hwci: retry flash+boot on the AMYboard's intermittent early-boot flake

[bwhitman]

Problem

The AMYboard HW CI intermittently fails with [boot] board MIDI never appeared (didn't boot?). The board panic-loops in the ESP-IDF 2nd-stage bootloader (Guru Meditation StoreProhibited, EXCVADDR 0xb0, registers full of 0xaaaaaaaa) right after entry 0x403c8890, before USB-MIDI enumerates — so a single transient glitch fails the whole run.

This is a bench flake, not a firmware bug (proven)

Investigated under shorepine/amy#748. The same amy SHA built via the amyboard-amypr-<N> (dispatch) path booted + passed, while the amyboard-pr-<N> (pull_request) build of that identical SHA boot-looped. esptool image_info + cmp of the two full images over the bootloader region shows the bootloader machine code is byte-identical — the only differences are 4 bytes of the embedded __TIME__ build-time string (the two compiled at different wall-clock times) and the derived checksum/SHA-256. Identical code can't be a deterministic code bug, so it's a transient flash-read / power / reset glitch after the ~61 s USB flash.

Fix

Wrap flash+boot in a retry loop (--boot-attempts, default 3):

• On MIDI never appeared, stop the attempt's boot logger (freeing the dongle so esptool can drive the reset lines), re-flash — which forces ROM download mode, a clean known-good reset that clears the glitch — and wait again.
• A genuinely broken build still fails every attempt, so real regressions aren't masked.
• Each failed attempt's crash log is retained in the combined serial log (labeled [boot N/M]), so a persistent failure is still fully debuggable.
• Happy path is unchanged: boots first try → one flash, no extra time. CI passes no new flags (default 3).

wait_for_board() now returns None on timeout instead of sys.exit.

Testing

py_compile clean. Verified the loop with mocked flash/wait_for_board:

• all attempts fail → exactly 3 re-flashes then a clear exit;
• boot succeeds on attempt 2 → 2 flashes, then proceeds into the test.

🤖 Generated with Claude Code

[github-actions]

🔌 AMYboard PR preview

Editor + flasher: https://amyboard-pr-1029.vercel.app/editor/

This preview bundles this PR's firmware — its flasher only flashes this build (not the release). Rebuilt on every push; removed when the PR closes.

The hardware CI has been kicked off and should return within a few minutes, stand by!

[github-actions]

🎛️ HW CI (physical bench)

AMYboard (USB-MIDI + AMY zP → audio): ✅ PASS — flashed this PR’s firmware; all checks matched the references.

Tulip (TULIP4_R11; serial-REPL audio + WiFi screenshot): ✅ PASS — flashed this PR’s firmware; all checks matched the references.

⬇️ Artifacts: recordings · screenshot · serial logs · run logs

_{Self-hosted bench. Audio spectral-compared to ref/hwci_basic.wav + ref/tulip_basic.wav; Tulip screenshot pixel-compared to ref/tulip_screenshot.png. Both analog outs share one capture card, so the tests run sequentially.}