GitLab

BUG=internal b/19071054

Review URL: https://codereview.chromium.org/972393002

Cr-Commit-Position: refs/heads/master@{#318962}

The keyboard detection code is heuristic based as we don't have a reliable way to detect if a keyboard
is attached. We use the Setup API's for keyboard class GUID's and look for specific prefixes in the device names like
ACPI\PNP*, ACPI\MSF* and HID\VID*. If we find one of these we assume that a keyboard is present.

Fixed a bug in the keyboard detection function where in the order of params for the StartsWith helper was flipped.

This works on the Surface and Samsung touch screen laptops in my testing.

BUG=335735

Review URL: https://codereview.chromium.org/969293002

Cr-Commit-Position: refs/heads/master@{#318961}

This patch re-lands the original change plus the fixes I
had to introduce the 'gn_all' and 'gyp_remaining' targets
in the GYP build. This patch is thus reverting
378e940e (#318636).

The code was actually correct at #318636 but was failing
on the iOS bots due to problems w/ their recipes that have
since been fixed.

TBR=brettw@chromium.org
BUG=461019

Review URL: https://codereview.chromium.org/977703003

Cr-Commit-Position: refs/heads/master@{#318960}

To support dynamically added/removed content scripts in <webview>, we store
these content scripts in the same shared memory where user scripts from
declarative content API are stored, but the decision for injection will
be made in the render once the url pattern is matched. Since render doesn't
need to send IPCs to ask a decision from browser, it makes a precise time
injection possible.

Special to <webview>, once the content scripts are updated (added/removed),
browser will only notify the render process of the given <webview>, rather than
all of the renders (which is the case for user scripts).

BUG=437566

Committed: https://crrev.com/f853287c95cf818d0b5b02d7f4a6588761096d1c
Cr-Commit-Position: refs/heads/master@{#318774}

Review URL: https://codereview.chromium.org/906493004

Cr-Commit-Position: refs/heads/master@{#318959}

When stress-testing desktop resize (by running a CrOS Ash desktop in a
window), a problem with the capture frame size calculations was
revealed: The max frame size was being updated for each change in
desktop size.  This meant that once width or height were reduced, they
could never be increased again.  When the width or height were reduced
to zero, Chrome would crash.

This change fixes the size calculation (the max size is made constant
from construction), adds a few extra sanity-checks to prevent crashes on
OOM or zero frame sizes, and also adds re-scaling of the rendered mouse
cursor bitmap whenever the desktop size has changed.

BUG=462799
TEST=Resized CrOS Ash-desktop-in-a-window on my local dev workstation to confirm crash is resolved.

Review URL: https://codereview.chromium.org/965123002

Cr-Commit-Position: refs/heads/master@{#318958}

them.

   "enable_truncated_connection_ids": "true",
   "load_server_info_time_to_srtt": "0.25"

Deleted "load_server_info_timeout" field trial, because
"load_server_info_time_to_srtt" field trial accomplishes the same in a
better manner.

R=rch@chromium.org

Review URL: https://codereview.chromium.org/975613002

Cr-Commit-Position: refs/heads/master@{#318957}

Fixes the surfaceless renderer since it now has a specfic API for the
creation of the GLSurface.

BUG=none
TEST=Run ./ozone_demo --ozone-platform=gbm --ozone-use-surfaceless

Review URL: https://codereview.chromium.org/975733004

Cr-Commit-Position: refs/heads/master@{#318956}

Also, change name of {S,G}etAutofillServerProfiles to {S,G}etServerProfiles to better match {S,G}etServerCreditCards

BUG=459741

Review URL: https://codereview.chromium.org/969103003

Cr-Commit-Position: refs/heads/master@{#318955}

This change has three main goals: First, to be able to pass extra
information associated with each VideoFrame from the capture source to
the downstream consumers (see bugs for details).  Second, to eliminate
redundancies in the current MediaStreamVideoSink API; specifically,
media::VideoFrame contains most of the same properties as
media::VideoCaptureFormat.  Third, to fully support the separate
VideoFrame concepts of "coded size" versus "visible size" in the capture
pipeline, rather than force all producers/consumers to deal with packed
data.  (Using packed frame sizes can be suboptimal for performance in
some use cases.)

