- Added a method to send failure notifications to Flutter when the background service fails to enter the foreground.
- Implemented a broadcast receiver to handle service failure notifications and notify Flutter about the failure.
- Enhanced the persistent streaming service to attempt recovery for failed streams.
- Introduced heartbeat monitoring for SSE streams to detect stale connections and trigger recovery actions.
- Added a method to ignore specific errors related to RenderFlex overflow.
- Refactored the error display layout to use a centered scrollable view.
- Improved visual presentation with a gradient background for the error icon and updated button styles for retry functionality.
Hide archived assistant variants in the linear chat view and track
previous assistant as versions so regenerated responses do not
duplicate or lose history. When regenerating, mark the previous assistant
message with an archivedVariant flag for the UI and keep it in server
history. Add a ChatMessageVersion model and a versions field to
ChatMessage to store prior generated variants. Implement
archiveLastAssistantAsVersion in chat providers to snapshot the last
assistant message into versions and reset the message for a fresh
streamed generation. Finalize flow updates to attach an adjacent archived
assistant as a version when needed so the UI can present a switcher
between current and past variants. These changes prevent duplicate
messages, preserve previous responses, and enable variant switching.
Add Riverpod-aware image header and cache manager support for network
images used in avatar and markdown widgets. Convert AvatarImage to a
ConsumerWidget and read a self-signed cache manager and HTTP headers
from Riverpod so CachedNetworkImage can send auth/custom headers and use
the provided cache manager. In Markdown image builder, obtain headers
and cache manager from a ProviderContainer (via ProviderScope.containerOf)
to enable the same authenticated loading in non-consumer contexts.
Introduce image_header_utils.dart to centralize building Authorization
and custom headers from auth/api providers, with helpers for Ref,
WidgetRef, and ProviderContainer. Add dependency adjustments in
pubspec.lock for flutter_cache_manager and http marked as direct main.
These changes ensure protected images (self-signed or auth-required)
load correctly across the app and reuse the configured cache manager.
Introduce server TTS support and engine selection while keeping
device TTS as the default.
- Add new persistence keys for storing TTS engine and selected
server voice (ttsEngine, ttsServerVoiceId, ttsServerVoiceName).
- Extend TextToSpeechService to support two engines:
TtsEngine.device (FlutterTts) and TtsEngine.server (remote audio).
- Wire in an AudioPlayer and optional ApiService to fetch raw
audio bytes from the server and play them, with event hooks
mapped to existing lifecycle callbacks.
- Implement fallback to device TTS on server errors or empty
responses, and ensure player lifecycle (pause/stop/dispose)
is handled when using server engine.
- Allow engine and preferred voice to be configured before
initialization and updated at runtime via updateSettings.
This enables selecting a server-side voice and using a remote
TTS provider while preserving compatibility with the existing
device TTS implementation.
- Updated multiple instances of the 'monospace' font family to use AppTypography.monospaceFontFamily for consistency across the application.
- This change enhances maintainability and ensures a unified typography approach throughout the app.
- Introduced new regex patterns to remove thinking and reasoning blocks from markdown input.
- Added functionality to strip emojis from the text, improving clarity for TTS.
- Implemented HTML entity decoding to ensure proper text representation.
- Replaced the existing sanitization method with a more comprehensive markdown-to-text conversion approach, enhancing performance and maintainability.
- Integrated autofocus management for the composer in both chat and onboarding contexts to improve user experience.
- Added error handling for focus management to ensure smooth onboarding transitions.
- Updated modal bottom sheet handling to restore autofocus state after onboarding completion, enhancing usability across the app.
- Updated the file attachment service to utilize a new LocalAttachment class, improving the management of file metadata such as display names.
- Refactored methods for picking and uploading files to accommodate the new LocalAttachment structure, ensuring consistent handling of file attributes.
- Improved the chat page to validate and manage file attachments more effectively, enhancing user experience during file uploads.
- Added functionality for image previews in the file attachment widget, allowing users to see selected images before sending.
- Introduced a remove button for attachments, improving usability by enabling users to easily discard unwanted files.
- Replaced references to AppColorPalettes with TweakcnThemes across various files to standardize theme usage.
