bpo-46503: Prevent an assert from firing when parsing some invalid \N…

… sequences in f-strings. (GH-30865) * bpo-46503: Prevent an assert from firing. Also fix one nearby tiny PEP-7 nit. * Added blurb.
python · Jan 25, 2022 · 0daf721 · 0daf721
1 parent 7c26472
commit 0daf721
Show file tree

Hide file tree

Showing 3 changed files with 19 additions and 2 deletions.
diff --git a/Lib/test/test_fstring.py b/Lib/test/test_fstring.py
@@ -746,12 +746,16 @@ def test_misformed_unicode_character_name(self):
         # differently inside f-strings.
         self.assertAllRaise(SyntaxError, r"\(unicode error\) 'unicodeescape' codec can't decode bytes in position .*: malformed \\N character escape",
                             [r"f'\N'",
+                             r"f'\N '",
+                             r"f'\N  '",  # See bpo-46503.
                              r"f'\N{'",
                              r"f'\N{GREEK CAPITAL LETTER DELTA'",
 
                              # Here are the non-f-string versions,
                              #  which should give the same errors.
                              r"'\N'",
+                             r"'\N '",
+                             r"'\N  '",
                              r"'\N{'",
                              r"'\N{GREEK CAPITAL LETTER DELTA'",
                              ])

diff --git a/Misc/NEWS.d/next/Core and Builtins/2022-01-24-21-24-41.bpo-46503.4UrPsE.rst b/Misc/NEWS.d/next/Core and Builtins/2022-01-24-21-24-41.bpo-46503.4UrPsE.rst
@@ -0,0 +1 @@
+Fix an assert when parsing some invalid \N escape sequences in f-strings.
diff --git a/Parser/string_parser.c b/Parser/string_parser.c
@@ -442,12 +442,23 @@ fstring_find_literal(Parser *p, const char **str, const char *end, int raw,
         if (!raw && ch == '\\' && s < end) {
             ch = *s++;
             if (ch == 'N') {
+                /* We need to look at and skip matching braces for "\N{name}"
+                   sequences because otherwise we'll think the opening '{'
+                   starts an expression, which is not the case with "\N".
+                   Keep looking for either a matched '{' '}' pair, or the end
+                   of the string. */
+
                 if (s < end && *s++ == '{') {
                     while (s < end && *s++ != '}') {
                     }
                     continue;
                 }
-                break;
+
+                /* This is an invalid "\N" sequence, since it's a "\N" not
+                   followed by a "{".  Just keep parsing this literal.  This
+                   error will be caught later by
+                   decode_unicode_with_escapes(). */
+                continue;
             }
             if (ch == '{' && warn_invalid_escape_sequence(p, ch, t) < 0) {
                 return -1;
@@ -491,7 +502,8 @@ fstring_find_literal(Parser *p, const char **str, const char *end, int raw,
             *literal = PyUnicode_DecodeUTF8Stateful(literal_start,
                                                     s - literal_start,
                                                     NULL, NULL);
-        } else {
+        }
+        else {
             *literal = decode_unicode_with_escapes(p, literal_start,
                                                    s - literal_start, t);
         }