[CT-400] Event module should handle "warn on error" behavior.

[nathaniel-may]

There are some places in the code that conditionally fire otherwise identical warning events and error events. For example see these lines in the printer task. Because the logged message is identical in both of these events, they should not be two separate events.

The proposed solution:

1. respect the flag --warn-error within the event module
2. add a default named parameter to fire_event for adhoc error-on-warn behavior. (i.e. - error_on_warn=False)
3. for all events that have a warning and error variant remove the error variant and use the added flag to fire_event at the call sites.

[jtcohen6]

This is eerily similar to (but indeed separate from) our warn_or_error behavior. That's also something we should look to rework, when we bring exceptions into the same structured framework: #4836 (comment)

[jtcohen6]

Realized that #5424 was another instance of this tech debt (and an actual bug, since it's not working correctly today)

[stu-k]

Having looked through the event types in events/types.py and where they are emitted, I believe the only events with identical error/warning variants are as follows:

• task/printer.py: RunResultWarning, RunResultFailure (conditionally fired)
• task/freshness.py PrintHookEndErrorLine, PrintHookEndErrorStaleLine, PrintEndHookWarnLine (conditionally fired)
• events/adapter_endpoint.py: AdapterEventWarning, AdapterEventError (not conditionally fired, but has identical event structure)

Additionally, it would seem that our events are defined as whatever level they are (i.e. test, debug, info, warn, error) via inheritance, so emitting one event as both warn and error would require a refactor of how our events are defined.

Given this information, I'm not sure it makes sense to have the fire_event function optionally turn events from warnings into errors given it would introduce unnecessary complexity into the events to be able to be transformed from one type into another.

I will close this for now in favor of #5424. Please feel free to reopen should my assessment be incorrect.