Greg Ames wrote:

>[EMAIL PROTECTED] wrote:
>
>>  +AP_DECLARE(int) ap_rgetline(char **s, int n, request_rec *r, int fold)
>>
>
>>  -        last_char = pos + length - 1;
>>  -        if (last_char < beyond_buff) {
>>  +        if (length - 1 < n) {
>>
>
>how do we keep track of buffer overrun conditions when there are
>multiple reads?  unlikely, I know, but...
>

Thanks for catching that...the new check only works properly on the
first read, where the invariant beyond_buff==pos+n is still true.
I'll fix it in a minute.

--Brian


Reply via email to