Revert Replace ASCII control chars with Unicode Control Pictures

[oli-obk]

reverts #127528 as that was a perf regression

r? @estebank

[rustbot]

These commits modify the Cargo.lock file. Unintentional changes to Cargo.lock can be introduced when switching branches and rebasing PRs.

If this was unintentional then you should revert the changes before this PR is merged.
Otherwise, you can ignore this comment.

[oli-obk]

@bors rollup=never

[compiler-errors]

@bors r+

I'd like to see this code re-landed in a way that's more efficient, since the improvement to diagnostics is quite nice. This is worth probably a spin-off discussion about how diagnostics formatting code affects the good path in the case of lints and warnings.

[bors]

📌 Commit 6679c62 has been approved by compiler-errors

It is now in the queue for this repository.

[estebank]

I'm honestly a bit shocked that it had a perf regression (as it only affected rendering...)

Edit: I'm assuming it is some second order effect on early_lint_checks?

[compiler-errors]

See the discussion at #128169 (comment) for more info, if you haven't already

[estebank]

@compiler-errors thanks, catching up

[estebank]

I suspect that the culprit is https://github.com/rust-lang/rust/pull/128200/files#diff-a35029c589a6b5b002ecf598efb24d174232e719ecbc6f741a58eee8509cd6a5L2612-L2614

It wasn't noticeable before because the list was short enough, but big-O likely dominates now over constant times.

[compiler-errors]

@bors r-

Let's give this a tiny bit of time to let Esteban fix the perf w/o a full revert.

[bors]

☔ The latest upstream changes (presumably #125443) made this pull request unmergeable. Please resolve the merge conflicts.

[lqd]

Closing in favor of #128200 which gains much of the performance back and should land soon.