Ok, main problem with all fresh git pulls for both kstars and stellarsolver 10min ago - kstars crashes on image pass off to stellarsolver.

More detail:
Rpi4 4G
StellarSolver Test install on desktop so all the relevant libs should be there when I build kstars.

Interesting Result:
When I run StellarSolver Test app from the desktop it will crash if the Process button is set to Solve,
works fine for Sextract or Sextract with HFR.

Conclusion:
This could be whats causing kstars to crash as well.
Am I missing a Process selection in the align options?

Thanks.

Ok after working on it a bunch today, I have now split up the run method in stellarsolver to support the changes to make StellarSovler not a thread itself.
Now the solvers it creates are all threads, but it is not a thread itself anymore.

With the changes I made it should fix the freezing behavior you guys noticed in the last day or so. Sorry it wasn't working for about 24 hours or so, but it did take me a few hours to fix the issues.

I would still like to address a logging issue, for now you might want to select "log none" still.

Is there something special I need to do to see the new solver in KStars? Install a new package or so? Because I am on the nightly build of Thursday (October and I don’t see the new solver. Also, we tried solving with a DSLR camera on a RedCat telescope, which worked before but now not anymore. When plate solving either with source-extractor or Python, only noise gets identified as stars and the real stars are not detected as such. In hindsight it was a miracle that polar alignment actually worked.

One of the images used can be found here

wvreeven.stackstorage.com/s/4XZCOo1PyQg47iCt

It was converted from CR2 to FITS in PixInsight without any stretching or other processing. When I use the stable release of KStars on my Mac, it solved within seconds but solving times out on my Raspberry Pi 4 running Ubuntu 20.04 and that nightly build.

Thanks for any help getting this working again and sorry if this is not related to the new solver!

Before I forget, here are the settings used:







And yes, the focal length of the RedCat and the DSLR sensor data were entered correctly. Like I said, plate solving worked really well with a nightly build of two weeks ago.


Wouter

Works fine for me after first try! Pretty fast.
But it would be great if it would give some feedback if the solve fails. Instead, the log just gets stuck at 'Image received'. See screenshot below:



Feedback is only provided if the solver is successful, but not if it fails. But it would be at least as important to know if and why the solver has failed (although in the above image it is obvious).

Jo

wvreeven, I am not sure it can be related to the new solver, since your screenshots are not of the new version of kstars. If you were using the new build, they would not look like that, the options are all different.

Jim, I think I have fixed the issues you were dealing with.

Jo, Yes I understand. I wanted to get everything working again first and then I will work on improving logging.

jackbak, sorry, it was broken when you were testing, but should be fixed pretty good now. Please test again.

Cerro Torre, please test again with astrometry logging set to "log none". I am hoping to work on this issue next. Astrometry is very very chatty.

Rob, I understand, I was not complaining. This was merely feedback. I am using SEP, ASTAP and option 4- parallel small scale or something like that. That worked great out of the box. Very fast, also for PA.

Great job!

Jo



Once the solver can see half an image frame, it solves successfully.

To be honest, I don't think I have tried solving any pictures of trees. . .it wouldn't occur to me to do so. Thank you for testing that

You are most welcome, Rob! Now, if your solver could tell me what kind of tree that is, I would be REALLY impressed. Maybe you could have a go at that next?

Nope, I think you need a daytime solver for that one... actually there is PlantSnap, that might do it

Hi,

i'm playing with the polar aligment routine (with indi simulator). Sometime that work, sometime not (kstars crash). I played with some configurations (essentially with 1-FastSolving and 7-MidSizedStars).

I attached the console capture (with the crash) and the separate astrometry log file.

I updated the stellarsolver lib this morning and kstars yesterday.


Thank.