a3a4c97ec5
Addresses the review findings on the per-turn PR attribution and adds the model and task-category surfaces. Correctness: - Cache migration: the 5 -> 6 session-cache bump now adopts the prior v5 file's expired-source PR entries instead of abandoning them, and the claude scan preserves and surfaces PR-bearing orphans, so a session whose transcript was deleted still appears as a legacy even-split instead of vanishing. The daily cache is untouched. - Date-range carry: the parser captures the PR set active at the start of the in-range turn slice and seeds the state machine with it, so a PR referenced before the window still owns its later, in-range, ref-less turns instead of the session falling back to a whole-session approx split. - Calls are split across a multi-PR turn by largest-remainder, keeping per-PR counts whole (a 1-call, 2-PR turn no longer renders as 2 calls). - The CLI and app footers reconcile to the rounded row values actually shown. - Distinct sessions are keyed by project + sessionId, not sessionId alone. - The app tolerates an older by-reference payload (no attributedCost): it keeps the old non-summable footer and never renders NaN. Features: - Each PR contribution records the models of its calls and the turn's task category (split by the same share on a multi-PR turn); legacy even-split rows carry the model union but no category breakdown. - Payload rows gain models (short names, cost-desc) and categories (label + cost, cost-desc, omitted when empty); the payload gains otherPrCount/otherPrCost for the PRs beyond the sent top 20. - CLI --by-pr gains a Models column. The desktop table is now full-width with a Models column and click-to-expand rows showing a per-category cost breakdown with proportional bars, keyboard accessible, with an "Other (N more PRs)" row when capped. Tests: state-machine seed/models/categories/largest-remainder/dedup, the prRefs round-trip through the real incremental-append path (continuation + straddle), v5 adoption of an expired PR session, payload round-trip, and the desktop expansion/models/old-payload cases. |
||
|---|---|---|
| .. | ||
| build | ||
| electron | ||
| renderer | ||
| scripts | ||
| .gitignore | ||
| DISTRIBUTION.md | ||
| package-lock.json | ||
| package.json | ||
| README.md | ||
| tsconfig.electron.json | ||
| tsconfig.json | ||
| vite.config.ts | ||
| vitest.config.ts | ||
CodeBurn Desktop
Electron desktop shell for CodeBurn's local-first usage views. M1 runs as a developer app and reads data by spawning the installed codeburn CLI; it does not run a daemon or HTTP server.
Development
npm --prefix app install
npm --prefix app run dev
Validation:
npm --prefix app run test
npm --prefix app run typecheck
CLI Dependency
Packaged builds ship their own version-matched copy of the codeburn CLI and require nothing installed — the app spawns the bundled CLI with Electron's own binary as Node (ELECTRON_RUN_AS_NODE). In development (Vite dev server) the app uses the repo's freshly-built CLI, and either can be overridden with CODEBURN_BIN or a persisted path file. See DISTRIBUTION.md for the bundling and resolution details. Electron resolves and spawns the CLI from the main process, then sends decoded JSON through the secure preload bridge into the renderer.
This follows the menubar pattern:
contextIsolation: true,nodeIntegration: false, andsandbox: true.- Renderer code calls
window.codeburnonly throughapp/renderer/lib/ipc.ts. - Main process handlers return JSON envelopes so structured CLI errors survive IPC.
- Missing CLI, bad JSON, timeout, and nonzero exits are surfaced as honest UI states.
- The renderer never imports CodeBurn engine code from
src/; the data contract is spawn CLI, decode JSON, poll.
Data Contract
Current bridge calls:
- Overview:
codeburn status --format menubar-json --period <period> [--provider <provider>] - Plans:
codeburn status --format json --period <period> - Models:
codeburn models --format json --period <period> [--provider <provider>] [--by-task] - Optimize:
codeburn yield --format json --period <period> - Spend flow:
codeburn spend --format flow-json --period <period> [--provider <provider>] - Devices:
codeburn devices --format json --period <period> - Device scan:
codeburn devices scan --format json - Share status:
codeburn share status --format json - Identity:
codeburn identity --format json
Supported M1 periods are today, week, 30days, month, and all. Provider filtering is passed through where the CLI command supports it.
Sections
- Overview: daily spend, spend stats, waste summary, and expensive sessions from
menubar-json. - Spend: project/activity/tool/MCP/subagent lenses plus model-to-project flow.
- Optimize: waste findings from
menubar-jsonand reverted/abandoned yield data. - Models: model and task tables from
models --format json. - Plans: plan pacing from
status --format json. - Settings: device identity, nearby scan results, paired-device usage, and M2 visual affordances.
Packaging
npm run package produces an ad-hoc-signed macOS .dmg/.zip (arm64 and x64) via electron-builder, no paid Apple Developer account required. Packaging rebuilds the root CLI and bundles it into the app (Resources/cli), so installs need nothing on the target machine. See DISTRIBUTION.md for build instructions, the bundled-CLI mechanism, artifact locations, and the Gatekeeper first-open story.
M2 Backlog
- Add Electron
autoUpdater(the app already bundles its own version-matched CLI, so end-user installs need nothing on the machine; auto-update is the remaining piece). - Keep npm as a separate CLI-user channel at the same version as the desktop app.
- Add macOS code signing with a paid Developer ID and notarization (ad-hoc packaging exists today; see
DISTRIBUTION.md). - Add a
codeburn desktoplauncher subcommand. - Implement in-app pairing, approve, pull, and visibility mutations currently shown as M2 affordances.
- Build the Models Compare sheet.
- Add light theme support.
- Expand
codeburn optimize --format jsonwith evidence and fix commands so Optimize can show richer actionable fixes.