fix: Markdown escape characters left in titles on import (#8102)

server/commands/documentImporter.ts

@@ -1,5 +1,4 @@
 import emojiRegex from "emoji-regex";
-import escapeRegExp from "lodash/escapeRegExp";
 import truncate from "lodash/truncate";
 import parseTitle from "@shared/utils/parseTitle";
 import { DocumentValidation } from "@shared/validations";
@@ -51,10 +50,7 @@ async function documentImporter({
   if (text.trim().startsWith("# ")) {
     const result = parseTitle(text);
     title = result.title;
-    text = text
-      .trim()
-      .replace(new RegExp(`#\\s+${escapeRegExp(title)}`), "")
-      .trimStart();
+    text = text.replace(/^.+(\n|$)/, "");
   }
 
   // Replace any <br> generated by the turndown plugin with escaped newlines

server/utils/turndown/index.ts

@@ -1,5 +1,6 @@
 import { taskListItems, strikethrough } from "@joplin/turndown-plugin-gfm";
 import TurndownService from "turndown";
+import { escape } from "@shared/utils/markdown";
 import breaks from "./breaks";
 import emptyLists from "./emptyLists";
 import emptyParagraph from "./emptyParagraph";
@@ -41,36 +42,6 @@ const service = new TurndownService({
   .use(breaks)
   .use(emptyLists);
 
-const escapes: [RegExp, string][] = [
-  [/\\/g, "\\\\"],
-  [/\*/g, "\\*"],
-  [/^-/g, "\\-"],
-  [/^\+ /g, "\\+ "],
-  [/^(=+)/g, "\\$1"],
-  [/^(#{1,6}) /g, "\\$1 "],
-  [/`/g, "\\`"],
-  [/^~~~/g, "\\~~~"],
-  [/\[/g, "\\["],
-  [/\]/g, "\\]"],
-  [/\(/g, "\\("], // OLN-91
-  [/\)/g, "\\)"], // OLN-91
-  [/^>/g, "\\>"],
-  [/_/g, "\\_"],
-  [/^(\d+)\. /g, "$1\\. "],
-  [/\$/g, "\\$"],
-];
-
-/**
- * Overrides the Markdown escaping, as documented here:
- * https://github.com/mixmark-io/turndown/blob/4499b5c313d30a3189a58fdd74fc4ed4b2428afd/README.md#overriding-turndownserviceprototypeescape
- *
- * @param text The string to escape
- * @returns A string with Markdown syntax escaped
- */
-service.escape = function (text) {
-  return escapes.reduce(function (accumulator, escape) {
-    return accumulator.replace(escape[0], escape[1]);
-  }, text);
-};
+service.escape = escape;
 
 export default service;

shared/utils/markdown.ts

@@ -0,0 +1,40 @@
+const escapes: [RegExp, string][] = [
+  [/\\/g, "\\\\"],
+  [/\*/g, "\\*"],
+  [/^-/g, "\\-"],
+  [/^\+ /g, "\\+ "],
+  [/^(=+)/g, "\\$1"],
+  [/^(#{1,6}) /g, "\\$1 "],
+  [/`/g, "\\`"],
+  [/^~~~/g, "\\~~~"],
+  [/\[/g, "\\["],
+  [/\]/g, "\\]"],
+  [/\(/g, "\\("], // OLN-91
+  [/\)/g, "\\)"], // OLN-91
+  [/^>/g, "\\>"],
+  [/_/g, "\\_"],
+  [/^(\d+)\. /g, "$1\\. "],
+  [/\$/g, "\\$"],
+];
+
+/**
+ * Escape markdown characters in a string
+ *
+ * @param text - The text to escape
+ * @returns The escaped text
+ */
+export const escape = function (text: string) {
+  return escapes.reduce(function (accumulator, esc) {
+    return accumulator.replace(esc[0], esc[1]);
+  }, text);
+};
+
+/**
+ * Unescape markdown characters in a string
+ *
+ * @param text - The text to unescape
+ * @returns The unescaped text
+ */
+export const unescape = function (text: string) {
+  return text.replace(/\\([\\*+-\d.])/g, "$1");
+};

shared/utils/parseTitle.test.ts

@@ -3,14 +3,23 @@ import parseTitle from "./parseTitle";
 it("should trim the title", () => {
   expect(parseTitle(`#    Lots of space     `).title).toBe("Lots of space");
 });
+
+it("should remove escaped characters", () => {
+  expect(parseTitle(`# 1\\. Title`).title).toBe("1. Title");
+  expect(parseTitle(`# 12\\. Title`).title).toBe("12. Title");
+  expect(parseTitle(`# 1\\.2 Title`).title).toBe("1.2 Title");
+});
+
 it("should extract first title", () => {
   expect(parseTitle(`# Title one\n# Title two`).title).toBe("Title one");
 });
+
 it("should parse emoji if first character", () => {
   const parsed = parseTitle(`# 😀 Title`);
   expect(parsed.title).toBe("😀 Title");
   expect(parsed.emoji).toBe("😀");
 });
+
 it("should not parse emoji if not first character", () => {
   const parsed = parseTitle(`# Title 🌈`);
   expect(parsed.title).toBe("Title 🌈");

shared/utils/parseTitle.ts

@@ -1,11 +1,12 @@
 import emojiRegex from "emoji-regex";
+import { unescape } from "./markdown";
 
 export default function parseTitle(text = "") {
   const regex = emojiRegex();
 
   // find and extract title
   const firstLine = text.trim().split(/\r?\n/)[0];
-  const title = firstLine.replace(/^#/, "").trim();
+  const title = unescape(firstLine.replace(/^#/, "").trim());
 
   // find and extract first emoji
   const matches = regex.exec(title);