Note the comment in #144 - while removing the system-ca-certs=true line might work, any changes you make to the network connection in Network Manager thereafter will re-add the line!
So the current workflow for connecting to a PEAP WIFI network is currently:
1. Make the connection. It will fail, but it will create an entry in /etc/NetworkManager/system-connections.
2. Edit that connection as root, remove the system-ca-certs=true line.
3. Never make any changes to that connection again.
One more small note - changing the line to read system-ca-certs=false causes the line to disappear when you save the file. I have no idea what is changing the file on write, but you can basically write the change and immediately cat the file, but instead of a line that says system-ca-certs=false, the line has vanished completely. Perhaps NetworkManager tracks the connection files in realtime, or some upstart job is submitting the changes? I don't know. I just found it interesting that the correct behaviour is actually the default behaviour and that this version causing this bug is actively changing that default behaviour.
Note the comment in #144 - while removing the system- ca-certs= true line might work, any changes you make to the network connection in Network Manager thereafter will re-add the line!
So the current workflow for connecting to a PEAP WIFI network is currently: ager/system- connections. ca-certs= true line.
1. Make the connection. It will fail, but it will create an entry in /etc/NetworkMan
2. Edit that connection as root, remove the system-
3. Never make any changes to that connection again.
One more small note - changing the line to read system- ca-certs= false causes the line to disappear when you save the file. I have no idea what is changing the file on write, but you can basically write the change and immediately cat the file, but instead of a line that says system- ca-certs= false, the line has vanished completely. Perhaps NetworkManager tracks the connection files in realtime, or some upstart job is submitting the changes? I don't know. I just found it interesting that the correct behaviour is actually the default behaviour and that this version causing this bug is actively changing that default behaviour.