Re: D-Bus interface of gvfs
- From: Alexander Larsson <alexl redhat com>
- To: Michael Albinus <michael albinus gmx de>
- Cc: David Zeuthen <david fubar dk>, gvfs-list gnome org
- Subject: Re: D-Bus interface of gvfs
- Date: Tue, 16 Dec 2008 09:49:27 +0100
On Mon, 2008-12-01 at 06:15 +0100, Michael Albinus wrote:
> David Zeuthen <david fubar dk> writes:
>
> > On Thu, 2008-11-27 at 06:41 +0100, Michael Albinus wrote:
> >> Hi,
> >>
> >> I'm looking for D-Bus introspection data of gvfs. The gvfs 1.0.2
> >> sources, I've grepped through, do not provide it, as far as I can see.
> >>
> >> Where shall I go?
> >
> > First of all you should use the gvfs-list instead. I've Cc'ed it for
> > follow up questions (I think you also need to subscribe to the list to
> > post messages).
>
> Thanks. Somehow, http://mail.gnome.org/mailman/listinfo and
> http://mail.gnome.org/mailman/listinfo/gvfs-list do not respond, so it
> is not so simple to search and to subscribe.
>
> Furthermore, it would be nice, if gvfs-list would be available at gmane
> (gnome-vfs-list exists there). By this, it would be more simple to read
> it for people like me.
>
> > Second, the D-Bus interfaces of GVfs are purely private implementation
> > details subject to change at any time. So you shouldn't use them at all;
> > they may change at any time and I don't think there are any plans to
> > provide any public API in that area.
> >
> > (In fact, all of GVfs right now is a giant implementation detail; the
> > only public API is GIO. That's subject to change when we make the GVfs
> > backend interface stable and public.)
> >
> > But I'm curious.. why do you need to know about the D-Bus interfaces?
>
> I'm planning to write an Emacs/Tramp client for gvfs. Tramp is an Emacs
> package, which allows access to files on remote hosts.
>
> Since the package is written in ELisp, I would prefer to communicate to
> gvfsd via D-Bus. An alternative could be the gvs-* commands, but this
> might decrease performance.
The dbus use of gvfs is very complicated, and quite internal, so I don't
think this is a good idea. For instance, for actual file data transfer
we use a custom binary protocol over socketpair fds that we pass over an
extra fd that is paired with each the client-to-backend dbus connection.
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]