GitLab

Revert of Roll buildtools aa47d9773d..924d50ecae (patchset #1 id:1 of https://codereview.chromium.org/2138013004/ )

Reason for revert:
Broke uses of 'clang=1 llvm_force_head_revision=1' on clang tot waterfall:
https://build.chromium.org/p/chromium.fyi/console?category=clang%20tot

Original issue's description:
> Roll buildtools aa47d9773d..924d50ecae
>
>   In order to roll GN b9999d69..4828a631 (r403866:r404813) and pick up
>   the following changes:
>
>   5662a808 GN: Don't define argument overrides globally
>   45a841b0 Suggest `gn help` for internally set arguments.
>   bf462ea0 [GN docs] Add missing word in gn actions help string
>   8273381e Fix typos in GN error message
>
> TBR=dpranke@chromium.org
>
> Committed: https://crrev.com/932788e86f1f59cdff52df8ae78784af6f8a17e6
> Cr-Commit-Position: refs/heads/master@{#405054}

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

Review-Url: https://codereview.chromium.org/2146023002
Cr-Commit-Position: refs/heads/master@{#405169}

I moved this code out of VisualViewport in r404359 but forgot this one check
and early-out. The crashes started happening right after that so it's likely the
culprit.

BUG=627442

Review-Url: https://codereview.chromium.org/2140263002
Cr-Commit-Position: refs/heads/master@{#405168}

https://chromium.googlesource.com/skia.git/+log/e3aea10428d1..9199a9fef989

$ git log e3aea1042..9199a9fef --date=short --no-merges --format='%ad %ae %s'
2016-07-13 robertphillips Move GrDrawTarget::clear logic into GrDrawContext

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

Review-Url: https://codereview.chromium.org/2147853004
Cr-Commit-Position: refs/heads/master@{#405167}

BUG=588956

Review-Url: https://codereview.chromium.org/2133383002
Cr-Commit-Position: refs/heads/master@{#405166}

Revert of arc: Use the new InstanceHolder for unittests (patchset #4 id:60001 of https://codereview.chromium.org/2138513002/ )

Reason for revert:
Caused unit_tests failure on Linux ChromiumOS Tests (1):

ArcAppModelBuilderTest.RemoveAppCleanUpFolder

Original issue's description:
> arc: Use the new InstanceHolder for unittests
>
> Instead of having unit tests go through Mojo and jump through hoops to
> ensure that all messages have propagated to all the threads, we can use
> the new arc::ArcBridgeService::InstanceHolder<T>::SetInstance() to
> directly set the concrete fake (and optional version). This simplifies
> testing quite a bit.
>
> BUG=626695
> TEST=trybots
>
> Committed: https://crrev.com/c5fffc64c14a0464f8870bbe91df7e945f42c3ca
> Cr-Commit-Position: refs/heads/master@{#405138}

TBR=hidehiko@chromium.org,stevenjb@chromium.org,lhchavez@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=626695

Review-Url: https://codereview.chromium.org/2146573005
Cr-Commit-Position: refs/heads/master@{#405165}

Disable Full-WPO for the x64 official builds (this doesn't
affect the PGO builds).

This should help me to get an official
"non-PGO & non-WPO" build that I can compare against an official PGO
build.

Also I'm not sure if it's worth keeping WPO turned on for the official
builds at the moment ? We don't ship these builds (we ship the PGO ones)
and so this only slowdown the builders that build an official x64 build
for testing purposes (especially considering that they need to link all the unittests with Full-WPO)

BUG=490934, 617982

Review-Url: https://codereview.chromium.org/2142603002
Cr-Commit-Position: refs/heads/master@{#405164}

The underlying issues with Flash fullscreen and OOPIFs were already
fixed by https://codereview.chromium.org/1973813002, and this follows
up with a test.  Note that the test doesn't actually turn on
--site-per-process so that we get coverage both with OOPIFs (on Site
Isolation FYI bots) and without (on regular bots).

BUG=593522
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation

Review-Url: https://codereview.chromium.org/2123093005
Cr-Commit-Position: refs/heads/master@{#405163}

BaseFeatureProvider is currently the only implementation of
FeatureProvider (despite its name implying there should be more). As we
move to generated code for Features, pulling the code to vend features
away from the code to populate from a JSON file is necessary.

Pull the parsing/population code out of BaseFeatureProvider into a
JSONFeatureProvider class, inheriting from BaseFeatureProvider (which
maintains the code for returning features). Generated FeatureProviders
will also inherit from BaseFeatureProvider.

