bpo-46820: Fix a SyntaxError in a numeric literal followed by "not in" (GH-31479) (GH-31493)

Fix parsing a numeric literal immediately (without spaces) followed by
"not in" keywords, like in "1not in x". Now the parser only emits
a warning, not a syntax error.
(cherry picked from commit 090e5c4)

Co-authored-by: Serhiy Storchaka <[email protected]>

Author: miss-islington

Lib/test/test_grammar.py

@@ -251,6 +251,15 @@ def check_error(test):
         check("1e3in x")
         check("1jin x")
 
+        check("0xfnot in x")
+        check("0o7not in x")
+        check("0b1not in x")
+        check("9not in x")
+        check("0not in x")
+        check("1.not in x")
+        check("1e3not in x")
+        check("1jnot in x")
+
         with warnings.catch_warnings():
             warnings.simplefilter('ignore', SyntaxWarning)
             check("0xfis x")

Misc/NEWS.d/next/Core and Builtins/2022-02-21-21-55-23.bpo-46820.4RfUZh.rst

@@ -0,0 +1,3 @@
+Fix parsing a numeric literal immediately (without spaces) followed by "not
+in" keywords, like in ``1not in x``. Now the parser only emits a warning,
+not a syntax error.

Parser/tokenizer.c

@@ -1251,6 +1251,9 @@ verify_end_of_number(struct tok_state *tok, int c, const char *kind)
     else if (c == 'o') {
         r = lookahead(tok, "r");
     }
+    else if (c == 'n') {
+        r = lookahead(tok, "ot");
+    }
     if (r) {
         tok_backup(tok, c);
         if (parser_warn(tok, "invalid %s literal", kind)) {