fix: Improve prettify_macro_expansion()

[A4-Tacks]

• Add basic tests for prettify_macro_expansion

Example

const _: () = {
    loop {break}
    foo()
};

Before this PR

const _:() = {
    loop{
        break
    }foo()
};

After this PR

const _: () = {
    loop {
        break
    }
    foo()
};

---

const _: () = {
    match 2 {
        tmp => foo!(),
    };
};

Before this PR

const _:() = {
    match 2 {
        tmp => foo!(),

        };
};

After this PR

const _: () = {
    match 2 {
        tmp => foo!(),
    };
};

[rustbot]

This PR was rebased onto a different master commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[ChayimFriedman2]

Is there a real reason to improve this function? Where important, we pass its output to rustfmt. We only need it so that tokens won't be combined.

[A4-Tacks]

Is there a real reason to improve this function?

Yes, when I use inline_macro, the old output often looks terrible

[ChayimFriedman2]

Then we just should pass it to rustfmt there too.

[A4-Tacks]

No, most of the time I use inline_macro to debug macros,
and rustfmt has poor support for formatting nested macros
And when developing macros, the expanded result may not be syntax correct, while rustfmt can only handle correct syntax