# Jason Thorpe 2006-10-28:
> On Oct 14, 2006, at 5:49 PM, Jachym Holecek wrote:
> >I think with current state of autoconf(9) the code is fairly OK. We'd
> >need to move relevant members of <foo>_attach_args to locators and  
> >maybe
> >replace cf_loc with property lists to get somewhat better.
> >
> >So, I'd wait for Jason to comment on this as I don't exactly recall  
> >his
> >concerns.
> 
> My main concerns are:
> 
> - Format of the data -- We now have plists, and should use them.  The  
> format send down to user space should be the same format as device  
> drivers (eventually) get for their attachment information.

Fine with me. I can go with prop_dictionary_{send,recv}_ioctl() to
pass data.

In the long term, I'd prefer to use read(2) to retrieve events (ioctl()
seems ugly for this purpose, IMO) -- what would you think about eventually
extending proplib to support internalize/externalize across buffer
boundaries?

> - Related, I'm concerned about plugging into the "print" hook for this.

Yeah, I'd like to avoid dvprint_*() altogether, though autoconf(9)
doesn't really help me with that.

Anyway, I hope I can get back to this sooner then later... ;-)

	-- Jachym