Timeout argument for ._create_job(), .enqueue_at(), enqueue_in()

[lechup]

Scheduler._create_job()

• now accepts also timeout argument
• added test for above
• cleanup of scheduler methods that were setting timeout not by _create_job

Scheduler.enqueue_at() + Scheduler.enqueue_in():

• added timeout argument that is popped from kwargs

[selwin]

@lechup thanks for the PR. Do you mind adding timeout argument to enqueue_in and enqueue_at methods as well?

[lechup]

@selwin I've prepared You asked here, but I'm not sure regarding how we can change arguments of enqueue_at/in I think that all Scheduler methods should have same logic of passing args/kwargs but it would be backward incompatible change. I've tried with addition of third argument but ended up with

======================================================================
FAIL: test_args_kwargs_are_passed_correctly (test_scheduler.TestScheduler)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/lechup/github/rq-scheduler/tests/test_scheduler.py", line 278, in test_args_kwargs_are_passed_correctly
    self.assertEqual(job.args, (1, 1, 1))
AssertionError: Tuples differ: (1, 1) != (1, 1, 1)

I've ended up with poping timeout from kwargs...

What I proposed is not clean solution either, if someone is passing timeout kwarg it won't be passed anymore and will end in unexpected behaviour.

I need Your help/decision here...

[lechup]

Oh and the commit is here: 316fdd9

[selwin]

I'm fine with popping timeout from kwargs because that's also what happens in RQ's queue.enqueue. Thanks for working on this!

[lechup]

Hm... But still I think all Scheduler public methods should provide same way of passing args and kwargs for jobs... Place for another PR? ;)

I think we should go the way queue.enqueue is doing with arguments... Right?

[selwin]

I don't mind introducing a slight backward incompatibility. To me it's more important that schedule, enqueue_at and enqueue_in share the same behavior. We can also make a note of this in the release notes. Thanks for your consistent contribution to RQ related libraries :)