data node is changing its name during scan
working on tests in new_data_before_new_node-branch I got the impression that the DataNode (and LimaDataNode) has a different name on creation ('NEW_NODE' event) and during 'NEW_DATA_IN_CHANNEL' events. This seems also to be the reason for #992 (closed)
NEW_NODE EVENT:
node.name image
node.db_name test_alias:tmp:scans:8_loopscan:timer:lima_simulator:image
NEW_DATA_IN_CHANNEL EVENT:
node.name: lima_simulator:image
node.db_name: test_alias:tmp:scans:8_loopscan:timer:lima_simulator:image
- is this something that should be allowed?
- looking at the test this does not seem to happen all the time but in about 1/4 of all cases (at least on my machine)
I couldn't reproduce it anywhere else but in the external writer script but there we get this sort of error:
*** KeyError: 'timer1:epoch'
(Pdb) p self.channel_indices
{'timer1:elapsed_time': 1, 'epoch': 0, 'simulation_diode_controller:diode': 0, 'timer2:elapsed_time': 1, 'timer2:epoch': 1, 'dummy_device:pi': 1, 'dummy_device:nb': 0}
(Pdb) self.channel_indices
{'elapsed_time': 0, 'timer1:epoch': 0, 'simulation_diode_controller:diode': 0, 'timer2:elapsed_time': 0, 'timer2:epoch': 0, 'dummy_device:pi': 0, 'dummy_device:nb': 0}
(Pdb) p node.name
'timer1:elapsed_time'
which points into the same direction (e.g. node.name epoch
instead of timer1:epoch
on the two events.)
Edited by Linus Pithan