Adam,
Could you try this out and report the results? I'm guessing that an EOS bucket was
being
encountered and removed from the brigade, leaving an empty brigade. APR_BRIGADE_FIRST
then
returned a bogus bucket and we segfaulted when trying to read from that bogus bucket.
I think this code is more correct than the code it replaces. Only guessing that it will
prevent the segfault though.
Bill
----- Original Message -----
From: <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, January 02, 2002 2:12 PM
Subject: cvs commit: httpd-2.0/modules/proxy proxy_util.c
> stoddard 02/01/02 11:12:40
>
> Modified: modules/proxy proxy_util.c
> Log:
> Change the return code from APR_TIMEUP to APR_ECONNABORTED, which seems
> to be a bit more descriptive. Move the check to inside the inner while()
> loop and add an additional check for eos. If we get an EOS bucket, there
> is no point in going further. Hopefully this will fix the last seg fault
> in the function.
>
> Revision Changes Path
> 1.76 +4 -5 httpd-2.0/modules/proxy/proxy_util.c
>
> Index: proxy_util.c
> ===================================================================
> RCS file: /home/cvs/httpd-2.0/modules/proxy/proxy_util.c,v
> retrieving revision 1.75
> retrieving revision 1.76
> diff -u -r1.75 -r1.76
> --- proxy_util.c 31 Dec 2001 20:43:59 -0000 1.75
> +++ proxy_util.c 2 Jan 2002 19:12:40 -0000 1.76
> @@ -1020,13 +1020,12 @@
> &zero /* readline */))) {
> return rv;
> }
> - if (APR_BRIGADE_EMPTY(bb)) {
> - /* The connection aborted or timed out */
> - return APR_TIMEUP;
> - }
> -
> /* loop through each bucket */
> while (!found) {
> + if (*eos || APR_BRIGADE_EMPTY(bb)) {
> + /* The connection aborted or timed out */
> + return APR_ECONNABORTED;
> + }
> e = APR_BRIGADE_FIRST(bb);
> if (APR_BUCKET_IS_EOS(e)) {
> *eos = 1;
>
>
>
>