BUG=280286

Review-Url: https://codereview.chromium.org/2133783002
Cr-Commit-Position: refs/heads/master@{#405162}

BUG=619086

Review-Url: https://codereview.chromium.org/2140793002
Cr-Commit-Position: refs/heads/master@{#405161}

The UI calls MostVisitedSites::RecordOpenedMostVisitedItem when an item is clicked. Before this CL, MostVisitedSites would get some info from its |current_suggestions_|, but those might have changed in the meantime. Now, all required info is passed into RecordOpenedMostVisitedItem instead.

BUG=612160

Review-Url: https://codereview.chromium.org/2136503002
Cr-Commit-Position: refs/heads/master@{#405160}

Reland of Make components source_sets on Android. (patchset #1 id:1 of https://codereview.chromium.org/2151523002/ )

Reason for revert:
Didn't help.

Original issue's description:
> Revert of Make components source_sets on Android. (patchset #1 id:1 of https://codereview.chromium.org/2137023003/ )
>
> Reason for revert:
> Speculative, might've broke llvm_force_head_revision=true builds (see comment on the original review).
>
> Original issue's description:
> > Make components source_sets on Android.
> >
> > There was a perf regression reported when components were changed from source sets to static libraries in the non-component build. This patch switches back to source sets on Android only to see what happens to thread_times.key_silk_cases and the build performance.
> >
> > BUG=619593
> >
> > Committed: https://crrev.com/89248dc70996f9927292a7fee6911715464e1643
> > Cr-Commit-Position: refs/heads/master@{#405056}
>
> TBR=dpranke@chromium.org,brettw@chromium.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=619593
>
> Committed: https://crrev.com/aa5466df6e58dd54bc1dd1a5192f2c52ebf6ff29
> Cr-Commit-Position: refs/heads/master@{#405152}

TBR=dpranke@chromium.org,brettw@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=619593

Review-Url: https://codereview.chromium.org/2145063002
Cr-Commit-Position: refs/heads/master@{#405159}

GCCC_ERROR_ACCESSDENIED is a reason value, not a return value.

GetRealUserId() doesn't need to be called until it's needed, and if it
fails, only FindChromeTicket(kUserTicket) needs to be skipped, not the
whole rest of the function.

Review-Url: https://codereview.chromium.org/2151503002
Cr-Commit-Position: refs/heads/master@{#405158}

Before, Sandwich was modifying local cache directories by issuing
cachetool commands sequentially. But it was painfully slow.

This CL implement the cachetool's batch mode that process sequence
of command received from the stdin and return the commands results
in the stdout. This way, Sandwich just have to launch cachetool once
in the background and push to its stdin commands to process.

On cache directories populated by big webpages such as cnn.com's
home page, the Sandwich cache processing for NoState-Prefetch
emulation was taking 491s when spawning one cachetool process
for each commands versus 3s with this new online mode.

BUG=582080

Review-Url: https://codereview.chromium.org/2114933002
Cr-Commit-Position: refs/heads/master@{#405157}

When merging before submitting, this change did not make it into
codereview.chromium.org/2070633002.

As it stands, this CL reduces .dex size by 127kb and saves us 27kb of
dirty memory per process.

See paste.googleplex.com/6304692032241664 for what was removed.

BUG=619937

Review-Url: https://codereview.chromium.org/2147583003
Cr-Commit-Position: refs/heads/master@{#405156}

BUG=277199

Review-Url: https://codereview.chromium.org/2136073002
Cr-Commit-Position: refs/heads/master@{#405155}

Also makes the filter bar thinner.

Screenshots: https://imgur.com/a/CClua

First round of reviews here: https://codereview.chromium.org/2105023004

BUG=606897
TEST=open network panel, padding for selected filter (XHR, JS, CSS)
should be even (see screenshots)

Review-Url: https://codereview.chromium.org/2150503002
Cr-Commit-Position: refs/heads/master@{#405154}

https://chromium.googlesource.com/angle/angle.git/+log/9585a78..8495498

BUG=621031,627417

TBR=geofflang@chromium.org

TEST=bots

CQ_INCLUDE_TRYBOTS=tryserver.chromium.win:win_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.linux:linux_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2148803002
Cr-Commit-Position: refs/heads/master@{#405153}

Revert of Make components source_sets on Android. (patchset #1 id:1 of https://codereview.chromium.org/2137023003/ )

Reason for revert:
Speculative, might've broke llvm_force_head_revision=true builds (see comment on the original review).

