GitLab

TBR=machenbach@chromium.org

Review URL: https://codereview.chromium.org/600673002

Cr-Commit-Position: refs/heads/master@{#296365}

expected.

BUG=405413

Review URL: https://codereview.chromium.org/488533002

Cr-Commit-Position: refs/heads/master@{#296364}

We used to just leak everything on Mac. Since r295194, we clean up
references to Profiles being deleted, in Chrome's AppListViewDelegate.
This has caught out the Mac app list trying to access stuff it shouldn't
if Cocoa is still drawing it while Chrome is shutting down. Tests
started failing flakily.

This CL adds AppListViewDelegateObserver::OnShutdown to that Chrome's
AppListViewDelegate can notify the AppListView that it's shutting down.

On Views platforms, the Widget is just closed (which also happens just
after the chrome::NOTIFICATION_APP_TERMINATING broadcast). However, (a)
Cocoa doesn't have a `CloseAllSecondaryWidgets` call to trigger this and
(b) just closing the NSWindow isn't usually enough on Cocoa to avoid
accessing C++ objects due to reference counting.

We could also go through the platform-specific AppListService (e.g. by
passing in the AppListService that creates a Chrome
AppListViewDelegate), but this seemed neater.

This de-flakes a bunch of applist tests on Mac. The Mac-specific
ShowAppListUsingShim test is also augmented with extra checks for the
OnShutdown stuff.

BUG=415264
TEST=AppListServiceMacInteractiveTest.ShowAppListUsingShim,
AppListServiceInteractiveTest.*

Review URL: https://codereview.chromium.org/593563002

Cr-Commit-Position: refs/heads/master@{#296363}

Becuase base/strings/string_split.h defines:
typedef std::vector<std::pair<std::string, std::string> > StringPairs;

BUG=412250

Review URL: https://codereview.chromium.org/599463002

Cr-Commit-Position: refs/heads/master@{#296362}

Revert of DevTools socket tunnel. (patchset #11 id:300001 of https://codereview.chromium.org/517233002/)

Reason for revert:
Fails to compile on Android: https://build.chromium.org/p/chromium.linux/builders/Android%20Arm64%20Builder%20(dbg)/builds/2163

devtools_bridge/android/javatests/src/org/chromium/components/devtools_bridge/SocketTunnelServerTest.java:291: cannot find symbol
symbol: method sleep()
                sleep();
                ^

Original issue's description:
> Implementation of DevTools socket tunneling service for remote debugging.
>
> Test APK in addition to instrumentation tests contains launching activity and foreground service what let duplicate ChromeShell debug socket to a differently named socket. If a device with running ChromeShell and DevToolsBridgeTest's service connect to a Desktop chrome (with USB debugging enabled) then chrome://inspect#devices will show 2 ChromeChell items. Both are debuggable (https://codereview.chromium.org/521573002/ should be applied to Desktop chrome to make the process less surprising). It's suitable for manual tests.
>
> This CL doesn't care of a few known issues:
> 1. Data channel buffer overflow. Data channel automatically closed in this case.
> 2. Uncontrolled number of pending sockets. DevTools throttles opening sockets and and keeps sockets that exceeds a threshold hanging.
> 3. Uncontrolled number of threads (it's actually not a problem if #2 solved, thanks to DevTools throttling).
>
> To keep this CL reasonably simple this issues will be addressed later.
>
> TEST=org.chromium.components.devtools_bridge.SocketTunnelServerTest, see description fro manual testing.
> BUG=383418
>
> Committed: https://crrev.com/15044609893faf9b2c2a533aa9a136f288ea20a1
> Cr-Commit-Position: refs/heads/master@{#296360}

TBR=mnaganov@chromium.org,tedchoc@chromium.org,erikwright@chromium.org,serya@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=383418

Review URL: https://codereview.chromium.org/597063002

Cr-Commit-Position: refs/heads/master@{#296361}

Test APK in addition to instrumentation tests contains launching activity and foreground service what let duplicate ChromeShell debug socket to a differently named socket. If a device with running ChromeShell and DevToolsBridgeTest's service connect to a Desktop chrome (with USB debugging enabled) then chrome://inspect#devices will show 2 ChromeChell items. Both are debuggable (https://codereview.chromium.org/521573002/ should be applied to Desktop chrome to make the process less surprising). It's suitable for manual tests.

