On 2/16/08, Truong Thanh Quang <quangtt at softend.vn> wrote:
> In fact, the stream was destroyed by PJSUA.
Ah yes, of course, what was I thinking!
> Following lines is call stack when I set break point in
> on_call_media_state() function:
>
The callstack looks okay so far.
> Step into inv_negotiate_sdp():
> static pj_status_t inv_negotiate_sdp( pjsip_inv_session *inv )
> {
> pj_status_t status;
>
> PJ_ASSERT_RETURN(pjmedia_sdp_neg_get_state(inv->neg)
> ==
> PJMEDIA_SDP_NEG_STATE_WAIT_NEGO,
> PJMEDIA_SDPNEG_EINSTATE);
>
> status = pjmedia_sdp_neg_negotiate(inv->pool,
> inv->neg, 0);
>
> PJ_LOG(5,(inv->obj_name, "SDP negotiation done, status=%d", status));
>
> if (mod_inv.cb.on_media_update && inv->notify)
> (*mod_inv.cb.on_media_update)(inv, status);
>
> return status;
> }
>
> The stream was destroyed by pjmedia_sdp_neg_negotiate() before callback
> function was called.
Ah I see. So the stream gets destroyed by PJSUA-LIB before application
has a chance to remove it from its media flow, is that right?
> About solution, I think it is difficult.
> 1) Create another callback function in PJSUA (called
> pjsua_call_on_media_updating or something) to notify app about media is
> about to re-create new one
Yeah we could do that, by installing a callback in
stop_media_session() function in pjsua_media.c before media session is
destroyed. I'm still not sure about adding the callback in the API,
since anyway doing high volume call with PJSUA-LIB is not supported.
The best way is to redesign everything to handle more calls, but this
will take some time of course.
cheers,
-benny
> 2) Handle call by app (the best way as Q&A suggests)
>
cheers,
-benny
> Thanks for your support!
> -quang
>
>
>
>
>
> Help me how to handle this case?
> Thanks,
> -quang
>
>
>
> _______________________________________________
> Visit our blog: http://blog.pjsip.org
>
> pjsip mailing list
> pjsip at lists.pjsip.org
> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org
>
>
>
>
>
> _______________________________________________
> Visit our blog: http://blog.pjsip.org
>
> pjsip mailing list
> pjsip at lists.pjsip.org
> http://lists.pjsip.org/mailman/listinfo/pjsip_lists.pjsip.org
>
>
