cylc-insert: prevent insertion at invalid cycle point. #2107
Conversation
oliver-sanders
force-pushed
the
1239.dont-insert-at-ungraphed-cycle-points
branch
2 times, most recently
from
467456e
to
7908ed2
Compare
| if not isinstance(items, list): | ||
| items = [items] | ||
| if stop_point_string == "None": | ||
| stop_point_string = None | ||
| return self._put("insert_tasks", (items,), | ||
| {"stop_point_string": stop_point_string}) | ||
| {"stop_point_string": stop_point_string, | ||
| "no_check": True if no_check == 'True' else False}) |
There was a problem hiding this comment.
Should be sufficient to do: "no_check": no_check == 'True'.
But then why would no_check be a string?
There was a problem hiding this comment.
Objects serialised for transmission on the comms layer? See line 154 in the diff above where stop_point_string is treated in a similar way.
There was a problem hiding this comment.
OK. I forgot it was done by urllib.urlencode not json.dumps.
oliver-sanders
force-pushed
the
1239.dont-insert-at-ungraphed-cycle-points
branch
from
7908ed2
to
8daa468
Compare
oliver-sanders
force-pushed
the
1239.dont-insert-at-ungraphed-cycle-points
branch
from
8daa468
to
525208d
Compare
|
(I guess you just rebased the branch; I'll wait on Travis CI before merging...) |
|
Huh, Travis CI doesn't like |
|
|
|
Passes in my environment. |
|
(I have bundled a change in #2100 to improve the stability of |
|
In that case, merging now... |
Close #1239
Don't insert a task unless its cycle point is on one of the tasks sequences. Over-ridable by the
--no-checkoption.