In some situations dedicated profile support does not work well. This includes a
handful of linux distributions which always install different application
versions to different locations, some application sandboxing systems as well as
enterprise deployments. This environment variable provides a way to opt out of
dedicated profiles for these cases.
Differential Revision: https://phabricator.services.mozilla.com/D35249
--HG--
extra : moz-landing-system : lando
This changes CreateClippedDrawTarget so that instead of taking
a max size and a transform it just takes a user space rect of
the desired bounds.
This change allows the caller to not worry about the computing
a max size based on the current clip. Instead this responsibility
is lowered into the specific backends.
The main motivation for this work is to allow blob recoordination
to create recordings that don't depend on the current clip.
Some additional benefits are that the API is easier to use and
as can be seen simplifies the SVG masking code because it doesn't
need to track surface offsets manually.
It's also an important step towards removing all the uses of
gfxContext::GetClipExtents which will let us get rid of the separate
clipping stack in gfxContext and help us move off of gfxContext
completely.
Most backend implementations of CreateClippedDrawTarget are relatively
simple. DrawTargetCapture is modified to track the current clip rect
so that it can create a new DrawTargetCapture of the appropriate size
without needing to worry about lazy resolution.
Differential Revision: https://phabricator.services.mozilla.com/D33363
--HG--
extra : moz-landing-system : lando
Our Restarter Service would have issues on Android Q:
It would run on a separate process that allowed it to kill the app's process
but when trying to restart the app it would fail because it would've been a
background Service and as such it would've been restricted from starting other
Android Application components.
To make it work on Android Q we would need to migrate it to a foreground
service or piggy-back the SYSTEM_ALERT_WINDOW permission given by the user for
the Tab Queue functionality.
The chosen solution is simpler and not affected by any of those limitations.
Depends on D35402
Differential Revision: https://phabricator.services.mozilla.com/D35403
--HG--
extra : moz-landing-system : lando
Fix for the following scenario:
- "Don't keep activities" checked
- Change system language
Upon detecting a locale change in BrowserApp's onCreate() we would call
finish() and then return before calling super.onCreate().
This patch introdues the call to super.onCreate() before returning and ensures
that the app will die cleanly so that upon restarting Gecko can reinit add-ons
that touch the UI.
Differential Revision: https://phabricator.services.mozilla.com/D35402
--HG--
extra : moz-landing-system : lando
Requiring callers to manually cleanup all DOM references that they create runs
counter to the usual JavaScript model of automatic garbage collection, where
developers usually do not need to think about manual deallocation. It's bound
to lead to leaks before long.
This patch adds a finalization witness bound to the element of any DOM
reference that we create. When that element is destroyed, the finalization
witness also removes its entry from the ID map.
Since the mappings are already stored as weak references, this shouldn't
result in a change to the behavior that callers see, only in the underlying
memory management. It essentially makes this behave as a true weak value map.
Differential Revision: https://phabricator.services.mozilla.com/D35354
--HG--
extra : rebase_source : baab8063668594df452697aae6e2115adfc43ab3
UUIDs are expensive to create, to use as map keys, and to transfer between
processes. They are also completely unnecessary for this use case, since the
number returned by Math.random() has the same characteristics that we're
depending on the UUID generator for, only with better performance.
That said, even Math.random() is probably overkill for this use case, when we
can just use continually incremented serial number and get the same behavior,
but I decided to continue using randomized IDs mainly to minimize the
magnitude of the change.
--HG--
extra : rebase_source : 1536569eca7442f1ae4dcd894cac6a7c39edbb2c
This is required so that crashes on import don't block updating tests.
It makes that change from 1539449 only apply to non-wpt suites, which is
not ideal but no worse than the previous setup.
Differential Revision: https://phabricator.services.mozilla.com/D35218
Automatic update from web-platform-tests
Don't show |cursor: pointer| on unclickable <area>s
The HTML spec says:
"The href attribute on a and area elements is not
required; when those elements do not have href
attributes they do not create hyperlinks."
Background:
Chrome already excludes unclickable <area>s, <area>s
without a href attribute, from the tab order.
Problem:
Unclickable <area>s still show a "click me" hint.
Solution:
Mimick Firefox. Only show the pointer for actual
<area> links.
Note:
Found during crrev.com/c/1632231. We don't want
unclickable <area>s to act navigable.
Bug: 972556
Change-Id: I9214171290659b8b800db3e462478e2a1327c832
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1653009
Commit-Queue: Hugo Holgersson <hholgersson@fb.com>
Reviewed-by: Fredrik Söderquist <fs@opera.com>
Cr-Commit-Position: refs/heads/master@{#668409}
--
wp5At-commits: dcac708930b6e73bf3e7629908b6467e7644cff9
wpt-pr: 17267
Automatic update from web-platform-tests
NG/DL: Implement size containment (and display lock) for NG fieldset algo
This patch adds size containment implementation for fieldsets in
LayoutNG.
