You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The old "repeating asynchronous" task type was removed from cylc-6, due to lack of compatibility with the new ISO8601 based cycling framework - #989. I still maintain that repeating workflows based on arbitrary IDs determined at run time (rather than a regular cycle) could be useful though. It has been proposed that we may be able to restore this capability using the new integer cycling. This should be considered, and if it can be made to work, documented.
The text was updated successfully, but these errors were encountered:
Initial thoughts: successive satellite passes and their associated arbitrary labels need to be associated with successive integer cycle points, so that each task in the processing tree knows to process the right dataset (and multiple passes might need to be processed in parallel). With integer cycling it should be easy enough to handle this manually - without direct support for it in cylc. E.g. the uppermost task in the graph polls for a new satellite dataset and, when it finds one, moves it to a new location determined by its (the task's) integer cycle point. Then each subsequent task with the same cycle point would work in the same cycle point based location.
The old "repeating asynchronous" task type was removed from cylc-6, due to lack of compatibility with the new ISO8601 based cycling framework - #989. I still maintain that repeating workflows based on arbitrary IDs determined at run time (rather than a regular cycle) could be useful though. It has been proposed that we may be able to restore this capability using the new integer cycling. This should be considered, and if it can be made to work, documented.
The text was updated successfully, but these errors were encountered: