feat: escape special characters (#20)

Author: mdjastrzebski

.gitignore

@@ -32,3 +32,4 @@ buck-out/
 lib/
 
 /.idea
+coverage/

src/__tests__/compiler.test.tsx

@@ -1,5 +1,5 @@
 import { buildPattern, buildRegex } from '../compiler';
-import { oneOrMore, optionally, one, zeroOrMore } from '../quantifiers/base';
+import { one, oneOrMore, optionally, zeroOrMore } from '../quantifiers/base';
 import { repeat } from '../quantifiers/repeat';
 
 test('basic quantifies', () => {
@@ -28,3 +28,31 @@ test('regex constructor', () => {
   expect(buildRegex('a').test('a')).toBeTruthy();
   expect(buildRegex('a').test('b')).toBeFalsy();
 });
+
+test('"buildPattern" escapes special characters', () => {
+  expect(buildPattern('.')).toBe('\\.');
+  expect(buildPattern('*')).toBe('\\*');
+  expect(buildPattern('+')).toBe('\\+');
+  expect(buildPattern('?')).toBe('\\?');
+  expect(buildPattern('^')).toBe('\\^');
+  expect(buildPattern('$')).toBe('\\$');
+  expect(buildPattern('{')).toBe('\\{');
+  expect(buildPattern('}')).toBe('\\}');
+  expect(buildPattern('|')).toBe('\\|');
+  expect(buildPattern('[')).toBe('\\[');
+  expect(buildPattern(']')).toBe('\\]');
+  expect(buildPattern('\\')).toBe('\\\\');
+
+  expect(buildPattern('*.*')).toBe('\\*\\.\\*');
+
+  expect(buildPattern(oneOrMore('.*'), zeroOrMore('[]{}'))).toBe(
+    '(?:\\.\\*)+(?:\\[\\]\\{\\})*'
+  );
+});
+
+test('buildRegex throws error on unknown element', () => {
+  expect(() =>
+    // @ts-expect-error intentionally passing incorrect object
+    buildRegex({ type: 'unknown' })
+  ).toThrowErrorMatchingInlineSnapshot(`"Unknown elements type unknown"`);
+});

src/compiler.ts

@@ -2,6 +2,7 @@ import type { RegexElement } from './types';
 import { characterClasses, isCharacterClass } from './character-classes';
 import { baseQuantifiers, isBaseQuantifier } from './quantifiers/base';
 import { compileRepeat } from './quantifiers/repeat';
+import { escapeText } from './utils';
 
 /**
  * Generate RegExp object for elements.
@@ -31,20 +32,20 @@ function compileList(elements: RegexElement[]): string {
 
 function compileSingle(element: RegexElement): string {
   if (typeof element === 'string') {
-    return element;
+    return escapeText(element);
   }
 
   if (isCharacterClass(element)) {
     return characterClasses[element.type];
   }
 
-  const compiledChildren = compileList(element.children);
-
   if (element.type === 'repeat') {
+    const compiledChildren = compileList(element.children);
     return compileRepeat(element.config, compiledChildren);
   }
 
   if (isBaseQuantifier(element)) {
+    const compiledChildren = compileList(element.children);
     const compiler = baseQuantifiers[element.type];
     return compiler(compiledChildren);
   }

src/utils.ts

@@ -7,3 +7,8 @@
 export function wrapGroup(regex: string): string {
   return regex.length === 1 ? regex : `(?:${regex})`;
 }
+
+// Source: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Regular_expressions#escaping
+export function escapeText(text: string) {
+  return text.replace(/[.*+?^${}()|[\]\\]/g, '\\$&'); // $& means the whole matched string
+}