On 2/26/08, Truong Thanh Quang <quangtt at softend.vn> wrote:
>
> Hi all,
>
> I think this is a bug in PJSUA when allocating call_id
>
> r1820, pjsua_call.c:218
> static pjsua_call_id alloc_call_id(void)
> {
> pjsua_call_id cid;
> #if 1
>
>
> /* New algorithm: round-robin */
> if (pjsua_var.next_call_id >= (int)pjsua_var.ua_cfg.max_calls ||
> pjsua_var.next_call_id < 0)
> {
> cid = 0;
> }
>
> for (cid=pjsua_var.next_call_id;
> cid<(int)pjsua_var.ua_cfg.max_calls;
> ++cid)
> {
> if (pjsua_var.calls[cid].inv == NULL) {
> ++pjsua_var.next_call_id;
> return cid;
> }
> }
>
> for (cid=0; cid < pjsua_var.next_call_id; ++cid) {
> if (pjsua_var.calls[cid].inv == NULL) {
> ++pjsua_var.next_call_id;
> return cid;
> }
> }
>
> #else
>
> when pjsua_var.next_call_id > (int)pjsua_var.ua_cfg.max_calls the next for
> loop will never be executed and from now on, allcating call_id will be the
> same with old algorithm.
> I think line 227 should be: pjsua_var.next_call_id = ; (instead of cid =
> 0;)
You're right, and that was a silly bug! Thanks for spotting it, and
it's been fixed in http://trac.pjsip.org/repos/ticket/455#comment:3
cheers,
-benny
