Issue #26581: Use the first coding cookie on a line, not the last one.

8 files changed, 10 insertions, 6 deletions

diff --git a/Lib/idlelib/IOBinding.py b/Lib/idlelib/IOBinding.py
index a4cc205..84f39a2 100644
--- a/Lib/idlelib/IOBinding.py
+++ b/Lib/idlelib/IOBinding.py
@@ -62,7 +62,7 @@ locale_encoding = locale_encoding.lower()
 encoding = locale_encoding  ### KBK 07Sep07  This is used all over IDLE, check!
                             ### 'encoding' is used below in encode(), check!
 
-coding_re = re.compile(r'^[ \t\f]*#.*coding[:=][ \t]*([-\w.]+)', re.ASCII)
+coding_re = re.compile(r'^[ \t\f]*#.*?coding[:=][ \t]*([-\w.]+)', re.ASCII)
 blank_re = re.compile(r'^[ \t\f]*(?:[#\r\n]|$)', re.ASCII)
 
 def coding_spec(data):
diff --git a/Lib/lib2to3/pgen2/tokenize.py b/Lib/lib2to3/pgen2/tokenize.py
index 1ff1c61..d14db60 100644
--- a/Lib/lib2to3/pgen2/tokenize.py
+++ b/Lib/lib2to3/pgen2/tokenize.py
@@ -236,7 +236,7 @@ class Untokenizer:
                 startline = False
             toks_append(tokval)
 
-cookie_re = re.compile(r'^[ \t\f]*#.*coding[:=][ \t]*([-\w.]+)', re.ASCII)
+cookie_re = re.compile(r'^[ \t\f]*#.*?coding[:=][ \t]*([-\w.]+)', re.ASCII)
 blank_re = re.compile(br'^[ \t\f]*(?:[#\r\n]|$)', re.ASCII)
 
 def _get_normal_name(orig_enc):
diff --git a/Lib/test/test_importlib/source/test_source_encoding.py b/Lib/test/test_importlib/source/test_source_encoding.py
index b604afb..1e0771b 100644
--- a/Lib/test/test_importlib/source/test_source_encoding.py
+++ b/Lib/test/test_importlib/source/test_source_encoding.py
@@ -14,7 +14,7 @@ import unittest
 import warnings
 
 
-CODING_RE = re.compile(r'^[ \t\f]*#.*coding[:=][ \t]*([-\w.]+)', re.ASCII)
+CODING_RE = re.compile(r'^[ \t\f]*#.*?coding[:=][ \t]*([-\w.]+)', re.ASCII)
 
 
 class EncodingTest:
diff --git a/Lib/test/test_source_encoding.py b/Lib/test/test_source_encoding.py
index 7979c82..3873400 100644
--- a/Lib/test/test_source_encoding.py
+++ b/Lib/test/test_source_encoding.py
@@ -178,7 +178,7 @@ class AbstractSourceEncodingTest:
     def test_double_coding_same_line(self):
         src = (b'#coding:iso8859-15 coding:latin1\n'
                b'print(ascii("\xc3\xa4"))\n')
-        self.check_script_output(src, br"'\xc3\xa4'")
+        self.check_script_output(src, br"'\xc3\u20ac'")
 
     def test_first_non_utf8_coding_line(self):
         src = (b'#coding:iso-8859-15 \xa4\n'
diff --git a/Lib/tokenize.py b/Lib/tokenize.py
index 7a00358..ec79ec8 100644
--- a/Lib/tokenize.py
+++ b/Lib/tokenize.py
@@ -34,7 +34,7 @@ import re
 import sys
 from token import *
 
-cookie_re = re.compile(r'^[ \t\f]*#.*coding[:=][ \t]*([-\w.]+)', re.ASCII)
+cookie_re = re.compile(r'^[ \t\f]*#.*?coding[:=][ \t]*([-\w.]+)', re.ASCII)
 blank_re = re.compile(br'^[ \t\f]*(?:[#\r\n]|$)', re.ASCII)
 
 import token
diff --git a/Misc/NEWS b/Misc/NEWS
index ae0a100..6f5c7ab 100644
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -10,6 +10,9 @@ Release date: tba
 Core and Builtins
 -----------------
 
+- Issue #26581: If coding cookie is specified multiple times on a line in
+  Python source code file, only the first one is taken to account.
+
 - Issue #26563: Debug hooks on Python memory allocators now raise a fatal
   error if functions of the :c:func:`PyMem_Malloc` family are called without
   holding the GIL.
diff --git a/Parser/tokenizer.c b/Parser/tokenizer.c
index be7cf49..c653121 100644
--- a/Parser/tokenizer.c
+++ b/Parser/tokenizer.c
@@ -275,6 +275,7 @@ get_coding_spec(const char *s, char **spec, Py_ssize_t size, struct tok_state *t
                         return 0;
                 }
                 *spec = r;
+                break;
             }
         }
     }
diff --git a/Tools/scripts/findnocoding.py b/Tools/scripts/findnocoding.py
index 5f3795e..6c16b1c 100755
--- a/Tools/scripts/findnocoding.py
+++ b/Tools/scripts/findnocoding.py
@@ -32,7 +32,7 @@ except ImportError:
                          "no sophisticated Python source file search will be done.", file=sys.stderr)
 
 
-decl_re = re.compile(rb'^[ \t\f]*#.*coding[:=][ \t]*([-\w.]+)')
+decl_re = re.compile(rb'^[ \t\f]*#.*?coding[:=][ \t]*([-\w.]+)')
 blank_re = re.compile(rb'^[ \t\f]*(?:[#\r\n]|$)')
 
 def get_declaration(line):