I haven't been able to debug why my change exactly causes this, because
both before and after my patch the test is blank on the emulator's
runner, and you can't just run the test on GVE...
My best guess is that a subtle layout change (the select is now 0.5px
taller) causes the radio button to end up in a subpixel position, and
the origin of 0, 0 just happens to get rounded down... Or an interaction
with our pointer event retargetting (the fat finger stuff), which is
only enabled on Android.
In any case I'm pretty sure my change is not the root cause of this
intermittent, this fixes it and is a reasonable thing to do.
Differential Revision: https://phabricator.services.mozilla.com/D203150
Using `isort` for import sorting in Python has been replaced by
`ruff`. Update the formatters list to remove `isort`. We will add
`ruff` at a later date once we have asserted it is appropriate to
add to this formatters list.
Differential Revision: https://phabricator.services.mozilla.com/D203127
* Flip the component pref to true by default for nightly builds only
* Move the pref check and initialization to a startup idle task
* And be a bit smarter about when we get and disable the addon
* Fix a bug where we try to communicate with the overlay after the window actor is destroyed when
the component pref gets flipped off during use
Differential Revision: https://phabricator.services.mozilla.com/D196888
The datetime input is inconsistent with other inputs when using spoof
English: its placeholder is not translated, unlike the default values
and texts of all the other inputs.
Differential Revision: https://phabricator.services.mozilla.com/D201726
This patch makes the `mNumTemporalLayers` in `VP9Specific` generated in
`VideoEncoderConfigInternal::ToEncoderConfig` match the number of
temporal layers in the given scalability mode.
Differential Revision: https://phabricator.services.mozilla.com/D202885
Added
- `--space-medium`
- `--space-large`
Changed
- `--space-xlarge` has a new value and the old one is now `--space-xxlarge`
Differential Revision: https://phabricator.services.mozilla.com/D202739
The 3 signifiers of the minimization being complete are:
1) The sizemodechange event
2) The occlusionstatechange event (on some platforms)
3) The browsing context becoming inactive
Ideally these would happen in 1-2-3 order, but they aren't consistent.
Specifically, since the browsing context is updated by an
occlusionstatechange event listener, tests that add a
occlusionstatechange event listener will fire before the browsing
context has been updated (since they fire in LIFO order). That makes it
very timing dependent on whether or not a check for browsing context
inactive will succeed, and that's the source of the intermittent
failures here. The safest thing to do is to turn the browsing context
active check into another promise and await the sizemodechange and the
active promise together, and that's what this patch does.
Differential Revision: https://phabricator.services.mozilla.com/D202609
Both methods now run on the PBackground thread, there's no need to have these as
separate methods anymore. The extra state can be now removed as well. This opens
the door for some further optimizations.
Differential Revision: https://phabricator.services.mozilla.com/D189757
FactoryOp::Open now contains no code which can only run on the main thread, so
it can be safely switched to running on the PBackground thread.
Differential Revision: https://phabricator.services.mozilla.com/D189756
InitializeQuotaManager already initializes the storage service on the main
thread, so the initialization in FactoryOp::Open can be just removed.
Differential Revision: https://phabricator.services.mozilla.com/D189755
mLocale is currently initialized in IndexedDatabaseManager::Init which is
called from IndexedDatabaseManager::GetOrCreate if the manager doesn't exist.
This can be a problem when IndexedDatabaseManager::GetOrCreate is called very
early,for example in nsLayoutStatics::Initialize in the parent process.
This is a preparation for moving IndexedDatabaseManager::GetOrCreate from
FactoryOp::Open to InitializeQuotaManager.
Differential Revision: https://phabricator.services.mozilla.com/D189891
FactoryOp::CheckPermission now contains only permission related code and it
always returns PermissionValue::kPermissionAllowed so it can be safely removed.
Differential Revision: https://phabricator.services.mozilla.com/D189754
Both Chrome and Safari override the sites autofill style.
Given that its crucial for users to be able to clearly see what Firefox autofills for
them, after discussing with UX, we decide also do the same.
Differential Revision: https://phabricator.services.mozilla.com/D202410
Some consumer is passing in integer as expected, most pass a numeric string
that is the result of a getAttribute. This causes a mismatch as that id is used
as key in a Map.
While it would be possible to fix all the consumers, the mistake seems so
common and trivial, that is worth just adapting the API to accept both.
Differential Revision: https://phabricator.services.mozilla.com/D203019
