This must have happened during one of the most recent updates. I noticed it with a build from source 2 days ago and the bug is also present in the nightly version I updated today.
Basically, I am using the Luminance filter to plate solve and when I start my sequence afterwards, which starts with S-II, refocusing does not happen, although this is clearly selected for every filter in the focus module.

Has anyone else noticed that?

I then went back to a previous build from ~10 days ago and there the refocusing routine on filter change is working as it always has.

Jo

Hi Jo,

at least during the sequence it did work nicely for me using a build from Dec 14 (kstars-3.5.1-1773_gdd0d33f0b). Didn't have a filter change on start, so I'm not sure if it behaves different there. I'll do a new compile and have a look again tonight (if clouds and wind stay away...).

Thanks, Peter, I will be trying that commit as well then.

This continues to stump me. I have a build from approximately 3 weeks ago where refocusing on filter change works nicely as it should. But every single build since then fails to execute refocusing after filter change. And that is with completely fresh build folders, no updated ones.
The nightly install also fails to refocus.
I cannot figure out what would be causing this.
Does anyone have any idea?
All the settings are faithfully transferred between the different builds. Just refocusing, although set in the focuser module for every filter, is not being executed, except for the older build.

I just tested this and cannot reproduce any issues, it works reliably well. At least with simulators. How do your filter settings look like?

Focus module should also at least run once successfully before all of this, but I presume you already done that?

Jo, is this (only) for the first change, or have you verified it also doesn't focus on changes within a series? And I assume it's with the scheduler running? Does it also fail if you only run the sequence (which is what I usually do). Really weird that is....

Here are my focus module settings:



The thing is, I tested Hy's latest guide module change in which capture is delayed until RMS < [a user definable value]. I still have that build on my Pi, and THAT build works. It unfailingly will refocus on filter change. So it's not like I am out of business, I have a fully functional Kstars I can use to take pictures. However, I can no longer update it.
When I switch now to ANY other build, either the installed Nightly, or any other recent build I have installed completely independently on my SSD in a separate folder, the refocusing on filter change will no longer work.
All the settings are retained when I switch between builds, so it's not like the focus module filter settings are suddenly deselected. It is just that they are being ignored.
Since nobody else has that problem, I suspect that has something to do with me running that other build, which may have changed SOME file that is used by the installed Nightly as well as any of the independent builds from source.
Question is, which file is that and how can I change it, if that were indeed the problem.
Sorry, grasping for straws here. And I LOVE all the changes Hy has made to Kstars, it is fun testing all of that, so please no one should look at this as criticism, just simple troubleshooting. I can always go back to a backup, not a big deal. But that would beat the purpose.
Jo

Hooray, success (I hope)! I reinstalled Kstars and performed a full-upgrade. Now refocusing on filter change seems to work again in Nightly. At least the first change went smoothly as expected.
I am letting it run through the night, will recheck in the morning. Will also check whether this translates to independent builds as well.
I hope that settles it. This has been driving me crazy for the last 10 days or so.

After the purge, reinstallation of Nightly and full upgrade the problem seems to have disappeared. I don't know whether that was the result of an updated file in Nightly or a purge of a misconfigured one, but the system appears to be back to normal.
Thanks to Hy, Peter and Jasem for looking into it.
Jo

PS: Source build as of this morning also works again. Still wonder what the problem was, though.