[haiku-commits] Re: r35085 - in haiku/trunk/src: kits/tracker system/libroot/os
- From: "Axel Dörfler" <axeld@xxxxxxxxxxxxxxxx>
- To: haiku-commits@xxxxxxxxxxxxx
- Date: Sun, 17 Jan 2010 15:11:10 +0100
Stephan Assmus <superstippi@xxxxxx> wrote:
> On 2010-01-17 at 13:53:08 [+0100], Axel Dörfler <axeld@xxxxxxxxxxxxxxxx>
> > wrote:
> > I would prefer lower case for the trash.
> > BTW this is the FreeDesktop.org's POV:
> > http://www.ramendik.ru/docs/trashspec.html
> This draft fails to mention why, when a "$topdir/.Trash" directory is
> absent, it should not be created and instead "$topdir/.Trash-$uid" is
> to be
> created. I don't follow that at all. I could understand that
> "$topdir/.Trash-$uid" is to be created in the case when "$topdir/.
> Trash" is
> there, but cannot be used for some reason.
>
> Also, they seem to consider both Gnome and KDE being installed and
> either
> one being used by different users. What they don't consider is two or
> more
> operating systems being installed on the same computer. In that case,
> it is
> more likely that users with the same "name" are configured in each
> system,
> but it's rather unlikely that they would have the same uid. Maybe
> this
> situation in itself is considerd problematic and an invalid setup
> anyway.
>
> Also, regarding networks, I have various computers here which are all
> networked. Most of them have users by the same name, but I doubt that
> they
> would have the same uid. Network login seems to work fine, though, by
> user
> name. While a particular user is then probably logged in with the
> correct
> uid on the remote system, on the local system (where .Trash/$uid
> would be
> referred) he/she may have a different uid. It would not be a problem
> if the
> user name was to be used.
>
> On top of that, when I want to admisiter my system(s), it would be
> much
> easier for me to read user names instead of numerical uids.
>
> Finally, we have file system attributes, so why should we use those
> info
> files when it then prevents us from listing interesting attributes in
> the
> Trash folder window (like sorting files by trashed time)? On non-BFS
> volumes, I could agree with that.
FWIW I have had similar feelings while reading it. Although we could
try to comply with it at least partially (as in creating a top-level .
Trash directory if absent) for foreign file systems like FAT. However,
in the case of FAT, being compatible to Windows would be even more
worthwhile.
Bye,
Axel.
Other related posts:
