[ClutchplateDude]

Summary

Two fixes for guide-pulse handling, from a user report about guiding clients (not board-specific):

1. Guide pulses arriving during an active pulse were silently dropped

V1.13.14 (#261) added an isSlewingTRK() guard to guidePulse() to refuse pulses when the mount is not tracking (e.g. after the RA tracking limit auto-stop). But slewStatus() reports NOT_SLEWING for the entire duration of any active guide pulse, so the guard also rejected every pulse that arrived while another one was still running — on either axis.

• Clients that wait for pulse completion before sending the next one (PHD2 via the ASCOM driver polling :GIG#) never overlap pulses, so they are unaffected.
• Blind senders (INDI/EKOS) fire pulses back-to-back: with typical 500–3000 ms pulses on a 1–2 s cadence, most corrections were dropped and guiding was effectively dead. Both :MG…# replies are empty, so the client cannot detect the drop.

Fix: test the STATUS_TRACKING status flag directly via a new isTracking() accessor. The intended refusal (limit reached, tracking off, parked) is preserved; overlapping pulses get the pre-V1.13.14 last-one-wins semantics again.

2. Info display renders during guide pulses delayed the pulse end

V1.13.7 (#248) added an updateInfoDisplay() call to the else-branch of the guiding block in loop(). A full SSD1306 frame (~1 KB over the default 100 kHz I2C — no Wire.setClock() is configured) blocks loop() for roughly 100 ms, and guiding also bypasses the slew-based 5 Hz refresh throttle (slewStatus() is masked while guiding, so the idle 30 Hz rate applies). The pulse end-time check only runs at the top of each loop() iteration, so every pulse could overshoot its requested duration by up to one render (~100 ms worst case) — real extra guide displacement, since the stepper ISR keeps stepping during the render.

Fix: skip the render while a pulse is running. The display is not left stale — guidePulse() renders once at each pulse start, and the normal end-of-loop() refresh covers the gaps between pulses.

Testing

• pio run -e ramps green with both changes (-Werror)
• esp32 fails locally on both pristine develop and this branch — pre-existing local-config issue (Configuration_local_esp32dev.hpp), not from these changes; CI builds clean without local configs
• Unit tests don't compile Mount.cpp (native env builds core/ports/adapters only); CI covers it

Follow-up idea (not in this PR): Wire.setClock(400000) would cut render time ~4x for all display paths.

🤖 Generated with Claude Code