Original issue's description:
> Make components source_sets on Android.
>
> There was a perf regression reported when components were changed from source sets to static libraries in the non-component build. This patch switches back to source sets on Android only to see what happens to thread_times.key_silk_cases and the build performance.
>
> BUG=619593
>
> Committed: https://crrev.com/89248dc70996f9927292a7fee6911715464e1643
> Cr-Commit-Position: refs/heads/master@{#405056}

TBR=dpranke@chromium.org,brettw@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=619593

Review-Url: https://codereview.chromium.org/2151523002
Cr-Commit-Position: refs/heads/master@{#405152}

BUG=626435
R=yukishiino@chromium.org, mek@chromium.org

Review-Url: https://codereview.chromium.org/2141083002
Cr-Commit-Position: refs/heads/master@{#405151}

R=brettw@chromium.org
BUG=596283
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_site_isolation

Review-Url: https://codereview.chromium.org/2131423003
Cr-Commit-Position: refs/heads/master@{#405150}

The ContentSuggestionsService retrieves suggestions from the
NTPSnippetsService as a provider. The latter currently delivers its
snippets to the chrome://snippets-internals/ page for debugging.
Before the NTP UI will be changed to retrieve suggestions from the
ContentSuggestionsService instead of directly from the
NTPSnippetsService, this adds the output of the
ContentSuggestionsService to the snippets-internals page to allow
debugging both and compare the data converted data to the original.
This also adds controls for clearing cached/discarded suggestions.

BUG=619560
CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:closure_compilation

Review-Url: https://codereview.chromium.org/2145563002
Cr-Commit-Position: refs/heads/master@{#405149}

(1) Remove redundant checks for bundled vs. component-updated Flash:
    This distinction is handled at registration time by the DCI.
(2) Update PathService after calling register. These two events are
    already racy, and it's more convenient to have the registration
    finished by the time the component updater notifies that the
    component has been updated.
(3) No need to explicitly pick up the bundled version of Flash for non-
    Linux: the component updater will manage registration of bundled
    Flash the same way it does component-updated Flash.

BUG=624067
TEST=https://docs.google.com/document/d/1iTQiaqjuHsKV4cPqSOet-eJKWb2SsJLp2ieDj_Mul4s/edit?pref=2&pli=1

Review-Url: https://codereview.chromium.org/2130803003
Cr-Commit-Position: refs/heads/master@{#405148}

BUG=none
R=pfeldman

Review-Url: https://codereview.chromium.org/2020803002
Cr-Commit-Position: refs/heads/master@{#405147}

When building with "gn", there is no way to perform the substitution
using shell commands to remove underscore from BUNDLE_ID_TEST_NAME,
so instead use :rfc1034identifier modifier implemented by the scripts
used both gyp and gn to perform substitutions in Info.plist files.

BUG=625875

Review-Url: https://codereview.chromium.org/2146483002
Cr-Commit-Position: refs/heads/master@{#405146}

It was changed wrongly by crrev.com/395903. This patch fixes this.

BUG=none

Review-Url: https://codereview.chromium.org/2139193002
Cr-Commit-Position: refs/heads/master@{#405145}

These DCHECKS sometimes cause crashes in tests.
A request may be destroyed before receiving a reply (e.g., if the
document is reloaded in the middle of a reply)

BUG=626893

Review-Url: https://codereview.chromium.org/2151443002
Cr-Commit-Position: refs/heads/master@{#405144}

This was missed in https://codereview.chromium.org/2100013002/

BUG=623158

Review-Url: https://codereview.chromium.org/2140173002
Cr-Commit-Position: refs/heads/master@{#405143}

The same histograms already exist on desktop, see ntp_user_data_logger.cc

BUG=625153

Review-Url: https://codereview.chromium.org/2126263002
Cr-Commit-Position: refs/heads/master@{#405142}

https://chromium.googlesource.com/external/github.com/catapult-project/catapult.git/+log/65c2d9df0d06..8b7bf3c29a56

$ git log 65c2d9df0..8b7bf3c29 --date=short --no-merges --format='%ad %ae %s'

BUG=627221

TBR=catapult-sheriff@chromium.org

Review-Url: https://codereview.chromium.org/2152433002
Cr-Commit-Position: refs/heads/master@{#405141}

NOTRY=true

Review-Url: https://codereview.chromium.org/2094583004
Cr-Commit-Position: refs/heads/master@{#405140}

This patch removes locks from ParsedChunkQueue, and use
of crossThreadBind in HTMLDocumentParser/BackgroundHTMLParser,
which introduces unecessary copies if we're not actually crossing
threads.

BUG=627198

Review-Url: https://codereview.chromium.org/2134323002
Cr-Commit-Position: refs/heads/master@{#405139}

Instead of having unit tests go through Mojo and jump through hoops to
ensure that all messages have propagated to all the threads, we can use
the new arc::ArcBridgeService::InstanceHolder<T>::SetInstance() to
directly set the concrete fake (and optional version). This simplifies
testing quite a bit.

BUG=626695
TEST=trybots

Review-Url: https://codereview.chromium.org/2138513002
Cr-Commit-Position: refs/heads/master@{#405138}

https://chromium.googlesource.com/skia.git/+log/ba3880fa6d47..e3aea10428d1

$ git log ba3880fa6..e3aea1042 --date=short --no-merges --format='%ad %ae %s'
2016-07-13 bungeman Remove user specified typeface id.