The metadata is stored in a base::DictionaryValue owned by
media::VideoFrame to allow for structured data passing.  DictionaryValue
is a great choice since an efficient IPC (de)serialization
implementation already exists, and the metadata can be easily
pretty-printed for logging where needed.  Also, it's logical for
VideoFrame to own the metadata, as both need to be passed/shared
together across threads without copying.

Finally, this change includes one new use of the metadata as a
motivation for its existence: The tab/desktop capture code now includes
capture timing information, which will allow Cast streaming sessions to
be analyzed for user experience improvements.  In the future, some of
the special-use-case data members in VideoFrame should be moved into
the metadata.

BUG=461116,462101

Review URL: https://codereview.chromium.org/955253002

Cr-Commit-Position: refs/heads/master@{#318954}

CL https://crrev.com/902763002/ enables parallel telemetry unit tests. This CL will test how it works on FYI bot "Android Tests (K N5, user build)(dbg)" http://build.chromium.org/p/chromium.fyi/builders/Android%20Tests%20(K%20N5%2C%20user%20build)(dbg)

BUG=379378

Review URL: https://codereview.chromium.org/941793002

Cr-Commit-Position: refs/heads/master@{#318953}

This picks up:

    Roll the GN binaries to #318791.

    Roll the GN binaries from 317120:318791 (from git 74add9fc
    to 3d7a60db).

    This fixes 'gn clean' and picks up some new features for
    introspections (see #317159).

TBR=brettw@chromium.org
BUG=430215

Review URL: https://codereview.chromium.org/965963002

Cr-Commit-Position: refs/heads/master@{#318952}

Enhances accessibility of throttled plugins.

This enables these steps:
1. Using ChromeVox, focus throttled plugin using Alt+Shift+Arrow Keys.
2. Press ENTER or SPACE.

BUG=458285

Review URL: https://codereview.chromium.org/977653002

Cr-Commit-Position: refs/heads/master@{#318951}

BUG=450711

Review URL: https://codereview.chromium.org/967403005

Cr-Commit-Position: refs/heads/master@{#318950}

this may be casuing an increase in rendererexit code.

BUG=463646

This reverts commit 0197d5df.

Merge branch 'master' of https://chromium.googlesource.com/a/chromium/src

This is causing a huge amounts of renderer crashes because of RESULT_CODE_KILLED_BAD_MESSAGE.  Reverting this from master.

Revert "<webview>: Removed ResolveURL"

This reverts commit c2fc6144.

Cr-Commit-Position: refs/heads/master@{#318949}

Fix for MIME handler bug. The bug resulted from a previous bugfix (https://codereview.chromium.org/939443002/) that added a conversion from logical to physical units to fix a similar sizing bug with other guestviews. However, MIME handler guests never use logical pixels, so this conversion broke their sizing when they are created in a zoomed embedder.

This patch adds a more robust method of checking whether sizes are provided in logical or physical units, and all cases are handled appropriately.

BUG=462194

Review URL: https://codereview.chromium.org/972193003

Cr-Commit-Position: refs/heads/master@{#318948}

Revert of Temporarily move android_rel_tests_recipe to 100% experiment to unblock CQ (patchset #1 id:1 of https://codereview.chromium.org/973063002/)

Reason for revert:
clang roll got reverted, this should no longer be needed

Original issue's description:
> Temporarily move android_rel_tests_recipe to 100% experiment to unblock CQ
>
> TBR=machenbach
>
> BUG=463436
>
> Committed: https://chromium.googlesource.com/chromium/src/+/064e04615cf8557e63c10aa7917ea23341116fa0

TBR=machenbach@chromium.org,phajdan.jr@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=463436

Review URL: https://codereview.chromium.org/975093002

Cr-Commit-Position: refs/heads/master@{#318947}

This patch is just a small refactor from
https://codereview.chromium.org/959283004/

R=danakj, enne

Review URL: https://codereview.chromium.org/977533002

Cr-Commit-Position: refs/heads/master@{#318946}

- Submit on tab
- Do nothing when field is empty

BUG=457585

Review URL: https://codereview.chromium.org/960863002

Cr-Commit-Position: refs/heads/master@{#318945}

The name of the file for OzoneGpuTestHelper was incorrect.

