This commit contains the Gecko-side changes from WebRender PR#3277:
- Dedicated DirtyRect type.
- Separate the blob image APIs from regular image ones.
Differential Revision: https://phabricator.services.mozilla.com/D12463
--HG--
extra : moz-landing-system : lando
This allows us to use the same Java code for any native platform,
enabling a "fat" AAR.
Differential Revision: https://phabricator.services.mozilla.com/D11497
--HG--
extra : moz-landing-system : lando
This also uses idiomatic packaging of the native libraries, which
will allow easier downstream consumption of GeckoView.
Differential Revision: https://phabricator.services.mozilla.com/D10775
--HG--
extra : moz-landing-system : lando
This lets us avoid having to put usize in the exported signatures from bindings.rs
It also avoids a heap allocation when defining a clip chain.
Differential Revision: https://phabricator.services.mozilla.com/D12785
--HG--
extra : moz-landing-system : lando
Generally, this switches Services.locale.requestedLocales calls to use
Services.locale.appLocalesAsBCP47.
Differential Revision: https://phabricator.services.mozilla.com/D10980
--HG--
extra : moz-landing-system : lando
JaCoCo chokes on default implementations in interfaces, so remove those
for now.
Differential Revision: https://phabricator.services.mozilla.com/D12607
--HG--
extra : moz-landing-system : lando
Even if the GeckoApp activity is destroyed, the user's tabs remain open as long
as the app process itself isn't killed.
Therefore, for users who aren't automatically restoring their previous session's
tabs on startup, there's no point in renaming the current session store file so
it appears under "Recently closed" instead of being automatically restored,
unless we're really doing a cold startup (and hence mIsRestoringActivity is
false).
This also better matches what we do when we *are* automatically restoring tabs:
We only do it when !mIsRestoringActivity, too.
Differential Revision: https://phabricator.services.mozilla.com/D11800
--HG--
extra : moz-landing-system : lando
Currently selectors that generate the taskgraph receive a list of task labels
back. While this works, it isn't very robust and limits the selectors in what
they can accomplish.
For example, it means we need to use regexes to map test suites to tasks. If
we had the full taskgraph, we could use filter functions instead.
There are also new selectors in the works which will need this (like ./mach
try chooser). Finally the try syntax selector would also need this if we
ever migrate it to use the 'try_task_config.json' format.
Differential Revision: https://phabricator.services.mozilla.com/D12570
--HG--
extra : moz-landing-system : lando
We no longer set up a listener on a per-window basis for category preference changes.
Differential Revision: https://phabricator.services.mozilla.com/D12667
--HG--
extra : moz-landing-system : lando
Bug 828240 switched the children only transform on an outer svg from applying to
each of the anonymous child's children to applying directly to the anonymous
child instead. So now when the viewBox changes on an outer svg, we need to
update (just) the overflow of the anonymous child. The children only transform
on an inner svg still applies to the children of the inner svg, so we continue
updating those children in that case.
Hit testing uses overflows as part of the testing process, so was broken by the
lack of overflow updates.
Differential Revision: https://phabricator.services.mozilla.com/D5668
--HG--
extra : moz-landing-system : lando
Updated to the new path and found a filter that gives us the exact colours we had previously in both light and dark mode.
Differential Revision: https://phabricator.services.mozilla.com/D12660
--HG--
extra : moz-landing-system : lando