chore(markdown): update the table fix to current main

Keep the contributor's table regression tests and source adaptation while
bringing in current browser tooling for supported CI runners.

Co-authored-by: Okan YILDIRIM <63511519+ragokan@users.noreply.github.com>
Co-authored-by: thsrite <thsrite@gmail.com>
Co-authored-by: Woosuk Nam <me@suk.kr>

packages/app/src/components/markdown/renderer.test.ts

@@ -130,6 +130,186 @@ it("preserves empty table cells while escaping spoilers in cell text", () => {
   );
 });
 
+it("renders a table directly after a paragraph through the library parser", async () => {
+  const source = "**Notes**\n| Item | Detail |\n| --- | --- |\n| A | one |\n| B | two |";
+  const separated = "**Notes**\n\n| Item | Detail |\n| --- | --- |\n| A | one |\n| B | two |";
+  const flags = { hardSoftBreaks: true, latexMath: false };
+  const expected = await parseMarkdown(separated, flags);
+  expect(expected.children?.map((node) => node.type)).toEqual(["Paragraph", "Table"]);
+  expect(await parseMarkdown(prepareMarkdownSource(source).render(), flags)).toEqual(expected);
+});
+
+it.each([
+  [
+    "a quote",
+    "> Notes\n> | Item |\n> | --- |\n> | A |",
+    "> Notes\n>\n> | Item |\n> | --- |\n> | A |",
+    ["Document", ["Blockquote", "Paragraph", "Table"]],
+  ],
+  [
+    "a list item",
+    "- Notes\n  | Item |\n  | --- |\n  | A |\n- Next",
+    "- Notes\n\n  | Item |\n  | --- |\n  | A |\n- Next",
+    ["Document", ["UnorderedList", ["ListItem", "Paragraph", "Table"], ["ListItem", "Paragraph"]]],
+  ],
+  [
+    "a quoted list item",
+    "> - Notes\n>   | Item |\n>   | --- |\n>   | A |",
+    "> - Notes\n>\n>   | Item |\n>   | --- |\n>   | A |",
+    ["Document", ["Blockquote", ["UnorderedList", ["ListItem", "Paragraph", "Table"]]]],
+  ],
+  [
+    "a quote inside a list item",
+    "- > Notes\n  > | Item |\n  > | --- |\n  > | A |",
+    "- > Notes\n  >\n  > | Item |\n  > | --- |\n  > | A |",
+    ["Document", ["UnorderedList", ["ListItem", ["Blockquote", "Paragraph", "Table"]]]],
+  ],
+  [
+    "aligned columns without outer pipes",
+    "Notes\nItem | Detail\n:--- | ---:\nA | one",
+    "Notes\n\nItem | Detail\n:--- | ---:\nA | one",
+    ["Document", "Paragraph", "Table"],
+  ],
+  [
+    "an entity at the start of a header without outer pipes",
+    "Notes\nAT&T | Detail\n--- | ---\nA | one",
+    "Notes\n\nAT\\&T | Detail\n--- | ---\nA | one",
+    ["Document", "Paragraph", "Table"],
+  ],
+  [
+    "a local image at the start of a header without outer pipes",
+    "Notes\n![Local](assets/a.png) | Detail\n--- | ---\nA | one",
+    "Notes\n\nLocal | Detail\n--- | ---\nA | one",
+    ["Document", "Paragraph", "Table"],
+  ],
+  [
+    "CRLF lines",
+    "Notes\r\n| Item |\r\n| --- |\r\n| A |",
+    "Notes\r\n\r\n| Item |\r\n| --- |\r\n| A |",
+    ["Document", "Paragraph", "Table"],
+  ],
+  [
+    "CR lines",
+    "Notes\r| Item |\r| --- |\r| A |",
+    "Notes\r\r| Item |\r| --- |\r| A |",
+    ["Document", "Paragraph", "Table"],
+  ],
+  [
+    "a quote with CRLF lines",
+    "> Notes\r\n> | Item |\r\n> | --- |\r\n> | A |",
+    "> Notes\r\n>\r\n> | Item |\r\n> | --- |\r\n> | A |",
+    ["Document", ["Blockquote", "Paragraph", "Table"]],
+  ],
+  [
+    "a list item with CRLF lines",
+    "- Notes\r\n  | Item |\r\n  | --- |\r\n  | A |\r\n- Next",
+    "- Notes\r\n\r\n  | Item |\r\n  | --- |\r\n  | A |\r\n- Next",
+    ["Document", ["UnorderedList", ["ListItem", "Paragraph", "Table"], ["ListItem", "Paragraph"]]],
+  ],
+  [
+    "a quote with CR lines",
+    "> Notes\r> | Item |\r> | --- |\r> | A |",
+    "> Notes\r>\r> | Item |\r> | --- |\r> | A |",
+    ["Document", ["Blockquote", "Paragraph", "Table"]],
+  ],
+  [
+    "a list item with CR lines",
+    "- Notes\r  | Item |\r  | --- |\r  | A |\r- Next",
