[maxidragon]

Problem

reset_solve_state clears the station's own solve state, so the screen correctly goes back to waiting for a competitor. It doesn't clear the globals that the once-a-second live state packet reads (INSPECTION_TIME, CURRENT_COMPETITOR, CURRENT_GROUP_ID, CURRENT_SESSION_ID). After a solve is submitted, cancelled or rejected, an empty station keeps reporting the previous competitor and their inspection time, and the live solving view shows a stale competitor.

Only hard_state_reset cleared them.

Change

Clear those globals in reset_solve_state, the same way hard_state_reset does.

Every caller means "this competitor is done or cancelled": solve submitted, card rejected or error with reset, the reset button, a delegate decision, and the e2e reset. Nothing reads these globals except the live state packet.

CURRENT_TIME is left as is. The stackmat loop overwrites it on every frame, and inspection_start only starts inspection when it reads 0.

CI fix

The toolchain is unpinned nightly, and a newer clippy added the needless_bool lint, which failed CI under -D warnings in should_skip_other_actions and the logger's enabled. Both now return the condition directly. Behaviour is unchanged. master hits the same failure until this is merged.

Testing

• CI passes: fmt, plus build and clippy for v3 and v4.
• Tested on a v4 station with a dev build of this branch (v4,sleep): once a solve is done, the live state no longer reports the previous competitor or their inspection time.

[maxidragon]

Closing in favour of #34. On the refresh branch the live state packet reads directly from the solve state, and reset_solve replaces the whole solve state and clears the timer value, so an empty station no longer reports the previous competitor. #34's CI also passes on the current nightly without the clippy changes here.