Editors have more control over publication notifications

[joemull]

Closes #4089. Also closes #1186.

This pull request improves the usability of the prepub notification modal, as well as upgrading a few underlying utilities and forms.

• Adds an email form that includes the To field, so that you can create a formset with different recipients for each form, passed to the formset in initial, a list of dicts. Without this, we'd have to modify each form in the formset based on its index value before calling send_email, which would be prone to bugs and not declarative enough.
• Upgrades the logging and display of To, CC, and BCC fields so that editors who check the logs can be sure they are not violating anonymous review when sending emails to peer reviewers.
• Moves the latest effective function for migrating translatable setting values into migration_utils and tests it.
• Lets editors view and modify the full emails that will be sent out to authors, section editors, and peer reviewers, and prefills the right email addresses based on new settings.

[mauromsl]

Thanks Joe, this I think also provides a nice alternative for #3685
I believe the current implementation won't cascade the form data properly onto the events framework, so that needs to be added.

A similar refactor was done in f8edb28 and 8fc2ea1 . You can follow the "email_data" argument to see how it is implemented.

[mauromsl]

These are all very welcome changes!
I added one small comment inline.

More generally, I'm a bit scared by the fact this relatively large refactor hasn't required updating tests.

Would it be possible to add a TestCase that confirms emails are being sent to the relevant parties according to configured settings?

[mauromsl]

Even if we de-register the send_author_publication_notification, we should properly deprecate an event like this one before deletion as others might depend on it.

I'd recommend adding such warning on events.logic.Event.register_for_event perhaps with a set of deprecated events:

class Event():
DEPRECATED_EVENTS = {
   ON_AUTHOR_PUBLICATION,
}
def register_for_event(event, *functions):
   if event in self.DEPRECATED_EVENTS:
       # raise DeprecationWarning
[...]

[joemull]

Good call, but if we are trying to add a deprecation warning for this event because we think people might be using it in plugins, we need to make it work with all the supporting functions.

I've added back the full set of functions to make this old event work the way it used to, and I've added the deprecation warning in raise_event.

[joemull]

More generally, I'm a bit scared by the fact this relatively large refactor hasn't required updating tests.

That would be because I didn't run the tests 😮

Would it be possible to add a TestCase that confirms emails are being sent to the relevant parties according to configured settings?

Yes, for sure.

[joemull]

This now has 7 failing tests, but I believe they are unrelated and some may be fixed in #4206.

[mauromsl]

good changes!

[ajrbyers]

WWMD.