Might make life easier if things were more isolated from existing filesystem. Of course depends on use case.
There was a syncthing-fuse project that someone worked on.
It’s not maintained anymore however.
So, do you think there are valid arguments against it?
If not, I might try to poke at it.
not sure what you mean by against it.
I guess if your question is how you think it should be integrated, I think you should start by resurrecting that separate application and get it to a working state. Given the feature is only usable on a unix, and doesn’t need half of the bells and whistles that exist in syncthing, so it makes sense to start externally.
Once it’s mature, I guess it could be moved to syncthing itself.
There is at least one library that does FUSE functionality on windows. Whether it uses FUSE itself, I don’t recall.
Let me try to dig up the project you were talking about and I’ll ask for directions if I can’t find it.
This guy here, right?
yeah that’s the one I’m refering to. There is no fuse for windows, there is Dokan for windows, but that’s not fuse, requires a custom kernel driver, etc, hence not very usable.
Yeah, in addition to Dokan there is winfsp which has the same drawback. I should read more.