Sharing from another app listed Element X once, with no way to pick a
conversation: the Direct Share row never showed any.
The app already publishes a long-lived, conversation-categorised shortcut per
room, which is most of what the sharing shortcuts API needs. The part that was
missing is the declaration that ties those shortcuts to the share intent: the
launcher activity had no `android.app.shortcuts` meta-data, so there was no
`<share-target>` for the system to match them against.
It now ships one, matching the activity that already receives ACTION_SEND, and
the shortcuts carry the category it names.
The system offers these from Android 10 onwards, where sharing shortcuts are
handled natively. Reaching further back would additionally need the
androidx.sharetarget library, which is not added here.
Part of #3835
Upgrading over an older install crashed before the login screen with
`SQLiteException: no such column: SessionData.position`. The session, cache
and push databases were opened at their new schema version but their tables
were never migrated.
`SqlCipherDriverFactory` overrides `AndroidSqliteDriver.Callback.onUpgrade` so
the database key can be replaced, and never called through to the base
implementation - which is the only thing that runs `schema.migrate`. Every
SQLDelight migration in the app was skipped as a result.
The override now calls `super.onUpgrade` after the re-key, so a database that
crosses the re-key version is still re-encrypted before its tables are
migrated. It moved out of an anonymous object into a named `MigrationCallback`
so the behaviour can be tested.
Part of #7509
The room list preview prefixes the sender's name unless the room is a
direct message. A room that was created as a group and later converted to
a direct message is not marked as one by the server, so its preview kept
naming the only other person in it on every message, while iOS did not.
The prefix now depends on how many people are actually in the room rather
than on the direct message flag, which is the behaviour asked for in the
issue: no name while only two people can be talking, the name back as
soon as a third joins.
Part of #1548
The HTML sanitiser in toHtmlDocument() drops every tag outside its safelist,
which does not include img. A formatted body whose only content is an image —
a custom emoji, or a bot posting an image inline — therefore cleans down to an
empty document, and because that document is not null the ?: body fallbacks in
the timeline item factory, the room list preview, the notification resolver and
plainText are all dead. The message renders as an empty bubble.
Return null from toHtmlDocument() when no text survives the sanitisation, so the
existing fallbacks take the plain text body, which for these messages carries the
alt text the sender provided.
Every @BindingContainer now lives in a di subpackage, in a file named after
it. Containers that were declared inside a shared file (SimplePlayer.kt,
LockScreenConfig.kt, VoiceMessagePresenter.kt, TimelineItemPresenterFactories.kt,
MediaItemPresenterFactories.kt) are extracted into their own file.
Method names are made consistent too:
- @Binds functions are named bind<ConcreteParameterType>, fixing three that
named something else (notably RoomAliasResolverBindingContainer, whose
provider was still called providesJoinRoomPresenterFactory).
- @Provides functions are named after what they provide, and all use the
provides prefix rather than a mix of provide and provides.
ANDROID_KEYSTORE moves back to the cryptography impl package, next to
KeyStoreSecretKeyRepository which uses it; it is not DI specific and would
otherwise be unreachable from there.
No behaviour change: Metro discovers containers via @ContributesTo and does
not care about method names, so the only call site to follow was in
DefaultVoiceRecorderTest.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Metro's @BindingContainer replaced Dagger's @Module, but the declarations
kept the old Module suffix. Rename them (and their dedicated files) so the
type name matches the annotation, matching UserStatusBindingContainer which
already followed this convention.
No behaviour change: Metro discovers containers via @ContributesTo, so there
were only two call sites to update, both in tests.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The 174 commits develop gained since this branch was cut touch preview data
in three ways, and each is resolved so the flipped konsist rule still holds
across the merged tree.
Two screens the branch had renamed were deleted upstream by "Remove the
account provider search screen and homeserver resolver": the
changeaccountprovider and searchaccountprovider packages. Their deletion
wins, so ChangeAccountProviderStatePreviewParam, SearchAccountProviderStatePreviewParam
and the two views that used them go with them rather than coming back
under the new name.
