Skip to content

GitLab

T
tangled
Switch branch/tag
  1. 19 Aug, 2020 29 commits
  3. 18 Aug, 2020 11 commits
    • Tommy Li's avatar
      [omnibox] ZeroSuggestProvider::TypeOfResultToRun w/ AutocompleteInput · 26a9f582
      Tommy Li authored 
      This CL refactors ZeroSuggestProvider::TypeOfResultToRun to take an
AutocompleteInput as a parameter instead of just the current page URL.

We need this because differentiating the on-focus vs. on-clobber
ZeroSuggest modes requires access to AutocompleteInput::focus_type().

This CL also makes the method 'static', for three reasons:

 1. The function is more easily testable when it only depends on its
    parameters, rather than any state in the member variables.

 2. The member variables store data that's also in the |input|
    parameter, and I didn't want to introduce ambiguity as to which
    is used.

 3. We plan to split MostVisited into its own provider, and we will
    want to share some logic from TypeOfResultToRun. Making this
    method static is a step towards that future.

For reviewability, this CL actually has no functional changes, and
just moves things around.

The followup CL will actually add the on-clobber vs. on-focus logic to
TypeOfResultToRun.

Bug: 1106096
Change-Id: I69c8f2f08a182796be9ce1c66611fe6788a6158a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2360562Reviewed-by: default avatarTomasz Wiszkowski <ender@google.com>
Commit-Queue: Tommy Li <tommycli@chromium.org>
Cr-Commit-Position: refs/heads/master@{#799394}
      26a9f582
    • Steven Bennetts's avatar
      CrOS: Do not set an empty policy hostname · 1c0816a5
      Steven Bennetts authored 
      This is in preparation for adding UI to allow users
to set a custom DHCP hostname. Currently the policy
code is always setting the hostname, even if
kDeviceHostnameTemplate is not set by policy,
which overrides any other custom set hostname.

Bug: 1103272
Change-Id: Ia99ee58c59ae1f38e8fe10ed693997dead861c4b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2320991
Commit-Queue: Steven Bennetts <stevenjb@chromium.org>
Reviewed-by: default avatarKyle Horimoto <khorimoto@chromium.org>
Reviewed-by: default avatarOmar Morsi <omorsi@google.com>
Reviewed-by: default avatarDenis Kuznetsov [CET] <antrim@chromium.org>
Reviewed-by: default avatarRoman Sorokin [CET] <rsorokin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#799393}
      1c0816a5
    • Sammie Quon's avatar
      dark mode: Flip base colors for ink drop. · 64f8e172
      Sammie Quon authored 
      Original thought the dark ripple was for the dark background.

Test: manual
Bug: 1117203
Change-Id: I3221271f8bb1d1d32a8832c75a46c0b8f726426e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2363997Reviewed-by: default avatarXiyuan Xia <xiyuan@chromium.org>
Commit-Queue: Sammie Quon <sammiequon@chromium.org>
Cr-Commit-Position: refs/heads/master@{#799392}
      64f8e172
    • John Palmer's avatar
      Change spacing for undo window · c3ee7e10
      John Palmer authored 
      Should be 8px, old number worked out to 16px

BUG=1117379

Change-Id: Ic6cef89a6c9b93fab24b538f13d6c07fa4f06647
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2362062
Auto-Submit: John Palmer <jopalmer@chromium.org>
Commit-Queue: Keith Lee <keithlee@chromium.org>
Reviewed-by: default avatarKeith Lee <keithlee@chromium.org>
Cr-Commit-Position: refs/heads/master@{#799391}
      c3ee7e10
    • David Bokan's avatar
      Cleanup FakeLayerTreeHost's initialization of Impl · 613ec9bd
      David Bokan authored 
      As part of splitting out input from LayerTreeHostImpl I found some
issues in tests that use FakeLayerTreeHost. This class also initializes
a FakeLayerTreeHostImpl which is used in some tests.

However, some tests also call InitializeSingleThreaded to use the
FakeLayerTreeHost with a Proxy. Initializing a Proxy will call
LayerTreeHost::CreateLayerTreeHostImpl which currently creates a (real)
LayerTreeHostImpl that's owned by the Proxy. These tests are incorrectly
running with multiple LTHI objects.

This CL overrides CreateLayerTreeHostImpl in the fake so that this
method creates a fake Impl object. We also avoid manually creating a
fake Impl inside the FakeLayerTreeHost and instead rely on the call
to CreateLayerTreeHostImpl to store a pointer to the fake impl.

For tests that don't use a Proxy, we now manually call a new
CreateFakeLayerTreeHostImpl method that does the same thing and the
FakeLayerTreeHost takes ownership of the Impl instead.

Bug: 915926
Change-Id: I63202e1717d2eab026688d69c8d0ef75939c7c47
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2363247Reviewed-by: default avatarXianzhu Wang <wangxianzhu@chromium.org>
Commit-Queue: David Bokan <bokan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#799390}
      613ec9bd
    • Miguel Casas's avatar
      media/VaapiVideoDecoder: refactor failures UMA · 993d2e9e
      Miguel Casas authored 
      ToT VaapiWrapper has a simple report-to-uma callback that clients have
to implement and that usually just points to "error happened" in either
the normal or the Vpp VaapiWrapper.

This CL changes that:
- VaapiWrapper provides a function for this: ReportVaapiErrorToUMA().
- Clients bind their histogram (eg. Media.VaapiVideoDecoder.VAAPIError)
to that new ReportVaapiErrorToUMA() method.
- A new VaapiFunctions enumeration (with a matching UMA enum) is added
to vaapi_wrapper.cc to identify which concrete vaBla() function failed,
and is logged to UMA.

