I have that covered here:
indilib.org/forum/mounts/2938-dome-telescope-lock-policies.html

So this has been broken for over 2 years now

afraid so.

As a late follow-up on this topic, I am about to complete my clamshell KStars/EKOS/INDI controlled observatory (based on Gonzothegreat), and wanted to test the appropriate mount and dome policies to make sure that no mount and dome collision would happen. I used the Dome Scripting Gateway driver.

As the mount unpark was working correctly when both mount and dome were parked, I only tested dome park when mount was not parked. I positionned the mount close but not at its park position such as if the dome was closed (parked) no collision would occur.

I tested the following 2 conditions:

1. Mount Locks - Dome Both

After issuing a dome park command, the dome parked first and the mount parked after the dome was closed. So, unless your mount is in a position that would avoid collision, there will likely be damage to mount and scope, as mentionned in Gonzo's case study.

2. Mount Locks - Dome Locks

After issuing a dome park command, the dome parked first and the mount did not park after the dome was closed, it was still tracking. as in the previous case, unless your mount is in a position that would avoid collision, there will likely be damage to mount and scope, but furthermore, as the mount is still tracking, damage will ultimately occur.

I understand the mount should explicitely be parked before the dome is but I assume the safest sequence should be that if a dome park command is issued, it should make sure that the mount is parked first before parking the dome, and if not, it should tell the mount driver to park the mount before it parks the dome. I realize the the python scripts may need to do the appropriate checking but I thought that INDI was taking care of that.

Hope this helps.

INDI drivers should take care of that. I figured someone would take a look at the code and propose a fix for this but this hasn't happened yet. I added this to my TODO list so we'll see.

Is the code related to parking policy in the specific mount and dome drivers (i.e. lx200_OnStep and Dome Scripting Gateway in my case) or somewhere else in the indiserver? Had a look in dome_script.cpp and lx200_OnStep.cpp but did not find anything obvious regarding parking policy.

No, it's in the parent classes INDI::Telescope and INDI::Dome

Ok, I think we need to come up with a matrix of all possible actions and what needs to be done between the dome/mount so that it can be coded safely to handle all possible conditions . Maybe a google sheet?

I think that if we want to come up with a matrix of all permitted combinations, we need to also specify a dome/roof type as, depending of the type, parking will either likely cause damage (roof, ROR or clamshell) or not (hemispherical dome, where there is no mount movement restriction) and as such parking policy combinations may be allowed or not (e.g. for a roof the combination "Dome Locks - Ignore Telescope" shall not be allowed). That also means that the DOME driver will need to have a DOME type option added. Does that make sense to you?

Well, that's possible as well. I have a roll-off but it's perfectly safe for the mount to be in any position. At any rate, once we have the matrix then it's possible to account for all possible permutations safely.

Sorry I did not see your message earlier. From your picture, I can see that clearance is not a problem for your mount/roof combination.

I can try to come up with a matrix of the different possibilities for domes and other types of mounts but as I may not be aware of all the subtleties, I may be missing a few or more cases. I hope to come up with something soon.

Hello Jasem, I started making videos of mount and roof exemplifying the different mount and roof/dome parking policies, as presented in Gonzo's case study. I will be able to prepare a matrix of the different cases as well as provide the videos if you see them fit. The expected behavior, from my knowledge, will be included in the matrix.