From 10cf62c2b41d975115f996e13ba4bfe005ffaa4b Mon Sep 17 00:00:00 2001
From: Chen <99816898+donteatfriedrice@users.noreply.github.com>
Date: Fri, 23 May 2025 17:56:31 +0800
Subject: [PATCH] fix(editor): handle polynomial regular expression used on
 uncontrolled data issue

---
 .../adapters/preprocessor.unit.spec.ts        |   6 +
 .../src/adapters/markdown/preprocessor.ts     | 120 ++++++++++++++++--
 2 files changed, 115 insertions(+), 11 deletions(-)

diff --git a/blocksuite/affine/inlines/footnote/src/__tests__/adapters/preprocessor.unit.spec.ts b/blocksuite/affine/inlines/footnote/src/__tests__/adapters/preprocessor.unit.spec.ts
index 5ef0f62cea..fa909f37a7 100644
--- a/blocksuite/affine/inlines/footnote/src/__tests__/adapters/preprocessor.unit.spec.ts
+++ b/blocksuite/affine/inlines/footnote/src/__tests__/adapters/preprocessor.unit.spec.ts
@@ -33,6 +33,12 @@ describe('FootnoteReferenceMarkdownPreprocessorExtension', () => {
     expect(preprocessFootnoteReference(content)).toBe(expected);
   });
 
+  it('should not add space when there is not a footnote reference after URL', () => {
+    const content = 'https://example.com[^label';
+    const expected = 'https://example.com[^label';
+    expect(preprocessFootnoteReference(content)).toBe(expected);
+  });
+
   it('should handle multiple footnote references with mixed URL and non-URL text', () => {
     const content = 'https://example.com[^1]normal text[^2]http://test.com[^3]';
     const expected =
diff --git a/blocksuite/affine/inlines/footnote/src/adapters/markdown/preprocessor.ts b/blocksuite/affine/inlines/footnote/src/adapters/markdown/preprocessor.ts
index 4db7e06a35..676963860a 100644
--- a/blocksuite/affine/inlines/footnote/src/adapters/markdown/preprocessor.ts
+++ b/blocksuite/affine/inlines/footnote/src/adapters/markdown/preprocessor.ts
@@ -17,6 +17,113 @@ function isUrl(str: string): boolean {
   }
 }
 
+// Types for footnote parser tokens
+type Token = {
+  type: 'TEXT' | 'FOOTNOTE_REF' | 'SPACE';
+  value: string;
+};
+
+class FootnoteReferenceParser {
+  private pos: number = 0;
+  private input: string = '';
+  private tokens: Token[] = [];
+
+  // Lexer: Convert input string into tokens
+  private tokenize(input: string): Token[] {
+    this.input = input;
+    this.pos = 0;
+    this.tokens = [];
+
+    while (this.pos < this.input.length) {
+      const char = this.input[this.pos];
+
+      // Handle spaces
+      if (char === ' ') {
+        this.tokens.push({ type: 'SPACE', value: ' ' });
+        this.pos++;
+        continue;
+      }
+
+      // Match footnote reference [^...]
+      if (char === '[' && this.input[this.pos + 1] === '^') {
+        let footnoteRef = '[^';
+        this.pos += 2;
+
+        while (this.pos < this.input.length && this.input[this.pos] !== ']') {
+          footnoteRef += this.input[this.pos];
+          this.pos++;
+        }
+
+        if (this.pos < this.input.length) {
+          footnoteRef += ']';
+          this.pos++;
+          // Only add as footnote reference if it's not followed by ':'
+          if (this.pos >= this.input.length || this.input[this.pos] !== ':') {
+            this.tokens.push({ type: 'FOOTNOTE_REF', value: footnoteRef });
+          } else {
+            this.tokens.push({ type: 'TEXT', value: footnoteRef });
+          }
+        } else {
+          // If we didn't find a closing bracket, treat it as regular text
+          this.tokens.push({ type: 'TEXT', value: footnoteRef });
+        }
+        continue;
+      }
+
+      // Handle regular text
+      let text = '';
+      while (
+        this.pos < this.input.length &&
+        this.input[this.pos] !== ' ' &&
+        !(this.input[this.pos] === '[' && this.input[this.pos + 1] === '^')
+      ) {
+        text += this.input[this.pos];
+        this.pos++;
+      }
+      if (text) {
+        this.tokens.push({ type: 'TEXT', value: text });
+      }
+    }
+
+    return this.tokens;
+  }
+
+  // Process tokens to add spaces after URLs
+  private processTokens(tokens: Token[]): string {
+    let result = '';
+    let i = 0;
+
+    while (i < tokens.length) {
+      const token = tokens[i];
+
+      if (token.type === 'TEXT') {
+        result += token.value;
+
+        // Check if next token is a footnote reference
+        if (
+          i + 1 < tokens.length &&
+          tokens[i + 1].type === 'FOOTNOTE_REF' &&
+          isUrl(token.value)
+        ) {
+          result += ' ';
+        }
+      } else {
+        result += token.value;
+      }
+
+      i++;
+    }
+
+    return result;
+  }
+
+  // Main processing function
+  public process(input: string): string {
+    const tokens = this.tokenize(input);
+    return this.processTokens(tokens);
+  }
+}
+
 /**
  * Preprocess footnote references to avoid markdown link parsing
  * Only add space when footnote reference follows a URL
@@ -29,17 +136,8 @@ function isUrl(str: string): boolean {
  * ```
  */
 export function preprocessFootnoteReference(content: string) {
-  return content.replace(
-    /([^\s]+?)(\[\^[^\]]+\])(?!:)/g,
-    (match, prevText, footnoteRef) => {
-      // Only add space if the previous text is a URL
-      if (isUrl(prevText)) {
-        return prevText + ' ' + footnoteRef;
-      }
-      // Otherwise return the original match
-      return match;
-    }
-  );
+  const parser = new FootnoteReferenceParser();
+  return parser.process(content);
 }
 
 const footnoteReferencePreprocessor: MarkdownAdapterPreprocessor = {