Skip to content

GitLab

T
tangled
Switch branch/tag
  1. 11 Jul, 2017 12 commits
    • Jeff Carpenter's avatar
      Add 4 TestExpecations from WebKit Linux Trusty Leak · 0cbeffae
      Jeff Carpenter authored 
      After landing the CL enabling WPT tests that don't necessarily
correspond to real files (crrev.com/c/461722/), 4 Timeout 
TestExpectations from that CL turned up as Failures on WebKit Linux 
Trusty Leak. 

Bug: 703837
TBR: qyearsley@chromium.org
Change-Id: Ideb2d49c866f04d5105c0ad30446b6c92124f0f2
Reviewed-on: https://chromium-review.googlesource.com/566019
Commit-Queue: Jeff Carpenter <jeffcarp@chromium.org>
Reviewed-by: default avatarQuinten Yearsley <qyearsley@chromium.org>
Reviewed-by: default avatarBret Sepulveda <bsep@chromium.org>
Cr-Commit-Position: refs/heads/master@{#485464}
      0cbeffae
    • pdfium-deps-roller@chromium.org's avatar
      Roll src/third_party/pdfium/ 92e2276a8..00c3cfdba (1 commit) · 802e1f7a
      pdfium-deps-roller@chromium.org authored 
      https://pdfium.googlesource.com/pdfium.git/+log/92e2276a8be4..00c3cfdbae07

$ git log 92e2276a8..00c3cfdba --date=short --no-merges --format='%ad %ae %s'
2017-07-10 npm Roll FreeType to c56d8851ea987023cc73981a70d261b3f6427545

Created with:
  roll-dep src/third_party/pdfium


Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, see:
http://www.chromium.org/developers/tree-sheriffs/sheriff-details-chromium#TOC-Failures-due-to-DEPS-rolls


TBR=dsinclair@chromium.org

Change-Id: Id8dfcbbe74abfc5aec38128e39f5df4876d7a31b
Reviewed-on: https://chromium-review.googlesource.com/565943
Reviewed-by: <pdfium-deps-roller@chromium.org>
Commit-Queue: <pdfium-deps-roller@chromium.org>
Cr-Commit-Position: refs/heads/master@{#485463}
      802e1f7a
    • dougt's avatar
      Move two BrowserAccessibility role predicates to ax_role_properties. · 37f7a151
      dougt authored 
      BUG=703369

Review-Url: https://codereview.chromium.org/2969603002
Cr-Commit-Position: refs/heads/master@{#485462}
      37f7a151
    • lukasza's avatar
      Remove DOMAutomationController::automation_id_ · c7d6bd3d
      lukasza authored 
      Main change in this CL
======================

For a long time now the DOMAutomationController::automation_id_ value has
not been used for its original purpose (associating sends with
particular "jobs" / "automations" / render_views).  Right now, the
argument of domAutomationController::send is always routed back to
WebContents (not to any more granular/specific destination).

Right now the DOMAutomationController::automation_id_ field is still
used to enable/disable domAutomationController::send - the send only
works if setAutomationId was called previously (with a value different
than MSG_ROUTING_NONE).  In fact - in almost all cases, the scripts
say:
  domAutomationController.setAutomationId(0)
  domAutomationController.send(...)

This CL removes DOMAutomationController::automation_id_ field and
makes domAutomationController.send work unconditionally.

After this CL lands, we will be able to remove all calls to
setAutomationId (after this CL they will be no-ops).

Other changes needed by this CL
===============================

The CL needed to make the following, additional fixes:

1. ExecuteScript (in content/public/test/browser_test_utils.h / .cc)
   appends "; domAutomationController.send(0);" to the script provided
   by the caller.  This extra value would be usually ignored before
   this CL, but gets in the way after this CL (e.g. can get confused
   with other calls to domAutomationController.send(...)).  To account
   for this a new ExecuteScriptAsync function has been provided in
   this CL.

   Ideally ExecuteScript would become ExecuteScriptAsync, but
   there are multiple tests that depend on the extra pumping of the
   message loop that happens in ExecuteScript and not in
   ExecuteScriptAsync.  Therefore, ExecuteScriptAsync is
   a lesser evil, that minimizes amount of changes needed right now.

   To make sure that ExecuteScript stops waiting after the call to
   |domAutomationController.send(...)| that is added by ExecuteScript
   (and not by some other, accidental send) this CL sends a GUID that
   is unique in each call to ExecuteScript and verifies that the same
   GUID has been received back.  This change has flushed out a few
   additional cases where ExecuteScriptAsync needs to be used
   as well as cases where a call to domAutomationController.send
   can be removed.

