Hilary James Oliver

A final small thing: `cylc scan` shouldn't get completed with IDs.

> The completion cannot replace foo// with foo/run1// because that would involve replacing existing characters. That's not what I meant. In principle the server could figure out that `runN` is...

in that case ... :boom:

From project meeting: - `remove` is not a good command name in Cylc 8, it requires users to understand the scheduler task pool concept (the content of which is no...

@wxtim - not sure I understand the problem. Can we not just do this?: 1. manually trigger the next task(s) (i.e., carry on despite events) - (or use set_outputs to...

> manually set the state of switch tasks (at an optional branch point) to direct the flow one way That seems like a different problem? 1. use set-outputs to spawn...

@dpmatthews - I've started working on this, so thinking about some of the details in more depth. > Use --flow to specify a specific flow or a new flow (but...

> Use --no-spawn (?) to prevent spawning of tasks. This provides a way to set "off-flow prerequisites" without starting a flow. Is there a better way to handle this case?...

I know how to proceed here, will try to write it up succinctly for review soon. Just a couple of comments for now: The scheduler "task pool" **is**, to all...

Closing as superseded by: - cylc/cylc-flow#5054 - cylc/cylc-ui#1104