fix: handle CR and CRLF in no-missing-atx-heading-space

[lumirlumir]

Prerequisites checklist

• I have read the contributing guidelines.

What is the purpose of this pull request?

Which language are you using?

CommonMark and GFM.

What did you do?

I expect the no-missing-atx-heading-space rule to correctly recognize CRLF and CR line endings, but it currently does not.

What did you expect to happen?

I expect the no-missing-atx-heading-space rule to correctly recognize CRLF and CR line endings.

Link to minimal reproducible Example

Adding the following test cases to tests/rules/no-missing-atx-heading-space.test.js would help identify the problem:

	{
		code: "Text before\r\n#Heading with ``` code markers\r\nText after",
		output: "Text before\r\n# Heading with ``` code markers\r\nText after",
		errors: [
			{
				messageId: "missingSpace",
				data: { position: "after" },
				line: 2,
				column: 1,
				endLine: 2,
				endColumn: 3,
			},
		],
	},
	{
		code: "Text before\r#Heading with ``` code markers\rText after",
		output: "Text before\r# Heading with ``` code markers\rText after",
		errors: [
			{
				messageId: "missingSpace",
				data: { position: "after" },
				line: 2,
				column: 1,
				endLine: 2,
				endColumn: 3,
			},
		],
	},

What changes did you make? (Give an overview)

This PR originated from #493. While working on that, I found the scope had become too large, so I split this fix into a separate PR.

In this PR, I've fixed an issue where the no-missing-atx-heading-space rule did not recognize CRLF and CR line endings.

The previous logic relied on a custom line breaker pattern using const newLinePattern = /\r?\n/u;, but this is no longer necessary now that the getLocFromIndex method has been added. So, I've removed it.

Also, the previous logic required a lot of calculations to convert "offset" into "line, column", but that's no longer necessary, so I refactored the code. (The larger git diff was unavoidable.)

I consistently used the pattern let match; while ((match = pattern.exec(text)) !== null) when visiting Paragraph and Heading nodes. Since this pattern is used in over 10 places throughout the Markdown rule implementations, using it consistently would make the codebase easier to understand.

Notable changes:

• Use named capture groups in the regex patterns.
• Add the m flag to leadingAtxHeadingHashPattern to avoid custom line break logic.
• Move the findMissingSpaceBeforeClosingHash helper function into the Paragraph visitor, since this logic was only needed before there was no direct "offset" to "line, column" conversion.
• Use the let match; while ((match = pattern.exec(text)) !== null) pattern consistently when visiting Paragraph and Heading nodes.

Related Issues

Ref: #493, eslint/css#280

Is there anything you'd like reviewers to focus on?

I'll update the other rules and processors to recognize CR in #493. This PR is separate from that one to keep the scope smaller.

[lumirlumir]

Pending until #554 is merged.