   The following callers of ExecuteScript had to be switched to
   call ExecuteScriptAsync instead:

   - ExecuteWebUIResourceTest (consumes messages via DOMMessageQueue
     and relies on the fact that this queue is not stumped upon).
     Additionally, it turned out that client_renderer.js and
     media_internals.js (used by WebUIResourceBrowserTest tests) were
     always erroring out before this CL (this has gone unnoticed,
     because the "FAILURE" string would have been eaten by
     ExecuteScript).

   - SecurityExploitBrowserTest.InterstitialCommandFromUnderlyingContent
     (otherwise the artificual "0" ends up in interstitial->last_command()).

   - captive_portal_browsertest.cc (otherwise we trigger a CHECK in
     captive_portal_blocking_page.cc: Command 0 isn't handled by the
     captive portal interstitial).

     Note that CaptivePortalBrowserTest /
     ShowCaptivePortalInterstitialOnCertError was relying on message
     pumping done by ExecuteScript.  This was fragile - the timing of
     exiting the message pump was not related to the timing of the event
     the test needs to wait for.  This was broken after going through
     ExecuteScriptAsync.  This is fixed by introducing an explicit
     TabActivationWaiter.

   - ExecuteScriptInBackgroundPageNoWait (otherwise DOMMessageQueue
     used by ProcessManagerBrowserTest.ExtensionProcessReuse would
     have unexpectedly gotten a value sent by ExecuteScript, rather
     than the value send by the script itself;  additionally, calling
     Execute*Unmodified*Script here makes sense in a function named
     "NoWait" - ExecuteScript would have waited).

   - saml_browsertest.cc - there was undesired interaction with
     listening to the value sent by SetupAuthFlowChangeListener.

   - ChromeSitePerProcessTest.PopupWindowFocus - this would conflict
     with domAutomationController.send from
     chrome/test/data/page_with_focus_events.html

   The following callers of ExecuteScript were needlessly calling
   domAutomationController.send from the script:

   - ProxyAuthOnUserBoardScreenTest.ProxyAuthDialogOnUserBoardScreen
   - WizardControllerTest::JSExecute
   - NavigationControllerBrowserTest.PostInSubframe
   - SitePerProcessBrowserTest.NavigateRemoteFrameToBlankAndDataURLs

2. autofill_interactive_uitest.cc used to call
   domAutomationController.send(true) from onfocus DOM event handlers.
   This would confuse expectations of ExecuteScriptAndExtractInt running
   "domAutomationController.send(42)" (it would sometimes receive a
   boolean rather than int back).  This is fixed by using a one-time
   onfocus handlers (rather than permanent handlers) from
   FocusFieldByName method.

3. After the changes some GeolocationBrowserTest would receive
   "geoposition-updated" message twice - once from the
   geoSuccessCallbackWithResponse callback (from
   chrome/test/data/geolocation/basic_geolocation.js)
   and once from a checkIfGeopositionUpdated function.

   There is no need to resolve the race via checkIfGeopositionUpdated if
   we start listening (via DOMMessageQueue) *before* trigerring an
   update in the geo position.  This way we can keep just the callback
   and remove the no longer needed checkIfGeopositionUpdated function.

4. content::IsWebcamAvailableOnSystem (via kHasVideoInputDeviceOnSystem)
   would call domAutomationController.send(...) twice before this CL.
   After this CL it is only called once.

Other, opportunistic changes
============================

I've noticed that DOMMessageQueue::message_loop_runner_ can be null
and therefore uses of this field need to be guarded.  This required
remembering result of DOMMessageQueue::RenderProcessGone, so that
it still has an effect when there was no runner when the renderer
crash notification happened.

BUG=662543
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_site_isolation

Review-Url: https://codereview.chromium.org/2478803003
Cr-Commit-Position: refs/heads/master@{#485461}
      c7d6bd3d
    • Mike Dougherty's avatar
      Remove Javascript POST workaround when running on iOS 11 or later · fc015703
      Mike Dougherty authored 
      BUG=581021

Change-Id: Id6d7ab0cabc525cd575f6140281fdf5dc213a2b9
Reviewed-on: https://chromium-review.googlesource.com/565786Reviewed-by: default avatarEugene But <eugenebut@chromium.org>
Commit-Queue: Mike Dougherty <michaeldo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#485460}
      fc015703
    • Mark Mentovai's avatar
      Update GTM to ec72a2bc500a716369c383837bffdc7d2a22855b · 2b0d6d14
      Mark Mentovai authored 
      This is needed for 10.13 SDK compatibility. See
https://github.com/google/google-toolbox-for-mac/commit/ec72a2bc500a716369c383837bffdc7d2a22855b.

