We have shipped them for more than 3 months and we don't have issue
right now. Chromium also removed the flags, so it should be fine to drop
them to make the code simpler.
Differential Revision: https://phabricator.services.mozilla.com/D205615
The asserts added in the first patch turned up this existing problem.
AllocChars takes a context so is assumed to report an error on failure.
Differential Revision: https://phabricator.services.mozilla.com/D205553
Jan announced changes to the module in Dec 2023:
https://groups.google.com/a/mozilla.org/g/governance/c/pXQkIm4V8so/m/FNBN0ForAAAJ
I'm reflecting these changes for Jan of:
- Jari is an IDB peer and so is added to mots.yaml
- Bevis is moved to an emeritus peer and this was the last active use of his
anchor his identifier so I removed it.
- I see from the docs that anchors can be used in the emeritus list but this
seems rarely used and the docs provide no guidance on when this is
appropriate and I didn't see this until after I'd made the changes.
- :baku is moved to emeritus by name.
- He actually is used by anchor in other places?
Differential Revision: https://phabricator.services.mozilla.com/D205643
Substitute
use-system-python: false
by
use-python: default
or if one wants to use a different version (available by taskcluster,
i.e. 3.11 as of this commit)
use-python: X.Y
Differential Revision: https://phabricator.services.mozilla.com/D204951
When running `mach try fuzzy` locally, notarization jobs end up in the
full graph, because level isn't forced to 1. As part of try's decision
task though, they're filtered out in the mac_notarization transforms.
Adding them to UNCOMMON_TRY_TASK_LABELS means they won't be selected
at least in the non-`--full` case.
Differential Revision: https://phabricator.services.mozilla.com/D205693
The android-gradle-dependencies toolchain task already downloads gradle,
so we can reuse it instead of going through the wrapper.
Differential Revision: https://phabricator.services.mozilla.com/D205333
Changed
* `--color-background-*` are now part of the group of `--background-color` tokens
* `--box-background-color` is now part of the group of `--background-color` tokens as `--background-color-box`
* `--color-canvas` is now part of the group of `--background-color` tokens as `--background-color-canvas`
* `--border-interactive-*` are now part of the group of `--border-color` tokens and consequently now start with `--border-color`
* `--color-error-outline` is now an outline-specific token as `--outline-color-error`
* Comment headings were cleaned up as to follow the existing standard
* Better documentation behind the difference of "Base", "Application", and "Component" tiers in the design tokens documentation
* Tokens files are alphabetized
Removed
* `--outline-color-error` had duplicates in both high contrast mode media queries, so its value under `forced-colors` was removed
* "Base" and "Application" headings in tokens files were removed in order to lean in on the alphabetical order organization of token groups and promote a cleaner file and less confusion
* Documentation was edited down to not incude any guidance that we don't adhere to
* Remove duplicate interactive border color high contrast mode tokens in `tokens-brand.css` in favor of `tokens-shared.css`
* Remove `--button-border-color-primary` duplicate in `tokens-brand.css`
* Remove duplicate `--button-text-color` in `tokens-shared.css`
Differential Revision: https://phabricator.services.mozilla.com/D202767
This fixes the alignment on linux. On windows and macOS, due to the
implementation specializing this case, there is no difference in
alignment. This also sets a fixed, identical width for both buttons to
be more visually appealing (on all platforms).
Differential Revision: https://phabricator.services.mozilla.com/D205616
This replaces the fix to bug 1883144 to so that we don't throw here on failure
rather than trying to clean up after the fact.
It seems there are a bunch of places where we can set a unique ID where we
already have a pending exception making that approach fragile.
This is also more efficient since we don't have to set the exception in the
first place.
Differential Revision: https://phabricator.services.mozilla.com/D205436
Split window configuration (mGdkWindow setup) and compositor resume and use mutex to protect window config only.
That ensures we always have valid mGdkWindow but we don't deadlock if compositor resume is called from map/unmap during
forced EGLSurface update.
Differential Revision: https://phabricator.services.mozilla.com/D205633
