I think what you mean here is the above error appeared *until* the
above patch was applied, and then went away. Ok.
> root@freescale /dev$
> wvdial
> --> WvDial: Internet dialer version
> 1.61
> --> Warning: section [Dialer Defaults] does not exist in
> wvdial.conf.
> --> Cannot get information for serial
> port.
> --> Initializing
> modem.
> --> Sending:
> ATZ
> ATZ
> OK
> --> Modem
> initialized.
> --> Configuration does not specify a valid phone
> number.
> --> Configuration does not specify a valid login
> name.
> --> Configuration does not specify a valid password.
Please send the following information:
- the contents and filename of your wvdial.conf file (please replace
the password first before sending to the list)
- the output of 'strace -e open wvdial'
Thanks,
Avery
Hmm, that does look pretty wrong, all right. Can you try the
following changes to your wvdial.conf?
- Change [Dialer defaults] to [Dialer Defaults]
- Completely remove the lines starting with #
Thanks,
Avery
--
You received this message because you are subscribed to the Google Groups "wvstreams-devel" group.
To post to this group, send email to wvstrea...@googlegroups.com.
To unsubscribe from this group, send email to wvstreams-dev...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/wvstreams-devel?hl=en.
As far as I understand, the only problem people have with running
wvdial on ARM systems has been with uClibc, which is missing the
setcontext() stuff, so we have to switch wvstreams to using
setjmp()/longjmp() on embedded ARM platforms that use uClibc. (Debian
systems all use glibc, which doesn't have this problem.) People have
reported that it works on both systems, though.
All that said, I'm a little stumped about this wvdial.conf problem.
Obviously *that* isn't happening to everybody, and the
configfile-reading stuff has nothing to do with setjmp/longjmp/etc.
Have fun,
Avery