- Updated the AppTheme and AppColorTokens to utilize TweakcnThemeDefinition for improved theme management.
- Adjusted UI components in ChatPage, ChatsDrawer, AppCustomizationPage, and ProfilePage to align with the new theme structure, ensuring consistent styling and color application.
- Removed the deprecated color_palettes.dart file to streamline the theme architecture.
- Introduced new preference keys for TTS settings: voice, speech rate, pitch, and volume.
- Updated SettingsService to handle TTS settings and persist them.
- Enhanced AppSettings to include TTS-related properties.
- Implemented TTS settings UI in AppCustomizationPage, allowing users to select voice and adjust speech parameters.
- Added localization support for TTS settings in multiple languages.
- Updated the Deliverfile to submit for review and skip metadata and screenshots.
- Removed outdated build lane from Fastfile.
- Refactored API service to eliminate fallback to latest assistant message, addressing duplication issues in multi-turn conversations.
This commit refactors the connectivity service to be more robust and less prone to UI flicker.
Key changes:
- Any successful API response now briefly suppresses the offline warning. This prevents the UI from flashing an offline message between regular connectivity checks.
- The threshold for showing the offline warning is increased from 2 to 3 consecutive failed health checks.
- The timeout for health checks is increased to better handle slow networks.
- The offline warning is now suppressed if there are active data streams to avoid interrupting the user.
- A custom JSON converter is added for conversation metadata to handle potential type mismatches from local storage.
This commit refactors the `modelToolsAutoSelectionProvider` to more robustly handle the automatic selection of tools when the model changes or the list of available tools is updated.
Key changes include:
- Triggering the auto-selection logic not only when the selected model changes, but also when the `toolsListProvider` finishes loading. This ensures tools are correctly applied even if they load after the model is selected.
- Restructuring the logic to handle asynchronous tool loading more gracefully, preventing race conditions.
- Using `Future.microtask` to schedule the tool application, ensuring state updates are handled correctly within the provider lifecycle.
- Clearing selected tools if the new model has no associated tools or if previously selected tools become invalid.
- Integrated BackgroundTasks framework in iOS to manage background processing for audio streams.
- Added methods to register, schedule, and handle background tasks, allowing streams to continue for extended periods.
- Enhanced the BackgroundStreamingHandler to support background task notifications and keep-alive signals.
- Updated Info.plist to permit background task identifiers, ensuring compliance with iOS requirements.
- Improved the PersistentStreamingService to handle background task extensions and keep-alive signals effectively, enhancing overall streaming reliability.
- Added a new `toolIds` field to the `Model` class to support tool identification.
- Implemented `modelToolsAutoSelectionProvider` to automatically apply model-specific tools when the selected model changes.
- Enhanced the logic to filter and set valid tool IDs based on available tools, improving user experience by ensuring relevant tools are automatically selected.
- Updated the default model auto-selection provider to initialize the new tools auto-selection feature, ensuring seamless integration within the app's lifecycle.
- Deleted the SelfSignedCertificateManager and its associated files to simplify the certificate management process.
- Updated ApiService and ConnectivityService to include self-signed certificate configuration directly, enhancing clarity and maintainability.
- Adjusted comments to reflect the new approach to handling self-signed certificates, ensuring better understanding of security considerations.
- Improved the application startup sequence by deferring unnecessary initializations, contributing to a more efficient first paint performance.
- Removed the immediate initialization of SelfSignedCertificateManager to improve the app's first paint performance.
- Introduced lazy initialization of the certificate manager after the first frame, enhancing startup efficiency.
- Implemented a fast path for migration checks in PersistenceMigrator to skip unnecessary operations if already completed in the current session, improving overall performance during app startup.
- Updated comments for clarity on the changes made to the startup sequence and migration process.
- Updated AuthStateManager to clear the active server ID upon logout, ensuring a proper return to the server connection page.
- Enhanced RouterNotifier to allow users to stay on the server connection page if authenticated, improving navigation flow.
- Modified AuthenticationPage to navigate back to the server connection page instead of popping the navigation stack, enhancing user experience during server setup.
