This was added as a temporary measure to allow matrix.org users to use Element X before it supported MAS, but it hasn't been needed for a very long time
The GITHUB_MAX_ISSUE_BODY_SIZE constant added with the description truncation
is a five-digit literal, above detekt's acceptableLength of 4, so
:features:rageshake:impl:detekt failed. Write it as 65_536, matching
MAX_LOG_LINES_SIZE in RageshakeConfig.
The timeline action sheet offered Forward on a live location share, but
forwarding one always failed: Mode.Live maps to the SDK's
MsgLikeKind.LiveLocation, which RoomContentForwarder rejects outright. So
the action was shown for exactly the set of events on which it could
never work.
canBeForwarded() now returns true for a location event only when its mode
is Static. Ended live shares keep Mode.Live and fail for the same reason,
so they are excluded too. ActionListPresenter needs no change: its check
is the function's only call site in the repo, so the single gate covers
the timeline sheet, the threaded timeline and the pinned messages list.
Hiding the action mirrors Element Web and the sticker exclusion already
present in the same function. Converting a live share into a static
m.location on forward would be a product decision and is out of scope.
Relates to https://github.com/element-hq/element-x-android/issues/7190
The whole crash call stack was appended to the bug report description, and the
bug report server copies that description verbatim into the GitHub issue it
opens. GitHub refuses an issue body over 65536 characters, so the server's save
step failed and answered "Could not save report". The stored crash data is only
cleared after a successful send, so every retry failed the same way.
Truncate the description before sending it. The user's own text and the GitHub
issue link come first, so only the tail of the crash stack is dropped. The
uncaught exception handler already writes the full stack through Timber, so it
stays available in the attached log files.
The 40 KB budget leaves room for the other reported fields, mainly the list of
omitted log file names, which is not otherwise bounded.
* Expand the emoji picker to full height when the keyboard is shown
The picker content is rendered with fillMaxSize, so the sheet measures itself to
the full window height. That makes the partially expanded anchor available, and
the modal sheet auto-shows at that anchor, half way down the screen.
The sheet reserves the IME inset through its content window insets, but applies
that padding at its own bottom edge, which is off-screen while it rests half way
down. The band that is both on-screen and above the keyboard is then only about
a tenth of the screen, which is why the matching emoji end up behind it.
Expand the sheet while the keyboard is visible so its bottom edge returns to the
screen edge and the inset it already reserves lands on top of the keyboard.
Modifier.imePadding on the sheet content is not an option: the sheet applies the
insets with windowInsetsPadding, which consumes them, so a nested imePadding
reads zero.
* Cover the keyboard driven sheet expansion with a test
The IME reaction moves into ExpandSheetWhenImeIsVisible so a Robolectric test can
drive it: dispatching window insets to the composition flips
WindowInsets.isImeVisible, and the sheet state is observable from the test. Both
the shown and the hidden case are asserted, which is the patch coverage this
change was missing.
* Do not leave a blank strip below the expanded emoji search
The 10dp of bottom padding on the search bar separates the collapsed
field from the category tabs underneath it. It stayed applied once the
search bar expanded to fill the picker, so the picker stopped 10dp short
of its own bottom edge.
That strip used to sit below the visible part of a half height sheet.
Now that the sheet expands against the keyboard it reads as a gap
between the results and the keyboard, so only pad while collapsed.
heroes is populated for any room whose name is derived from its members,
not only for direct messages, so a two person room that is not a DM
showed a member's Matrix ID as its subtitle.
Carry isDm on SelectRoomInfo and read the hero only when it is set. The
canonical alias still takes precedence, so nothing changes for a room
that has one.
Restore the snapshot backed set the store started with and declare
allowedEventIds as SnapshotStateSet so the interface says out loud that
reads of it are observable.
TimelineProtectionPresenter reads the set inside derivedStateOf, so a
plain set silently stops the timeline from re-rendering when an event is
revealed. Naming the type keeps that requirement from being mistaken for
an implementation detail.
* Show a Message not found placeholder when a replied to message cannot be loaded
When the replied to event could not be fetched, the reply preview rendered the
raw SDK error string in critical styling, and tapping it emitted a focus
request that could never succeed, so the user got a dead end dialog.
Render the localised error_message_not_found string in the neutral italic style
already used for redacted and undecryptable replies, and drop the click handler
for that case. The existing key is reused rather than adding a new one; a "you
do not have access" wording was rejected because this state covers any failure
to load the event, not only a permission problem.
