Fix #1679: "0x" was taken as a valid integer literal.
Fixes the tokenizer, tokenize.py and int() to reject this.
Patches by Malte Helmert.
diff --git a/Python/mystrtoul.c b/Python/mystrtoul.c
index f007057..a02992f 100644
--- a/Python/mystrtoul.c
+++ b/Python/mystrtoul.c
@@ -112,27 +112,40 @@
/* check for leading 0 or 0x for auto-base or base 16 */
switch (base) {
- case 0: /* look for leading 0, 0x or 0X */
- if (*str == '0') {
- ++str;
- if (*str == 'x' || *str == 'X') {
- ++str;
- base = 16;
+ case 0: /* look for leading 0, 0x or 0X */
+ if (*str == '0') {
+ ++str;
+ if (*str == 'x' || *str == 'X') {
+ /* there must be at least one digit after 0x */
+ if (_PyLong_DigitValue[Py_CHARMASK(str[1])] >= 16) {
+ if (ptr)
+ *ptr = str;
+ return 0;
}
- else
- base = 8;
+ ++str;
+ base = 16;
}
else
- base = 10;
- break;
+ base = 8;
+ }
+ else
+ base = 10;
+ break;
- case 16: /* skip leading 0x or 0X */
- if (*str == '0') {
+ case 16: /* skip leading 0x or 0X */
+ if (*str == '0') {
+ ++str;
+ if (*str == 'x' || *str == 'X') {
+ /* there must be at least one digit after 0x */
+ if (_PyLong_DigitValue[Py_CHARMASK(str[1])] >= 16) {
+ if (ptr)
+ *ptr = str;
+ return 0;
+ }
++str;
- if (*str == 'x' || *str == 'X')
- ++str;
}
- break;
+ }
+ break;
}
/* catch silly bases */