Platform agnostic line endings

[DecimalTurn]

Description

This PR should make all remaining regex patterns follow our new approach of using platform-agnostic line endings as discussed here: #6417 (comment)

Here are the replacements made:

1. Standalone¹ \r?\n with (?:\r?\n|\r)
2. Standalone¹ \n with (?:\r?\n|\r)
3. (\s|\n) with [\s\r\n]
4. [^\n] with [^\r\n]

I also replaced (^|\n) with just ^ since matching with the previous line ending or not doesn't affect the outcome.

(Checklist removed as it doesn't apply).

Footnotes

1. In the sense that they are not part of a character set. ↩ ↩²

[Alhadis]

Shit, sorry I missed your review request, @lildude. 😞 Been having trouble getting myself in front of a workstation lately.

3. (\s|\n) with [\s\r\n]

I'm not sure why or where (\s|\n) was used in the first place, but this should've simply been \s; that construct is equivalent to [ \t\r\n\f\v] in Ruby. Ergo, it should've been simplified to just \s.

[lildude]

Ooo, I've just noticed this PR introduced a regex warning in the test output:

/home/runner/work/linguist/linguist/lib/linguist/heuristics.rb:91: warning: character class has duplicated range: /^[=-]+[\s\r\n]|\{\{[A-Za-z]/

This is for the AsciiDoc regex.

If you've got a mo @DecimalTurn could you please address this.

[Alhadis]

@lildude The /^[=-]+[\s\r\n] part of the offending expression is an example of the redundancy I mentioned above.

To fix:

-/^[=-]+[\s\r\n]|\{\{[A-Za-z]/
+/^[=-]+\s|\{\{[A-Za-z]/

[DecimalTurn]

@Alhadis Thanks, for the catch.

@lildude Should I create a new PR for this fix?

[lildude]

@DecimalTurn yes please. 🙇

[DecimalTurn]

Done.