* Update screenshots
* Trigger CI after recording screenshots
---------
Co-authored-by: ElementBot <android@element.io>
The store no longer needs a mutableStateSetOf: an ordinary Set kept in a
mutableStateOf does the same job, and matches DefaultMediaPreviewConfigStateStore.
The Compose state itself has to stay. TimelineProtectionPresenter derives
ProtectionState from the store inside a derivedStateOf, so the set must be a
snapshot object for a reveal to invalidate it and reach the timeline. Backing it
with a plain mutableSetOf leaves the presenter with nothing to observe and it
stops emitting after ShowContent.
The pinned events timeline is built with an exclusion filter rather than an
allowlist, so an event of a type the app has no renderer for reaches both
pinned surfaces. The banner falls through to the formatter catch-all and reads
"Unsupported event", and the list renders an unknown content tile.
Drop unknown content and message-like or state events that failed to parse from
both surfaces, along with any day divider left without an event to introduce,
and show the empty state when nothing renderable remains. The spec makes this
the client's call. Such an event can still be unpinned from the room timeline,
where it is still rendered.
Undecryptable and redacted events are untouched: they carry information the
user needs.
The set of revealed event ids lived on TimelineProtectionPresenter, which is
scoped to the room graph. That graph is rebuilt every time the room is opened,
so the set was empty on re-entry and every revealed image was hidden again.
Move it into a session scoped TimelineProtectionStore. The store is in memory
rather than persisted: revealing media is a privacy control and there is no
hide action today, so keeping the reveals for the session leaves restarting the
app as a way to undo one, and the set is dropped on logout.
The account provider search treated its whole input as a host, so a full user
ID was probed as https://@alice:example.org, which never resolves, and the
permissive URL fallback then offered that invalid address as a provider.
Recognise a complete user ID and resolve the homeserver from its server name,
preserving a port when there is one. Input that starts with @ but is not yet a
complete user ID resolves nothing, so no candidate is produced while the user
is still typing. Detection reuses MatrixPatterns.isUserId and UserId.domainName
rather than a new pattern.
When the replied to event could not be fetched, the reply preview rendered the
raw SDK error string in critical styling, and tapping it emitted a focus
request that could never succeed, so the user got a dead end dialog.
Render the localised error_message_not_found string in the neutral italic style
already used for redacted and undecryptable replies, and drop the click handler
for that case. The existing key is reused rather than adding a new one; a "you
do not have access" wording was rejected because this state covers any failure
to load the event, not only a permission problem.
The blocked users list displayed each Matrix ID but offered no way to get it
out of the app, which is what someone who blocked a user for abuse needs in
order to report them onwards.
Long press now copies the Matrix ID and confirms with a snackbar, reusing the
ClipboardHelper and SnackbarDispatcher idiom from the user profile screen. A
short tap still opens the unblock confirmation. The row declares a long click
label so the action is announced by a screen reader. No new string is needed.
After leaving a room listed in a space, the room stayed in the child list with
a Join button, even though it can only be entered with a new invitation. It
disappeared only after leaving and reopening the space, because the server
omits it from the refetched hierarchy.
Apply the same rule locally in the presenter. The filter keys on the LEFT
membership combined with an invite join rule rather than on the derived
visibility, which collapses knock and unknown join rules into private and
would wrongly remove the Join button from knockable rooms.
Two of the three modes on the room notification settings screen had
screen-scoped keys; the third reused common_mute, which also labels the Mute
action button on the room details screen. A translator could not fix one
occurrence without breaking the other.
Add screen_room_notification_settings_mode_mute and use it at both places on
that screen. The English value is unchanged, so nothing renders differently.
The action button keeps common_mute, which is the one context where the
generic verb is correct.
A DM whose other member has left still offered the call button. Starting
that call put the user alone in a room nobody could join, with no way out
of the call screen.
The presenter now clears canStartCall for a DM with no other active
member. activeMembersCount counts joined and invited members, so a fresh
DM whose peer has been invited but has not joined yet can still start a
call. Joining an already ongoing call is untouched, so a call started
before the other member left stays reachable.
Relates to https://github.com/element-hq/element-x-android/issues/6189
Benoit Marty
Benoit Marty
Jorge Martín
hayaksi1
ganfra
ganfra
bmarty
Benoit Marty
Eklavya Jandial