I submitted the pull request to Jamie for the changes to the Script.
I submitted a pull request to home-brew/science to get them to fix wcslib.
And now I am playing with Travis CI to see if I can get it to work. Don't worry, I forked the kstars-on-osx script after I submitted my pull request to make sure it doesn't mess it up somehow.
After a bit of coaxing (25 attempts or so. . .). I managed to get the script to run nicely in Travis CI. The only thing that you really need to do to get it to work is to create a small configuration file called .travis.yml This file is extremely crucial however, and is the cause of my having to repeat the build multiple times. Of course there is still an error in craft, but I was just hoping to get the script to go as far as it does on my computer, which it does. Later we can work out any other issues. The options on Travis CI support building on OS 10.11 and 10.12. There is currently support for 10.10, but that will be removed shortly. And anyway, I think we need xcode8 in order to avoid the clock monotonic build failure in the 3rd party drivers. You will note that right now I am using the brew qt5 (and without dbus!). It took too long to build the dbus version and I don't know how to get the normal qt installed from the command line. We will probably need to change that later. Also note that I turned off the announce parameter because that stopped the script from running.
Here are the contents of .travis.yml that worked for me:
The instructions are above if you want to test plate solving and gsc related stuff, but the dmg and the KStars and Ekos therein should work right out of the box. You just need to right click to open the app the first time and click ok to the unidentified developer prompt. The dialog boxes should guide you through the rest and then it should be up and running.
I am testing it right now and the only thing it's not working is the ZWO Driver.
"2017-01-17T22:59:01 Unable to establish remote device:
+ ZWO CCD Please ensure remote device name corresponds to actual device name."
I had this problem before with my Atk 420 and it was a device name problem.
It does not even show the device tab. On the device list there are 2 ZWO Drivers listed. None works.
Gphoto, EQmod and Moonlite drivers work ok. Some more testing.
- Kstars crashes defining custom paths for executables and using Apply button, no problem with OK Button (ex, solve-field, wcs-info, cfg location)
- Using offline solver with internal executables works but fails to solve, I had remove the -no-fits2fits keyword:
Thank you for testing it out! I believe you are the first person to try KStars installed from a DMG with real equipment on a Mac and have it working! Of course I have connected my equipment to my raspberry pi and connected from my Mac, but of course it is full of development stuff that could make it work even if it shouldn't. I have used it with the simulators on my wife's computer, but I haven't yet tried real equipment from her computer. I have been focusing on getting the build script in order and the weather has been bad.
As to the ZWO issue, I don't know if that is a Mac issue or a driver thing.
For GPhoto Eqmod and Moonlite, were you using them directly connected to your Mac or via a remote computer like a Pi?
I can look into the custom paths issue, since that feature will be necessary in the future in case somebody wants to run a different astrometry version or something. But there should be working astrometry libraries built into the app. Did you try plate solving with the internal solver? I put instructions in this thread a few days ago.
I think the ZWO thing is really a name issue, I had the same problem months ago with an Atik and was solved.
All my equipment is remote, connected to an Atom x64 PC.
Just tried local astrometry but it fails. I just update my last post with some info on it.
If it fails immediately, usually it is either a configuration issue like python, pyfits, or netpbm not being installed or there is an extra space in your arguments. I don't think the second one is the problem.
Well it solves on command line with no problems at all, just checked, have pyfits and netpbm both installed.
The same command running on ekos immediately gives a solver failed message. It seems to me that maybe it's not finding the index files!
I was trying to define the logs files but it also crashes doing that when I press ok. At least it kepts the options after the crash, lets see if the