fix(rules): avoid processing strings with case-less Letter category s…

…ymbols in `subject-case`

Fixes conventional-changelog#3585

@commitlint/rules/src/subject-case.test.ts

@@ -7,6 +7,7 @@ const messages = {
 	lowercase: 'test: subject',
 	lowercase_unicode: 'test: тема', // Bulgarian for `subject`
 	mixedcase: 'test: sUbJeCt',
+	caseless: 'test: 这是一次提交', // Chinese for `this is a commit`
 	uppercase: 'test: SUBJECT',
 	uppercase_unicode: 'test: ÛNDERWERP', // Frisian for `SUBJECT`
 	camelcase: 'test: subJect',
@@ -29,6 +30,7 @@ const parsed = {
 	lowercase: parse(messages.lowercase),
 	lowercase_unicode: parse(messages.lowercase_unicode),
 	mixedcase: parse(messages.mixedcase),
+	caseless: parse(messages.caseless),
 	uppercase: parse(messages.uppercase),
 	uppercase_unicode: parse(messages.uppercase_unicode),
 	camelcase: parse(messages.camelcase),
@@ -115,6 +117,24 @@ test('with mixedcase subject should fail for "always uppercase"', async () => {
 	expect(actual).toEqual(expected);
 });
 
+test('with caseless subject should succeed for "never sentensecase"', async () => {
+	const [actual] = subjectCase(await parsed.caseless, 'never', 'sentense-case');
+	const expected = true;
+	expect(actual).toEqual(expected);
+});
+
+test('with caseless subject should succeed for "never uppercase"', async () => {
+	const [actual] = subjectCase(await parsed.caseless, 'never', 'upper-case');
+	const expected = true;
+	expect(actual).toEqual(expected);
+});
+
+test('with caseless subject should succeed for "always uppercase"', async () => {
+	const [actual] = subjectCase(await parsed.caseless, 'always', 'upper-case');
+	const expected = true;
+	expect(actual).toEqual(expected);
+});
+
 test('with uppercase subject should fail for "never uppercase"', async () => {
 	const [actual] = subjectCase(await parsed.uppercase, 'never', 'uppercase');
 	const expected = false;

@commitlint/rules/src/subject-case.ts

@@ -4,20 +4,20 @@ import {TargetCaseType, SyncRule} from '@commitlint/types';
 
 /**
  * Since the rule requires first symbol of a subject to be a letter, use
- * combination of Unicode `Cased_Letter` and `Other_Letter` categories now to
- * allow non-Latin alphabets as well.
+ * Unicode `Cased_Letter` category now to allow non-Latin alphabets as well.
  *
  * Do not use `Letter` category directly to avoid capturing `Modifier_Letter`
- * (which just modifiers letters, so we probably shouldn't anyway) and to stay
- * close to previous implementation.
+ * (which just modifiers letters, so we probably shouldn't anyway) and
+ * `Other_Letter` (they actually are case-less, so they can't be validated)
+ * categories, and to stay close to previous implementation.
  *
  * Also, typescript does not seem to support almost any longhand category name
  * (and even short for `Cased_Letter` too) so list all required letter
  * categories manually just to prevent it from complaining about unknown stuff.
  *
  * @see [Unicode Categories]{@link https://www.regular-expressions.info/unicode.html}
  */
-const startsWithLetterRegex = /^[\p{Ll}\p{Lu}\p{Lt}\p{Lo}]/iu;
+const startsWithLetterRegex = /^[\p{Ll}\p{Lu}\p{Lt}]/iu;
 
 const negated = (when?: string) => when === 'never';