dangernoodle-io / breadboard / 30164363049

97%

push

github

test: extend live route-schema walk to bb_system reboot route (B1-1214) (#1079)

- remove bb_system.h's ESP_PLATFORM fence entirely around
  bb_system_routes_reserve/bb_system_routes_init: the fence guarded only
  bb_http_server.h's include plus those two declarations, all fully
  portable (opaque bb_http_handle_t, no platform-specific content), and
  their sole definition (bb_system_routes.c) is already host-compiled via
  the bbtool-scaffold-hint -- the fence only blocked the declaration, not
  the definition, so nothing was left fenced. Transitively exposing
  bb_http_server.h to host TUs (bb_system_reboot_record.c,
  bb_system_snap.c, bb_system_reboot_budget.c,
  bb_system_safeguard_reboot.c, bb_system_boot_fail.c,
  bb_system_boot_banner_format.c, platform/host/bb_system/bb_system_host.c,
  and several test_host files) for the first time is safe: components/
  bb_system/CMakeLists.txt already lists bb_http_server in REQUIRES
  unconditionally (not gated on CONFIG_BB_SYSTEM_ROUTES_AUTOREGISTER), so
  no new component dependency is created
- extend seed_live_routes() in test_route_schema_literals_live.c to call
  bb_system_routes_init(), retiring the now-stale comment explaining why
  it was deliberately skipped
- assert presence of POST /api/reboot in the walked registry, following
  PR1's presence-over-count pattern

Prefix is test: (not feat:) because the header edit only lifts a
declaration fence around code that was already compiled and reachable on
ESP-IDF -- the sole externally observable effect is new host-test
coverage over POST /api/reboot's request schema literal.

7668 of 8265 branches covered (92.78%)

Branch coverage included in aggregate %.

13116 of 13274 relevant lines covered (98.81%)

2588.48 hits per line