This CL doesn't care of a few known issues:
1. Data channel buffer overflow. Data channel automatically closed in this case.
2. Uncontrolled number of pending sockets. DevTools throttles opening sockets and and keeps sockets that exceeds a threshold hanging.
3. Uncontrolled number of threads (it's actually not a problem if #2 solved, thanks to DevTools throttling).

To keep this CL reasonably simple this issues will be addressed later.

TEST=org.chromium.components.devtools_bridge.SocketTunnelServerTest, see description fro manual testing.
BUG=383418

Review URL: https://codereview.chromium.org/517233002

Cr-Commit-Position: refs/heads/master@{#296360}

BUG=402356
TEST=browser_tests --gtest_filter=SyncFileSystemTest.AuthorizationTest

Review URL: https://codereview.chromium.org/493043002

Cr-Commit-Position: refs/heads/master@{#296359}

Cr-Commit-Position: refs/heads/master@{#296358}

Revert of EventSender: Added checks for missing arguments (patchset #4 id:60001 of https://codereview.chromium.org/584203003/)

Reason for revert:
This makes the fast/events/hit-test-counts.html layout test fail. Reverting for now.

http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fast%2Fevents%2Fhit-test-counts.html&testType=layout-tests

Original issue's description:
> Added checks for missing arguments in event_sender.cc. Also simplified the uses of x & y in EventSender::GestureEvent.
>
> BUG=415970
>
> Committed: https://crrev.com/302907463636d2ddfe25c5381bd1f2b64016d1c6
> Cr-Commit-Position: refs/heads/master@{#296256}

TBR=dpranke@chromium.org,rbyers@chromium.org,earthdok@chromium.org,mustaq@chromium.org
NOTREECHECKS=true
NOTRY=true
BUG=415970

Review URL: https://codereview.chromium.org/600733002

Cr-Commit-Position: refs/heads/master@{#296357}

BUG=344769

Review URL: https://codereview.chromium.org/554213002

Cr-Commit-Position: refs/heads/master@{#296356}

The Android link disambiguation popup is rendered by PopupZoomer on the Java side. This CL tries to recreate that UI in Aura using a BubbleDelegate.

This part works just like the Android code:

When Blink detects an ambiguous gesture it calls RenderViewImpl::didTapMultipleTargets(), which uses paintCompositedDeprecated() in the WebWidget to get a zoomed bitmap region containing the ambiguous links the gesture could possibly indicate. RenderViewImpl then sends a ShowDisambiguationPopup message via IPC which ultimately ends up in RenderWidgetHostImpl::OnShowDisambiguationPopup(), which marshals the bitmap and then then calls ShowDisambiguationPopup on the view_.

This part is specific to aura:

After a series of delegation steps the callback ends up in WebContentsViewAura::OnLinkDisambiguationPopupRequested, which converts the target rectangle from pixels to DIPs and then requests the popup be rendered by asking its WebContentsViewDelegate, ChromeWebContentsViewDelegateViews, which owns a LinkDisambiguationPopup and asks it to render.

The LinkDisambiguationPopup uses BubbleDelegate to show the zoomed image. Any GestureEvents it receives converted into the coordinate space of the content view and then are relayed via callback back to WebContentsViewAura, which then converts them in to Blink events and sends them on to the RenderWidgetHostImpl for relaying back to Blink.

jam: content/ parts
sky: everything else?

BUG=385249

Review URL: https://codereview.chromium.org/351683002

Cr-Commit-Position: refs/heads/master@{#296355}

BUG=380213, 416961, 408734, 370175
R=stevenjb@chromium.org
TEST=manually

Review URL: https://codereview.chromium.org/593273002

Cr-Commit-Position: refs/heads/master@{#296354}

This is the preparation for introducing other types of file system supported by
external file system backend into the drive protocol handler.

The CL does not change the format of drive URL.

BUG=367027
TEST=Open HTML, MHTML, PDF, GDOC, OGG, MP4 files on the drive volume.

Review URL: https://codereview.chromium.org/560313002

Cr-Commit-Position: refs/heads/master@{#296353}

This patch will fixed the build errors which is due to while notification flag is
enabled and extension flag is disabled.

BUG=None

Review URL: https://codereview.chromium.org/582673002

Cr-Commit-Position: refs/heads/master@{#296352}

Adding split view divider widget which could be dragged to exit the split view.

BUG=403207,408691

Review URL: https://codereview.chromium.org/545393002

Cr-Commit-Position: refs/heads/master@{#296351}

Previously, this test would crash if the output wasn't quoted properly
(since it assumes that cmd_string.length() >= program_string.length() +
2). Instead, directly compare the string with the expected output. This
also guards against the possibility of mis-constructing the expected
string.

