dangernoodle-io / breadboard / 29388725544

96%

push

github

refactor: relocate factory-reset to /api/diag/factory-reset on bb_storage_http (B1-960) (#878)

- new bb_storage_vtable_t.erase_all (optional, NULL -> BB_ERR_UNSUPPORTED,
  fail-closed exactly like erase_namespace) + bb_storage_erase_all(backend)
  facade fn -- broader than erase_namespace: wipes a whole backend/partition
  rather than one ns_or_dir, since factory-reset is a whole-NVS-partition
  erase, not a namespace-scoped one
- bb_storage_nvs implements erase_all with the nvs_flash_erase() call moved
  VERBATIM from bb_nv_config_factory_reset() (whole-partition blast radius
  preserved unchanged -- deliberately NOT narrowed to a per-namespace loop,
  which would miss namespaces and change behavior); left NULL for ram/rtc
- bb_storage_http gains POST /api/diag/factory-reset (own bbtool:init
  marker, 1:1 with the route) alongside the existing DELETE
  /api/diag/storage; request schema {"confirm":"factory-reset"} preserved
  exactly, RTC-mirror-clear + reboot-record-save + deferred esp_restart()
  moved verbatim from the old handler
- delete platform/espidf/bb_nv/bb_nv_routes.c, bb_nv_config_factory_reset(),
  and the CONFIG_BB_NV_FACTORY_RESET Kconfig option/route-only deps from
  bb_nv -- grep-confirmed no callers of bb_nv_config_factory_reset() besides
  the deleted route and its own tests
- fan-out caveat: this renames a PUBLIC bbtool:init marker
  (bb_nv_factory_reset_routes_init -> bb_storage_http_factory_reset_routes_init).
  Any downstream repo composing bb_nv for this route will get a
  MissingProviderError on codegen regen until it switches to bb_storage_http


Claude-Session: https://claude.ai/code/session_01V5JXJNaYWPwjrcS8nk183m

Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>

6972 of 7603 branches covered (91.7%)

Branch coverage included in aggregate %.

9 of 9 new or added lines in 1 file covered. (100.0%)

11686 of 11938 relevant lines covered (97.89%)

2427.13 hits per line