We set the ContentBlockingSettings directly on the GeckoRuntime now to
improve the startup of the engine.
This change has requirements from Android Components and GeckoView, so
we would only see the full perf benefits in Nightly as the changes ride
the train, although we might start to see some of them as we're updating
the GeckoProvider for the `geckoBeta` variant as well.
Co-authored-by: Arturo Mejia <arturomejiamarmol@gmail.com>
* Extract controller into it's own class. Implement find dupes and filter based on username.
Create edit login controller. Add text watchers and check for duplicates.
Edit controller test
* Find duplicates and save to store
* Retrieve duplicates from AC and check list on username text changed
Move duplicates logic into the controller
* Add glean pings for delete and edit. Move logic for login manipulation into the datastore.
* Use correct threads in controller. Enable save button when applicable.
Save enabled in datastore.
Move login data to datastore
Rebase with password error states
Update metrics to be more specific for edit
* Create logins controller for AC calls
* Interactor and controller methods for edit login. Add edit view to separate out some layout manipulation.
Inflate view in edit fragment. Double layout showing up.
Edit view
Controller tests
Controller tests passing
Interactor tests
Lint and detekt cleanup
* Remove datastore and use storage controller for all logins calls to password storage.
Addressed comments
Lint
:
Rebase - 1
* For 11657: add LP attribute for tracking protection
* For #11704: added tracking_protection_enabled attribute
* Added docs for the new attributes, linking to data-review to the mma.md
* Rename null to none when no ETP is enabled
Stricter synchronization by always using the same "loadedSearchEngines"
variable.
With "loadedSearchEngines" calling "refreshAsync()" we also get the fallback
engines to contain reddit and youtube (which are programatically added) and
also now we properly remember and display the engines added by user.
We have two search engine types:
- one based on MLS reported region,
- one based only on Locale.
There are multiple steps involved in returning the default search engine for
example and though at each step we could verify if a certain operation is
completed we are still exposed to concurrency issues.
Simplest and most effective way to make sure the MLS engines do not mix with
Locale based engines is to use the same type of engines for the entire duration
of the app. At the next cold start we'll verify again which engines to use.
Using the Locale based engines (fallbacks) is expected to only happen once, at
the first run of the application after being installed.
* for #11830 added new metric for collecting startup method
move all source startup telemetry into its own logic and added an UNKOWN state
* switched back to onNewIntent solution
* renamed the metric
When you remove a search engine from our list, an undo snackbar that outlives the fragment is shown. If you go back to the search fragment and press undo, now the shortcut list gets updated
Tiger Oakes