Enable rendering of rich exception objects in traceback

[pradyunsg]

Type of changes

• Bug fix
• New feature
• Documentation / docstrings
• Tests
• Other

Checklist

• I've run the latest black with default args on new code.
• I've updated CHANGELOG.md and CONTRIBUTORS.md where appropriate.
• I've added tests for new code.
• I accept that @willmcgugan may be pedantic in the code review.

Description

See #2717 for discussion. This enables exceptions to be rich objects and be rendered via the rich rendering pipeline, in tracebacks. This can be tested manually by running python -m rich.traceback whose example now uses a rich object instead of the NameError.

[pradyunsg]

This is currently a draft since I'm not sure how exactly exceptions within the rich object rendering should be handled -- I'm leaning toward trying to catch them and presenting a fallback like "render failed" or similar. Not quite sure how to do that, so figured I'd open this as a draft PR in case y'all have any opinions/suggestions on how to achieve that.

[pradyunsg]

OK, I think I figured out something that works! 🎉

Using the Console object to eagerly render the exception and iterating through to collect all the segments, before yielding those segments forward. This enables this logic to catch any rendering-related issues inline, rather than needing to handle them at some other level (eg: console.render).

Marking this as ready for review, since I think everything that is necessary to land this is in place now¹ -- code, tests, documentation, changelog, contributing.md. :)

Footnotes

1. Assuming of course that there are no concerns with the implementation/approach. 😉 ↩

[pradyunsg]

sys was unused at this point, and there's already a top-level import sys.

[pradyunsg]

Oh, and here's an SVG showing that this works!

[pradyunsg]

These prints are useful for diagnosing failures, and pytest will capture this output appropriately. :)