Closed ispielma closed 8 months ago
I haven't had a chance to look at the changes to the other components yet, but is there a reason why you chose to store the lyse server in the HDF5 file rather than just sending it along with the ZMQ request to BLACS?
On first glance, storing it in the HDF5 file seems like it would pose problems for rerunning a shot (e.g. drag-drop old HDF5 file into BLACS) if the lyse hostname changed. But I'm guessing there are reasons you chose to store it in the file.
This family of pull requests could be implemented in a number of ways. Now that I have actually implemented it, one way, I think that the better solution would be to have blacs send a message back to runmanager when the shot is done and allow runmanager to dispatch to lyse. This would be a good addition anyway because I have been wanting to add a just in time compilation option to runmanager and this signaling is also needed for that.
In the interest of not keeping PRs around for a long time (it's approaching 2 years!) I'm going to close this. I believe we're all on the same page that, even if some new functionality is needed around support for multiple lyse instances (which it probably is), such a change doesn't need to involve labscript (this repository).
This is part of a series of pull requests to labscript, blacs and runmanager that move the Lyse server address into runmanager. The reason for this is we are setting up a lab where people from different computers have the reasonable expectation to submit shots from their computer and the the results back to their copy of lyse.
The changes to labscript are pretty simple.