unconfigurable limits

[Kay Sievers]

On Tue, 2004-08-17 at 11:44 +0200, David Zeuthen wrote:
> On Tue, 2004-08-17 at 01:47 +0200, Kay Sievers wrote:
> > The hal-device-manager the python application using D-BUS is still
> > unusable, cause of the hard coded D-BUS limits.
> 
> This is simply because the high-level python bindings needs to expose
> some API to subscribe to signals without specifying an object path.
> 
> Hence, for every device object exported by the hal daemon, h-d-m adds
> two rules because we are interested in two signals PropertyModified and
> DeviceCondition. Remember, h-d-m is just a crude hack meant for
> development, it's not production quality (though some distributions
> actually include it).

Hey, I really need something to make screenshots :)

> The idea is that if a desktop environment, GNOME,
> KDE, whatever, wants to use, they should write their own GUI for
> handling hal signals.
> 
> I'm sure Seth will accept some patch for the high-level python bindings
> so perhaps one patch would make this possible
> 
>   bus.add_signal_receiver(device_changed,                 # local method
>                           "PropertyModified",             # signal
>                           "org.freedesktop.Hal.Device",   # interface
>                           "org.freedesktop.Hal",          # service
>                           None)                           # object
> 
> so we can match all objects.

This would be nice.
Here is the crude hack from the last experiments. It acceepts 'None' in
all fields, but needs to find the match by iteration instead of the
current simple lookup. Any idea how to do it in sane way is welcome.

Thanks,
Kay
-------------- next part --------------
A non-text attachment was scrubbed...
Name: dbus-python-None-01.patch
Type: text/x-patch
Size: 5104 bytes
Desc: not available
Url : http://freedesktop.org/pipermail/dbus/attachments/20040817/7aff6b4a/dbus-python-None-01-0001.bin