Re: apr_vformatter mishandles "%n" (patch included)

I've applied this one. Looks good to me.

Cheers,
-g

On Mon, Oct 30, 2000 at 02:44:56PM -0600, Karl Fogel wrote:
> The call
>
> apr_psprintf (pool, "%u%n", random_number, &random_portion_width);
>
> was producing a string that had garbage bytes following the correctly
> formatted number. Turns out that even for "%n", apr_psprintf() was
> printing into the string. The bug was in apr_vformatter().
>
> The patch below is a pretty quick job; if someone else has a more
> elegant fix, that's great too.
>
> -Karl
>
> 2000-10-30 Karl Fogel <kfogel@collab.net>
>
> * apr_snprintf.c (apr_vformatter): don't print anything for "%n".
>
> --- apr_snprintf.c Mon Oct 30 14:23:52 2000
> +++ apr_snprintf.c Mon Oct 30 14:30:06 2000
> @@ -708,6 +708,7 @@
> /*
> * Default variable settings
> */
> + boolean_e print_something = YES;
> adjust = RIGHT;
> alternate_form = print_sign = print_blank = NO;
> pad_char = ' ';
> @@ -1000,6 +1001,7 @@
> *(va_arg(ap, short *)) = cc;
> else
> *(va_arg(ap, int *)) = cc;
> + print_something = NO;
> break;
>
> /*
> @@ -1131,10 +1133,12 @@
> /*
> * Print the string s.
> */
> - for (i = s_len; i != 0; i--) {
> - INS_CHAR(*s, sp, bep, cc);
> - s++;
> - }
> + if (print_something == YES) {
> + for (i = s_len; i != 0; i--) {
> + INS_CHAR(*s, sp, bep, cc);
> + s++;
> + }
> + }
>
> if (adjust_width && adjust == LEFT && min_width > s_len)
> PAD(min_width, s_len, pad_char);

-- 
Greg Stein, http://www.lyra.org/