|
| 1 | +# Debug Builder Integration Plan |
| 2 | + |
| 3 | +## Overview |
| 4 | + |
| 5 | +Integrate the debugger directly into the Test Builder page instead of having a separate debug page. This allows users to: |
| 6 | +- See test execution state directly in the task list/graph they're familiar with |
| 7 | +- Modify, add, or remove future tasks while paused |
| 8 | +- Set breakpoints by clicking on tasks |
| 9 | +- View variables and frame state in context |
| 10 | + |
| 11 | +## Current Architecture |
| 12 | + |
| 13 | +``` |
| 14 | +BuilderPage |
| 15 | +├── BuilderToolbar (save, run, debug buttons) |
| 16 | +├── SplitPane |
| 17 | +│ ├── Left: TaskListEditor / TaskGraphEditor (switchable) |
| 18 | +│ └── Right: TaskConfigEditor (selected task config) |
| 19 | +└── TestRunModal (shows test execution logs) |
| 20 | +``` |
| 21 | + |
| 22 | +## Target Architecture |
| 23 | + |
| 24 | +``` |
| 25 | +BuilderPage |
| 26 | +├── BuilderToolbar (+ debug controls when debugging) |
| 27 | +├── DebugBar (frame selector, connection status - only when debugging) |
| 28 | +├── SplitPane |
| 29 | +│ ├── Left: TaskListEditor / TaskGraphEditor |
| 30 | +│ │ └── Enhanced with execution state indicators |
| 31 | +│ └── Right: TaskConfigEditor OR DebugInspector (contextual) |
| 32 | +└── DebugBottomPanel (variables, expression evaluator - collapsible) |
| 33 | +``` |
| 34 | + |
| 35 | +## Phase 1: Debug State in Builder |
| 36 | + |
| 37 | +### 1.1 Builder Debug Mode State |
| 38 | + |
| 39 | +Add debug mode state to the builder: |
| 40 | + |
| 41 | +```typescript |
| 42 | +// In BuilderPage or a new useBuilderDebug hook |
| 43 | +interface BuilderDebugState { |
| 44 | + isDebugging: boolean; |
| 45 | + runId: number | null; |
| 46 | + sessionId: string | null; |
| 47 | +} |
| 48 | +``` |
| 49 | + |
| 50 | +When "Debug" is clicked: |
| 51 | +1. Save test (existing) |
| 52 | +2. Schedule debug test (existing) |
| 53 | +3. **Stay on builder page** (change from current redirect) |
| 54 | +4. Connect to SSE events |
| 55 | +5. Enter debug mode UI |
| 56 | + |
| 57 | +### 1.2 Debug Controls in Toolbar |
| 58 | + |
| 59 | +When debugging, BuilderToolbar shows: |
| 60 | +- **Stop Debug** button (red) - aborts session, exits debug mode |
| 61 | +- **Continue** (F8) - resume execution |
| 62 | +- **Step Over** (F10) - next task at same depth |
| 63 | +- **Step Into** (F11) - step into nested tasks |
| 64 | +- **Step Out** (Shift+F11) - step out of current scope |
| 65 | +- **Skip** (F9) - skip current task |
| 66 | +- **Pause All** - pause all running frames |
| 67 | + |
| 68 | +Hide/disable normal toolbar items that don't apply during debug: |
| 69 | +- "Run Test" - hidden (already debugging) |
| 70 | +- "Debug" - hidden (already debugging) |
| 71 | +- "Save" - could remain enabled for live modifications |
| 72 | + |
| 73 | +### 1.3 Frame Selector Bar |
| 74 | + |
| 75 | +New component below toolbar (only visible when debugging): |
| 76 | + |
| 77 | +``` |
| 78 | +┌─────────────────────────────────────────────────────────────┐ |
| 79 | +│ [main ●] [background ○] [matrix[0] ●] [matrix[1] ○] [SSE ●] │ |
| 80 | +└─────────────────────────────────────────────────────────────┘ |
| 81 | +● = paused, ○ = running, ◌ = completed |
| 82 | +``` |
| 83 | + |
| 84 | +- Shows all execution frames as tabs |
| 85 | +- Click to switch active frame |
| 86 | +- Visual indicator for paused/running/completed |
| 87 | +- Connection status indicator |
| 88 | + |
| 89 | +## Phase 2: Task Execution Visualization |
| 90 | + |
| 91 | +### 2.1 Task List Execution State |
| 92 | + |
| 93 | +Enhance TaskListEditor to show execution state: |
| 94 | + |
| 95 | +``` |
