From 8ad59ebbc09964d0cbcb900c8a051cdbf75a2e15 Mon Sep 17 00:00:00 2001
From: "Gregory P. Smith" <greg@krypto.org>
Date: Thu, 22 Oct 2026 05:22:19 +0000
Subject: [PATCH] gh-161092: Require an octal mode in the uu codec "begin" line

Skip lines starting with "begin" that are not followed by an octal mode,
as email.message._decode_uu() does, instead of taking them as the header.
---
 Lib/encodings/uu_codec.py                     | 13 ++++++++++--
 Lib/test/test_codecs.py                       | 21 +++++++++++++++++++
 ...-10-22-05-20-02.gh-issue-161092.OtC5l6.rst |  4 ++++
 3 files changed, 36 insertions(+), 2 deletions(-)
 create mode 100644 Misc/NEWS.d/next/Library/2026-10-22-05-20-02.gh-issue-161092.OtC5l6.rst

diff --git a/Lib/encodings/uu_codec.py b/Lib/encodings/uu_codec.py
index 4f8704016e2131..56ea394a7806de 100644
--- a/Lib/encodings/uu_codec.py
+++ b/Lib/encodings/uu_codec.py
@@ -46,8 +46,17 @@ def uu_decode(input, errors='strict'):
         s = readline()
         if not s:
             raise ValueError('Missing "begin" line in input data')
-        if s[:5] == b'begin':
-            break
+        # Only a "begin <octal mode>" line, optionally followed by a
+        # filename, starts the encoded data (as in
+        # email.message._decode_uu()); anything else is leading text.
+        if s[:6] != b'begin ':
+            continue
+        mode = s[6:].split(b' ', 1)[0]
+        try:
+            int(mode, 8)
+        except ValueError:
+            continue
+        break
 
     # Decode
     while True:
diff --git a/Lib/test/test_codecs.py b/Lib/test/test_codecs.py
index 37ac57063c3539..3e174fb0130c26 100644
--- a/Lib/test/test_codecs.py
+++ b/Lib/test/test_codecs.py
@@ -3251,6 +3251,27 @@ def test_uu_invalid(self):
         # Missing "begin" line
         self.assertRaises(ValueError, codecs.decode, b"", "uu-codec")
 
+    def test_uu_begin_line(self):
+        # Only a "begin <octal mode> [<filename>]" line starts the encoded
+        # data; other lines starting with "begin" are leading text.
+        block = codecs.encode(b"data", "uu-codec")
+        other = codecs.encode(b"other", "uu-codec").partition(b"\n")[2]
+        for line in (b"begin 666 name", b"begin 0 name with spaces",
+                     b"begin 644 name\r", b"begin 666"):
+            with self.subTest(line=line):
+                encoded = line + b"\n" + other
+                self.assertEqual(codecs.decode(encoded + block, "uu-codec"),
+                                 b"other")
+        for line in (b"beginning of the text", b"begin", b"begin ",
+                     b"begin mode name", b"begin 888 name",
+                     b"beginx 666 name", b"begin-base64 666 name"):
+            with self.subTest(line=line):
+                encoded = line + b"\n" + other
+                self.assertRaises(ValueError, codecs.decode,
+                                  encoded, "uu-codec")
+                self.assertEqual(codecs.decode(encoded + block, "uu-codec"),
+                                 b"data")
+
 
 # The codec system tries to add notes to exceptions in order to ensure
 # the error mentions the operation being performed and the codec involved.
diff --git a/Misc/NEWS.d/next/Library/2026-10-22-05-20-02.gh-issue-161092.OtC5l6.rst b/Misc/NEWS.d/next/Library/2026-10-22-05-20-02.gh-issue-161092.OtC5l6.rst
new file mode 100644
index 00000000000000..abdc68c7f7a5db
--- /dev/null
+++ b/Misc/NEWS.d/next/Library/2026-10-22-05-20-02.gh-issue-161092.OtC5l6.rst
@@ -0,0 +1,4 @@
+Fix the ``uu`` codec decoder taking any line that starts with ``begin``, such
+as ``beginning of the report``, as the header of the encoded data. Like the
+:mod:`email` package, it now requires ``begin`` followed by an octal mode and
+skips other lines as leading text.