Fix retry when status of job is aborting

[medihack]

Closes #1131

Successful PR Checklist:

• Tests
  • (not applicable?)
• Documentation
  • (not applicable?)

PR label(s):

• PR type: breaking 💥 Contains breaking changes
• PR type: feature ⭐️ Contains new features
• PR type: bugfix 🕵️ Contains bug fix
• PR type: miscellaneous 👾 Contains misc changes
• PR type: dependencies 🤖 Contains only dependencies updates
• PR type: documentation 📚 Contains documentation updates

[github-actions]

Coverage report

This PR does not seem to contain any modification to coverable code.

[ewjoachim]

Code is great but we need to decide whether we want to retry or not. I'm not convinced yet that it makes sense to retry.

Here is my rationale. The following is a timeline.

-> job is deferred

If the job is cancelled here, it is immediately discarded

-> job starts

If the job is cancelled here, it is just marked for abortion. It may not be aborted at all. of course, given the job has started, it makes sense that it is treated differently

-> job ends with an error, is enqueued for retrying (potentially with a delay)

If the job is cancelled here, it is immediately discarder

-> job starts again

What seems strange is that if the job is nearing its end (or error), if you wait for it to end with an error and then cancel it, it will be cancelled sooner that if you cancel it immediately. Similarily, if you cancel it again while it's enqueued for retrying, it will be discarded immediately. A process that requires the user to startegically wait before giving an order for it to be executed sooner or requires the user to request it multiple times seems like procrastinate wouldn't be trying to follow it, I think?

That said, I'm not 100% sure. At all.