Modernise 'cylc dump -t' print format.

[hjoliver]

This has been annoying me forever.

Master 🤮:

$ cylc dump -t bug --flows
foo, 100, running, not-held, not-queued, not-runahead, [1]
foo, 101, waiting, not-held, not-queued, runahead, [1]
foo, 96, running, not-held, not-queued, not-runahead, [1]
foo, 97, running, not-held, not-queued, not-runahead, [1]
foo, 98, running, not-held, not-queued, not-runahead, [1]
foo, 99, running, not-held, not-queued, not-runahead, [1]

This branch:

$ cylc dump -t bug --flows
91/foo:running flows=[1]
92/foo:running flows=[1]
93/foo:running flows=[1]
94/foo:running flows=[1]
95/foo:running flows=[1]
96/foo:waiting (runahead) flows=[1]

QUESTION: is it OK to make the new format the default?
(As I have done, with -l/--legacy to generate the old format)

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.cfg (and conda-environment.yml if present).
• Tests are included (or explain why tests are not needed).
• Changelog entry included if this is a change that can affect users
• Cylc-Doc pull request opened if required at cylc/cylc-doc/pull/XXXX.
• If this is a bug fix, PR should be raised against the relevant ?.?.x branch.

[oliver-sanders]

I think this is fine (especially as you've added a legacy format option).

If we're preserving the old format long-term we might want to call it --porcelain or something like that (the simple table is much easier for things like awk to parse).

[oliver-sanders]

LGTM

[dwsutherland]

Looks good.. We adding a legacy format to avoid changing all the tests?
I guess someone might like it (or use it) 😆

[hjoliver]

Looks good.. We adding a legacy format to avoid changing all the tests? I guess someone might like it (or use it) 😆

Not so much adding it, but keeping it just in case anyone is relying on it. (Not trying to avoid changing the tests, honest!)

[dwsutherland]

I don't know how the patch coverage has been calculated? I wouldn't have thought there would be many indirect losses from this change...

[hjoliver]

From a quick look on Codecov, I don't get it:

The yellow ("partial") lines are covered by my integration test, and the red (not-covered) lines are not changed on this branch.

[oliver-sanders]

To turn those yellow lines green we need a test that makes the if fall in either direction, e.g. at least one task which is held and one task which is not held.

[hjoliver]

At first glance I thought it was odd because those if blocks don't have an else clause. But the do-nothing option does affect the following logic of course - so that explains it, thanks! I'll extend the tests...

[hjoliver]

Patch coverage now 100% 🎉