@dvnrrs unfortunately that's the hardest part of the change
We are going to be sprinting next week and I've added this topic to the agenda. I'd like to consider a plan of moving ~/snap to ~/.snapdata. Unfortunately any such move is extremely complex as it involves running systems, data migration for users who are not logged in but have encrypted or remote home directory, semantics for rollback of snapd/core which brings older snapd and old semantics, security review and a host of other considerations.
The part where the user can give it a custom name only comes after that, where a new FUSE filesystem would be mounted in user session, exposing a subset of data from $HOME/.snapdata to $HOME/Snap/$SNAP_NAME/... The magic is that at this stage "Snap" is arbitrary because it no longer matters at runtime. Applications would use $HOME/.snapdata/ automatically.
So unfortunately this bug is as hard as it was before. The thing that has improved is the overall health of the mount code so that we have more trust in our abilities to develop changes here, reliably.
@dvnrrs unfortunately that's the hardest part of the change
We are going to be sprinting next week and I've added this topic to the agenda. I'd like to consider a plan of moving ~/snap to ~/.snapdata. Unfortunately any such move is extremely complex as it involves running systems, data migration for users who are not logged in but have encrypted or remote home directory, semantics for rollback of snapd/core which brings older snapd and old semantics, security review and a host of other considerations.
The part where the user can give it a custom name only comes after that, where a new FUSE filesystem would be mounted in user session, exposing a subset of data from $HOME/.snapdata to $HOME/Snap/ $SNAP_NAME/ ... The magic is that at this stage "Snap" is arbitrary because it no longer matters at runtime. Applications would use $HOME/.snapdata/ automatically.
So unfortunately this bug is as hard as it was before. The thing that has improved is the overall health of the mount code so that we have more trust in our abilities to develop changes here, reliably.