snapshot of https://github.com/ueberdosis/tiptap.git (main, 2026-09-15; blobs > 131072B omitted)

packages/markdown/tests/integration/block-token-blank-line.test.ts

@@ -0,0 +1,88 @@
+import { Blockquote } from '@tiptap/extension-blockquote'
+import { Document } from '@tiptap/extension-document'
+import { Heading } from '@tiptap/extension-heading'
+import { BulletList, ListItem, OrderedList } from '@tiptap/extension-list'
+import { Paragraph } from '@tiptap/extension-paragraph'
+import { Table, TableCell, TableHeader, TableRow } from '@tiptap/extension-table'
+import { Text } from '@tiptap/extension-text'
+import { MarkdownManager } from '@tiptap/markdown'
+import { beforeEach, describe, expect, it } from 'vite-plus/test'
+
+describe('Blank lines after a block token', () => {
+  let markdownManager: MarkdownManager
+
+  beforeEach(() => {
+    markdownManager = new MarkdownManager({
+      extensions: [
+        Document,
+        Paragraph,
+        Text,
+        Heading,
+        Blockquote,
+        BulletList,
+        OrderedList,
+        ListItem,
+        Table,
+        TableRow,
+        TableHeader,
+        TableCell,
+      ],
+    })
+  })
+
+  // Heading tokenizer absorbs trailing blank lines into the token's raw,
+  // so we recover them here to prevent losing a blank line per parse cycle.
+  it('preserves a blank line between a heading and the following text', () => {
+    const doc = markdownManager.parse('# Title\n\n\n\nBody')
+
+    expect(doc.content).toEqual([
+      { type: 'heading', attrs: { level: 1 }, content: [{ type: 'text', text: 'Title' }] },
+      { type: 'paragraph', content: [] },
+      { type: 'paragraph', content: [{ type: 'text', text: 'Body' }] },
+    ])
+  })
+
+  it('does not create an empty paragraph when a heading is directly followed by text', () => {
+    const doc = markdownManager.parse('# Title\n\nBody')
+
+    expect(doc.content).toEqual([
+      { type: 'heading', attrs: { level: 1 }, content: [{ type: 'text', text: 'Title' }] },
+      { type: 'paragraph', content: [{ type: 'text', text: 'Body' }] },
+    ])
+  })
+
+  it('round-trips blank lines after a heading without eroding them', () => {
+    const markdown = '# Title\n\n\n\nBody'
+
+    const once = markdownManager.serialize(markdownManager.parse(markdown))
+    const twice = markdownManager.serialize(markdownManager.parse(once))
+
+    expect(twice).toBe(once)
+  })
+
+  it('normalizes an explicit   paragraph without eroding the blank line', () => {
+    const once = markdownManager.serialize(markdownManager.parse('# Title\n\n \n\nBody'))
+    const twice = markdownManager.serialize(markdownManager.parse(once))
+
+    expect(once).toBe('# Title\n\n\n\nBody')
+    expect(twice).toBe(once)
+  })
+
+  // Tables absorb trailing blank lines the same way.
+  it('preserves a blank line between a table and the following text', () => {
+    const doc = markdownManager.parse('| a | b |\n| --- | --- |\n| 1 | 2 |\n\n\n\nAfter')
+    const kinds = (doc.content ?? []).map(node => node.type)
+
+    expect(kinds).toEqual(['table', 'paragraph', 'paragraph'])
+    expect(doc.content?.[1]).toEqual({ type: 'paragraph', content: [] })
+  })
+
+  // HTML blocks absorb trailing blank lines the same way.
+  it('preserves a blank line after a raw HTML block (parsed as paragraph)', () => {
+    const doc = markdownManager.parse('<div>test</div>\n\n\n\nAfter')
+    const kinds = (doc.content ?? []).map(node => node.type)
+
+    expect(kinds).toEqual(['paragraph', 'paragraph', 'paragraph'])
+    expect(doc.content?.[1]).toEqual({ type: 'paragraph', content: [] })
+  })
+})