+    "- Notes\r\r  | Item |\r  | --- |\r  | A |\r- Next",
+    ["Document", ["UnorderedList", ["ListItem", "Paragraph", "Table"], ["ListItem", "Paragraph"]]],
+  ],
+  [
+    "a non-interrupting ordered marker with tildes",
+    "Intro\n2. ~~~\n   Notes\n   | A |\n   | --- |\n   | x |",
+    "Intro\n2. ~~~\n   Notes\n\n   | A |\n   | --- |\n   | x |",
+    ["Document", "Paragraph", "Table"],
+  ],
+  [
+    "a non-interrupting ordered marker with backticks",
+    "Notes\n2. ```\n   | Item |\n   | --- |\n   | A |",
+    "Notes\n2. ```\n\n   | Item |\n   | --- |\n   | A |",
+    ["Document", "Paragraph", "Table"],
+  ],
+])("preserves tables after paragraphs in %s", async (_name, source, separated, expected) => {
+  const rendered = prepareMarkdownSource(source).render();
+  expect(rendered).toEqual(separated);
+  const flags = { hardSoftBreaks: true, latexMath: false };
+  function blocks(node: ASTNode): unknown {
+    if (node.type === "Paragraph" || node.type === "Table") return node.type;
+    return [node.type, ...(node.children ?? []).map(blocks)];
+  }
+  expect(blocks(await parseMarkdown(rendered, flags))).toEqual(expected);
+});
+
+it.each([
+  [
+    "a literal backslash",
+    "Notes\\\n| Item |\n| --- |\n| A |",
+    "Notes\\\n\n| Item |\n| --- |\n| A |",
+    "Notes\\",
+  ],
+  [
+    "an entity and literal backslash with LF lines",
+    "AT&T\\\n| Item |\n| --- |\n| A |",
+    "AT\\&T\\\n\n| Item |\n| --- |\n| A |",
+    "AT&T\\",
+  ],
+  [
+    "an entity and literal backslash with CRLF lines",
+    "AT&T\\\r\n| Item |\r\n| --- |\r\n| A |",
+    "AT\\&T\\\r\n\r\n| Item |\r\n| --- |\r\n| A |",
+    "AT&T\\",
+  ],
+  [
+    "an entity and literal backslash with CR lines",
+    "AT&T\\\r| Item |\r| --- |\r| A |",
+    "AT\\&T\\\r\r| Item |\r| --- |\r| A |",
+    "AT&T\\",
+  ],
+])(
+  "preserves paragraph text before an interrupting table in %s",
+  async (_name, source, separated, text) => {
+    const rendered = prepareMarkdownSource(source).render();
+    expect(rendered).toEqual(separated);
+    const tree = await parseMarkdown(rendered, { hardSoftBreaks: true, latexMath: false });
+    expect(tree.children?.map((node) => node.type)).toEqual(["Paragraph", "Table"]);
+    expect(tree.children?.[0]?.children?.map((node) => node.content).join("")).toEqual(text);
+  },
+);
+
+it.each([
+  "```yaml\n---\nkey: value\n---\n```\n",
+  "~~~md\n| Item |\n| --- |\n| A |\n~~~\n",
+  "````md\n```\n| Item |\n| --- |\n```\n`````\n",
+  "> ```md\n> | Item |\n> | --- |\n> ```\n",
+  "> ```md\n> Code\n",
+  "- ```md\n  Code\n",
+  "    ```\n\n",
+  "```not`a-fence\n",
+])("renders a table after a fence without changing its code: %s", async (prefix) => {
+  const table = "| Item |\n| --- |\n| A |";
+  const rendered = prepareMarkdownSource(`${prefix}Notes\n${table}`).render();
+  expect(rendered).toEqual(`${prefix}Notes\n\n${table}`);
+  const tree = await parseMarkdown(rendered, { hardSoftBreaks: true, latexMath: false });
+  expect(tree.children?.slice(-2).map((node) => node.type)).toEqual(["Paragraph", "Table"]);
+});
+
+it.each([
+  "| Item | Detail |\n| --- | --- |\n" + "| A | one |\n".repeat(100),
+  "Notes\n\n| Item | Detail |\n| --- | --- |\n" + "| A | one |\n".repeat(100),
+  "Notes\n\n| Item |\n| --- |\n| A |",
+  "| Item |\n| --- |\n| A |",
+  "```md\nNotes\n| Item |\n| --- |\n| A |\n```",
+  "```yaml\n---\nkey: value\n---\n```",
+  "```yaml\n---\nkey: value",
+  "~~~md\n| Item |\n| --- |\n| A |\n~~~",
+  "````md\n```\n| Item |\n| --- |\n```\n````",
+  "    Notes\n    | Item |\n    | --- |\n    | A |",
+  "Notes\n| Item | Detail |\n| --- |\n| A |",
+])("leaves separated tables and table-like text unchanged: %s", (source) => {
+  expect(prepareMarkdownSource(source).render()).toEqual(source);
+});
+
 it.each([
   "src/components/message.tsx",
   "C:\\repo\\file.ts",

