Re: Bug in src/nm-gsm-device.c

[Pablo Martí Gamboa <pmarti warp es>]

2009/5/21 Dan Williams <dcbw redhat com>

On Thu, 2009-05-21 at 09:15 +0200, Pablo Martí Gamboa wrote:
>
>
> 2009/5/21 Dan Williams <dcbw redhat com>
>         On Wed, 2009-05-20 at 10:23 +0200, Pablo Martí Gamboa wrote:
>         > Hi all,
>         >
>         > I've just been bitten by a small bug in src/nm-gsm-device.c
>         +443:
>         >
>         > const char *responses[] = { "+CREG: 0,0", "+CREG: 0,1",
>         "+CREG: 0,2",
>         > "+CREG: 0,3", "+CREG: 0,5", NULL };
>         >
>         >
>         > If an external implementation has send '+CREG=1' those
>         terminators
>         > won't work. I think that you should just check the last part
>         of the
>         > +CREG response.
>
>
>         What would the response & terminator look like in that case?
>          Not quite
>         sure I follow.
>
> Something along the lines of what is done in ModemManager:
> if (g_str_has_prefix (response->str, "+CREG: ")) {
>     int unsolicited, stat;
>     sscanf (response->str + 7, "%d,%d", &unsolicited, &stat);
>     ....
>     switch(stat) {
>         case 5: break;
>         case 4: break;
>         ....
>     }
> }

Right, but I meant what is the exact response that's screwing up NM here
on 0.7 branch?  In 0.7 we do need to match the full response, otherwise
we might match some other unsolicited response the modem sends, instead
of CREG.

It does not recognizes responses like +CREG: 1,1. For now my workaround has been to disable +CREG notifications right before connecting via NM DBus[0], otherwise it won't work.

AFAICT this wasn't present in the NM shipped with Fedora 10, OpenSUSE 11.1 or Ubuntu 8.10.

[0] http://public.warp.es/wader/browser/node/core/trunk/wader/common/dialers/nm_pppd.py#L84

Dan

-- 
Pablo Martí
http://www.linkedin.com/in/pmarti || http://www.warp.es
python -c "print '706d6172746940776172702e6573'.decode('hex')"