Trash spec 0.2, technical questions

C. Gatzemeier c.gatzemeier at
Wed Sep 1 13:23:58 EEST 2004

Am Tuesday 31 August 2004 17:14 schrieb Jerry Haltom:

> What about permissions of this top-level .Trash directory?
> There is a lot of things which I think are more complex than are warranted
> by making subdirectory's vs Nautilus' .Trash-${user} design.

Yes, part of why I think it is important not to put in some extra trash 
permissions but strictly stick to what is provided and used on the system 
wether plain unix permissions or ACLs.

As written before, different from what the current draft spec says better move 
stuff aside into the same full path under trash. Don't change and try to 
"resort" permissions by throwing files into arbitrary file pooling trash 

> Again, what any of these cases bring to the table to admins is "make sure
> to set up permissions right for your trash cans" which would be a bit
> annoying. ;)

The "mv $topdir/A/B/C/file $topdir/Trash/A/B/C/file[timestamp]" approach is 
permission neutral and shlould work regardless of the location as far as I 
can think now. It should work just fine executed with only regular user 
permission in home directories. This will already cover most of the uses 
cases without any sysadmin intervention.

Creating new base (trash) directories outside of $HOME should always need 
elevated permissions thus needing some filesystem, kernel, sysadmin or deamon 

It's probably best to let the permissions of ./Trash equal those of the dir 
where ./Trash resides.


More information about the xdg mailing list