On Tue, Mar 30, 2010 at 07:32:14PM +0200, Yann Droneaud wrote:
> Le mardi 30 mars 2010 à 19:12 +0200, Karel Zak a écrit :
> > On Tue, Mar 30, 2010 at 05:17:55PM +0200, Yann Droneaud wrote:
> > > You've made little mistakes in your fixes to mount and mkswap:
> > > - one must check that endptr doesn't match the string pointer to convert
> >
> > if ((endptr && *endptr) ||
> >
> > is enough, because endptr == str means that endptr is not empty
> >
>
> See the output of this little test function call with an empty string:
>
> int test(const char *s)
> {
> char *endptr = NULL;
>
> unsigned long l;
>
> printf("s = '%s'\n", s);
>
> errno = 0;
> l = strtoul(s, &endptr, 0);
>
> printf(" errno = %d\n", errno);
> printf(" l = %lu 0x%08lX\n", l, l);
> printf(" str = %p\n", s);
> printf(" endptr = %p\n", endptr);
> if (endptr != NULL) {
> printf(" *endptr = '%c' (0x%02x)\n", *endptr, *endptr);
> }
>
> return 0;
> }
>
> s = ''
> errno = 0
> l = 0 0x00000000
> str = 0x804865c
> endptr = 0x804865c
> *endptr = '' (0x00)
Ah, empty string. OK. Thanks. I'll fix it.
> endptr is not NULL and *endptr point to '\0'.
>
> > > otherwise empty string will be accepted as correct (eg *str == '\0', so
> > > if endptr == str, *endptr == '\0').
> > > - when using strtoll(), overflow must be check against LLONG_MIN and LLONG_MAX.
> >
> > xnum < 0)
> >
> > is enough, because LLONG_MIN is smaller than zero.
> >
>
> True, but you still have to correct the ULLONG_MAX use instead of
> LLONG_MAX.
We are talking about strtoll() (in mkswap).
Karel
--
Karel Zak <kzak@xxxxxxxxxx>
--
To unsubscribe from this list: send the line "unsubscribe util-linux-ng" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
