Will Kahn-Greene [:willkg] ET needinfo? me Assignee
	Description • 6 years ago

The processor has layers and layers of error handling and abstraction reminiscent of Gormenghast Castle. Some of the layers are partially-generalized and some are specific. Some handle things that no longer exist in the code-base. Sometimes you discover the reason something is "working" is because one of the layers is swallowing the errors indicating it's not working.

The result of all these layers is that it's really hard to have any confidence about correctness. Errors may not propagate upwards and get surfaced in Sentry or other tools (covered by bug #1362434). Errors may get swallowed so tests pass that really shouldn't. Crash reports may get through processing coming out "looking weird". Some of the error handling layers are written in a way that doesn't work in Python 3 which is blocking migration.

After talking about this with Lonnen, we decided we should fix this before continuing with Python 3 migration.

This tracker bug covers removing/reducing/simplifying error handling in the processor. This includes the transaction executor (bug #1503434), _transform in the processor app (bug #1426257), action in Rule, and probably some other things.

	Will Kahn-Greene [:willkg] ET needinfo? me Assignee
	Comment 1 • 5 years ago

All the work has been done now. Marking as FIXED.