Josh McKinney

One possible idea might be to give each pipeline element an id of its own and pull the span code up into the code which runs the transform / ......

> Node ids are UUIDs lazily generated from the node content and path, this means that when chunked, ids change. Currently besides the path there is no way to trace...

Yeah, perhaps splitting the id you currently have into assigned id (e.g. a v7 uuid) and a calculated cache key (perhaps just a hash, no uuid) could be worth considering....

Thanks for the report. If I had to guess, the problem here is in the width calls (using unicode-width) in the following code (or elsewhere in either Buffer or the...

This is a bit of an odd place where the fix would be found. If I understand that code right, what should be happening there is that we'd be drawing...

This issue seems like it would be a good inclusion for 0.30 if we can find time to dig into it a bit more.

Some missing context that may help understanding that suggestion is that maintainer time is finite and often oversubscribed. A "known good enough" solution often beats an "unknown but perfect" solution...

I'd be fine with a PR for this. Add `dep:crossterm?/serde` and check the other backends for the same.

yep on the way you suggested. termwiz has a use_serde feature

Fixed in https://github.com/ratatui/ratatui/pull/1812