Improve > handling in blockquote in Markdown (#19802)

Co-authored-by: codex <codex@openai.com>
Co-authored-by: fisker Cheung <lionkay@gmail.com>
Co-authored-by: fisker <lionkay@gmail.com>

changelog_unreleased/markdown/19802.md

@@ -0,0 +1,13 @@
+#### Fix Markdown blockquote containing `>` characters (#19802 by @seiyab)
+
+<!-- prettier-ignore -->
+```md
+<!-- Input -->
+> **test**\>
+
+<!-- Prettier stable -->
+> **test**>\>
+
+<!-- Prettier main -->
+> **test**\>
+```

src/language-markdown/print/preprocess.js

@@ -1,5 +1,6 @@
 import htmlWhitespace from "../../utilities/html-whitespace.js";
 import { getOrderedListItemInfo, mapAst, splitText } from "../utilities.js";
+import { getBlockquoteRawText } from "../utilities/get-blockquote-raw-text.js";
 
 // 0x0 ~ 0x10ffff
 const isSingleCharRegex = /^\\?.$/su;
@@ -248,19 +249,6 @@ function splitTextIntoSentences(ast) {
   }
 }
 
-function getBlockquoteRawText(text, node) {
-  const angleBracketsRegex = /^([ \t]*>[ \t]*)*/;
-  const rawLines = text.split("\n");
-  const valueLines = node.value.split("\n");
-  const resultLines = rawLines.map((rawLine, index) => {
-    const valueLine = valueLines[index] ?? "";
-    const leadingTextAngleBrackets =
-      valueLine.match(angleBracketsRegex)[0] ?? "";
-    return rawLine.replace(angleBracketsRegex, leadingTextAngleBrackets);
-  });
-  return resultLines.join("\n");
-}
-
 function transformIndentedCodeblock(ast, options) {
   return mapAst(ast, (node) => {
     if (node.type !== "code") {

src/language-markdown/utilities/get-blockquote-raw-text.js

@@ -0,0 +1,67 @@
+/**
+ * @param line {string}
+ * @returns {number}
+ */
+function countValueLeadingGreaterThan(line) {
+  let count = 0;
+
+  for (const character of line) {
+    if (character === ">") {
+      count++;
+    } else if (character !== " " && character !== "\t") {
+      break;
+    }
+  }
+
+  return count;
+}
+
+/**
+ * @param line {string}
+ * @param greaterThanToKeep {number}
+ * @returns {string}
+ */
+function removeBlockquoteMarkers(line, greaterThanToKeep) {
+  const rawBlockquoteRegexp =
+    /[ \t]*(?:>|\\>|>|>|&#0*62;|&#[xX]0*3[eE];)[ \t]*/y;
+  /** @type {Array<number>} */
+  const greaterThanIndexes = [];
+  let { lastIndex } = rawBlockquoteRegexp;
+
+  let match;
+  while ((match = rawBlockquoteRegexp.exec(line))) {
+    greaterThanIndexes.push(match.index);
+    lastIndex = rawBlockquoteRegexp.lastIndex;
+  }
+
+  if (greaterThanToKeep === 0) {
+    return line.slice(lastIndex);
+  }
+  const firstGreaterThanToKeep = greaterThanIndexes.at(-greaterThanToKeep);
+  /* c8 ignore next 3 */
+  if (firstGreaterThanToKeep === null) {
+    return line;
+  }
+  return line.slice(firstGreaterThanToKeep);
+}
+
+const newLineRegexp = /(?<!\\)(?:\\\\)*(?:&NewLine;|&#0*10;|&#[Xx]0*[Aa];)/g;
+function getBlockquoteRawText(text, node) {
+  const rawLines = text.split("\n");
+  const valueLines = node.value.split("\n");
+  let valueIndex = 0;
+  const resultLines = rawLines.map((rawLine, index) => {
+    const valueLine = valueLines[valueIndex++];
+    const decodedNewlines = rawLine.matchAll(newLineRegexp);
+    valueIndex += [...decodedNewlines].length;
+
+    if (index === 0) {
+      return rawLine;
+    }
+    const valueMarkerCount = countValueLeadingGreaterThan(valueLine);
+    return removeBlockquoteMarkers(rawLine, valueMarkerCount);
+  });
+  return resultLines.join("\n");
+}
+
+export { getBlockquoteRawText };

tests/format/markdown/blockquote/prose-wrap-always/__snapshots__/format.test.js.snap

