Athom build & publish lessons
- Socket hang up: the upload succeeds; Athom-side processing fails later. Failure rate grows with manifest (app.json) size; parallel uploads and file count are not the cause. Retry policy v2: socket hang up 4 retries, other transient 2, AggregateError 2 then bisect, unknown none. Stop waiting once Athom marks a build failed.
- AggregateError / processing_failed: found by bisecting commits between the last good and first bad build, then fixed and guarded in CI.
- Size: keep research data out of the bundle via .homeyignore (backups/, tmp/, data/kb/, data/leads/).
Detailed notes:
- GitHub Actions — mesure et optimisation (10/10/2026)
- Athom AggregateError on master — bisect result (2026-10-10)
- Athom / Homey publish limits — what we know (2026-10-10, Paris)
- New driver ids rejected by Athom (AggregateError) — offline analysis 2026-10-11 (Paris)
- Athom "processing failed": diagnosis (20/10/2026, P2902)
- Athom
processing_failed | socket hang up— evidence, conclusion, fixes (2026-10-19) - Athom "socket hang up" / processing_failed / AggregateError — root-cause study (P2930, 21/10/2026 Paris)
- Autonomy map — assistant routines vs GitHub Actions (P3097, 2026-10-31)
- Disabled workflows (P2916, 20/10/2026)
- GitHub Actions hardening P2900 (20/10/2026, Paris time)
- GHA_REPORTS_INVENTORY.md
- Gledopto GL-SPI-206P (
led_controller_spi_tuya) — runtime fix + Athom status (2026-10-10) - P2924 — "socket hang up": where it happens and what we changed (21/10/2026)
- P3036: Athom publish policy (green CI, spacing, one retry, bisect exit)
- led_controller_spi_tuya — Athom AggregateError bisect (19/10/2026)