fix(BulletList): remove preserveWhitespace: true parseHTML override

This fixes a regression with prosemirror-model 1.25.4 which changed DOM
parsing so that newlines in whitespace-preserved text nodes are
converted into a hard break instead of being collapsed to spaces.
The `preserveWhitespace: true` setting never really had an effect: The
newline in `* first\n  second` already was converted to a space in
a full Markdown - editor - Markdown round-trip before the regression.

Also add several regression tests.

Fixes: #8775

Signed-off-by: Jonas <jonas@freesources.org>

Author: mejo-

src/nodes/BulletList.ts

@@ -12,12 +12,6 @@ import { listInputRule, toggleListCommand } from '../commands'
  * Only there we know the user is not trying to create a task list.
  */
 const BulletList = TiptapBulletList.extend({
-	parseHTML() {
-		return this.parent?.()?.map((rule) =>
-			Object.assign(rule, { preserveWhitespace: true }),
-		)
-	},
-
 	addAttributes() {
 		return {
 			...this.parent?.(),

src/tests/markdown.spec.js

@@ -52,6 +52,26 @@ describe('Markdown though editor', () => {
 		expect(markdownThroughEditor('- foo\n- bar')).toBe('- foo\n- bar')
 		expect(markdownThroughEditor('- foo\n\n- bar')).toBe('- foo\n- bar')
 		expect(markdownThroughEditor('- foo\n\n\n- bar')).toBe('- foo\n- bar')
+		expect(markdownThroughEditor('- foo\n  - bar')).toBe('- foo\n  - bar')
+		expect(markdownThroughEditor('- foo\n  - bar\n- baz')).toBe(
+			'- foo\n  - bar\n- baz',
+		)
+		expect(markdownThroughEditor('- foo  \n  bar  \n  baz')).toBe(
+			'- foo  \n  bar  \n  baz',
+		)
+	})
+	test('ul - no suprious linebreak in nested lists (#8775)', () => {
+		const source = '- foo\n   - bar\n- baz'
+		const tiptap = createRichEditor()
+		tiptap.commands.setContent(markdownit.render(source))
+		// Walk the doc: no hardBreak node should exist in outer list item's paragraph
+		let hasHardBreak = false
+		tiptap.state.doc.descendants((node) => {
+			if (node.type.name === 'hardBreak') {
+				hasHardBreak = true
+			}
+		})
+		expect(hasHardBreak).toBe(false)
 	})
 	test('ol', () => {
 		expect(markdownThroughEditor('1. foo\n2. bar')).toBe('1. foo\n2. bar')