• Tweaked some background poller logs
• Refactored the `BackgroundFetch` handling to use a `DispatchSourceTimer` instead of `NSTimer` and use a specific `DispatchQueue` to avoid race conditions
• Refactored the BackgroundTaskManager to use `DispatchQueue` and `DispatchSourceTimer` and removed unused code (was seeing background tasks incorrectly running too long so wanted to clean it up)
• Fixed an issue where pollers would incorrectly be released during background polling
• Fixed an issue where the background poller wouldn't update the app notification badge count
• Fixed an issue where the community pollers 'timeSinceLastPoll' was incorrectly being given both an epoch timestamp as well as a duration since the last poll (resulting in always just refetching recent messages)
• Fixed an issue where the community poller wasn't updating the last poll timestamp (also renamed some functions to make them clearer)
• Fixed an issue where pollers could incorrectly be started in the background (eg. when receiving a PN)
• Added a 'requireAllRequestsSucceed' flag to the ConfigurationSyncJob (so it'll report a failure if any individual request fails)
• Fixed an issue where a number of 'response' types weren't encoding correctly (only impacted unit tests)
• Fixed an issue where the logger wasn't correctly respecting the log level settings
• Fixed an issue where the path status indicator wouldn't default to unknown
• Fixed an issue where the generic database error didn't replace the 'app_name' variable
• Fixed an issue where notification content might not be shown correctly
• Fixed an issue where a group could be partially created due to one of it's configs failing to be stored (we now consider that a failure so the user can try again)
• Fixed an issue where processing a config message in the BackgroundPoller could result in attempting to fetch from communities after the process completed
• Fixed a crash where a database query could incorrectly be interrupted after it completed if both happened at just the right time
• Fixed broken unit tests
• Updated the string linter logic to error when incorrect parameter names are given (or missed)
• Updated the code based on the latest string changes
• Fixed an issue where the string linter wasn't working
• Added code to schedule any missing recurring jobs (so we no longer need to worry that the jobs have been lost or the migrations that added them ran correctly in their final state)
• Added the 'FailedGroupInvitesAndPromotionsJob' to flag invites/promotions which weren't sent before the app closed as failed
• Updated to the latest libweb
• Updated to the latest libSession
• Updated the config sync job to delay marking itself as failed if the network is not connected (it'll now observer the network status and trigger the failure callback when reconnected, which will result in another sync attempt shortly after - this will prevent a disabled network from building up the failure count excessively causing sync delays)
• Fixed a bad memory race condition which could occur with the new `Storage.performOperation` logic
• Updated the SyncPushTokensJob to wait for a bit for paths to build before failing
• Updated the PushNotification service to be selectable via the dev settings
• Updated the database timeout code to be a little more developer friendly
• Updated the code to stop blocked contacts from appearing in the main conversation list
• Removed the invalid "push-testnet" push server
• Removed the logic to configure the APNS push service based on the service network (was incorrect)
• Removed the 'Int' raw type constraint for the 'FeatureOption'
• Fixed an issue where the initial conversation query for groups would fail due to an invalid join
• Fixed an issue where the initial conversation query wouldn't include the 'markedUnread' flag (meaning the conversation wouldn't correctly get marked as read)
• Fixed a rare bad memory crash
• Fixed an issue where the modal wouldn't be dismissed after updating the group display name
• Fixed an issue where the "Recreate Group" button was the wrong height
• Reworked synchronous database access to prevent reentrancy issues
• Updated the JobExecuter to take a "Scheduler" rather than a "DispatchQueue" (allows for a potentially refactor to use "OperationQueue" instead if we want more fine-tuned control over the number of concurrent executions which can occur)
• Reworked the database rentrancy fixes
• Added a timeout to prevent deadlocks with our custom synchronous database actions
• Fixed a case where we were incorrectly accessing the database in a reentrant way
• Added a 2000 limit to the input field
• Added an error when opening the share extension before completing onboarding
• Updated the code to treat an empty `adminSignature` as null (in case a client sends it incorrectly)
• Updated the code to include updated group control messages when determining unread
• Updated code to perform a supplemental key rotation (if needed) when resending invites to members flagged as supplemental (to handle a case where the initial invite request failed)
• Updated the code to schedule group config update control messages when making the local changes but make their sending contingent on the next config sync succeeding
• Updated the code to only enable the Share and Notification extensions once the onboarding process has been completed
• Updated the code to always enqueue config syncs for all configs on launch (if there are no pending changes then it won't do anything, but if there are then this will increase the likelihood that they will be synced)
• Fixed an issue where the default contacts query was based on thread records instead of contact records
• Fixed an issue where creating a conversation by tapping on a group member or avatar wasn't correctly creating the conversation in the draft state
• Fixed an issue where incoming messages could be added to the "Note to Self" conversation if the "syncTarget" was incorrectly set on them
• Added some localised copy
• Added the helpdesk article URL for the deprecated groups banner
• Updated mentions to include the current user in groups, 1-1 and NTS conversations (also made the current user appear as "you")
• Fixed a layout issue with the deleted message view
• Fixed an issue where the "mark as unread" action would incorrectly appear for the Note to Self conversation
• Fixed an issue where we were incorrectly sending an ExpirationTimerUpdate in some cases
• Added logic to trigger a full refetch when getting promoted to admin within a group
• Added a timeout to display picture uploads
• Updated the code to inject the DisplayPictureManager via dependencies
• Updated the resend invite functionality to have a blocking loading indicator as well
• Refactored some closure-based logic for display picture uploads to use Combine instead
• Refactored the SnodeReceivedMessageInfo (with a migration) so it's no longer using a weird constructed key (now have individual queryable columns for the values)
• Fixed a crash due to our ThreadSafe code using a struct instead of a class (rendering it non thread safe)
• Fixed a crash which could occur on the home screen if the data loaded before the UI finished loading
• (Hopefully) Fixed a crash which could occur when the OS optimised async execution to run immediately within an existing database transaction (potentially resulting in re-entrant database access)
• Fixed an issue where the database read/write publishers weren't checking for a valid database state before actual query execution (only during the creation of the stream)
• Updated to the latest Lucide version
• Updated the bin icons to use the new Lucide icon
• Fixed an issue where flagging a members messages to be deleted when kicking them wasn't using the standard message removal process
• Fixed an issue where you couldn't delete a group after it was destroyed
• Fixed an issue where being promoted to admin would fail because you weren't already an admin
• Fixed an issue where the share extension wasn't populating it's conversation list
• Fixed an issue where the share extension had a translucent header
• Fixed an issue where messages could incorrectly be considered disappearing messages (with no duration)
• Fixed an issue where non-admins couldn't locally delete messages in community conversations
• Fixed an issue where deleting Note to Self messages for all devices was incorrectly leaving an artifact
• Fixed an issue where emoji reacts added to messages which have been deleted locally would still get added to the deleted artifact
• Fixed a layout issue in groups/communities with the "this message was deleted" artifact
• Fixed an issue where highlighting and unhighlighting a "danger" action in the context menu would result in it changing to a non-danger colour
• Fixed an issue where the community poller always thought it was the initial poll
Morgan Pretty
Morgan Pretty
mpretty-cyro
Ryan ZHAO
ThomasSession
Bilb
stfsession