1124813003

Cr-Commit-Position: refs/heads/master@{#329807}

BUG=487505
TEST=GYP_GENERATORS=ninja tools/gyp/gyp --depth . ui/file_manager/file_manager/background/js/compiled_resources.gyp && ninja -C out/Default

Review URL: https://codereview.chromium.org/1140873002

Cr-Commit-Position: refs/heads/master@{#329806}

It's been off by default since Chrome 40 with a policy opt-in. The policy was
documented to stop working after Chrome 43, so remove all logic for SSL 3.0.
Chrome will no longer honor a --ssl-version-min=ssl3 flag and the option is
removed from about:flags.

This CL also removes the SSL 3.0 support from the extensions' TLS socket API.

BUG=487730

Review URL: https://codereview.chromium.org/1139013002

Cr-Commit-Position: refs/heads/master@{#329805}

TEST=Tested manually in open dialog, and standalone Files app.
BUG=487905

Review URL: https://codereview.chromium.org/1134213006

Cr-Commit-Position: refs/heads/master@{#329804}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/7aa732b..0f5ff62

TBR=alancutter@chromium.org,jsbell@chromium.org

Review URL: https://codereview.chromium.org/1138953005

Cr-Commit-Position: refs/heads/master@{#329803}

R=mkwst@chromium.org
BUG=478250

Review URL: https://codereview.chromium.org/1142463002

Cr-Commit-Position: refs/heads/master@{#329802}

The ResourceRecords must contain the ServiceWorker's main script.
So the size of it must be bigger than 1.

If the method returns STATUS_ERROR_CORRUPTED, DeleteAndStartOver() will be executed.

BUG=485900,478732

Review URL: https://codereview.chromium.org/1135743002

Cr-Commit-Position: refs/heads/master@{#329801}