lazy umount for fuse mounts

[cehteh]

Would adding the -z option (lazy unmounting) make any sense? if something accesses the fuse mounts and one kills the daemon there are still lingering mounts left behind, preventing a restart.

The only problem I see would be that it could shadow other bugs leaving the filesystem mounted, but all other umount command either do their job, block or report an error.

[whyrusleeping]

Yeah... fuse is tricky that way. I'd really like to move towards running the fuse mount as a separate process, independent of the daemon (although it wouldnt function properly without the daemon running).

[linas]

Seems to be a duplicate of #1248 which describes fuse messiness after unclean shutdown of ipfs.