> I have been debugging this more. Here are the diferences: > WiFi with internet connectivity: > pjsua_create() executes quite quick > REGISTER message is sent almost immediately on call to pjsua_acc_add > WiFI without internet connectivity: > pjsua_create() takes a while to execute but is successful > REGISTER message is not set out immediately on a call to > pjsua_acc_add as in above scenario however this function returns > success. But it turns out that we do get a REGISTER sent eventually > it takes 5 minutes for that to happen. > Anyone know what the problem might be? The first thing I'd check would be attempts to perform DNS operations. Whether the putative attempt is name-to-address or address-to-name remains to be determined. If that doesn't pan out, the second thing I'd check would be something somewhere trying to phone home for something. I have nothing but wild conjectures what those "something"s might be. In either case, the way I'd look for it would be to sniff all traffic the device generates, comparing the traces from "working" and "slow" cases to try to see what's different. Anything unexpected in either case should be at least considered for investigation. (If you don't have traffic-snooping capability on, or immediately upstream of, the phone, then it would be substantially more difficult. I'd try to find something functionally equivalent to tcpdump on the device; if no such thing exists or can be added, I might have to give up.) /~\ The ASCII Mouse \ / Ribbon Campaign X Against HTML mouse at rodents-montreal.org / \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B
