At this moment, we have two extension methods that have duplicate
functionality to construct search term groupings. One on `List<Tab>` and
one on `List<TabSessionState>`. The former is used for everything
related to tabs piped through the `TabsFeature` and the latter is for
consumers of `BrowserState` directly.
The bug occurs because our implementation of search groupings was
updated only on the former extension, but the `HeaderBinding`, that
observes the BrowserState and updates the title visibility, was using
the latter.
Ideally, we remove this duplication when we no longer have separate data
classes for consumers of `TabsFeature`, but this intermediary fix should
suffice.
Before this patch, this was the behavior - 'remove' button is clicked, we'd ask
the storage to remove metadata (on its IO thread), then navigate to Home
Screen.
This resulted in a race we could end-up on the Home Screen before delete
finishes, so the search groups do not appear to be removed (but,
refreshing the Home Screen again shows that they are removed).
This also resulted in an unnecessary navigation which felt very janky
(screen will "scroll" to the top) and was way more work than necessary.
After this patch, we:
- dispatch two actions (on browserstore, on homefragmentstore) which
remove the search groups from any relevant in-memory state; any UI bound to
this state will be automatically "refreshed"
- no longer navigate as part of the remove action, so the UI doesn't
move and removal happens "in-place"
It seems like we no longer need to use rotation for the chevron, since
we are now using two different icons within the `ic_chevon` that change
depending on the `state_activated`.
This helps identify file IO. Unfortunately, with this marker, it's
difficult to separate code we own from code we don't own. However, I
wasn't sure what the best implementation would be to address that
(e.g. ideally, we would ignore violations in code we don't own rather than
annotate the markers) so I thought we can land it this simple way and
improve it incrementally.
* For #21437 - Relocated Home-related settings to its dedicated sub screen
* For #21437 - Updated show top sites toggle text
* PR: Fixed lint warning. Reverted preference keys
* PR: added ignore for UI test
* PR: Added ignore for UI test
The bottom gray border of the header item from the Inactive Tabs section was correctly set when collapsing or expanding said section, but not on init. So if the section was initialized collapsed the gray border would not be present.
This regressed in our previous fix that made sure child tabs don't
mistakenly get moved out of the group if their parent is navigated
away, or in case the child tabs are redirected.
However, when a subsequent load occurs in any tab in the group the
search terms need to be cleared and the tab removed from the group
to prevent false positives.
* Fix description for `home_screen_displayed` metric
The current description appears to be incorrect by my reading of the source.
* Update app/metrics.yaml
Co-authored-by: Gabriel Luong <gabriel.luong@gmail.com>
Co-authored-by: Will Lachance <wlachance@mozilla.com>
Co-authored-by: Gabriel Luong <gabriel.luong@gmail.com>
* Docs only: Add extra context to `events.browser_menu_action`
This came up as a potential point of confusion in a discussion with DS. This is a docs only change.
* Update metrics.yaml
Co-authored-by: Will Lachance <wrlach@gmail.com>
* For #21360 - Added toggle for search term tab groups
* For #21360 - Lint cleanup
* PR: Added missing licenses and possibly fixed UI test
* PR: Added a "scrollTo" to potentially fix a UI test
* PR: Added potential fix for alwaysStartOnHomeTest
* PR: Added temporary ignore to alwaysStartOnHomeTest
* PR: added missing ignore comment
* For #21360 - Added missing feature flag driven visibility logic
Co-authored-by: Sebastian Kaspari <s.kaspari@gmail.com>
Co-authored-by: mergify[bot] <37929162+mergify[bot]@users.noreply.github.com>
Home screen isn't actually visible in case we're displaying awesomebar
search results. The navigation is thus unnecessary and actually causes visual
jankiness as we display home for a moment before covering it up with
search results.
The latest geckoview-omni package correctly declares its capabilities,
including the `glean-native` one.
Additionally it is able to pick geckoview-omni over glean-native in all
configurations.
* Issue mozilla-mobile#21319 - Moved inactive tabs to the top of the normal tabs tray.
* Issue mozilla-mobile#21319 - Added a delete icon to delete ALL inactive tabs.