This forces ToT UMA histograms (and associated enums) to be marked as
obsolete, namely:
-Media.VAJDA.DecoderFailure
-Media.VAJDA.VppFailure
-Media.VAJDAWorker.DecoderFailure
-Media.VAJEA.VppFailure
-Media.VAVDA.DecoderFailure
-Media.VAVEA.EncoderFailure

New ones are landed instead (with associated VaapiFunctions enum):
-Media.VaapiImageDecodeAcceleratorWorker.VAAPIError
-Media.VaapiJpegEncodeAccelerator.VAAPIError
-Media.VaapiJpegEncodeAccelerator.VAAPIVppError
-Media.VaapiMjpegDecodeAccelerator.VAAPIError
-Media.VaapiMjpegDecodeAccelerator.VAAPIVppError
-Media.VaapiVideoDecodeAccelerator.VAAPIError
-Media.VaapiVideoDecoder.VAAPIError
-Media.VaapiVideoEncodeAccelerator.VAAPIError
-Media.VaapiVideoEncodeAccelerator.VAAPIVppError

This CL also adds two new boolean histograms for VaapiVideoDecoder,
who had none.

Bug: b:162962069
Change-Id: I561e48183f61de86247598fea4f3b10cd590c2e0
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2339787
Commit-Queue: Miguel Casas <mcasas@chromium.org>
Reviewed-by: default avatarSteven Holte <holte@chromium.org>
Reviewed-by: default avatarJao-ke Chin-Lee <jchinlee@chromium.org>
Cr-Commit-Position: refs/heads/master@{#799389}
      993d2e9e
    • Becca Hughes's avatar
      Make content attributes optional · 5b4d8221
      Becca Hughes authored 
      For normal images contentAttributes are optional in
the spec but the converter requires them. This makes
them optional.

Change-Id: I4073de51009e6e7802667b2d79a7ad485d939978
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2363477
Auto-Submit: Becca Hughes <beccahughes@chromium.org>
Commit-Queue: Sam Bowen <sgbowen@google.com>
Reviewed-by: default avatarSam Bowen <sgbowen@google.com>
Cr-Commit-Position: refs/heads/master@{#799388}
      5b4d8221
    • Evan Stade's avatar
      Remove some KitKat specific resources. · f0fc9ffb
      Evan Stade authored 
      Bug: 1110930
Change-Id: I20914d3f88ad98944e0f5525421bc666cc79eb73
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2363696Reviewed-by: default avatarDavid Trainor <dtrainor@chromium.org>
Commit-Queue: Evan Stade <estade@chromium.org>
Cr-Commit-Position: refs/heads/master@{#799387}
      f0fc9ffb
    • James Cook's avatar
      lacros: Rich notification support · c83cc719
      James Cook authored 
      Add support for image, list, and progress notifications. This
involves serializing the RichNotificationData structure.

Move some existing crosapi::Bitmap serialization code into a
shared utility file. Add deserialization to SkBitmap. For now
the image format is hard-coded. crbug.com/1116652 tracks more
flexible image serialization.

Add unit tests for both ash side and lacros side.

Bug: 1113889
Test: unit_tests, manual with Notifications Galore chrome app

Change-Id: I27b1c74f828751f35d845e625052386530de9722
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2357749
Commit-Queue: James Cook <jamescook@chromium.org>
Reviewed-by: default avatarJun Mukai <mukai@chromium.org>
Reviewed-by: default avatarGreg Kerr <kerrnel@chromium.org>
Reviewed-by: default avatarDarin Fisher <darin@chromium.org>
Reviewed-by: default avatarErik Chen <erikchen@chromium.org>
Cr-Commit-Position: refs/heads/master@{#799386}
      c83cc719
    • Daniel Cheng's avatar
      Add escape hatch for plugins in NavigationClient::CommitNavigation(). · 612358a6
      Daniel Cheng authored 
      The current browser-side navigation code can ask the renderer to commit
a MIME type that it cannot handle. When this happens, the renderer
simply ignores the commit request, tripping the CHECKs added in
https://crrev.com/796508.

For now, allow these commits to silently fail and not commit a new
document...

Bug: 999255, 1115912, 1117282
Change-Id: I435831058db8e059b0389a8e2262f6bd72e875fa
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2360602
Commit-Queue: Daniel Cheng <dcheng@chromium.org>
Reviewed-by: default avatarAlex Moshchuk <alexmos@chromium.org>
Reviewed-by: default avatarDave Tapuska <dtapuska@chromium.org>
Cr-Commit-Position: refs/heads/master@{#799385}
      612358a6
    • Anupam Snigdha's avatar
      [Narrator] Fix VK show after narrator sets focus · de10ebb1
      Anupam Snigdha authored 
      Before this change we used to set the pointer type to touch only when
DoDefaultAction was called. This led to issues as sometimes I don't
see this API being called when narrator is active and the user double
taps on the edit control to set the selection. VK should show up when
caret starts blinking in the edit control, so we set the pointer type
to touch when we receive focus from the narrator and then in blink we
invoke a call to show the VK if the focus happened to be inside an
edit control. FWIW, I discussed this with narrator team and it is
hard to come up with a solution that is deterministic and reliable as
narrator eats up all the gesture events and VK is controlled by a
different component that relies on VK show call from Chromium when
input modality is touch and input policy is set as manual(Chromium
specific). This solution is just a heuristic.

Bug: 1106556
Change-Id: I12fcfcda52ba058ed5377b0312e6652988dacced
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2361045Reviewed-by: default avatarAvi Drissman <avi@chromium.org>
Reviewed-by: default avatarDominic Mazzoni <dmazzoni@chromium.org>
Commit-Queue: Anupam Snigdha <snianu@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#799384}
      de10ebb1