Re: [PATCH] Re: sscanf("-1", "%d", &i) fails, returns 0

Randy.Dunlap (rddunlap@osdl.org)
Fri, 8 Nov 2002 11:59:36 -0800 (PST)


On Fri, 8 Nov 2002, Richard B. Johnson wrote:

| On Fri, 8 Nov 2002, Randy.Dunlap wrote:
|
| > On Sat, 9 Nov 2002, Douglas Gilbert wrote:
| >
| > | In lk 2.5.46-bk3 the expression in the subject line
| > | fails to write into "i" and returns 0. Drop the minus
| > | sign and it works.
| >
| > Here's an unobstrusive patch to correct that.
| > Please apply.
| >
| > --
| > ~Randy
| > -
|
| I was thinking that if anybody ever had to change any of this
| stuff, it might be a good idea to do the indirection only once?
| All those "splats" over and over again are costly.

Sure, it looks cleaner that way, although gcc has already put <*dig>
in a local register; i.e., it's not pulled from memory for each test.
Here's a (tested) version that does that.

-- 
~Randy

--- ./lib/vsprintf.c%signed Mon Nov 4 14:30:49 2002 +++ ./lib/vsprintf.c Fri Nov 8 12:03:15 2002 @@ -517,6 +517,7 @@ { const char *str = buf; char *next; + char *dig, onedig; int num = 0; int qualifier; int base; @@ -638,12 +639,14 @@ while (isspace(*str)) str++;

- if (!*str - || (base == 16 && !isxdigit(*str)) - || (base == 10 && !isdigit(*str)) - || (base == 8 && (!isdigit(*str) || *str > '7')) - || (base == 0 && !isdigit(*str))) - break; + dig = (*str == '-') ? (str + 1) : str; + onedig = *dig; + if (!onedig + || (base == 16 && !isxdigit(onedig)) + || (base == 10 && !isdigit(onedig)) + || (base == 8 && (!isdigit(onedig) || onedig > '7')) + || (base == 0 && !isdigit(onedig))) + break;

switch(qualifier) { case 'h':

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/