Constrain window by final cycle point

[dwsutherland]

These changes close #4255

Constrain the data store n-window by the final cycle point (FCP)... (ICP doesn't appear to be an issue)

There would be two ways of doing this:

• change the taskdef.py function that generates children
• filter at the window expansion for children whose cycle point extends beyond the FCP.

I chose the latter, because we may wish to (or already do) record the workflow state of tasks beyond the FCP for the purpose of restarting the workflow with an extended FCP.

The issue only applies to those tasks in the [scheduling][[special tasks]]sequential section, not those defined in the graph as sequential.

The test tests/functional/graphql/01-workflow.t will fail with the filter commented out.

Requirements check-list

• I have read CONTRIBUTING.md and added my name as a Code Contributor.
• Contains logically grouped changes (else tidy your branch by rebase).
• Does not contain off-topic changes (use other PRs for other changes).
• Applied any dependency changes to both setup.py and
  conda-environment.yml.
• Appropriate tests are included (unit and/or functional).
• Appropriate change log entry included.
• No documentation update required.

[hjoliver]

The issue only applies to those tasks in the [scheduling][[special tasks]]sequential section, not those defined in the graph as sequential.

That doesn't seem to be true, e.g. @dpmatthews example from Element chat shows an n=1 task in 2000-01-04:

[scheduling]
    initial cycle point = 2000-01-01
    final cycle point = 2000-01-03
    [[graph]]
        T00 = "task1[-PT24H] => task1"
[runtime]
    [[task1]]
        script = sleep 10

Same again for a variation to avoid sequential tasks: task1[-PT24H] => task2 => task1

But ... this PR fixes it 🎉

[hjoliver]

looks good 👍

[dwsutherland]

The issue only applies to those tasks in the [scheduling][[special tasks]]sequential section, not those defined in the graph as sequential.

That doesn't seem to be true, e.g. @dpmatthews example from Element chat shows an n=1 task in 2000-01-04:

I was just looking on startup (cylc play --pause fox/run1) with ICP = FCP, but didn't test as the workflow progresses to a FCP further out... The fix, as you've seen, is agnostic to either case.