If we're only aligning enum discriminants that are "not too far apart
(length-wise)", then this works really well for enums with
consistently-long or consistently-short idents, but not for the mixed
ones.
However, consistently-long idents is somewhate of an uncommon case and
overlong idents may be allowed to be formatted suboptimally if that
makes mixed-length idents work better (and it does in this case).
The issue with the current code is that comments are collapsed, so comments like
the one from the test end up in a string like:
```
"// this is a single line comment\n/* block = */"
```
I chose to fix it by detecting whether we're in a block comment starting from
the end instead, and tested a single-line comment ended in `*/` just for sanity,
ensuring line breaks are not removed in that case, which would break the
formatting.
The right fix eventually is probably to lex the comments properly, but this does
the work for now, I guess :)
Fixes#3025
Lines that didn't start with a comment sigil were returned unchanged in
comment::rewrite_comment. Then these unchanged lines were indented in
MacroBranch::rewrite.
The leading whitespace of a multine string was taken into account when
computing the `min_prefix_space_width`, even if that line couldn't be
trimmed. The consequence is it was always shifting the macro's content
to the right.
I've dabbled recently in seeing how hard it would be to compile rustfmt to wasm
and then run it in a web browser, and it turns out that it's [not too
hard][wasm]! In addition to patching a few dependencies which already have a
number of patches out rustfmt also needed some modifications to get it to work,
namely avoiding the usage of `Instant::now()` on the "happy path" which doesn't
work on wasm (it just panics).
This commit is an attempt to add a support for this by avoiding using
`Instant::now()` on the wasm target, but panicking if the actual time elapsed is
requested (which doesn't happen unless verbosely logging I believe).
[wasm]: https://alexcrichton.github.io/rustfmt-wasm/
test::verify_check_works was failing in relase mode on my machine.
The problem was it would check for target/debug/rustfmt in release mode instead of
target/release/rustfmt and fail an assert.
This commit fixes it so that cargo check looks for target/debug/rustfmt and
cargo check --release looks for target/release/rustfmt