BUG=none
TEST=Built Ozone with a GN build

Review URL: https://codereview.chromium.org/978473002

Cr-Commit-Position: refs/heads/master@{#318944}

After  https://codereview.chromium.org/966283002,
https://codereview.chromium.org/926213002,
https://codereview.chromium.org/927763002, all the pages used
with raterize_and_record_micro do not issue any interactions
in RunPageInteractions method, should this change should
not cause any functional changes to how the pages get run.

BUG=418375

Review URL: https://codereview.chromium.org/973173002

Cr-Commit-Position: refs/heads/master@{#318943}

The NewApi warning complains when we use APIs added to Android after ICS
because these will causes crashes on older versions of Android. This
reenables this helpful warning and whitelists the files that currently
trigger the warning. Follow-up CLs will fix these whitelisted warnings
(i.e. by fixing the code if it's broken, or adding the @TargetApi
annotation)

BUG=266140
R=aurimas@chromium.org

Review URL: https://codereview.chromium.org/969213002

Cr-Commit-Position: refs/heads/master@{#318942}

For Chromecast we want to choose which media renderer to use based on
the types of input content streams. We will use CMA media renderer for
media types that are supported by our hardware (H264, AAC, etc) and
will use the default media renderer for audio streams other than AAC
or Vorbis. This will allow us support software decoding of FLAC and
Opus via the default Chrome audio path.

BUG=457959

Review URL: https://codereview.chromium.org/973633002

Cr-Commit-Position: refs/heads/master@{#318941}

It seems the roll broke Android tests, see crbug.com/463436.

Also, fix common.gypi to add a dependency on the Clang version when
host_clang is set. We think this might be the reason the breakage
didn't show on trybots when the roll landed: the host binaries
weren't recompiled with the new Clang version.

BUG=444089, 463436
NOTRY=true
NOPRESUBMIT=true
TBR=jln, jbauman

Review URL: https://codereview.chromium.org/976833002

Cr-Commit-Position: refs/heads/master@{#318940}

BUG=417532

Review URL: https://codereview.chromium.org/957183002

Cr-Commit-Position: refs/heads/master@{#318939}

The extensions::Dispatcher::is_extension_process flag is set for both the
standalone extensions renderer and single process mode. This caused
ChromeContentRendererClient to act differently in single process mode, like
returning true from ShouldFork for any navigation started by blink, which
would cause a new WebView to be created for each navigation.

BUG=462210

Review URL: https://codereview.chromium.org/956313002

Cr-Commit-Position: refs/heads/master@{#318938}

BUG=428842

Review URL: https://codereview.chromium.org/976493003

Cr-Commit-Position: refs/heads/master@{#318937}

BUG=458541

Review URL: https://codereview.chromium.org/974633002

Cr-Commit-Position: refs/heads/master@{#318936}

BUG=463552

Review URL: https://codereview.chromium.org/977583003

Cr-Commit-Position: refs/heads/master@{#318935}

Update LED state in kernel in InputControllerEvdev::SetCapsLockEnabled.
This is currently the only way to toggle caps lock.

BUG=463242
TEST=plug in external keyboard on link_freon & press caps lock. Light on
  keyboard should turn on.

