Trash specification, version 0.1

Mikhail Ramendik mr at ramendik.ru
Mon Aug 30 11:36:01 EEST 2004


Alexander Larsson wrote:

> > P.S.2.I can easily host the document on ramendik.ru at this point, but
> > of course it will need to go to freedesktop.org if it becomes accepted.
> > I'd suggest waiting for a stable version before putting it on
> > freedesktop.org.
> 
> I think its clear we should use $HOME/.Trash for the whole device that
> has $HOME on it, not only files inside $HOME. The spec doesn't seem to
> say this.

Agreed.

> > It may also choose to create .Trash directories in Top directories of 
> > some mounted file systems. How these directories are created (by the 
> > system administrator, a daemon, etc.) is determined by the 
> > implementation.
> 
> Not determined by the implementation. Determined by the spec, but not
> finished yet.

I'm not sure one should predefine this. There are different systems,
with different levels of system administrator involvement, and different
security requirements. On a tightly administered corporate network,
creation-by-sysadmin is best. On a loosely administered university
network, creation-by-daemon might work wonders. 

> The spec says "$topdir/.Trash/user". I'd prefer "$topdir/.Trash/$uid",
> since it contains no strange characters, and the filesystem permissions
> are tied to the uid, not the username which can change. (The uid is the
> numerical id of the user)

But might this create a problem with removable devices, and with some
cases of shared network resources? As far as I understand, $uid is
usually machine-specific. The same user can have different $uid values
on different computers, even in the same environment (especially if no
centralised logon is implemented).

Yours, Mikhail Ramendik






More information about the xdg mailing list