[patch] wireless patch, take 2

[David Zeuthen]

On Thu, 2004-05-27 at 12:20 -0400, Robert Love wrote:
> > I don't think it's so bad.  We still need callouts to do the distro-
> > specific stuff of doing DHCP, maybe saving the ESSID as a default
> > somewhere in YaST or /etc/sysconfig, but those things are unavoidable.
> > But the ESSID would be changed to whatever you set.
> 
> I am cool with this.
> 

Me too. 

Some rambling: we probably want to complicate the implementation a bit
more like this

 /etc/hal/methods.d/Net/Ethernet/80211/SetEssId

where SetEssId probably is a symlink and the client invokes the method
SetEssId on the interface org.freedesktop.Hal.Device.Net.Ethernet.80211.

This way we can reap the benefits of the D-BUS security framework.

Parameters to the D-BUS method are passed as positional parameters to
the script and the device properties are exported in the environment and
the UDI environment variable equals the UDI of the device. Should the
method be void or should it return the contents of stderr as a string?
If non-zero is returned the HAL daemon throws an exception much the same
way as the raise_* functions in hald_dbus.c.

One way it still feels a bit to me like feature-creep, but OTOH it's a
very nice way of providing infrastructure for using devices. We'd have
to be careful though not to abuse this though, e.g. it would be tempting
to provide a GetPhotos method on the org.freedesktop.org.Device.Camera
interface that returns the picture as a base64 encoded string :-). Oh
well.

Cheers,
David




_______________________________________________
hal mailing list
hal at freedesktop.org
http://freedesktop.org/mailman/listinfo/hal