GitLab

Introduce "Preferred" accelerators, which may be processed by fullscreen, but will otherwise be processed first.

BUG=398905
TEST=covered by unit tests

Review URL: https://codereview.chromium.org/582143004

Cr-Commit-Position: refs/heads/master@{#295823}

This will be needed to move the "paused message queue" into the
ChannelEndpoint, since transferring messages from the
LocalMessagePipeEndpoint needs to happen "atomically" with its
conversion to a ProxyMessagePipeEndpoint. (Before the MessagePipe lock
is released, the ProxyMessagePipeEndpoint needs to have a place to queue
messages, and any already-queued messages must be at the front of that
queue.)

R=darin@chromium.org

Review URL: https://codereview.chromium.org/585093002

Cr-Commit-Position: refs/heads/master@{#295822}

Summary of changes available at:
https://chromium.googlesource.com/chromium/blink/+log/686c9e7..5f95dbe

TBR=caseq@chromium.org,pdr@chromium.org,wjmaclean@chromium.org

Review URL: https://codereview.chromium.org/588773002

Cr-Commit-Position: refs/heads/master@{#295821}

BUG=388885

Review URL: https://codereview.chromium.org/585103002

Cr-Commit-Position: refs/heads/master@{#295820}

BUG=316383

Review URL: https://codereview.chromium.org/566643002

Cr-Commit-Position: refs/heads/master@{#295819}

Review URL: https://codereview.chromium.org/590613002

Cr-Commit-Position: refs/heads/master@{#295818}

BUG=

Review URL: https://codereview.chromium.org/581853002

Cr-Commit-Position: refs/heads/master@{#295817}

new api on ApplicationManager.

BUG=

Review URL: https://codereview.chromium.org/568173003

Cr-Commit-Position: refs/heads/master@{#295816}

Revert of Enable incremental linking for static_library too (patchset #2 id:20001 of https://codereview.chromium.org/502993005/)

Reason for revert:
Does not appear to work well for waterfall workload, see number of instances of "performing full link" in http://build.chromium.org/p/chromium.win/builders/Win%20Builder/builds/2780/steps/compile/logs/stdio

Might still be worth investigating if it's worthwhile on trybots where less things might change per-job? But doesn't seem that likely.

Original issue's description:
> Enable incremental linking for static_library too
>
> Timings on z620:
>
> Without:
>
> GYP_DEFINES=component=static_library chromium_win_pch=0 incremental_chrome_dll=0
>
> touch content\app\main.cc (causes chrome.dll, chrome_child.dll, and unit_tests.exe to relink)
> tim ninja -C out\Release unit_tests
> [19/19] LINK_EMBED unit_tests.exe
>
> real: 4m46.828s
>
> With:
>
> GYP_DEFINES=component=static_library branding=Chrome chromium_win_pch=0 incremental_chrome_dll=1
>
> touch content\app\content_main.cc && tim ninja -C out\Release unit_tests
> ninja: Entering directory `out\Release'
> [19/19] LINK_EMBED unit_tests.exe
>
> real: 0m43.625s
>
> This is basically best case, and the actual performance on bots will differ based on
> their config and the number of files that change from run-to-run, but it seems worth
> enabling to see what happens, at least.
>
> R=jam@chromium.org
> BUG=404809,402270
>
> Committed: https://crrev.com/6922bc1d4b0a2f4604598082a877d3add6d86d1c
> Cr-Commit-Position: refs/heads/master@{#295786}

TBR=jam@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=404809,402270

Review URL: https://codereview.chromium.org/591723003

Cr-Commit-Position: refs/heads/master@{#295815}

BUG=388885

Review URL: https://codereview.chromium.org/587833003

Cr-Commit-Position: refs/heads/master@{#295814}

This is a quick hack to make the initializeChildFrame message go through
RenderFrameProxy rather than RenderFrameImpl, since the RenderFrameHost
is deleted before it can receive the message.

We can clean up the hack by moving initializeChildFrame to
WebRemoteFrameClient and removing WebFrame::setIsRemote.

BUG=416102,407160
TEST=Simple OOPIF paints in --site-per-process.

Review URL: https://codereview.chromium.org/591503003

Cr-Commit-Position: refs/heads/master@{#295813}

R=estade@chromium.org
BUG=393873

Review URL: https://codereview.chromium.org/586783003

Cr-Commit-Position: refs/heads/master@{#295812}

This is a band-aid fix for crbug.com/415167, which I'd like to try to get into M38.
Apparently it's possible to have "temporary" entry IDs set as pending and then
thrown away. This confuses OverscrollNavigationOverlay into thinking that the
page was never loaded and the overlay never gets dismissed as a result, which
makes the webpage unusable.
Using URLs is not as reliable avoids this problem. It can introduce false
positives in some corner cases, however a false positive doesn't have severe
consequences - we simply dismiss the overlay too soon, and the use may see an
empty page or a flicker.

