新增: 设置页 UX 重构(master-detail 左导航 + 搜索 + 即时反馈 + 导入导出)

- SettingsNav 左导航 7 类 + 顶部搜索(索引匹配 label/desc 高亮定位)
- GeneralPanel 拆分 Appearance/Performance/Advanced Section
- SettingRow 统一行 + 即时反馈(已保存)+ 下轮生效 badge
- 统一组件库 settings.css(scoped 重复 class 抽全局)
- 导入导出(默认排除 keyring 密钥)
- activeCategory 持久化 + 窄屏折叠 tab

src/i18n/en/settings.ts

@@ -3,6 +3,25 @@ export default {
     // ===== Page title =====
     title: '⚙️ Settings',
 
+    // ===== Left nav categories (phase 3 master-detail) =====
+    navAppearance: 'Appearance',
+    navAiModel: 'AI Models',
+    navKnowledge: 'Knowledge',
+    navPerformance: 'Performance',
+    navSecurity: 'Security',
+    navAdvanced: 'Advanced',
+
+    // ===== Settings search (phase 4) =====
+    searchPlaceholder: 'Search settings…',
+    searchNoResult: 'No results',
+
+    // ===== Instant feedback (phase 5 SettingRow) =====
+    savedHint: '✓ Saved',
+    nextRoundBadge: 'Next message',
+
+    // ===== Connection mgmt (phase 7: experimental badge, not wired to AI yet) =====
+    experimentalBadge: 'Experimental · not wired to AI tools',
+
     // ===== AI Provider panel =====
     panelAiProvider: '🤖 AI Model Config',
     addProvider: '+ Add Provider',
@@ -73,6 +92,10 @@ export default {
 
     // ===== General settings panel =====
     panelGeneral: '🎨 General',
+    // 阶段2 UX 重构:GeneralPanel 拆 3 Section 后各 Section 标题
+    sectionAppearance: '🎨 Appearance',
+    sectionPerformance: '⚡ Performance',
+    sectionAdvanced: '⚙️ Advanced',
     labelTheme: 'Theme',
     descTheme: 'Dark / light',
     themeDark: '🌙 Dark',
@@ -97,6 +120,9 @@ export default {
     labelShowTokenUsage: 'Show token usage',
     descShowTokenUsage: 'Show token consumption of each reply in conversations',
 
+    labelStreamingMd: 'Streaming Markdown rendering',
+    descStreamingMd: 'Render code blocks/lists/headings during streaming (unclosed code blocks degrade to plain text to avoid flicker); when off, only plain text is shown during streaming and full formatting applies after generation completes',
+
     labelGlobalConcurrency: 'Global max concurrency',
     descGlobalConcurrency: 'Shared cap on simultaneous LLM requests across all conversations',
 
@@ -104,7 +130,8 @@ export default {
     descPerConvConcurrency: 'Concurrent LLM calls within one conversation (main loop / title / distill shared)',
 
     labelAgentMaxIterations: 'Agent max iterations',
-    descAgentMaxIterations: 'Max rounds of the agentic loop (stream → tool → feedback) (1-50, default 10); changes take effect on the next message',
+    descAgentMaxIterations: 'Max rounds of the agentic loop (stream → tool → feedback) (0=unlimited, 1-50, default 10); changes take effect on the next message',
+    labelUnlimited: 'Unlimited',
 
     labelAgentMaxRetries: 'Stream failure retries',
     descAgentMaxRetries: 'Auto-retry count on stream failure (0-10, default 3); only retries pre-stream failures (no token output yet); mid-stream failures are abandoned',
@@ -131,12 +158,9 @@ export default {
     labelTriggerMode: 'Distill trigger',
     descTriggerMode: 'When auto-distillation is triggered',
     triggerOnComplete: 'On conversation complete',
-    triggerOnIdle: 'On conversation idle',
     triggerManualOnly: 'Manual only',
     labelMinMessages: 'Min message count',
     descMinMessages: 'Distillation triggers only after this many messages (filters out small talk noise)',
-    labelIdleTimeout: 'Idle timeout (s)',
-    descIdleTimeout: 'Seconds to wait in idle-trigger mode',
     labelAutoInject: 'Auto-inject related knowledge',
     descAutoInject: 'Retrieve related knowledge and inject it into the conversation context during chat',
     labelVectorEnabled: 'Semantic retrieval (vector)',
@@ -159,6 +183,18 @@ export default {
     toastSetDefaultFail: 'Failed to set default: {msg}',
     toastConnIncomplete: 'Please fill in Name and Host',
 
+    // ===== Import / Export (phase 6) =====
+    export: 'Export',
+    import: 'Import',
+    exportHint: 'Export all preferences (appSettings + knowledge config) as a JSON file',
+    importHint: 'Import from a JSON file and overwrite current preferences',
+    importConfirm: 'Importing will overwrite all current preferences (appearance / performance / knowledge, etc.). Continue?',
+    importSuccess: 'Imported successfully (some items like concurrency / iterations take effect after restart or next message)',
+    importFail: 'Import failed: {msg}',
+    exportFail: 'Export failed: {msg}',
+    importInvalidJson: 'The file is not a valid DevFlow settings backup (wrong format)',
+    importScopeNote: 'Scope: appSettings + knowledge config (no provider keys)',
+
     // ===== Confirm dialogs =====
     confirmDeleteProvider: 'Delete provider "{name}"? This cannot be undone.',
     confirmDeleteConn: 'Delete connection "{name}"?',