Re: Translations of folder names - two proposals

[Alexander Larsson <alexl redhat com>]

On Fri, 2004-12-10 at 13:57 +0000, Damon Chaplin wrote:
> On Fri, 2004-12-10 at 13:13, James Henstridge wrote:
> > On 10/12/04 14:05, Damon Chaplin wrote:
> > 
> > >On Fri, 2004-12-10 at 11:36, James Henstridge wrote:
> > >
> > >  
> > >
> > >>Like the gconf solution, it isn't clear what a script should do if the 
> > >>~/.folders/XXXX symlink or gconf key is missing though.
> > >>    
> > >>
> > >
> > >We could provide a small utility, e.g.
> > >
> > > gnome-create-standard-folder "Templates"
> > >
> > >That would use the same API that we would provide for apps.
> > >It could even return the directory pathname, making scripts fairly
> > >painless. We could also provide a
> > >
> > > gnome-get-standard-folder "Templates"
> > >
> > >that returns the directory pathname or nothing if it doesn't exist.
> > >  
> > >
> > If the script is going to need to use one of these programs if the 
> > symlink is missing, then is it ever correct for them to blindly use the 
> > ~/.folders/XXXX directories?  If not, then that kills off some of the 
> > benefits of the symlink approach.
> 
> I don't know. I'm not sure we even need the symlinks if we provide a C
> API and utility apps. We just need an FDO standard for storing the
> pathnames somewhere.

But why would you want to store the mapping in a way that makes it much
harder to use from scripts and suchlike (needing to be in c, or
requiring gnome-specific apps), when there is a way to store it that
makes it very easy to use from any unix language (symlinks).

If the directory/symlink doesn't exist its likely that the user didn't
want that directory, so perhaps we shouldn't create it. This only works
if we create all such directories on the first login of course.