Fix some pretty printing tests

[petrochenkov]

Many pretty-printing tests are un-ignored.
Some issues in classification of comments (trailing/isolated) and blank line counting are fixed.
Some comments are printed more carefully.
Some minor refactoring in pprust.rs
no-pretty-expanded annotations are removed because this is the default now.
pretty-expanded annotations are removed from compile-fail tests, they are not tested with pretty-printer.

Closes #23623 in favor of more specific #37201 and #37199
r? @nrc

[petrochenkov]

@nrc
Some more general questions:
Are there any plans regarding the rustc pretty-printer, considering that rustfmt is actively developed?
Printing of some small fragments of AST/HIR is still useful for diagnostics, but what about all the formatting stuff, comments?
Also, HIR pretty-printer could probably use some diet. It can't even print parens reliably, but at the same time it tries to print other surface stuff like comments. Do you think they are needed or they can be removed?

[nrc]

@petrochenkov

Are there any plans regarding the rustc pretty-printer, considering that rustfmt is actively developed?

It's been on my mind, but I don't have a plan. ATM, I think we leave it in place and maintain it to the minimal state necessary. If it becomes a burden, we might look to sharing code between Rustfmt and the pretty printer. I don't know exactly how this would work. This would probably be easier with Rustbuild and using an external crate.

Printing of some small fragments of AST/HIR is still useful for diagnostics, but what about all the formatting stuff, comments?

Agree. I think it is sometimes useful to have pretty printing of whole programs for debugging, and that requires keeping the formatting stuff. We could probably lose comments, but it doesn't seem like a burden.

Also, HIR pretty-printer could probably use some diet. It can't even print parens reliably, but at the same time it tries to print other surface stuff like comments. Do you think they are needed or they can be removed?

I'm not sure off the top of my head whether we use HIR pretty printer for error messages, etc. But it seems likely we do for types, etc. Again, it is sometimes useful for debugging to dump HIR as code. I'm honestly not sure whether it is worth spending effort on it.

[nrc]

@bors: r+

[bors]

📌 Commit 2cb8cdc has been approved by nrc

[bors]

☔ The latest upstream changes (presumably #36969) made this pull request unmergeable. Please resolve the merge conflicts.

[petrochenkov]

@bors r=nrc

[bors]

📌 Commit 4a91a80 has been approved by nrc