BUG=415167

Review URL: https://codereview.chromium.org/575203002

Cr-Commit-Position: refs/heads/master@{#295811}

R=asvitkine@chromium.org,lcwu@chromium.org
BUG=400925

Review URL: https://codereview.chromium.org/587853002

Cr-Commit-Position: refs/heads/master@{#295810}

- Expose GetPendingEntry to the public Java layer for Android

BUG=385490

Review URL: https://codereview.chromium.org/577353002

Cr-Commit-Position: refs/heads/master@{#295809}

There's no reason to close the tag. Well-formedness doesn't matter.

R=thestig@chromium.org
BUG=393873

Review URL: https://codereview.chromium.org/588673003

Cr-Commit-Position: refs/heads/master@{#295808}

Instead of bitwise ORing the interstitial options, we were simply
assigning new values. This CL changes the assignment to a proper bitwise
OR so all options are maintained.

R=meacer@chromium.org,
TBR=jam@chromium.org,creis@chromium.org
BUG=415256

Review URL: https://codereview.chromium.org/578373002

Cr-Commit-Position: refs/heads/master@{#295807}

Revert of Update V8 to version 3.29.75 (based on bleeding_edge revision r24018). (patchset #1 id:1 of https://codereview.chromium.org/579123002/)

Reason for revert:
This roll appears to be causing major crashing.
BUG=416185

Original issue's description:
> Update V8 to version 3.29.75 (based on bleeding_edge revision r24018).
>
> TBR=machenbach@chromium.org
>
> Committed: https://crrev.com/73d6505aac73a6382fb8596153a2ea6f289df620
> Cr-Commit-Position: refs/heads/master@{#295435}

TBR=machenbach@chromium.org,v8-autoroll@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/591753002

Cr-Commit-Position: refs/heads/master@{#295806}

Make the "Keep Awake" extension load its saved state in
response to chrome.runtime.onInstalled in addition to
onStartup. Otherwise, it releases its wake locks after being
upgraded and fails to reacquire them until the next time
Chrome is restarted.

BUG=none
TBR=miket@chromium.org

Review URL: https://codereview.chromium.org/584033003

Cr-Commit-Position: refs/heads/master@{#295805}

BUG=330264

Review URL: https://codereview.chromium.org/590523003

Cr-Commit-Position: refs/heads/master@{#295804}

Otherwise chrome:// are not recognized as the search result.

This CL also modifies the logic of URLSearchProvider's result
aggregation, because the SearchProvider doesn't emit result change
for URL-ish results.

BUG=416083
R=oshima@chromium.org
TBR=mpearson@chromium.org, agl@chromium.org
TEST=manually

Review URL: https://codereview.chromium.org/584263002

Cr-Commit-Position: refs/heads/master@{#295803}

BUG=415656

Review URL: https://codereview.chromium.org/583583005

Cr-Commit-Position: refs/heads/master@{#295802}

Revert of Make mojo/services pass in "gn check" (patchset #3 id:40001 of https://codereview.chromium.org/581273003/)

Reason for revert:
Broke component build

Original issue's description:
> Make mojo/services pass in "gn check"
>
> This mostly just adds the missing dependency identified by the checker.
>
> In some cases I looked at the target and added public_deps that I thought were appropriate given the nature of the target.
>
> This renames the generated mojo action names to be foo__generator instead of foo_generator since it is not clear from the current name that it is an internally-generated target and I kept trying to use it. The double-underscore matches the Android generated targets.
>
> In a few cases I changed ordering (sources come before deps in GN style) but I didn't want to make the changes difficult to diff so didn't do it for most cases.
>
> R=darin@chromium.org
>
> Committed: https://crrev.com/e2311613971cf63c66b60ebea38b4592b8520b8f
> Cr-Commit-Position: refs/heads/master@{#295797}

TBR=davemoore@chromium.org,darin@chromium.org
NOTREECHECKS=true
NOTRY=true

Review URL: https://codereview.chromium.org/583423002

Cr-Commit-Position: refs/heads/master@{#295801}

If the user cancels their sign-in, or if some other error occurs, we
currently display "An unknown error occurred", which is not very helpful.

Review URL: https://codereview.chromium.org/585913003

Cr-Commit-Position: refs/heads/master@{#295800}

This reverts commit bd0aa4dc.

TBR=darin@chromium.org
BUG=415059

Review URL: https://codereview.chromium.org/584173002

Cr-Commit-Position: refs/heads/master@{#295799}

This should fix the iOS builder include errors as the previous Mac package failed
to include the include/ dir.

BUG=400088
R=thakis@chromium.org

Review URL: https://codereview.chromium.org/586923002

Cr-Commit-Position: refs/heads/master@{#295798}

This mostly just adds the missing dependency identified by the checker.

In some cases I looked at the target and added public_deps that I thought were appropriate given the nature of the target.

