This ensures completed tasks have a higher rank than entries added by
eager-index, when a cached task is scheduled despite not changing index
path.
Differential Revision: https://phabricator.services.mozilla.com/D216915
This started with a fix for verifyCFRAfterBlockingTheCookieBanner which failed on the beta branch because of the website was loaded slowly.
To overcome this problem, I've increased the waiting time of the page load.
The UI test successfully passed 150x on Firebase ✅
Aaron suggested using the same approach for real website from our UI tests.
All UI tests successfully passed 1x on Firebase ✅
Differential Revision: https://phabricator.services.mozilla.com/D216809
Depends on D216814.
This patch enables the new attachment bundling/extraction logic for the intermediate certificate downloads.
In my local testing, this gives a significant improvement in the time it takes to download and unpack all the intermediate attachments.
Testing on my laptop (linux laptop on good wifi and gigiabit).
Before: 3.2-4.1 seconds to download and unpack
After: 0.75-0.95 seconds to download and unpack
Would love to pull before/after data for a variety of devices.
Differential Revision: https://phabricator.services.mozilla.com/D216815
This is to enhance our initial remote-settings sync processes for collections. We are able to bundle attachments for collections (targeting collections with many small attachments) to improve client performance.
Ground work that will allow remote-settings consumers to download an attachment bundle. This patch itself does not change any existing client behaviors.
Differential Revision: https://phabricator.services.mozilla.com/D216814
Analyzing if there's any other cases which would be interesting:
1. CoverInitializedName isn't supposed to be generated inside a class, so can't test for that.
2. Shorthand also cannot be created inside a class.
2. Field and FieldWithAccessor is handled in the block prior to this one.
That leaves Constructor and DerivedConstructor, which I think are handled elsewhere.
Differential Revision: https://phabricator.services.mozilla.com/D216868
Increasing the absolute number of http connections on Android from 40 to 128
Will monitor telemetry for indications that this was a bottleneck.
Telemetry landed in bug 1901855.
Differential Revision: https://phabricator.services.mozilla.com/D213475
When double clicking the tag name in the markup view, people can edit the tag,
but they might want to add attributes at the same time, and we didn't allow that.
This patch makes it possible, parsing the input to check for any attributes in
it, and in such case, setting them on the node before changing the tag name.
Differential Revision: https://phabricator.services.mozilla.com/D216384
When there's a mutation and the client sees that the currently selected node
isn't connected anymore, a detached-front event is emitted with a node we think
is the parent node of the removed node, and select it.
Unfortunately, we were picking the first node who was mutated as the one we
should select, without checking that this mutation was the one triggered by the
node removal.
The fix is pretty simple: only pick a parent node if the currently selected node
is in the list of removed nodes from the mutation.
We're not adding a specific test for it, the next patch on the queue (for Bug 1173057),
has a test which was failing intermittently without this fix.
Differential Revision: https://phabricator.services.mozilla.com/D216667
To avoid having to request and await various browser data early during the PDF Viewer initialization, we can include that data in the existing preference fetching instead. With other planned changes to the PDF Viewer, the current situation would only become worse over time.
*Note:* Technically this data aren't preference-values, however we're already including other non-prefs in this list (e.g. `isInAutomation`) and doing it this way simplifies the overall implementation.
Differential Revision: https://phabricator.services.mozilla.com/D216819
Depends on D216811
In most cases, the calls would fail early because of a missing channel argument.
Ideally for invalid tests, all the payload should be correct except for the invalid argument we are testing.
Differential Revision: https://phabricator.services.mozilla.com/D216865
