Use string-specific pretty_assertions displayer in tests

[neunenak]

The pretty_assertions crate that the test harness is already using has a string-specific comparison implementation that can properly display ANSI escapes, which are pretty frequent in just tests.

When running e.g. cargo test unicode_escape_invalid_character:

Before:

failures:

---- string::unicode_escape_invalid_character stdout ----
Bad stderr: Diff < left / right > :
<"\u{1b}[31merror:\u{1b}[0m unicode escape sequence value `BadBad` greater than maximum valid code point `10FFFF`\n   \u{1b}[38;5;246m╭\u{1b}[0m\u{1b}[38;5;246m─\u{1b}[0m\u{1b}[38;5;246m[\u{1b}[0mjustfile:1:6\u{1b}[38;5;246m]\u{1b}[0m\n   \u{1b}[38;5;246m│\u{1b}[0m\n \u{1b}[38;5;246m1 │\u{1b}[0m \u{1b}[38;5;249mx\u{1b}[0m\u{1b}[38;5;249m \u{1b}[0m\u{1b}[38;5;249m:\u{1b}[0m\u{1b}[38;5;249m=\u{1b}[0m\u{1b}[38;5;249m \u{1b}[0m\"\\u{BadBad}\"\n\u{1b}[38;5;246m───╯\u{1b}[0m\n"
>"error: unicode escape sequence value `BadBad` greater than maximum valid code point `10FFFF`\n ——▶ justfile:1:6\n  │\n1 │ x := \"\\u{BadBad}\"\n  │      ^^^^^^^^^^^^\n"

thread 'string::unicode_escape_invalid_character' panicked at tests/string.rs:509:6:
Output mismatch.
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

After:

failures:

---- string::unicode_escape_invalid_character stdout ----
Bad stderr: Diff < left / right > :
<error: unicode escape sequence value `BadBad` greater than maximum valid code point `10FFFF`
<   ╭─[justfile:1:6]
<   │
< 1 │ x := "\u{BadBad}"
<───╯
>error: unicode escape sequence value `BadBad` greater than maximum valid code point `10FFFF`
> ——▶ justfile:1:6
>  │
>1 │ x := "\u{BadBad}"
>  │      ^^^^^^^^^^^^


thread 'string::unicode_escape_invalid_character' panicked at tests/string.rs:509:6:
Output mismatch.
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

This should help with the test conversion that needs to happen in #2426

[casey]

Nice, this is much better!