| 96 | +┌──────────────────────────────────────────────────────────────┐ |
| 97 | +│ Tasks [+] │ |
| 98 | +├──────────────────────────────────────────────────────────────┤ |
| 99 | +│ ✓ 1. check_clients_are_healthy [success] │ |
| 100 | +│ ▶ 2. generate_transaction ← RUNNING │ |
| 101 | +│ ◉ 3. wait_for_inclusion ← BREAKPOINT (click to rm) │ |
| 102 | +│ 4. verify_transaction_receipt │ |
| 103 | +│ 5. cleanup_resources [cleanup] │ |
| 104 | +└──────────────────────────────────────────────────────────────┘ |
| 105 | +``` |
| 106 | + |
| 107 | +State indicators: |
| 108 | +- `✓` Green check - completed successfully |
| 109 | +- `✗` Red X - failed |
| 110 | +- `▶` Yellow arrow - currently running |
| 111 | +- `⏸` Blue pause - paused (hit breakpoint or step) |
| 112 | +- `◉` Red dot - has breakpoint |
| 113 | +- `○` Gray circle - pending/not executed |
| 114 | +- `⊘` Gray slash - skipped |
| 115 | + |
| 116 | +### 2.2 Task Graph Execution State |
| 117 | + |
| 118 | +Enhance TaskGraphEditor nodes: |
| 119 | + |
| 120 | +- **Completed nodes**: Green border, slightly faded |
| 121 | +- **Running node**: Yellow/amber border, pulsing glow |
| 122 | +- **Paused node**: Blue border, highlighted |
| 123 | +- **Failed node**: Red border |
| 124 | +- **Pending nodes**: Gray/default |
| 125 | +- **Breakpoint indicator**: Red dot in corner |
| 126 | + |
| 127 | +Add execution flow visualization: |
| 128 | +- Animate edge when task transitions |
| 129 | +- Show execution path taken so far |
| 130 | + |
| 131 | +### 2.3 Breakpoint Management |
| 132 | + |
| 133 | +Click on task to toggle breakpoint: |
| 134 | +- In list: Click on row left margin or dedicated column |
| 135 | +- In graph: Click on node corner or right-click menu |
| 136 | + |
| 137 | +Breakpoint types from UI: |
| 138 | +- **Task breakpoint** (click on task) - pause before this task |
| 139 | +- **Conditional breakpoint** (right-click > Add conditional) - pause if condition met |
| 140 | + |
| 141 | +Show breakpoints inline rather than in separate panel. |
| 142 | + |
| 143 | +## Phase 3: Debug Inspector Panel |
| 144 | + |
| 145 | +### 3.1 Contextual Right Panel |
| 146 | + |
| 147 | +When debugging and a task is paused, the right panel switches from TaskConfigEditor to DebugInspector: |
| 148 | + |
| 149 | +``` |
| 150 | +┌─────────────────────────────────────────────┐ |
| 151 | +│ Task: generate_transaction │ |
| 152 | +│ Status: Paused (breakpoint) │ |
| 153 | +├─────────────────────────────────────────────┤ |
| 154 | +│ [Variables] [Config] [Output] │ |
| 155 | +├─────────────────────────────────────────────┤ |
| 156 | +│ Variables: │ |
| 157 | +│ ├─ walletPrivKey: "0x1234..." │ |
| 158 | +│ ├─ targetAddr: "0xabcd..." │ |
| 159 | +│ └─ clients: [{...}, {...}] │ |
| 160 | +│ │ |
| 161 | +│ Config (resolved): │ |
| 162 | +│ ├─ privateKey: $.walletPrivKey │ |
| 163 | +│ │ → "0x1234..." │ |
| 164 | +│ └─ toAddress: $.targetAddr │ |
| 165 | +│ → "0xabcd..." │ |
| 166 | +└─────────────────────────────────────────────┘ |
| 167 | +``` |
| 168 | + |
| 169 | +Tabs: |
| 170 | +- **Variables**: Current scope variables (editable) |
| 171 | +- **Config**: Task config with resolved values shown |
| 172 | +- **Output**: Task logs/output (if running or completed) |
| 173 | + |
| 174 | +### 3.2 Bottom Panel (Collapsible) |
| 175 | + |
| 176 | +New collapsible panel at bottom for advanced debug features: |
| 177 | + |
| 178 | +``` |
| 179 | +┌─────────────────────────────────────────────────────────────┐ |
| 180 | +│ Debug Console [▼ Collapse]│ |
| 181 | +├─────────────────────────────────────────────────────────────┤ |
| 182 | +│ > $.clients | length │ |
| 183 | +│ 6 │ |
| 184 | +│ > $.walletPrivKey │ |
