horiba-ihr320: state changes on restart

[ddkohler]

User complaint

Sometimes when you restart the monochromator daemon, the settings change (e.g. slit position)

This has been problematic because such property changes are easily missed when restarting daemons. It's likely that the slit position changes to some hardware default position after a hard restart, and state is updated with the position it reads.

I am proposing, that, on daemon startup, we actively set slits/whatever else to positions in the daemon state (if state has valid positions). Of course open to other suggestions...

[ksunden]

the thing is, we do try to do that, soooo

[ksunden]

https://github.com/yaq-project/yaqd-horiba/blob/main/yaqd_horiba/_horiba_ihr.py#L28

[untzag]

We should loop and retry until correct. We do this in other daemons. Hardware sucks, make it work 😢

[usui-bluec]

Hello! I faced the same problem when i tried to operate micro-hr using this program.
I don't know if this works at all, but I was able to initialize it successfully by using

loop.run_until_complete(self._reset_position()) 

at the end of the home function in _horiba_mono.py to call the task.
I hope this works !

[untzag]

I'm not quite sure why you would want to do that @usui-bluec, but I'm willing to listen if it seems to be improving things for you.

We already call _reset_position at the end of home

I've never used run_until_complete, but it seems like it does the same thing it just waits for the coroutine is finished. Just the same as if you put an await in front of it?

I feel like I have a concrete, ugly solution to this problem. I'm going to make a PR.

[untzag]

@ksunden @kameyer226 @usui-bluec please checkout #19 and see what you think

[ksunden]

The difference between run until complete and await is that run until complete is called from a synchronous method (and spins up an event loop of one doesn't exist) await cannot be used directly in init

We use the alternative of create task (because we know that an event loop exists)