I originally started going down the rabbit hole of porting as much as possible to .netstandard before finding the fork that was further along but going to .net5. Going to .netstandard for some libraries (Notably the backends) would reduce the effort of merging develop repeatedly into the .net5 branch (the csproj files have to be hand resolved whenever anyone updates a dependency).
Note executables cannot be .netstandard and there are some libraries that cannot as well because of features they use. Technically you could target both runtimes for some of those libraries but there are code changes so you end up in a bit of conditional code hell.
So the mentality I’ve been using at the moment is to run headfirst into getting the .net5 branch mergeable. The trayicon on linux and autoupdater so far is the biggest problem I am seeing. The autoupdater there seems to be consensus that we can just ditch for the moment, while there is probably a desire to make it work the world doesn’t stop if its gone. The trayicon may be a harder sell. It would not surprise me if it is desirable to call the…net5 port Duplicati 3.0 to follow semantic versioning.
I can try and merge the mainline again after I get this stupid debian bug I’ve been chasing worked out.