Review URL: https://codereview.chromium.org/589393002

Cr-Commit-Position: refs/heads/master@{#296350}

If you click on the SSL interstitial's error code, the following is
now shown:
* subject
* issuer
* validity end date
* current date
* pem encoded chain

This CL adds a placeholder for debugging information to the security
interstitial. It's intentionally meant to be flexible, so that it could
support other debugging information in the future from other warnings.

BUG=413838

Review URL: https://codereview.chromium.org/584463005

Cr-Commit-Position: refs/heads/master@{#296349}

call OnTerminating in unit test.

BUG=None

Review URL: https://codereview.chromium.org/599683003

Cr-Commit-Position: refs/heads/master@{#296348}

App windows launched in windows ASH should not be messing with the AppUserModelId for the remote HWND.

This causes problems on Windows 7 Chrome OS mode where launching popup app windows in Chrome OS mode
causes the taskbar button for the remote HWND to change to the application button. On Windows 8
these calls silently fail.

BUG=409824

Review URL: https://codereview.chromium.org/589303002

Cr-Commit-Position: refs/heads/master@{#296347}

The close button was setting the AXDescription property, it somehow caused
VoiceOver to use it as the tab view's description. Change it to set the AXTitle
property instead to fix the issue, and to match Safari.

Also change the tab strips' AX children to be sorted from left to right, so
that VoiceOver's "Tab X of Y" order is always ascending left to right.

BUG=415570

Review URL: https://codereview.chromium.org/588393002

Cr-Commit-Position: refs/heads/master@{#296346}

This is to help unblock some of the Blink work to move local frame only
APIs off WebFrame. Strictly speaking, all that was required to unblock
this bug was to manually coerce the WebFrame to a WebLocalFrame for the
call to setIsolatedWorldHumanReadableName.

However, it never makes sense to inject script on a remote frame, so
I've updated ScriptInjection to just work with local frames where
possible. Code that uses ScriptInjection now does the type coercion at
the class boundary. As extensions code is gradually updated to
understand out-of-process iframes, these manual type coercions should
disappear.

BUG=416659

Review URL: https://codereview.chromium.org/594043002

Cr-Commit-Position: refs/heads/master@{#296345}

Cleanup: Remove the instances of "rm_binaries.py" from 'src/tools/gn/bin'. The GN binaries have moved to src/buildtools. So, this script was removing GN binaries from tools/gn/bin that aren't used anymore. It was kept with the same intent till july 2014 (expected), and now think so, it is the good time to remove this script.

BUG=

Review URL: https://codereview.chromium.org/598493002

Cr-Commit-Position: refs/heads/master@{#296344}

BUG=397019

Review URL: https://codereview.chromium.org/545373005

Cr-Commit-Position: refs/heads/master@{#296343}

They were accidentally enabled with the switch to recipes and they don't work
because the older builds don't necessarily work with the generated profile.

BUG=388885
TBR=dtu@chromium.org

Review URL: https://codereview.chromium.org/596783005

Cr-Commit-Position: refs/heads/master@{#296342}

Older versions of Chrome expanded thumbnails to 1024x1024 to work around
an IMG driver bug.  When these were processed through DecompressionTask,
they got corrupted because this method assumed the content size and
buffer size were the same.  This patch changes it to decompress into a
potentially larger buffer then copy if needed.

BUG=416437

Review URL: https://codereview.chromium.org/598593003

Cr-Commit-Position: refs/heads/master@{#296341}

This adds Java code for running junit tests, as well as gyp targets for
both runnable and non-runnable host-side JARs.

BUG=316383

Review URL: https://codereview.chromium.org/574433003

Cr-Commit-Position: refs/heads/master@{#296340}

Rebased on https://codereview.chromium.org/588893006/

Comments that look like standalone line comments but immediately follow
a previous suffix comment should be considered suffix as well so they
can be added to the same node.

R=brettw@chromium.org
BUG=348474

Review URL: https://codereview.chromium.org/595753002

Cr-Commit-Position: refs/heads/master@{#296339}

Looks like a typo from https://codereview.chromium.org/522553002.
NOTRY to green up the perf bots.

