Looking at Khalid's logs, I thought problem with a choice I made. Something I didn't do with the version code is prevent checking for extra focusers, in favor of detection with fast timeouts. If we do gate that, should it be with version == OnStepX or OnStepX & Unknown/Undetected? (Noting that doing the gating would probably break whatever compatability is left with TeenAstro, which I don't test against.)
However, That almost looks like a loss of communication in there on initialization? (It calls for a single char response, which times out (TTY_TIME_OUT = -4) and then fails to see the version (where it could before)
This admittedly isn't that well handled at that point, but that seems like an edge case: It connects then while initializing the connection drops? Aside from making sure everything is very careful, (eliminating some things which make the code a lot more readable, like sendOnStepCommand (which returns a bool, so doesn't give indication of errors, which is what most of the focuser/rotator detection is called with, Actually... I take that back it tries to, but returns error_type (I'd assume -4, timeout again) to a boolean, which would always return true. D'oh.) So we should probably go for singleCharResponse and check for <0, and not use/replace all occurrences of sendOnStepCommand. Still a rather limited case. But I'm 99% certain now it's a loss of communication.
(I've worked some on the checksummed commands, but that's more complicated than it should be and probably will end up with OnStep further drifting from the base lx200 (or modifying it and making it nasty for everyone else). Branch here if curious:
github.com/james-lan/indi/tree/onstep-2021-checksum-commands )
---
Unfortunately, I was a little confused by Alain's comment about focusers showing up in the panel, if it was in reference to his own (10.04e) log (showing 1 focuser detected, which should bring it up) or Khalid's showing 0 then 2, after what I think is the communication error (which would also on the 2nd detection, should also be brought up.) Is it detecting a focuser that's not there in that log? They shouldn't be showing up for boards without focusers based on someone else's change as I recall? ( And while typing this and looking the code had an aha moment: Though I'm not seeing it, so maybe I clobbered it by accident?)
Actually, I think I did, FI::updateProperties is up top of updateProperties, vs conditional like RI (rotator) is, when it really shouldn't be. (Not a huge deal, it just shows, and that's a regression, but not a huge one.) I or someone else can push that into the appropriate places. (I'll be a bit.)
Jasem: I don't think either of these really warrant a 0.93.1, unless that's already being prepared? Anyone (esp Alain, Khalid) disagree? If anyone does, I'll try extra to make sure a fix/merge request is done/tested tonight, unless someone else is faster. (That'd be ~12-16ish hours from now.)
