From 18cfb94c3eb4de7506a318ee6cbc701236f83388 Mon Sep 17 00:00:00 2001
From: Artem Sapegin <artem@sapegin.ru>
Date: Wed, 22 Jun 2022 09:41:07 +0200
Subject: [PATCH] fix: Improve matching of words near punctuation

Fixes #28
---
 index.js |  5 +++--
 test.js  | 26 ++++++++++++++++++++++++++
 2 files changed, 29 insertions(+), 2 deletions(-)

diff --git a/index.js b/index.js
index b22aee4..4fd6180 100644
--- a/index.js
+++ b/index.js
@@ -133,11 +133,12 @@ function readTermsFile(filepath) {
  * @param {string} pattern
  */
 function getExactMatchRegExp(pattern) {
+	const punctuation = '[\\.,;\'")]';
 	return new RegExp(
 		// 1. Beginning of the string, or any character that isn't "-" or alphanumeric
 		// 2. Exact match of the pattern
-		// 3. Space, ". ", "." at the end of the string, end of the string
-		`(?<=^|[^-\\w])\\b${pattern}\\b(?= |\\. |\\.$|$)`,
+		// 3. Space, punctuation + space, punctuation + punctuation, or punctuation at the end of the string, end of the string
+		`(?<=^|[^-\\w])\\b${pattern}\\b(?= |${punctuation} |${punctuation}${punctuation}|${punctuation}$|$)`,
 		'ig'
 	);
 }
diff --git a/test.js b/test.js
index 47b49e0..a94c039 100644
--- a/test.js
+++ b/test.js
@@ -119,6 +119,19 @@ describe('getMultipleWordRegExp', () => {
 		expect(result[0]).toBe('javascript');
 	});
 
+	it.each([
+		['Bad Javascript. Is it bad?'],
+		['Bad Javascript, is it bad?'],
+		['Bad Javascript; is it bad?'],
+		['Bad (Javascript) is it bad?'],
+		['Bad "Javascript" is it bad?'],
+		["Bad 'Javascript' is it bad?"],
+		['Bad "Javascript", is it bad?'],
+	])('should match a pattern regardless of punctuation: %s', string => {
+		const result = getMultipleWordRegExp(variants).exec(string);
+		expect(result).toBeTruthy();
+	});
+
 	it('should not match a pattern in as a part of a file name', () => {
 		const result = getMultipleWordRegExp(variants).exec('javascript.md');
 		expect(result).toBeFalsy();
@@ -148,6 +161,19 @@ describe('getExactMatchRegExp', () => {
 		expect(regexp.test('Webpack')).toBeTruthy();
 	});
 
+	it.each([
+		['Javascript.'],
+		['Javascript,'],
+		['Javascript;'],
+		['(Javascript)'],
+		['"Javascript"'],
+		["'Javascript'"],
+		['"Javascript",'],
+	])('should match a pattern regardless of punctuation: %s', string => {
+		const regexp = getExactMatchRegExp('javascript');
+		expect(regexp.test(string)).toBeTruthy();
+	});
+
 	it('returned RegExp should not match in the middle of the word', () => {
 		const regexp = getExactMatchRegExp('webpack');
 		expect(regexp.test(`FooWebpack`)).toBeFalsy();