* Reduced the button visibility timer from 1500ms to 500ms for quicker feedback.
* Decreased the trigger width from 80 to 60 for improved interaction.
* Updated the height calculation from 40% to 30% of the window height for better layout consistency.
* feat: code tools, editor, executor
CodeEditor & Preview
- CodeEditor: CodeMirror 6
- Switch to CodeEditor in the settings
- Support edit&save with a accurate diff&lookup strategy
- Use CodeEditor for editing MCP json configuration
- CodePreview: Original Shiki syntax highlighting
- Implemented using a custom Shiki stream tokenizer
- Remov code caching as it is incompatible with the current streaming code highlighting
- Add a webworker for shiki
- Other preview components
- Merge MermaidPopup and Mermaid to MermaidPreview, use local mermaidjs
- Show mermaid syntax error message on demand
- Rename PlantUML to PlantUmlPreview
- Rename SyntaxHighlighterProvider to CodeStyleProvider for clarity
- Both light and dark themes are preserved for convenience
CodeToolbar
- Top sticky toolbar provides quick tools (left) and core tools (right)
- Quick tools are hidden under the `More` button to avoid clutter, while core tools are always visible
- View&edit mode
- Allow switching between preview and edit modes
- Add a split view
Code execution
- Pyodide for executing Python scripts
- Add a webworker for Pyodide
* fix: migrate version and lint error
* refactor: use constants for defining tool specs
* refactor: add user-select, fix tool specs
* refactor: simplify some state changing
* fix: make sure editor tools registered after the editor is ready
---------
Co-authored-by: 自由的世界人 <3196812536@qq.com>
* upgrade electron to 35.2.0
* update electron to 32.3.3
* udpate electron-vite to 3.0.0
* upgrade to 35.2.2
* patch https://github.com/electron-userland/electron-builder/pull/9046
* add minimumSystemVersion config
* update electron vite
---------
Co-authored-by: beyondkmp <beyondkmkp@gmail.com>
* Updated padding in CustomCollapse for improved spacing.
* Modified Messages component to conditionally apply padding based on prompt visibility.
* Enhanced ModelListSearchBar by adding a flex display style for better alignment.
* Increased gap in ProviderSetting's HStack for improved layout consistency.
* Introduced new translations for "Tool Use Mode" and its options ("Function" and "Prompt") in English, Japanese, Russian, Simplified Chinese, and Traditional Chinese.
* Refactored settings components to replace the deprecated `enableToolUse` with `toolUseMode`, updating related logic and UI elements accordingly.
* Adjusted migration logic to ensure backward compatibility with previous settings.
* Changed the wording for max tokens settings in English, Japanese, Russian, Simplified Chinese, and Traditional Chinese to improve clarity.
* Adjusted confirmation messages to better reflect the functionality of setting maximum token limits.
* Removed EXTENDED_CONTEXT_LIMIT and EXTENDED_CONTEXT_STEP constants from the configuration.
* Eliminated enableMaxContexts state and associated logic from settings components.
* Updated contextCount handling to accommodate a maximum of 100, with a special case for 100 to represent an unlimited context.
* Cleaned up related translations for max contexts in multiple languages.
* Deleted the @cherry-studio/artifacts package including its README, package.json, and associated CSS files.
* Removed the @cherry-studio/database package along with its README, package.json, and source files for data handling.
* Cleaned up related yarn.lock files and installation states for both packages.
* fix: update current topic id for ChatFlowHistory to work
* refactor: set current topic id early in loadTopicMessagesThunk
* refactor: extract EmojiAvatar
* fix: style
* feat: Extended the fillable context length to 100. #5514
Update src/renderer/src/pages/home/Tabs/SettingsTab.tsx
fix: update context length settings and localization for multiple languages
* fix: adjust context count slider limits and behavior based on max contexts setting
* feat: implement event handling for max contexts changes in settings
* feat: update context count handling to use EXTENDED_CONTEXT_LIMIT in settings
* feat: update context count logic to use EXTENDED_CONTEXT_LIMIT when max contexts are enabled
* refactor: remove unused EXTENDED_CONTEXT_LIMIT import from AssistantService
* feat: simplify context count logic with validAndChangeContextCount function
* feat: 移除TokenCount组件中对最大计数为20的特殊处理，简化上下文计数显示逻辑
* refactor: simplify getContextCount logic by removing max context handling
* feat: adjust step value for context count input based on max contexts setting
---------
Co-authored-by: George Zhao <georgezhao@SKJLAB>
* fix: quickpanel scrollto changed to smart
* fix: add scrollTrigger as the replacement for scrollBlock
* fix: add a 'none' trigger to prevent accidental scrolling
* Updated the reset logic to conditionally handle model and modelId for grouped messages.
* Ensured that the original model is retained when regenerating responses for grouped messages.
* Updated the reset logic to conditionally handle model and modelId for grouped messages.
* Ensured that the original model is retained when regenerating responses for grouped messages.
