Trash specification regarding bind mounts

David Faure faure at
Wed Oct 15 15:14:31 PDT 2014

On Tuesday 23 September 2014 18:08:12 Colomban Wendling wrote:
> We have gathered basically three possible behavior in such cases:
> 1. Trash at the mount location, regardless of the file system root. This
> is what GIO's g_file_trash() currently does.
> 2. Trash at the file system root. This has implementation issues, like
> that a rename generally doesn't pass mount boundaries and trashing may
> then involve a copy. Likewise, the file system root may not be
> accessible for writing for an user even when the bind mount is.
> 3. Do not support trashing in such cases. This seems to currently be the
> preferred behavior for GIO maintainers.
> In my opinion, the expected behavior from an user point of view would be
> to follow solution 1 and trash at the mount location, regardless of
> whether it is a file system root or a bind mount.

Option 1 is fine with me.

It might be what my KDE code already does, since it doesn't do anything 
special for bind mounts, but I'm too lazy to actually test that :-)

David Faure, faure at,
Working on KDE Frameworks 5

More information about the xdg mailing list