I understand the reasoning, but I don't understand how the lines are drawn here (or how they could meaningfully be drawn).
Distributing two pieces of source code with different licenses is ok. Distributing them in a way that they can trivially be combined (e.g by compiling automatically on the receiving end during installation of the latest Ubuntu) can't be considered different from a distribution where the receiver takes more manual steps to combine them.
I'm curious why Canonical have to use binaries if they could easily just produce them on the clients machine and thereby not violate any copyrights? There has to be something else to it (or they just want to test the limits of the license conflict here, as a useful experiment).
Distributing two pieces of source code with different licenses is ok. Distributing them in a way that they can trivially be combined (e.g by compiling automatically on the receiving end during installation of the latest Ubuntu) can't be considered different from a distribution where the receiver takes more manual steps to combine them.
I'm curious why Canonical have to use binaries if they could easily just produce them on the clients machine and thereby not violate any copyrights? There has to be something else to it (or they just want to test the limits of the license conflict here, as a useful experiment).