Tested with LayoutNGFieldset feature flag.
Along with the new tests,
external/wpt/css/css-contain/contain-size-fieldset-001.html and
external/wpt/css/vendor-imports/mozilla/mozilla-central-reftests/contain/contain-size-fieldset-001.html
are also passing.
R=chrishtr@chromium.org, ikilpatrick@chromium.org, mstensho@chromium.org
Bug: 958975
Change-Id: I5e735adc65ed9450f84e4c1ab68cba69e3fd85ed
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1613844
Reviewed-by: Morten Stenshorne <mstensho@chromium.org>
Commit-Queue: vmpstr <vmpstr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#660866}
--
Add colliding ref names to lint.whitelist
See also https://github.com/web-platform-tests/wpt/pull/17289 which
was discovered when preparing this change.
--
wp5At-commits: 2d4a450b602564b5c7461c76e281552b60363602, 677d31b5fbbcfe9d3f7628a9b73c4a6066bbe76d
wpt-pr: 16858
Automatic update from web-platform-tests
Revert "Active Processing for ConvolverNode"
This reverts commit 499976c54a2b086bbde54ae137d0ac4adb3cd427.
Reason for revert:
Findit (https://goo.gl/kROfz5) identified CL at revision 668149 as the
culprit for flakes in the build cycles as shown on:
https://analysis.chromium.org/p/chromium/flake-portal/analysis/culprit?key=ag9zfmZpbmRpdC1mb3ItbWVyQwsSDEZsYWtlQ3VscHJpdCIxY2hyb21pdW0vNDk5OTc2YzU0YTJiMDg2YmJkZTU0YWUxMzdkMGFjNGFkYjNjZDQyNww
Sample Failed Build: https://ci.chromium.org/buildbot/chromium.mac/WebKit%20Mac10.13%20%28retina%29/15474
Sample Failed Step: webkit_layout_tests on ATI GPU on Mac Retina on Mac-10.13.6
Sample Flaky Test: external/wpt/webaudio/the-audio-api/the-convolvernode-interface/active-processing.https.html
Original change's description:
> Active Processing for ConvolverNode
>
> The ConvolverNode should output a single channel of silence when it is not
> actively processing. Active processing is already supported and just needed
> to have the convolver output a single channel of silence when it is
> constructed, before anything is connected to it.
>
> Bug: 971389
> Change-Id: Icc894cd7ea6dec60a1eccb0fa4a3aea68ac5b51f
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1644899
> Commit-Queue: Raymond Toy <rtoy@chromium.org>
> Reviewed-by: Hongchan Choi <hongchan@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#668149}
Change-Id: Ie7fedee332ca167c3dfd4fd99142de89234ff03a
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: 971389
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1653924
Cr-Commit-Position: refs/heads/master@{#668243}
--
wp5At-commits: 2477f79f3c0757225d06762aeb1d3b1eea6eb52a
wpt-pr: 17288
Automatic update from web-platform-tests
Worker: Modernize worker-interception.https.html
This CL modernizes worker-interception.https.html using new ES features like
async/await and arrow functions.
This is a preparation to add more test, for example, for shared workers.
Bug: n/a
Change-Id: I65ebe309358b8b680b9b173916d6aa1d477b1d95
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1654673
Commit-Queue: Hiroki Nakagawa <nhiroki@chromium.org>
Reviewed-by: Matt Falkenhagen <falken@chromium.org>
Cr-Commit-Position: refs/heads/master@{#668316}
--
wp5At-commits: 16f941927fe2ce5abb5bc54ab02af5f15c501c1a
wpt-pr: 17284
Automatic update from web-platform-tests
ChannelMergerNode supports active processing
When a ChannelMergerNode is not actively processing its inputs, the
number of channels in its output is one, instead of the number of
inputs.
Just need to disable outputs when the node is created. Connections to
the node will adjust the number of channels appripriately, and when
inputs are no longer actively processing, the number of output
channels becomes one, as required.
Bug: 925046
Test: the-channelmergernode-interface/active-processing.https.html
Change-Id: I08e5fd2fe9c99b75af177eb85212162ecf39efd4
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1645478
Reviewed-by: Hongchan Choi <hongchan@chromium.org>
Commit-Queue: Raymond Toy <rtoy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#668159}
--
wp5At-commits: bfb037131941940b9f7efa16cb1ec0829a2461cb
wpt-pr: 17273