Hello,
I'm playing with libpeer running on an ESP32 and I'm noticing some interesting behavior. I'm running Chrome and if I give the page permission to use the microphone, it's able to connect. It generates 8 candidates in the sdp answer, including IPV4 candidates, but it takes about 10 seconds to generate all of these candidates. But if I don't give it permission, it only generates 2 mdns candidates and 2 IPV6 candidates -- it takes very little time to generate the candidates, and it won't connect.
Digging into this some more, it has to do with mdns candidates... in ports.c in ports_resolve_mdns_host, it successfully gets the address, but returns -1. Fixing this by returning 0 the connection proceeds and I can send data. (nice! and great work btw, getting mdns working except for this tiny issue on esp32. :)
I'm curious if anyone is seeing the huge delay with generating the answer with the microphone permissions enabled. I've tried it on Chrome running in Linux and Windows and the delays are similarly long. I'm guessing it's a Chrome issue though.
Hello, I'm playing with libpeer running on an ESP32 and I'm noticing some interesting behavior. I'm running Chrome and if I give the page permission to use the microphone, it's able to connect. It generates 8 candidates in the sdp answer, including IPV4 candidates, but it takes about 10 seconds to generate all of these candidates. But if I don't give it permission, it only generates 2 mdns candidates and 2 IPV6 candidates -- it takes very little time to generate the candidates, and it won't connect.
Digging into this some more, it has to do with mdns candidates... in ports.c in ports_resolve_mdns_host, it successfully gets the address, but returns -1. Fixing this by returning 0 the connection proceeds and I can send data. (nice! and great work btw, getting mdns working except for this tiny issue on esp32. :)
I'm curious if anyone is seeing the huge delay with generating the answer with the microphone permissions enabled. I've tried it on Chrome running in Linux and Windows and the delays are similarly long. I'm guessing it's a Chrome issue though.
thanks :)