| 185 | +│ "0x1234..." │ |
| 186 | +│ > _ │ |
| 187 | +└─────────────────────────────────────────────────────────────┘ |
| 188 | +``` |
| 189 | + |
| 190 | +Features: |
| 191 | +- JQ expression evaluator (REPL) |
| 192 | +- Watch expressions |
| 193 | +- Variable diff view (what changed since last pause) |
| 194 | + |
| 195 | +Default state: collapsed (just a thin bar showing "Debug Console") |
| 196 | + |
| 197 | +## Phase 4: Live Task Modification |
| 198 | + |
| 199 | +### 4.1 Edit Future Tasks |
| 200 | + |
| 201 | +While paused, allow editing tasks that haven't executed yet: |
| 202 | +- Click on pending task to edit its config |
| 203 | +- Changes are staged and applied when execution reaches that task |
| 204 | +- Visual indicator for modified tasks |
| 205 | + |
| 206 | +### 4.2 Add/Remove Tasks |
| 207 | + |
| 208 | +While paused: |
| 209 | +- **Add task**: Insert new task at any position after current |
| 210 | +- **Remove task**: Remove pending task from execution queue |
| 211 | +- **Reorder**: Drag pending tasks to reorder |
| 212 | + |
| 213 | +Constraints: |
| 214 | +- Cannot modify completed tasks |
| 215 | +- Cannot remove currently paused task |
| 216 | +- Changes tracked as "pending modifications" |
| 217 | + |
| 218 | +### 4.3 Task Injection |
| 219 | + |
| 220 | +Quick inject common debug tasks: |
| 221 | +- "Log Variables" - inject a task that logs current variables |
| 222 | +- "Wait" - inject a sleep task |
| 223 | +- "Checkpoint" - inject a manual pause point |
| 224 | + |
| 225 | +## Phase 5: Multi-Frame Support |
| 226 | + |
| 227 | +### 5.1 Frame-Aware Task Display |
| 228 | + |
| 229 | +When test has multiple frames (concurrent, background, matrix): |
| 230 | + |
| 231 | +Option A: **Merged view with frame labels** |
| 232 | +``` |
| 233 | +│ [main] 1. setup_task ✓ │ |
| 234 | +│ [main] 2. run_background ✓ │ |
| 235 | +│ [bg] 2.1. background_monitor ▶ │ |
| 236 | +│ [main] 3. do_work ⏸ │ |
| 237 | +│ [bg] 2.2. check_status ○ │ |
| 238 | +``` |
| 239 | + |
| 240 | +Option B: **Frame tabs with separate task lists** |
| 241 | +``` |
| 242 | +[main ⏸] [background ▶] |
| 243 | +
|
| 244 | +│ 1. setup_task ✓ │ |
| 245 | +│ 2. run_background ✓ │ |
| 246 | +│ 3. do_work ⏸ │ ← current |
| 247 | +│ 4. cleanup ○ │ |
| 248 | +``` |
| 249 | + |
| 250 | +Recommendation: Option B for clarity, with visual connection indicators in graph view. |
| 251 | + |
| 252 | +### 5.2 Frame-Specific Actions |
| 253 | + |
| 254 | +Debug controls apply to active frame: |
| 255 | +- Continue/Step applies to selected frame |
| 256 | +- "Continue All" button for resuming all paused frames |
| 257 | +- Frame selector shows which frames are paused |
| 258 | + |
| 259 | +## Implementation Order |
| 260 | + |
| 261 | +### Sprint 1: Core Debug Mode (1-2 days) |
| 262 | +1. Add debug mode state to BuilderPage |
| 263 | +2. Keep user on builder page after starting debug |
| 264 | +3. Add debug controls to BuilderToolbar |
| 265 | +4. Add basic frame selector bar |
| 266 | +5. Connect to SSE debug events |
| 267 | + |
| 268 | +### Sprint 2: Execution Visualization (1-2 days) |
| 269 | +1. Add execution state to TaskListEditor |
| 270 | +2. Add execution state to TaskGraphEditor (basic) |
| 271 | +3. Breakpoint toggle on task click |
| 272 | +4. Breakpoint indicators in list/graph |
| 273 | + |
| 274 | +### Sprint 3: Debug Inspector (1-2 days) |
| 275 | +1. Create DebugInspector component |
| 276 | +2. Variables tab with tree view |
| 277 | +3. Config tab with resolved values |
| 278 | +4. Switch right panel contextually |
| 279 | + |
| 280 | +### Sprint 4: Bottom Panel & Polish (1 day) |
| 281 | +1. Collapsible debug console |