@@ -0,0 +1,162 @@
+// Jest Snapshot v1, https://jestjs.io/docs/snapshot-testing
+
+exports[`encoded-greater-than.md - {"proseWrap":"always"} format 1`] = `
+====================================options=====================================
+parsers: ["markdown"]
+proseWrap: "always"
+                                                      printWidth: 80 (default) |
+=====================================input======================================
+> > a
+> > b
+
+> > a
+> > b
+
+> > a
+> > b
+
+> > a
+> > b
+
+> > a
+> > b
+
+> > a
+> > b
+
+=====================================output=====================================
+> > a > b
+
+> > a > b
+
+> > a > b
+
+> > a > b
+
+> > a > b
+
+> > a > b
+
+================================================================================
+`;
+
+exports[`encoded-newline.md - {"proseWrap":"always"} format 1`] = `
+====================================options=====================================
+parsers: ["markdown"]
+proseWrap: "always"
+                                                      printWidth: 80 (default) |
+=====================================input======================================
+> a 
 b
+> > c
+
+> a 
 b
+> > c
+
+> a 
 b
+> \\> c
+
+> a 
 b
+    > c
+
+> a &newline; b
+> \\> c
+
+> a \\
 b
+> \\> c
+
+> a \\\\
 b
+> \\> c
+
+=====================================output=====================================
+> a 
 b > c
+
+> a 
 b > c
+
+> a 
 b \\> c
+
+> a 
 b > c
+
+> a &newline; b \\> c
+
+> a \\
 b \\> c
+
+> a \\\\
 b \\> c
+
+================================================================================
+`;
+
+exports[`escaped-greater-than.md - {"proseWrap":"always"} format 1`] = `
+====================================options=====================================
+parsers: ["markdown"]
+proseWrap: "always"
+                                                      printWidth: 80 (default) |
+=====================================input======================================
+> \\>
+> \\>
+
+> *text* > \\>
+
+> abc
+> \\> abc
+
+=====================================output=====================================
+> \\> \\>
+
+> _text_ > \\>
+
+> abc \\> abc
+
+================================================================================
+`;
+
+exports[`indented-greater-than.md - {"proseWrap":"always"} format 1`] = `
+====================================options=====================================
+parsers: ["markdown"]
+proseWrap: "always"
+                                                      printWidth: 80 (default) |
+=====================================input======================================
+> first
+    > second
+
+> first
+	> second
+
+> > first
+    > second
+
+> first
+    >> second
+
+=====================================output=====================================
+> first > second
+
+> first > second
+
+> > first > second
+
+> first >> second
+
+================================================================================
+`;
+
+exports[`issue-19528.md - {"proseWrap":"always"} format 1`] = `
+====================================options=====================================
+parsers: ["markdown"]
+proseWrap: "always"
+                                                      printWidth: 80 (default) |
+=====================================input======================================
+> test\\>
+
+> **test**\\>
+
+> \\>
+
+=====================================output=====================================
+> test\\>
+
+> **test**\\>
+
+> \\>
+
+================================================================================
+`;

tests/format/markdown/blockquote/prose-wrap-always/encoded-greater-than.md

@@ -0,0 +1,17 @@
+> > a
+> > b
+
+> > a
+> > b
+
+> > a
+> > b
+
+> > a
+> > b
+
+> > a
+> > b
+
+> > a
+> > b

tests/format/markdown/blockquote/prose-wrap-always/encoded-newline.md

@@ -0,0 +1,20 @@
+> a 
 b
+> > c
+
+> a 
 b
+> > c
+
+> a 
 b
+> \> c
+
+> a 
 b
+    > c
+
+> a &newline; b
+> \> c
+
+> a \
 b
+> \> c
+
+> a \\
 b
+> \> c

tests/format/markdown/blockquote/prose-wrap-always/escaped-greater-than.md

@@ -0,0 +1,7 @@
+> \>
+> \>
+
+> *text* > \>
+
+> abc
+> \> abc

tests/format/markdown/blockquote/prose-wrap-always/format.test.js

@@ -0,0 +1 @@
+runFormatTest(import.meta, ["markdown"], { proseWrap: "always" });

tests/format/markdown/blockquote/prose-wrap-always/indented-greater-than.md

@@ -0,0 +1,11 @@
+> first
+    > second
+
+> first
+	> second
+
+> > first
+    > second
+
+> first
+    >> second

tests/format/markdown/blockquote/prose-wrap-always/issue-19528.md

@@ -0,0 +1,5 @@
+> test\>
+
+> **test**\>
+
+> \>