Incorrect client stack value about fbns-react

[valga]

@crouchingtiger Here is a captured CONNECT thrift from 13.0.0.1.91:

1 (08): "deadbeef-dead-beef-0"
4/1 (06): 0
4/2 (08): "[FBAN/MQTT;FBAV/13.0.0.1.91;FBBV/69989691;FBDM/{density=3.0,width=1080,height=1920};FBLC/en_US;FBCR/Carrier;FBMF/Manufacturer;FBBD/Brand;FBPN/com.instagram.android;FBDV/Model;FBSV/7.0;FBBK/1;FBCA/armeabi-v7a:armeabi;]"
4/3 (06): 439
4/4 (06): 128
4/5 (05): 1
4/6 (01): true
4/7 (02): false
4/8 (08): ""
4/9 (02): false
4/10 (05): 1
4/11 (05): 0
4/12 (06): 1234567890
4/14 (05): [76, 80]
4/15 (08): "device_auth"
4/16 (06): 567310203415052
4/20 (08): ""
4/21 (03): 3
5 (08): ""

No need to send fields 22-25, at least for Instagram app. Although other apps may use them to use single network connection for multiple endpoints.

[crouchingtiger]

I'm also sending field ID's 22-25 in my library (based off the protocol definition from 10.32.0). Doesn't seem to make much of a difference, but might cause issues down the road in terms of detection if they're not present.

22 = FBNS connection key (long)
23 = FBNS connection secret (string)
24 = FBNS device ID (string)
25 = FBNS device secret (string)

Sorry that I don't have time to make a proper PR right now.

[valga]

@crouchingtiger Hmm, 10.29.0 has these fields in protocol definition, but it doesn't send them. I have no time to check recent versions at the moment.

By the way, what values are being sent by 10.32.0? Same as in username/password or just empty?

[crouchingtiger]

Looks good then. I've actually not yet tried to capture the MQTT packets so far and was going still off decompilation.

Are you using this thrift plugin to read these in Wireshark?

[valga]

Actually I don't want to tell anyone how I am capturing the packets, because I am doing it in the most embarrassing way. I am too lazy to use Wireshark.