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

Review-Url: https://codereview.chromium.org/2149573002
Cr-Commit-Position: refs/heads/master@{#405137}

## Description of the problem
On iOS, tapping on an empty omnibox triggers a request to the
ClipboardUrlProvider.
The ClipboardURLProvider then calls VerbatimMatchForURL with an empty |input| (if the omnibox was empty), which eventually triggers a
DCHECK_GT in |AutocompleteMatch::ClassifyLocationInString|:

  // Classifying an empty match makes no sense and will lead to validation
  // errors later.
  DCHECK_GT(match_length, 0U);

## Description of the solution
This CL makes it forbidden to pass an empty input to
|VerbatimMatchForURL|, and fixes the clipboard_url_provider.

BUG=625313

Review-Url: https://codereview.chromium.org/2124563002
Cr-Commit-Position: refs/heads/master@{#405136}

This fixes the order of destruction of two of the members of RenderProcessHostImpl.

BUG=627778

Review-Url: https://codereview.chromium.org/2143013003
Cr-Commit-Position: refs/heads/master@{#405135}

- Rename listBoxSelectItem to selectMultipleOptionsByPopup, and it accepts index
  vector.
- Rename valueChanged to selectOptionByPopup.

Single-selection menulist accepts only selectOptionByPopup().
Multiple-selection menulist accepts selectOptionByPopup(-1) and
selectMultipleOptionsByPopup().

Review-Url: https://codereview.chromium.org/2143863003
Cr-Commit-Position: refs/heads/master@{#405134}

ipc_fuzzer starts failing to link recently due to undefined symbols to
NaCl messages on NaCl-disabled build.
This CL fixes the error by excluding them on the problematic case.

R=tsepez@chromium.org

Review URL: https://codereview.chromium.org/2136943002 .

Cr-Commit-Position: refs/heads/master@{#405133}

In Webgl's transferToImageBitmap(), we create an ImageBitmap from the
drawingBuffer's mailbox. When calling transferToImageBitmap() twice, the
mailbox is invalid in the second call. In this case, we should return
an ImageBitmap that is the same size as the drawingBuffer but should be
transparent black.

Corresponding layout test has been updated to reflect this case.

BUG=627374
CQ_INCLUDE_TRYBOTS=tryserver.chromium.win:win_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2139353003
Cr-Commit-Position: refs/heads/master@{#405132}

Reland of WebGL: use storage texture for DrawingBuffer only if CMAA is enabled. (patchset #1 id:1 of https://codereview.chromium.org/2133253003/ )

Reason for revert:
ATI bot failed due to this CL.

ATI bot was fine because we disabled WebglConformance.conformance_textures_misc_tex_image_webgl
However, https://codereview.chromium.org/2141893002/ enabled the test again, so this CL must be reverted.

Original issue's description:
> Revert of WebGL: use storage texture for DrawingBuffer only if CMAA is enabled. (patchset #3 id:60001 of https://codereview.chromium.org/2131863003/ )
>
> Reason for revert:
> ATI GPU bots became green, no matter this CL. ATI GPU bot might be in bad shape for some reasons.
> So revert this redundant speculative fix.
>
> Original issue's description:
> > WebGL: use storage texture for DrawingBuffer only if CMAA is enabled.
> >
> > It's because Linux ATI bot fails
> > WebglConformance.conformance_textures_misc_tex_image_webgl
> > When finding the solution for ATI driver, rollback this patch.
> >
> > BUG=557848
> > TEST=WebglConformance.conformance_textures_misc_tex_image_webgl
> > CQ_INCLUDE_TRYBOTS=tryserver.chromium.linux:linux_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel;tryserver.chromium.win:win_optional_gpu_tests_rel
> >
> > Committed: https://crrev.com/65f8f6c737278a80acee2bcd407d1fdc15426fa8
> > Cr-Commit-Position: refs/heads/master@{#404514}
>
> TBR=piman@chromium.org,kbr@chromium.org,zmo@chromium.org
> # Skipping CQ checks because original CL landed less than 1 days ago.
> NOPRESUBMIT=true
> NOTREECHECKS=true
> NOTRY=true
> BUG=557848
>
> Committed: https://crrev.com/e33a44a7077c60edfa28bc0d79b0910ab43d58d1
> Cr-Commit-Position: refs/heads/master@{#404563}

TBR=piman@chromium.org,kbr@chromium.org,zmo@chromium.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=557848

Review-Url: https://codereview.chromium.org/2143423002
Cr-Commit-Position: refs/heads/master@{#405131}

R=haraken@chromium.org
BUG=

Review-Url: https://codereview.chromium.org/2145863003
Cr-Commit-Position: refs/heads/master@{#405130}