Continuous AP fixes and improvements

[vrs11]

Continuous AP fixes and improvements after proper testing

[hypery11]

Thanks for the follow-up work — reviewed the actual changes (diffed against your branch's merge-base so the #118 noise below is excluded) and they're good:

• ESP_driverBrakeApply fix (>=2 instead of !=0) — correct. 1 = Not_Applied was wrongly counting as "brake applied"; 2 = Driver_applying, 3 = Faulty/SNA. Treating >=2 as applied is right, and folding SNA(3) into "applied" fails the continuous-AP interlock closed, which is the safe direction. 👍
• Continuous-AP gear-sequence robustness (counter-freshness latch, send-result tracking, retry/delay split READY_WAIT/REENGAGE_DELAY/ATTEMPT_RESULT/RETRY_DELAY, longer GEAR_SEQUENCE_TIMEOUT) — sensible hardening for the 0x229 double-press path; nice that it's HW3/Legacy-scoped as before.

One blocker before I can merge: this branch predates #118 (merged earlier today), so as-is it reverts the #116 HW4-Highland byte0 AP-state fallback — the diff against current main drops the das_hw4_use_byte0 logic, the three fsd_state.h fields, the #116 host tests, and the changelog line. Could you rebase onto current main? The overlap is only in fsd_handle_das_status_hw4 (you simplified the byte0 block away because your base didn't have it) — after a rebase your changes apply cleanly alongside #116. Once rebased (host tests green + your on-car notes), I'll merge.

Happy to rebase it for you if you'd rather — just confirm and I'll take care of it. The work itself is solid; it just needs to sit on top of #118.

[vrs11]

Thanks for the review!
It's fixed

[hypery11]

Rebased cleanly — verified #116 is preserved, host tests green (265), FAP + ESP32 build. Merged. Thanks @vrs11, both the brake-state fix and the gear-sequence hardening are good improvements.