Two preview classes develop added still carried the old suffix and are
renamed here, along with the `@PreviewParameter` reference in each view:
GlobalSearchStateProvider -> GlobalSearchStatePreviewParam and
SendFailureDialogStateProvider -> SendFailureDialogStatePreviewParam.
Everything else merged textually. No production class extending
PreviewParameterProvider is left with a `Provider` name, and no reference
to any of the 194 renamed names survives outside the CHANGES.md entry that
records a past pull request title.
* Make PIP react to camera orientation changes in the other participant's device
Adds a `controls.onPipMediaOrientationUpdate` callback for the WebView that will trigger a new `pipParamsBuilder.setAspectRatioFromOrientation(orientation)` method
* Fix warning about `PictureInPictureParams.Builder.setAspectRatio` only being available on API 26+
RootFlowNodeTest was declaring its own no-op fake, while the oauth test
module already provides one. Use it, and move it out of the customtab
package, since it is not specific to the custom tabs.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The fakes were declared in RootFlowNodeTest, so other tests could not use
them. Move each of them next to the API it fakes, following the existing
convention of a createNodeResult lambda defaulting to lambdaError():
- FakeLoginEntryPoint in features/login/test
- FakeBugReportEntryPoint in features/rageshake/test
- FakeSignedOutEntryPoint in features/signedout/test (new module)
- FakeAccountSelectEntryPoint in libraries/accountselect/test (new module)
The login and rageshake test modules did not have Appyx on their
classpath, so add the architecture module to their dependencies.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Develop gained resolveImageMimeType, which reads the image header so a
share that declares a broad type is encoded as the format it really is.
Resolving the wildcard first would hand that helper a concrete image/jpeg
and make it return early, so a transparent PNG shared as image/* would be
flattened and an animated GIF would be recompressed. The header is now
read from the type the sender declared and the default subtype is applied
to whatever comes back, so the sniffing still runs and an undecodable file
no longer keeps a wildcard.
A non-image shared as image/* is therefore sent as image/jpeg rather than
image/*, which is the whole point of this branch, so the test covering
that case now expects the default subtype.
Tapping a room alias whose server is down left the resolver screen on a
spinner for as long as the screen stayed open, with no way forward other
than going back.
Bound the resolution with the ten second timeout already used when
joining a room by address, so the retry dialog is shown instead. The fake
client's resolve lambda becomes suspending so a server that never answers
can be simulated.
* Keep the send button when the composer is emptied while editing
rememberEndButtonParams checked whether the composer had sendable text
before it checked whether the user was editing a message, so clearing
the text of a message being edited replaced the send button with the
voice recorder button. Starting a recording from there abandons the
edit.
Check the editing branch first, as the formatting composer already does,
and let it render a disabled send button when there is nothing to send.
* Update screenshots
* Request media audio focus when playing a voice message
Voice message playback asked for audio focus with
USAGE_VOICE_COMMUNICATION, the same bucket as Element Call and voice
message recording, while the player itself renders on the media stream.
Audio policy layers act on the declared usage, so a car head unit
switches to its call audio context and routes the playback away while
the player keeps running.
Declare USAGE_MEDIA (and STREAM_MUSIC on the legacy path) for playback,
which is what the player actually renders. Recording and Element Call
keep asking for voice communication focus.
* Keep playing a voice message through a short sound from another app
A notification, a navigation prompt or any other app taking focus for a
moment sends AUDIOFOCUS_LOSS_TRANSIENT or AUDIOFOCUS_LOSS_TRANSIENT_CAN_DUCK,
and DefaultMediaPlayer pauses on it. The voice message stops part way
through and the user has to find it again and restart it.
Recording already ignores those two, because the capture path is not
affected by focus. Give voice message playback the same treatment, moving
the requester test into a named function rather than growing the inline
condition. A permanent AUDIOFOCUS_LOSS, a real phone call, still pauses,
and the media viewer is untouched.
Benoit Marty
hayaksi1
Benoit Marty
Benoit Marty
bxdxnn
Jenna Vassar
Jorge Martín