gh-130057: Pygettext: Support translator comments

Lib/test/test_tools/test_i18n.py

@@ -447,6 +447,34 @@ def test_extract_all_comments(self):
                 '''), args=(arg,), raw=True)
                 self.assertIn('#. Translator comment', data)
 
+    def test_comments_with_multiple_tags(self):
+        """
+        Test that multiple --add-comments tags can be specified.
+        """
+        for arg in ('--add-comments={}', '-c{}'):
+            with self.subTest(arg=arg):
+                args = (arg.format('foo:'), arg.format('bar:'))
+                data = self.extract_from_str(dedent('''\
+                # foo: comment
+                _("foo")
+
+                # bar: comment
+                _("bar")
+                '''), args=args, raw=True)
+                self.assertIn('#. foo: comment', data)
+                self.assertIn('#. bar: comment', data)
+
+    def test_comments_not_extracted_without_tags(self):
+        """
+        Test that translator comments are not extracted without
+        specifying --add-comments.
+        """
+        data = self.extract_from_str(dedent('''\
+        # Translator comment
+        _("foo")
+        '''), raw=True)
+        self.assertNotIn('#.', data)
+
 
 def update_POT_snapshots():
     for input_file in DATA_DIR.glob('*.py'):

Tools/i18n/pygettext.py

@@ -144,7 +144,6 @@
 import importlib.machinery
 import importlib.util
 import os
-import re
 import sys
 import time
 import tokenize
@@ -331,7 +330,7 @@ def get_source_comments(source):
     for token in tokenize.tokenize(BytesIO(source).readline):
         if token.type == tokenize.COMMENT:
             # Remove any leading combination of '#' and whitespace
-            comment = re.sub(r'^[#\s]+', '', token.string)
+            comment = token.string.lstrip('# \t')
             comments[token.start[0]] = comment
     return comments