Review URL: https://codereview.chromium.org/971753006

Cr-Commit-Position: refs/heads/master@{#318934}

CL https://codereview.chromium.org/950223002 did not move all the densities
of ic_warning.png to components/

BUG=458328
TBR=tedchoc@chromium.org

Review URL: https://codereview.chromium.org/970113003

Cr-Commit-Position: refs/heads/master@{#318933}

This file is not supposed to be part of libgtest but in its
own library called libgtest_main.a.  Its also fairly
useless right now in the context of PPAPI (all our gtest
examples currently use ppapi_simple and define their own
entry point).  In the long run we could re-enable this
if it becomes useful, but really we should probably just
remove the gtest libraries from the SDK completely since
shipping gtest in library form is not recommended practice.

Review URL: https://codereview.chromium.org/958233004

Cr-Commit-Position: refs/heads/master@{#318932}

This patch is just a small refactor with a few more comments explaining
what this function is meant to do. I've been confused in the past when
trying to remember what this function does, so hopefully this helps...

R=danakj, enne

Review URL: https://codereview.chromium.org/973713002

Cr-Commit-Position: refs/heads/master@{#318931}

BUG=450050

Review URL: https://codereview.chromium.org/975453003

Cr-Commit-Position: refs/heads/master@{#318930}

BUG=internal b/19212285
R=lcwu@chromium.org

Review URL: https://codereview.chromium.org/961993006

Cr-Commit-Position: refs/heads/master@{#318929}

This CL reduces unnecessary work in StatusBubbleMac and replaces it with some
DCHECKs to verify invariants.  (The invariant here is that we set
kWindowSizeDeterminedLater once, when state is set to kBubbleHidden, and the
size stays that way until the bubble transitions to a different state.  If that
invariant holds, then we don't need to constantly re-set the size for a hidden
bubble.)

BUG=None
TEST=No visible impact.  Status bubble should continue to show, expand, and hide properly.

Review URL: https://codereview.chromium.org/968263005

Cr-Commit-Position: refs/heads/master@{#318928}

The code in [DownloadItemController showDangerousWarning:] calls
[GTMUILocalizerAndLayoutTweaker sizeToFitFixedHeightTextField:minWidth:]
to wrap the warning string within the DownloadItemController's
dangerousDownloadLabel_ textField. Within the nib this textField is 22
points tall and holds two lines of text. Under Yosemite, however, the
font metrics have changed and the two lines are now 24 points tall.

This cl insets the dangerousDownloadLabel_'s frame height by -2 points
when running on Yosemite. dangerousDownloadLabel_ is not as tall as its
superview, so the new frame still fits comfortably within the superview's
frame.

BUG=454782
TEST=Follow instructions in bug 454782 and confirm that text is wrapping
properly on Yosemite.

Review URL: https://codereview.chromium.org/911303002

Cr-Commit-Position: refs/heads/master@{#318927}

A telemetry test that checks for "q=low" in the Chrome-Proxy
request header and verifies that the image received was less than 100 bytes.

BUG=449665

Review URL: https://codereview.chromium.org/942863004

Cr-Commit-Position: refs/heads/master@{#318926}

Have BluetoothAdapterProfileChromeOS remove itself from adapters when it
is deleted via it's deconstructor, to avoid references being left behind
after deletion.

BUG=457978

Review URL: https://codereview.chromium.org/935383003

Cr-Commit-Position: refs/heads/master@{#318925}

This is necessary because GN's exec_script only runs Python scripts.

This was requested by Nico in https://codereview.chromium.org/971593003/.

BUG=462972
TEST=run ./tools/clang/scripts/plugin_flags.py and compare the output
with the bash version. They should be exactly the same.
R=thakis@chromium.org

Review URL: https://codereview.chromium.org/966723003

Cr-Commit-Position: refs/heads/master@{#318924}

The flag for specifying the URL for the Data Saver secure proxy check
was changed in https://codereview.chromium.org/961983002/. This change
makes the relevant integration test use the new flag value, and also
makes the test restart after each page.

BUG=463210

Review URL: https://codereview.chromium.org/971073002

Cr-Commit-Position: refs/heads/master@{#318923}