Truncate long object representations in `to_repr`

[liblaf]

Type of changes

• [x] Bug fix
• [ ] New feature
• [ ] Documentation / docstrings
• [ ] Tests
• [ ] Other

Checklist

• [x] 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.
• [x] I accept that @willmcgugan may be pedantic in the code review.

Description

Hi Rich maintainers,

First off, thank you for creating and maintaining rich – it's an incredibly useful library that I use frequently!

This PR addresses the issue of very long object representations in tracebacks, especially when show_locals=True is enabled. Objects like large NumPy arrays, PyTorch tensors, or JAX arrays can produce extremely verbose repr() outputs, making tracebacks difficult to read and navigate.

Currently, to_repr truncates strings and bytes if they exceed max_string. This change extends similar truncation logic to the generic repr(obj) output. If max_string is set and the repr() of an object exceeds this length, it will be truncated with an ellipsis (...).

This is a partial solution that helps improve readability for general objects whose repr() is too long. It doesn't allow for type-specific custom representations, which has been discussed in issues like #3393 and PR #3394. I believe allowing external registration of __rich_repr__ or similar custom display functions would be a more comprehensive solution, as mentioned in discussion #1258. See discussion #3774.

However, I hope this small change can provide an immediate improvement for users dealing with overly long representations in tracebacks.

Thank you for your consideration and for all your work on Rich!