Canvas uses CSS pixels, so this patch encodes the information into the type
system. Some painting code still uses unknown sizes, so we need to use
`CSSIntSize::FromUnknownSize` and `ToUnknownSize()` to convert them.
This patch doesn't change behavior.
Differential Revision: https://phabricator.services.mozilla.com/D216463
This is a temporary measure. We would like to destroy the native window
here, but we want to see the effect of removal on crash volume.
Differential Revision: https://phabricator.services.mozilla.com/D216066
This change allows users to remove the opaque background from the
sidebar, enabling sidebar backgrounds to be fully transparent when used
alongside a userChrome.css that provides transparency.
This is purely an aesthetic option for use with custom themes and a
continuation of Bug 28354.
Differential Revision: https://phabricator.services.mozilla.com/D212385
Each provider has different fallback behavior:
* Windows/macOS: Fallback only on error
* Gpsd/Portal: Fallback only on timeout
* Geoclue: Fallback on both
To meet all the behaviors MLSFallback got FallbackReason/ShutdownReason and report via Glean based on that.
Differential Revision: https://phabricator.services.mozilla.com/D216085
This is to quickly compare the existing data while including stable/esr releases too.
Also setting expiry date for eventual full glean migration.
(Confusingly "opt-out" means "opt-in" for release channel data collection while "opt-in" means prerelease channel-only data collection)
Differential Revision: https://phabricator.services.mozilla.com/D216084
And also some log for provider selection on Linux. This gives some clue to see what's happening in geolocation when dealing with bug reports.
Differential Revision: https://phabricator.services.mozilla.com/D215906
Attempt to fix `IllegalStateException` in `TranslationsDialogFragment`.
* Added `navigateSafe` to `MenuNavigationMiddleware`
* Added `runIfFragmentIsAttached` to places where `findNavController().popBackStack()` was used in the translations area
* Added `runIfFragmentIsAttached` to guard `onCreateDialog` indirect use of `requireContext()`
Differential Revision: https://phabricator.services.mozilla.com/D216412
The test to check if scrolling is allowed in non-html documents times out for JSON,
because it depends on the `onload` event, which Gecko is not firing for JSON (Bug 1905106).
Additionally, other test cases in that file can time out on CI as well.
With this patch, the JSON test case is separated into its own file and is expected to `TIMEOUT`. The other test cases get a longer timeout so that they pass in all configurations.
Differential Revision: https://phabricator.services.mozilla.com/D216565
The issue was that when in the previous patch, the
accent_high_contrast_normal_theme color was changed,
other than the intended ones were changed as well.
In the previous patch, only the dialog's buttons color was
desired to be changed to resolve an accessibility issue.
In this patch we change back accent_high_contrast_normal_theme
to photonViolet40.
And to resolve the previous ticket that caused
the regression, the button text color of the dialog of the dark mode
was changed from accentHighContrast to accentUsedOnDarkBackground.
Before and after the changed of this patch:
{F7416870}
And the colors of the dialogs like the one for adding a new collection should remain like it is:
{F7416882}
Differential Revision: https://phabricator.services.mozilla.com/D216529
The issue was that the permissions prompt was using the accentHighContrast for the background color of the positive button.
The fix was to replace the color with the primary button colors as confirmed by the UX.
| Before | After |
| {F7416367} | {F7416397} |
Differential Revision: https://phabricator.services.mozilla.com/D216522
The `RefPtr<StringBuffer>` argument of the new JS APIs can have pretty significant performance
overhead due to the atomic refcount used for `StringBuffer`.
This patch adds separate APIs that only use `RefPtr` when we're going to use the buffer for
a new JS string. This is a more similar to the behavior with the old external-string code
and fixes most of the reported performance regression.
Differential Revision: https://phabricator.services.mozilla.com/D216285
Even though drawing a canvas to itself tries to use ExtractSubrect to get a limited
copy where applicable, the canvas still inadvertently forces the copy-on-write snapshot
to also force a copy due to SurfaceFromElementResult holding a subversive extra reference
to the snapshot, keeping it alive when it need not be.
RecordedTextureData also holds a reference to the snapshot that needs to be cleaned up
as well when the DT is modified.
Differential Revision: https://phabricator.services.mozilla.com/D216520
The flow wanted is that when the user clicks on the close button of
microsurvey bottom sheet. the prompt is closed as well.
The issue was that when the close button was clicked, the
recomposition was not happening until we moved on another page.
The fix is to add a new variable that will tell us if the microsurvey
prompt was dismissed by clicking either on the close button of the
prompt or of the bottom sheet. Depending on this variable we can
make the prompt visible or not.
Differential Revision: https://phabricator.services.mozilla.com/D216256
SVG text elements in general don't have a clearly-defined concept of "lines", so we can't
know where it would be appropriate to trim.
Differential Revision: https://phabricator.services.mozilla.com/D216506
Automatic update from web-platform-tests
[webdriver] Reuse the HTTP connection to the WebDriver remote end (#47085)
`HTTPWireProtocol` has an internal flag `_last_request_is_blocked` to
tell if the underlying HTTP connection is unusable and should be thrown
away (see #21889). However, the transport sets the flag too early, which
has the effect of transparently closing and reopening the connection
before every request.
This is wasteful and appears to sometimes exhaust the host machine's
usable ports in long test runs (https://crbug.com/343720396). Set the
flag after it's checked to avoid unnecessary connection closures.
--
wpt-commits: a7577845e1aceaebfc009a4ddbdc9a52070b91c9
wpt-pr: 47085
Automatic update from web-platform-tests
[wptrunner] Clean up `testharnessreport-content-shell.js` (#47083)
This is dead code because wptrunner no longer supports Chromium's
`content_shell` (#46658).
--
wpt-commits: 5c346949cf362a5264a2355f6470963df14da0da
wpt-pr: 47083
