markdown: remove unnecessary blockquote >

[shisama]

Fix #9099

This removes unnecessary blockquote > when the paragraph is made a single line.
I am not confident my implementation is good. Please tell me if you know a better way to fix this.

• I’ve added tests to confirm my change works.
• (If changing the API or CLI) I’ve documented the changes I’ve made (in the docs/ directory)
• (If the change is user-facing) I’ve added my changes to changelog_unreleased/*/pr-XXXX.md file following changelog_unreleased/TEMPLATE.md.
• I’ve read the contributing guidelines.

✨Try the playground for this PR✨

[thorn0]

This solution seems to be fighting the consequences, not the reason of the bug. > shouldn't end up in a node of the type word in the first place. There is the code in preprocess.js that takes the original remark AST and creates these word nodes. The bug must be somewhere in that code.

[kachkaev]

Once the right solution is found, it'd be also great to add changelog_unreleased/markdown/pr-9125.md file, similar to a markdown in #9078

[thorn0]

Lines shouldn't be joined when --prose-wrap is set to preserve. Also please make sure to test nested blockquotes.

Prettier pr-9125
Playground link

--parser markdown

Input:

> `1`
> `2`

Output:

> `1` `2`

Expected: same as input

[thorn0]

A solution that involves replacing > in a string can't be correct. If > has been already stringified, it's too late to meddle. It shouldn't be there. Why does it get there?

[shisama]

@thorn0 Thanks. I found the cause that > is included. options.originalText.slice() in the below code transforms \n \n> . So I fixed the condition of this. But I'm not confident if the solution is correct.
Please take a look. d43f64d

prettier/src/language-markdown/preprocess.js

Lines 47 to 57 in 810325d

	value:
	node.value !== "*" &&
	node.value !== "_" &&
	node.value !== "$" && // handle these cases in printer
	isSingleCharRegex.test(node.value) &&
	node.position.end.offset - node.position.start.offset !==
	node.value.length
	? options.originalText.slice(
	node.position.start.offset,
	node.position.end.offset
	)

And I added a test for nested blockquotes. 64f13fc

[sosukesuzuki]

This output is incorrect with our default options(--prose-wrap=preserve ).
Can you change this to

> `x`
> `y`

?

---

Prettier pr-9125
Playground link

--parser markdown

Input:

> `x`
> `y`

Output:

> `x`
> `y`

[sosukesuzuki]

Can you add tests for prose-wrap=never and prose-wrap=preserve with updating tests/markdown/jsfmt.spec.js?

[kachkaev]

👋 @shisama, would you be interested in updating this PR to help us close #9099? See suggestion from @sosukesuzuki above.