packages/app/src/components/markdown/source.ts

@@ -138,6 +138,10 @@ export function prepareMarkdownSource(
     !source.includes("![") &&
     !source.includes("||") &&
     !source.includes("&") &&
+    // Match content, header, and delimiter lines; quote-only lines count as blank.
+    !/^[\t >]*[^\t >\r\n][^\r\n]*(?:\r\n|\r|\n)[^\r\n]+(?:\r\n|\r|\n)(?=[\t >:|-]*\|)[\t >|]*:?-[-\t :|]*$/m.test(
+      source,
+    ) &&
     !(workspaceRoot && source.includes("]("))
   ) {
     return { images, ...metadata, render: () => source };
@@ -212,11 +216,25 @@ export function prepareMarkdownSource(
       if (lineEnd < end) split(lineEnd);
     }
   }
-  function visit(node: Nodes) {
-    const start = node.position?.start.offset;
-    const end = node.position?.end.offset;
-    if (start !== undefined && end !== undefined) {
+  function visit(node: Nodes, previous?: Nodes) {
+    const position = node.position;
+    if (position?.start.offset !== undefined && position.end.offset !== undefined) {
+      const start = position.start.offset;
+      const end = position.end.offset;
       if (node.type === "paragraph") splitStandaloneImages(node, start, end);
+      if (
+        node.type === "table" &&
+        previous?.type === "paragraph" &&
+        previous.position?.end.line === position.start.line - 1
+      ) {
+        // MD4C only recognizes a table header on the first line of a paragraph.
+        const lineStart =
+          Math.max(source.lastIndexOf("\n", start - 1), source.lastIndexOf("\r", start - 1)) + 1;
+        const newline = source.slice(lineStart - 2, lineStart).replace(/[^\r\n]/g, "");
+        const blank = source.slice(lineStart, start).replace(/[^>]/g, " ").trimEnd();
+        const boundary = lineStart - newline.length;
+        edits.push({ start: boundary, end: lineStart, text: `${newline}${blank}${newline}` });
+      }
       if (node.type === "link") {
         const title = workspaceHoverTitle(node, source.slice(start, end), workspaceRoot);
         if (title !== null)
@@ -244,7 +262,8 @@ export function prepareMarkdownSource(
         if (text !== source.slice(start, end)) edits.push({ start, end, text });
       }
     }
-    if ("children" in node) node.children.forEach(visit);
+    if ("children" in node)
+      node.children.forEach((child, index) => visit(child, node.children[index - 1]));
   }
   visit(getTree());
   return {