BUG=413637
TEST=None.
TBR=nednguyen
NOTRY=True

Review URL: https://codereview.chromium.org/601663002

Cr-Commit-Position: refs/heads/master@{#296338}

StartPage/EndPage from spooling code. Calling this from metafiles unreliable. For example it's not played if for empty page metafiles.
Removed multipage support from windows metafile. We never use that, but it cause issues with empty files.

BUG=412690

Review URL: https://codereview.chromium.org/599733002

Cr-Commit-Position: refs/heads/master@{#296337}

BUG=414781, 416859
TEST=Manual, see bug

Review URL: https://codereview.chromium.org/598603003

Cr-Commit-Position: refs/heads/master@{#296336}

Use sidestep to detour CertVerifyCertificateSignatureEx on Windows versions earlier than Vista (excluding XP SP3), adding in SHA-256 support by deferring to NSS.

The canonical path to supporting SHA-256 is to install XP SP3 or the appropriate hotfixes for XP x64 / Windows Server 2003. However, as not all users may do so, and there's enough of a usability hurdle, provide an interception hook until we fully drop support for these systems.

XP is already outside of MSFT EOL (April 2014). Windows Server 2003 is EOL'd July 2015.

BUG=401365

Review URL: https://codereview.chromium.org/561613002

Cr-Commit-Position: refs/heads/master@{#296335}

 - Request batch attestation instead of individual
 - Updates from U2F API revisions
 - Enumerate HID devieces based on usage page instead of vid/pid
 - Handle wrong keyHandle length and init messages
 - Smaller bug fixes and refactoring

BUG=416998
R=xiyuan,juanlang

Review URL: https://codereview.chromium.org/596083002

Cr-Commit-Position: refs/heads/master@{#296334}

content: Fix OOM error handling in GpuMemoryBufferImplSharedMemory::AllocateSharedMemoryForChildProcess.

We need to run the callback with an empty handle after failing
to allocated shared memory.

BUG=
TBR=piman@chromium.org

Review URL: https://codereview.chromium.org/595063004

Cr-Commit-Position: refs/heads/master@{#296333}

BUG=415422

Review URL: https://codereview.chromium.org/602483002

Cr-Commit-Position: refs/heads/master@{#296332}

(Split off of https://codereview.chromium.org/588813002)

The main change here is to restructure the ParseNode tree to include
comments. Comments are not included in the tree as normal nodes, but
instead attached subordinate to other |ParseNode|s. This is useful
for two reasons:
- it means neither the evaluator nor the main code of the parser need
  to get more complicated, as they do not need to constantly consider
  comments.
- it confers additional information about the relative location of
  of the comments to their surrounding code when (in the future)
  outputting the formatted code.

Comments are added in relation to the node they're attached to as
before, after, or suffix. This is so that formatting can keep them
attached to the correct thing when things are reflowed.

Timing data (Windows, Release):

Before:
Done. Wrote 1618 targets from 530 files in 2613ms
Done. Wrote 1618 targets from 530 files in 2952ms
Done. Wrote 1618 targets from 530 files in 3093ms

After:
Done. Wrote 1618 targets from 530 files in 2767ms
Done. Wrote 1618 targets from 530 files in 3120ms
Done. Wrote 1618 targets from 530 files in 2872ms

So, it seems within the noise at this point.

R=brettw@chromium.org
BUG=348474

Review URL: https://codereview.chromium.org/588893006

Cr-Commit-Position: refs/heads/master@{#296331}

Although passing SkBitmap by value doesn't copy the underlying pixels, it indeed
involves memcpy (~64 bytes to be precise) to replicate the settings from the
other bitmap in the copy constructor.

This overhead, albeit minor, can be avoided by passing SkBitmap as reference.

R=jamesr@chromium.org

Review URL: https://codereview.chromium.org/597553002

Cr-Commit-Position: refs/heads/master@{#296330}

BUG=417093,417119
NOTRY=true
TBR=groby@chromium.org

Review URL: https://codereview.chromium.org/600663003

Cr-Commit-Position: refs/heads/master@{#296329}

Review URL: https://codereview.chromium.org/598453004

Cr-Commit-Position: refs/heads/master@{#296328}

BUG=407376
TEST=no change
TBR=ben@chromium.org

Review URL: https://codereview.chromium.org/589403004

Cr-Commit-Position: refs/heads/master@{#296327}

Also reverses "OK" + "Cancel" buttons to match other overlays.

R=estade@chromium.org
BUG=407498

Review URL: https://codereview.chromium.org/591353002

Cr-Commit-Position: refs/heads/master@{#296326}