docs(perf): LOOP.md — warm-deploy agent loop spec#113
Conversation
Adds `LOOP.md` at the repo root: a spec for an agent-driven loop that benchmarks the warm rebuild + verify-skip + monitor-reconnect path against a 4 000 ms end-to-end budget, with a phase budget tree, five numbered test cases (T1–T5 + a T0 cold baseline), one-iteration procedure, decision rules for each failure mode, and cadence guidance for the `/loop` skill. Scope: - Targets the ESP32-S3 DevKit path because it is the largest warm-build offender documented in `docs/PERF_WARM_BUILD.md`. - Relies on existing instrumentation: `FBUILD_PERF_LOG=1` phase lines, the daemon's "verify-flash: … skipping write" log from `handlers::operations`, the `SharedSerialManager` retry schedule, and the `auto_reconnect` behaviour of the Python `SerialMonitor`. - No code changes — purely a runnable spec to drive follow-up optimization work. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
Warning Rate limit exceeded
Your organization is not enrolled in usage-based pricing. Contact your admin to enable usage-based pricing to continue reviews beyond the rate limit, or try again in 14 minutes and 13 seconds. ⌛ How to resolve this issue?After the wait time has elapsed, a review can be triggered using the We recommend that you space out your commits to avoid hitting the rate limit. 🚦 How do rate limits work?CodeRabbit enforces hourly rate limits for each developer per organization. Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout. Please see our FAQ for further information. ✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
LOOP.md is a per-session operator benchmark spec, not a stable doc. Committing it to main in #113 was a mistake — remove it from the tree and add `/LOOP.md` + `tasks/loop-runs/` to `.gitignore` so the file can still live in local working copies without being pushed again. Reverts the tracking of the file added in cacba6f (#113). The issue it was referenced from (#114) points at the file path as a local artefact, not a repo-tracked doc. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
LOOP.md is a per-session operator benchmark spec, not a stable doc. Committing it to main in #113 was a mistake — remove it from the tree and add `/LOOP.md` + `tasks/loop-runs/` to `.gitignore` so the file can still live in local working copies without being pushed again. Reverts the tracking of the file added in cacba6f (#113). The issue it was referenced from (#114) points at the file path as a local artefact, not a repo-tracked doc. Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
1 participant
Summary
LOOP.mdat the repo root: an agent-loop spec targeting the warm rebuild + verify-skip + monitor-reconnect path with a 4 000 ms end-to-end budget./loopcadence guidance, and concrete references todocs/PERF_WARM_BUILD.md,perf_log.rs,SharedSerialManager, and the PythonSerialMonitor.auto_reconnect.Test plan
FBUILD_PERF_LOG=1 FBUILD_USE_ESPFLASH_VERIFY=1 FBUILD_USE_ESPFLASH_WRITE=1against a real ESP32-S3 DevKit and confirm the T1/T2/T3 timings land inside or near the budgets on the first pass. Tune budget numbers in a follow-up if the real hardware disagrees./loop @./LOOP.mdinvocation drives a full iteration + append totasks/loop-runs/RESULTS.csv.🤖 Generated with Claude Code