Re: Fix EMS + concat message, cleanup gnokii.c

[Pawel Kot]

On Wed, 12 Jun 2002, Pavel Machek wrote:

> > > Bitmaps would have to return 4 things: number of bytes eaten in D,
> > > bytes to put in D, number of bytes eaten in E, bytes to put in E.
> >
> > No, I don't think so. Bitmaps, just encode itself. Packing into some
> > boundary is made by SMS layer - and it just needs bitmap size. Or am I
> > missing something?
>
> Okay, what bitmap does is just add some header then
> memcpy(foo, UDH, bitmap->data). Nothing too bitmap specific. It would
> be ugly to split memcpy from size computation. OTOH it might be quite
> acceptable to move whole EMS handling into gsm-sms.c, if you prefer that.

Well, I see your point since yesterday. And I'm thinking about smart and
nice solution.

> > > That seems ugly to me, and AddUDH does not seem that bad.
> >
> > It seems as it add loop in the dependencies.
>
> But that loop does not matter, C compiler can survive that.

But that loop is ugly. And AFAICS cause compile warning.

> > It's hard to call it an evening, but... ;-) I made some changes to CVS.
> > First of all I decided to separate all Nokia-specific SMS stuff to the
> > separate functions/files. It will clean up the sources a bit (but make code
> > paths a bit longer). I don't think these changes broke anything yet.
>
> Run make test and see for yourself ;-). Tests involving --ocat are expected
> to be broken as those are only working in my tree.

Will do. Next changes will break it for sure. I'll remove RawData in SMS
usage and layouts. I'll try to get back to life 7110 and AT driver. fake
is based on AT AFAIR.

BTW. Is it OK to remove testing dir in the tarball?

pkot
-- 
mailto:address@hidden :: mailto:address@hidden
http://kt.linuxnews.pl/ :: Kernel Traffic po polsku