* Issue mozilla-mobile#21319 - Changed default inactive time period to 14 days
* Issue mozilla-mobile#21319 - Hooked inactive tabs setting to UI code
Inactive tabs setting is also disabled when the user has selected the one day or week auto-close tab setting.
* Issue mozilla-mobile#21319 - File and Lint cleanup
* PR: Fixed bug causing grouped tabs to also show in "Other" when marked as inactive but inactive is OFF in Settings
* PR: Fixed lint warnings
* PR: Removed redundant feature check
* PR - Ignore test until search term tab groups switch is done
These double comparisons are easier to read and see the pattern of on one line
so I'd rather keep them on one line. Additionally, it's difficult to
test each change individually so I'd rather not make additional changes.
To do this, I suppressed the max line length warning.
This commit was generated primarily by a macro that:
- appends `== component &&`
- appends `== item`
- (if applicable) Skips to the ending brace
- Go down one line and move cursor to the front of the line to prep for repeat
My only intervention was to skip extra lines to line it up to run again
and specify how many times in a row it should run.
---
The `to` in this code is an infix function that calls instantiates a
Pair under the hood. Subjectively observed, when this method is called
it generally hits the else case so 35 Pairs are instantiated each call -
that's 560 bytes. This method is called frequently - for example, an estimated
4 times each time a letter is typed on the homescreen and a measured 116 times
in a simple navigation (see the issue). The latter generates an estimated
63.4 KiB.
It was straightforward to remove these allocations so that's what this
change does.
The primary risk from this change is that it's difficult to test each
case to ensure it's working.
- stick to one naming scheme: rename articles to stories and use this all
throughout the app.
- add some spacing above the new section (as per the current design)
We currently have a 15s buffer to match metadata to its corresponding
visit. However, a existing metadata record can be updated more than
15s after it was created e.g. when closing the tab and updating
the view time.
We discovered that in a tab restore scenario we were recording view time
observations that were wrong - we'd record time deltas as-if user was
looking at the page while the browser wasn't running.
This happens because when we record a viewTime observation, we compare
current time with lastAccess time of the tab. In a restore scenario,
that lastAccess time happens to be from when the browser was last
running - which could be days ago.
The simplest solution was to not record a viewTime observation if the
url for a tab didn't change during a load event. To achieve this, we
needed to change which action we were using as a proxy for "navigation
events" - UpdateUrlAction contains the new url, allowing us to compare
against the current tab url.
Alternative solutions would be to keep using loading actions, but
dispatch a lastAccess event before performing a metadata update. This
would have worked, but would result in two lastAccess events being
dispatched for each navigation event instead of just one.
We only instrument onCreate because it's the only one with an
implementation.
While declaring this as a function that accepts a lambda is less
fragile, we've previously had issues with it such as suspected memory
leaks when used for telemetry. Therefore, we go with the simpler
approach.
This is a speculative fix for the intermittent issue. Typically, these
intermittents are caused by mocked lambdas but there is no mocked lambda
here. If this doesn't work, one `any()` argument fills in for a lambda:
it's possible that's causing the failure.
Unfortunately, I can't verify this fix easily because the "run test until
failure" option was removed from Android Studio.
See b3f5c87585 for a prior attempt to
address the intermittents in this class.
This is a bug we noticed after landing search term grouping.
An adapter can submit an empty list of items to the `ConcatAdapter`
early. This has the side-effect of triggering our `observeFirstInsert`
too soon and therefore updating the visibility to show the empty tray
placeholder and never switches back.
Our solution is to keep a constant observer on the adapter so we can
perform the visibility check on every insert/remove.
Co-authored-by: Roger Yang <royang@mozilla.com>
* For #21236: Separate tabs with the same search term into a different section in tabs tray
* Issue #21236: Scroll to selected tab + various tab fixes for groupings
* Issue #21236: Fix failing test
Co-authored-by: Jonathan Almeida <jalmeida@mozilla.com>
The test as it exists relies on the robolectric lifecycle, which is hard
to predict, so it doesn't seem worth fixing the test. Writing the test
any other way would require excessive mocking, which also seems
impractical.
