Re: [iPAQ] bluetooth/pcmcia status?

So now the milion dollar question is: why is ttyname() returning such a
mangled buffer?

Jordan

On Thursday 03 January 2002 16:16, Alex C. Snoeren mentioned:
> That fixed it! I noticed pppd was getting the wrong tty, but figured
> that was because hciattach had bombed out, and leaving ppp attached to
> the original tty. So, yes, I am happy to report that BlueZ does in fact
> work on the iPAQs.
>
> Many thanks,
>
> - Alex
>
> Jordan Crouse wrote:
> > Ahh now I recall the problem... there is something broken with
> > "ttyname()" that returned the wrong length of the name buffer, so
> > something like /dev/pts/0 would actually be returned to PPP as
> > /dev/pts/0a or something. This drove me nuts until I found it (thank god
> > for printfs).
> >
> > So, in order to fix it, I used the following config on the Ipaq:
> >
> > na {
> > channel 1;
> > up {
> > ppp "/dev/pts/0 noauto 460800";
> > };
> > }
> >
> > Now, this is quite un-automagic, but it works. This is built against
> > skiff libc version 2.1.2 - I am not sure if the problem still exists in
> > later libc versions. Toss a printf into tty_process_extra_options() in
> > pppd/tty.c and see if the correct name is returned, or if it is mangled.
> >
> > Jordan
Received on Thu Jan 3 15:39:42 2002