2ef3625a9193 s/GTMAddressBook/GTMABAddressBook/
a86965088edd First pass removing all deprecated uses of
             GTM_IPHONE_USE_SENTEST
157a381522da Merge pull request #129 from thomasvl/sentest_cleanup
9656673748be Bump the pod version to do a release
857a3f0ac59d Merge pull request #131 from thomasvl/update
5207aee433c5 Change file type from UTF16 to ASCII
cc2054645ca7 Merge pull request #134 from sergiocampama/type
94d4ccfc1589 Change include to import for GTMDefines
ca53d4727a2a Merge pull request #135 from sergiocampama/copy
9822411a0fcb Remove trailing whitespaces
e0fc2bad1e98 Merge pull request #136 from sergiocampama/white
89055f58cb50 Support for tvOS (#139)
4ae69e6f0b00 Fixed PodSpec for tvOS (#140)
d49d04d53abe Added -Wreserved-id-macro and but exclude existing
             nonconforming macros (#141)
d9d3b228b735 Avoid error from reserved macro name for GTMRegex.h
7ae58d8e92ef Provides a testing entitlements files
c867eb4055d7 Add default_subspecs to help avoid testing code ended up in
             apps
ec72a2bc500a 10.13 SDK (and iOS 11 equivalent) compatibility for GTM

Bug: 740712
Change-Id: Ia795ff53e53c53067b560bcb9111063507ab2455
Reviewed-on: https://chromium-review.googlesource.com/565770Reviewed-by: default avatarTrent Apted <tapted@chromium.org>
Commit-Queue: Mark Mentovai <mark@chromium.org>
Cr-Commit-Position: refs/heads/master@{#485459}
      2b0d6d14
    • dougt's avatar
      Update AX Roles for AX_ROLE_MENU_LIST_POPUP and AX_ROLE_MENU_LIST_OPTION · ce47e57c
      dougt authored 
      This CL is based on the work tmoniuszko@opera.com did in CL
2962633004 which found a regression in a changed to ROLE_SYSTEM_LIST
(see https://codereview.chromium.org/2909853002).

BUG=703369

Review-Url: https://codereview.chromium.org/2969743002
Cr-Commit-Position: refs/heads/master@{#485458}
      ce47e57c
    • rdevlin.cronin's avatar
      [Extensions Bindings] Add activity logging of custom handling · 7ab1596c
      rdevlin.cronin authored 
      If a method is handled by a custom hook, we won't notify the browser
(where we normally would with an API request), and so the activity log
won't be updated. Add support for activity logging with native binding.

If a request is handled by a custom hook and a request is not sent to
the browser, notify the activity logger (which will send the information
along to the browser process).  This matches the logic we currently have
in the JS bindings.

Add unit tests for the same.  In addition to the unit tests, this fixes
ActivityLogApiTest.TriggerEvent with native bindings enabled.

BUG=653596

Review-Url: https://codereview.chromium.org/2962093002
Cr-Commit-Position: refs/heads/master@{#485457}
      7ab1596c
    • megjablon's avatar
      Don't show the Data Saver promo in incognito mode · 7e469cba
      megjablon authored 
      BUG=737926

Review-Url: https://codereview.chromium.org/2969843004
Cr-Commit-Position: refs/heads/master@{#485456}
      7e469cba
    • Emil A Eklund's avatar
      Move LayoutObject::CanRenderBorderImage to ComputedStyle · b76265c2
      Emil A Eklund authored 
      Move the CanRenderBorderImage method from LayoutObject to ComputedStyle.
It only depends on information from ComputedStyle and it is not virtual.

Change-Id: I175d4b21b8d1e4f2f8e5535fc063dc65be0b841f
Reviewed-on: https://chromium-review.googlesource.com/565655Reviewed-by: default avatarWalter Korman <wkorman@chromium.org>
Commit-Queue: Emil A Eklund <eae@chromium.org>
Cr-Commit-Position: refs/heads/master@{#485455}
      b76265c2
    • hejq's avatar
      Use EnablePlayStoreAppSearch feature flag. · 0f8982ef
      hejq authored 
      We apply EnablePlayStoreAppSearch to decide whether to:

- search for uninstalled apps from Play Store;
- show a proper badge in the Play Store app result;
- show the review score and price of every searched app.

Currently using the IsFullscreenAppListEnabled flag is incorrect.

BUG=736552

Review-Url: https://codereview.chromium.org/2961923002
Cr-Commit-Position: refs/heads/master@{#485454}
      0f8982ef
    • Quinten Yearsley's avatar
      Remove auto-rebaseline and the NeedsRebaseline keyword. · efac3d1c
      Quinten Yearsley authored 
      Bug: 692811
Change-Id: Ic1a061020c1dc502655646c7a6e3ae80aee14535
Reviewed-on: https://chromium-review.googlesource.com/564681Reviewed-by: default avatarDirk Pranke <dpranke@chromium.org>
Commit-Queue: Quinten Yearsley <qyearsley@chromium.org>
Cr-Commit-Position: refs/heads/master@{#485453}
      efac3d1c
  3. 10 Jul, 2017 28 commits