It seems that the project involve endless tweak with the installation process and I can see that jitsi team had already spent great effort on it.
However , a project in the state of experimental for nearly 2 years really deserve re-evaluation of the cost and workload .
Maybe it will be easier for everyone if this project is discarded and replaced with a native webrtc client in c/c++ ?
As long as a native webrtc client can talk to jitsi-meet successfully , people can do whatever they want with it . If they have problems with the native client , they will not ask here either . The workload become the interface maintenance or protocol translation and it may be simpler than jibri ?
Just my 2 cents , no offense 
Guofu Han