| 282 | +2. JQ expression evaluator |
| 283 | +3. Keyboard shortcuts |
| 284 | +4. Connection status handling |
| 285 | + |
| 286 | +### Sprint 5: Live Modification (2-3 days) |
| 287 | +1. Edit future task configs |
| 288 | +2. Add/remove pending tasks |
| 289 | +3. Task injection shortcuts |
| 290 | +4. Pending modifications tracking |
| 291 | + |
| 292 | +### Sprint 6: Multi-Frame (1-2 days) |
| 293 | +1. Frame tabs in task list |
| 294 | +2. Frame visualization in graph |
| 295 | +3. Frame-specific controls |
| 296 | + |
| 297 | +### Sprint 7: Remove Standalone Debug Page (0.5 day) |
| 298 | +1. Remove `/debug/:runId` route from App.tsx |
| 299 | +2. Delete `pages/TestDebugger.tsx` |
| 300 | +3. Remove standalone debug components no longer needed: |
| 301 | + - `components/debug/ShortcutsHelp.tsx` (integrate into builder help) |
| 302 | + - Any other components only used by TestDebugger |
| 303 | +4. Update BuilderToolbar to no longer redirect |
| 304 | +5. Remove lazy import for TestDebugger |
| 305 | +6. Clean up unused exports from `components/debug/index.ts` |
| 306 | +7. Update any documentation/links referencing `/debug/` |
| 307 | + |
| 308 | +## Component Changes Summary |
| 309 | + |
| 310 | +| Component | Changes | |
| 311 | +|-----------|---------| |
| 312 | +| `BuilderPage` | Add debug mode state, bottom panel | |
| 313 | +| `BuilderToolbar` | Add debug controls, hide/show based on mode | |
| 314 | +| `TaskListEditor` | Add execution state, breakpoint indicators | |
| 315 | +| `TaskGraphEditor` | Add execution state, visual enhancements | |
| 316 | +| `TaskConfigEditor` | Add "pending change" mode | |
| 317 | +| `DebugBar` (new) | Frame selector, connection status | |
| 318 | +| `DebugInspector` (new) | Variables, config, output tabs | |
| 319 | +| `DebugConsole` (new) | Bottom panel with REPL | |
| 320 | + |
| 321 | +## Hooks/Stores |
| 322 | + |
| 323 | +Reuse existing: |
| 324 | +- `useDebugStore` - debug session state |
| 325 | +- `useDebugEvents` - SSE connection |
| 326 | +- `useDebugApi` - API calls |
| 327 | +- `useDebugKeyboard` - keyboard shortcuts |
| 328 | + |
| 329 | +New: |
| 330 | +- `useBuilderDebug` - builder-specific debug state and actions |
| 331 | + |
| 332 | +## Migration from Separate Debug Page |
| 333 | + |
| 334 | +1. During development (Sprints 1-6): Keep `/debug/:runId` route for testing |
| 335 | +2. Default flow changes: debug from builder stays in builder |
| 336 | +3. After Sprint 6 verification: Remove standalone debug page entirely (Sprint 7) |
| 337 | +4. No "Open in Debug Page" option - all functionality in builder |
| 338 | + |
| 339 | +## Open Questions |
| 340 | + |
| 341 | +1. **Graph complexity**: How to show execution in complex nested/concurrent graphs? |
| 342 | +2. **Performance**: Real-time updates for many tasks - throttle/batch? |
| 343 | +3. **State persistence**: Save breakpoints with test? Session storage? |
| 344 | +4. **Mobile/small screens**: How to handle debug UI on narrow viewports? |
| 345 | + |
| 346 | +## Success Criteria |
| 347 | + |
| 348 | +- [ ] User can start debug session and stay on builder page |
| 349 | +- [ ] User can see which task is running/paused in task list |
| 350 | +- [ ] User can set breakpoints by clicking on tasks |
| 351 | +- [ ] User can see variables when paused |
| 352 | +- [ ] User can step through execution with keyboard shortcuts |
| 353 | +- [ ] User can modify future tasks while paused |
| 354 | +- [ ] Multi-frame tests show correct execution state |
| 355 | +- [ ] Standalone `/debug/` page removed - all debug features in builder |
| 356 | +- [ ] No dead code or unused debug components remain |
0 commit comments