This renames the generated mojo action names to be foo__generator instead of foo_generator since it is not clear from the current name that it is an internally-generated target and I kept trying to use it. The double-underscore matches the Android generated targets.

In a few cases I changed ordering (sources come before deps in GN style) but I didn't want to make the changes difficult to diff so didn't do it for most cases.

R=darin@chromium.org

Review URL: https://codereview.chromium.org/581273003

Cr-Commit-Position: refs/heads/master@{#295797}

<include src="..."/>
<include src="..." />

R=tbreisacher@chromium.org
BUG=393873
NOTRY=true

Review URL: https://codereview.chromium.org/585053003

Cr-Commit-Position: refs/heads/master@{#295796}

BUG=388885
NOTRY=True

Review URL: https://codereview.chromium.org/575313003

Cr-Commit-Position: refs/heads/master@{#295795}

Update ui::GuessWindowManager() to detect popular
EWMH-supporting tiling window managers. Include these in the
Linux.WindowManager UMA histogram.

Detect Enlightenment 0.17 and later (enabling the custom
frame by default for it); previously only 0.16 was
detected.

Remove support for detecting chromeos-wm, the old Chrome OS
window manager that went away after Chrome OS switched to
Aura.

BUG=345482

Review URL: https://codereview.chromium.org/585553003

Cr-Commit-Position: refs/heads/master@{#295794}

BUG=400088
TBR=hans

Review URL: https://codereview.chromium.org/589773002

Cr-Commit-Position: refs/heads/master@{#295793}

Try failures are unrelated to this CL.
NOTRY=True

Review URL: https://codereview.chromium.org/575693002

Cr-Commit-Position: refs/heads/master@{#295792}

Revert of Removing browser.Start() and instead starts the browser at the point of creation. (patchset #5 id:80001 of https://codereview.chromium.org/565803002/)

Reason for revert:
missed changing constructor for CrOS browser

Original issue's description:
> Removing browser.Start() and instead starts the browser at the point of creation.
>
> BUG=412126
>
> Committed: https://crrev.com/01c68e8765254d9971296b291ff718b4f6af3004
> Cr-Commit-Position: refs/heads/master@{#295591}

TBR=nednguyen@google.com,slamm@chromium.org,slamm@google.com
NOTREECHECKS=true
NOTRY=true
BUG=412126

Review URL: https://codereview.chromium.org/587853003

Cr-Commit-Position: refs/heads/master@{#295791}

Folders should only have their id set by the server, and the clients
should only set the remote id on clips they created themselves.

Additionally, EnhancedBookmarkModel now monitors the remote id field for bookmarks, and initiates a de-duping protocol whenever two (or more) nodes with the same id are detected.
BUG=413876

Review URL: https://codereview.chromium.org/563363002

Cr-Commit-Position: refs/heads/master@{#295790}

Also remove an unused header and add it where actually used.

Review URL: https://codereview.chromium.org/587453005

Cr-Commit-Position: refs/heads/master@{#295789}

(Convert all uses to AttachEndpoint(). This means that Channel will no
longer create ChannelEndpoints.)

R=darin@chromium.org

Review URL: https://codereview.chromium.org/591573002

Cr-Commit-Position: refs/heads/master@{#295788}

BUG=416124
TEST=none
R=jam@chromium.org
TBR=jam@chromium.org

Review URL: https://codereview.chromium.org/586813003

Cr-Commit-Position: refs/heads/master@{#295787}

Timings on z620:

Without:

GYP_DEFINES=component=static_library chromium_win_pch=0 incremental_chrome_dll=0

touch content\app\main.cc (causes chrome.dll, chrome_child.dll, and unit_tests.exe to relink)
tim ninja -C out\Release unit_tests
[19/19] LINK_EMBED unit_tests.exe

real: 4m46.828s

With:

GYP_DEFINES=component=static_library branding=Chrome chromium_win_pch=0 incremental_chrome_dll=1

touch content\app\content_main.cc && tim ninja -C out\Release unit_tests
ninja: Entering directory `out\Release'
[19/19] LINK_EMBED unit_tests.exe

real: 0m43.625s

This is basically best case, and the actual performance on bots will differ based on
their config and the number of files that change from run-to-run, but it seems worth
enabling to see what happens, at least.

R=jam@chromium.org
BUG=404809,402270

Review URL: https://codereview.chromium.org/502993005

Cr-Commit-Position: refs/heads/master@{#295786}

R=hclam@chromium.org
TBR=hclam@chromium.org

Review URL: https://codereview.chromium.org/587823002

Cr-Commit-Position: refs/heads/master@{#295785}

With some changes in libvpx and webrtc, plus this change, we should be able to enable webrtc in the GN build.

R=hclam@chromium.org

Review URL: https://codereview.chromium.org/583353002

Cr-Commit-Position: refs/heads/master@{#295784}