Issue #18873: The tokenize module, IDLE, 2to3, and the findnocoding.py script

now detect Python source code encoding only in comment lines.

2 files changed, 10 insertions, 3 deletions

diff --git a/Lib/test/test_importlib/source/test_source_encoding.py b/Lib/test/test_importlib/source/test_source_encoding.py
index 0ca5195..ba02b44 100644
--- a/Lib/test/test_importlib/source/test_source_encoding.py
+++ b/Lib/test/test_importlib/source/test_source_encoding.py
@@ -10,7 +10,7 @@ import unicodedata
 import unittest
 
 
-CODING_RE = re.compile(r'coding[:=]\s*([-\w.]+)')
+CODING_RE = re.compile(r'^[ \t\f]*#.*coding[:=][ \t]*([-\w.]+)', re.ASCII)
 
 
 class EncodingTest(unittest.TestCase):
@@ -41,7 +41,7 @@ class EncodingTest(unittest.TestCase):
 
     def create_source(self, encoding):
         encoding_line = "# coding={0}".format(encoding)
-        assert CODING_RE.search(encoding_line)
+        assert CODING_RE.match(encoding_line)
         source_lines = [encoding_line.encode('utf-8')]
         source_lines.append(self.source_line.encode(encoding))
         return b'\n'.join(source_lines)
@@ -50,7 +50,7 @@ class EncodingTest(unittest.TestCase):
         # Make sure that an encoding that has never been a standard one for
         # Python works.
         encoding_line = "# coding=koi8-r"
-        assert CODING_RE.search(encoding_line)
+        assert CODING_RE.match(encoding_line)
         source = "{0}\na=42\n".format(encoding_line).encode("koi8-r")
         self.run_test(source)
 
diff --git a/Lib/test/test_tokenize.py b/Lib/test/test_tokenize.py
index b4a58f0..1765085 100644
--- a/Lib/test/test_tokenize.py
+++ b/Lib/test/test_tokenize.py
@@ -946,6 +946,13 @@ class TestDetectEncoding(TestCase):
         readline = self.get_readline((b'# coding: bad\n',))
         self.assertRaises(SyntaxError, detect_encoding, readline)
 
+    def test_false_encoding(self):
+        # Issue 18873: "Encoding" detected in non-comment lines
+        readline = self.get_readline((b'print("#coding=fake")',))
+        encoding, consumed_lines = detect_encoding(readline)
+        self.assertEqual(encoding, 'utf-8')
+        self.assertEqual(consumed_lines, [b'print("#coding=fake")'])
+
     def test_open(self):
         filename = support.TESTFN + '.py'
         self.addCleanup(support.unlink, filename)