OK lets regroup, my last commit is not what was intended!!!
I did a merge and believed I made the correct changes, then committed and did a pull, at which time I was told I was up to date. So I assumed because my changes were correct at my end, then that is what was committed.
I have just realised this is not the case, and the only change was the version number!!!??? This was discovered whilst setting up an AstroBerry, and pulling the latest nightly, I noticed my changes were not there.
So I have two file on my development end, that need to be committed via a pull request?
I guess the steps are
1. Create a commit (should I do some sort of sync first, and if so what should I do?).
2. Create a Pull Request at the head end, and associate the above commit?
I won't do anything until I get some feedback, as this is how I started getting in this mess, and can't understand why only the version number was committed?
In terms of myfp2, the standard should be that commands are :xx and terminated with #, commands being 2 characters long (maybe followed by parameters of variable length). This shouldve been consistent from 2014 and I cannot really understand when and if it ever changed to less than 2 [in which case the firmware wouldve rejected the command]
You can rest assured this will not change.
There is as I understand it, no plans to add or change commands in the forseeable future. There should be no reason to change the test for firmware version (at 291) for any future release. At present there are also no plans to add additional features in the pipeline, for 2 reasons - we are running out of space, and effort is now redirected to use the esp8266/esp32 based chips (which solves issue1 running out of space).
Not quite that simple as a command can be a command or a command that also includes parameters which are variable lengths
Accepting that commands are two characters, then conversion to an int type value based on those two chars and thus switch(cmd) becomes easy
I had a proper play last night, though still quite cloudy, and could not get the issue with losing absolute and maximum position. Even powering everything down and starting again, all values were retained as expected.
Could you try with a completely fresh install, and again see if you can come up with a sequence that will reproduce the error you are seeing?
I should point out that I observe the same issue with the kstars/indi Windows build (Windows 10) as well as directly with the kstars/indi build on Stellarmate on the Raspberry Pi
Shutting down the client windows and opening them up again will sometimes *not* show a value in the fields (hi-lighted in red in my screen captures) then for the same session shutting down
the client windows and stopping/re-starting indi will sometimes show a value in the field.
I observe that shutting down kstars completely and then restarting is another way that I'm able to reproduce the issue (sometimes)
I've not yet worked out the exact pattern/scenario to reproduce this consistently but I'll stick with it until I do.
A fresh install is a good idea. I have a vanilla build of Stellarmate on another SD card which I can bring into play.
I'll see about starting from that sd card then fetch the nightly build and then see if I can reproduce the issue
I think that the only way you might to the bottom of this is if the serial log files are looked at,
Thant would determine if it is a serial timeout issue, a serial timing issue on connect, or if it is a problem in either the firmware or higher up the chain.