pda rules, visor and libusb

[Bastien Nocera]

On Fri, 2006-04-28 at 09:59 +0100, Matt Davey wrote:
> What's the current status of the HAL pda rules?
> 
> I'm a developer of gnome-pilot, and have been trying to test recently
> contributed code that aims to integrate gnome-pilot with HAL.  I'm
> on fc5, and find that the libhal .fdi file 10-usb-pda.fdi is not
> correctly matching my (sony clie) palm PDA.
> 
> It looks like HAL isn't picking up any of the ttyUSB serial stuff,
> so the match rule doesn't therefore pick it up as a 'pda'
> capable device with pda.platform = "palm".
> 
> Is this a known issue?  I'm not familiar with the inner workings of HAL,
> so don't know whether it's likely to be a hal coding bug, a change in
> the functionality (I noticed some discussion as to whether ttyS* devices
> only should be marked as serial), or a kernel/udev problem, or what.
> 
> I'm attaching a portion of my 'lshal' output, showing my connected Clie,
> but no 'pda' keys.
> 
> More generally, the pilot-link folks (who provide the layer that talks
> to palmos devices) are moving towards libusb support in preference to
> the original usbserial 'visor' module.  From a quick look, it seems that
> the libhal pda rules require visor support.

There's one thing missing in the kernel to be able to use HAL properly
for Palm devices.

Most Palms show 2 TTY devices on their USB interface. Depending on the
model of the Palm, only one is useful, but there's no way to check which
device the first or second interface would be.

David told me it was an easy patch to the kernel, and I believe he still
owes me that patch ;)
I could surely make the patch myself, given some pointers on what I'm
supposed to be doing to export that "index".

-- 
Bastien Nocera <hadess at hadess.net> 
How can you be so serious on a film where you are dodging explosions
and running away with [...] an eight-foot monkey on this side, and the
eight-foot monkey is the one flying the spaceship?       -- Mark Hamill