[3.10] gh-68966: Make mailcap refuse to match unsafe filenames/types/params (GH-91993) (GH-93543)

* gh-68966: Make mailcap refuse to match unsafe filenames/types/params (GH-91993)
(cherry picked from commit b9509ba)
* Add a What's New entry for 3.10.8.

Co-authored-by: Petr Viktorin <[email protected]>
Co-authored-by: Gregory P. Smith <[email protected]>

Author: 3 people

Doc/library/mailcap.rst

@@ -60,6 +60,18 @@ standard.  However, mailcap files are supported on most Unix systems.
    use) to determine whether or not the mailcap line applies.  :func:`findmatch`
    will automatically check such conditions and skip the entry if the check fails.
 
+   .. versionchanged:: 3.10.8
+
+      To prevent security issues with shell metacharacters (symbols that have
+      special effects in a shell command line), ``findmatch`` will refuse
+      to inject ASCII characters other than alphanumerics and ``@+=:,./-_``
+      into the returned command line.
+
+      If a disallowed character appears in *filename*, ``findmatch`` will always
+      return ``(None, None)`` as if no entry was found.
+      If such a character appears elsewhere (a value in *plist* or in *MIMEtype*),
+      ``findmatch`` will ignore all mailcap entries which use that value.
+      A :mod:`warning <warnings>` will be raised in either case.
 
 .. function:: getcaps()
 

Doc/whatsnew/3.10.rst

@@ -2338,3 +2338,11 @@ line flag, or :mod:`sys` APIs. See the :ref:`integer string conversion
 length limitation <int_max_str_digits>` documentation.  The default limit
 is 4300 digits in string form.
 
+Notable security feature in 3.10.8
+==================================
+
+The deprecated :mod:`mailcap` module now refuses to inject unsafe text
+(filenames, MIME types, parameters) into shell commands. Instead of using such
+text, it will warn and act as if a match was not found (or for test commands,
+as if the test failed).
+(Contributed by Petr Viktorin in :gh:`98966`.)

Lib/mailcap.py

@@ -2,6 +2,7 @@
 
 import os
 import warnings
+import re
 
 __all__ = ["getcaps","findmatch"]
 
@@ -13,6 +14,11 @@ def lineno_sort_key(entry):
     else:
         return 1, 0
 
+_find_unsafe = re.compile(r'[^\xa1-\U0010FFFF\w@+=:,./-]').search
+
+class UnsafeMailcapInput(Warning):
+    """Warning raised when refusing unsafe input"""
+
 
 # Part 1: top-level interface.
 
@@ -165,15 +171,22 @@ def findmatch(caps, MIMEtype, key='view', filename="/dev/null", plist=[]):
     entry to use.
 
     """
+    if _find_unsafe(filename):
+        msg = "Refusing to use mailcap with filename %r. Use a safe temporary filename." % (filename,)
+        warnings.warn(msg, UnsafeMailcapInput)
+        return None, None
     entries = lookup(caps, MIMEtype, key)
     # XXX This code should somehow check for the needsterminal flag.
     for e in entries:
         if 'test' in e:
             test = subst(e['test'], filename, plist)
+            if test is None:
+                continue
             if test and os.system(test) != 0:
                 continue
         command = subst(e[key], MIMEtype, filename, plist)
-        return command, e
+        if command is not None:
+            return command, e
     return None, None
 
 def lookup(caps, MIMEtype, key=None):
@@ -206,14 +219,23 @@ def subst(field, MIMEtype, filename, plist=[]):
             elif c == 's':
                 res = res + filename
             elif c == 't':
+                if _find_unsafe(MIMEtype):
+                    msg = "Refusing to substitute MIME type %r into a shell command." % (MIMEtype,)
+                    warnings.warn(msg, UnsafeMailcapInput)
+                    return None
                 res = res + MIMEtype
             elif c == '{':
                 start = i
                 while i < n and field[i] != '}':
                     i = i+1
                 name = field[start:i]
                 i = i+1
-                res = res + findparam(name, plist)
+                param = findparam(name, plist)
+                if _find_unsafe(param):
+                    msg = "Refusing to substitute parameter %r (%s) into a shell command" % (param, name)
+                    warnings.warn(msg, UnsafeMailcapInput)
+                    return None
+                res = res + param
             # XXX To do:
             # %n == number of parts if type is multipart/*
             # %F == list of alternating type and filename for parts

Lib/test/test_mailcap.py

@@ -123,7 +123,8 @@ def test_subst(self):
             (["", "audio/*", "foo.txt"], ""),
             (["echo foo", "audio/*", "foo.txt"], "echo foo"),
             (["echo %s", "audio/*", "foo.txt"], "echo foo.txt"),
-            (["echo %t", "audio/*", "foo.txt"], "echo audio/*"),
+            (["echo %t", "audio/*", "foo.txt"], None),
+            (["echo %t", "audio/wav", "foo.txt"], "echo audio/wav"),
             (["echo \\%t", "audio/*", "foo.txt"], "echo %t"),
             (["echo foo", "audio/*", "foo.txt", plist], "echo foo"),
             (["echo %{total}", "audio/*", "foo.txt", plist], "echo 3")
@@ -207,7 +208,10 @@ def test_findmatch(self):
              ('"An audio fragment"', audio_basic_entry)),
             ([c, "audio/*"],
              {"filename": fname},
-             ("/usr/local/bin/showaudio audio/*", audio_entry)),
+             (None, None)),
+            ([c, "audio/wav"],
+             {"filename": fname},
+             ("/usr/local/bin/showaudio audio/wav", audio_entry)),
             ([c, "message/external-body"],
              {"plist": plist},
              ("showexternal /dev/null default john python.org     /tmp foo bar", message_entry))

Misc/NEWS.d/next/Security/2022-04-27-18-25-30.gh-issue-68966.gjS8zs.rst

@@ -0,0 +1,4 @@
+The deprecated mailcap module now refuses to inject unsafe text (filenames,
+MIME types, parameters) into shell commands. Instead of using such text, it
+will warn and act as if a match was not found (or for test commands, as if
+the test failed).