Re: Motorola C350 patch 2

[Pawel Kot]

On Thu, 19 Feb 2004, Ron Yorston wrote:

> >There's above:
> >if (ms->status != GN_ERR_NONE) {
> >we can set there some local variable:
> >     has_memory_stats = false;
> >and
> >if (!has_memory_stats) {
> >     [your code]
> >}
> >
> >We'd kill that ugly fbus3810 variable as well. What do you think? Patch
> >against CVS follows. I should have Hugh's 3810 somewhere where so I'll
> >test the phone later.
>
> This doesn't work with the Motorola.  The atgen driver never returns an
> error in repsonse to a '+CPBS?' query.  The code in AT_GetMemoryStatus
> tries a '+CPBR=?' when '+CPBS?' fails, but ReplyMemoryStatus doesn't do
> anything with the response and returns GN_ERR_NONE.

So we don't enter the branch I thought we do, right?

> If I remove the '+CPBR=?' query from AT_GetMemoryStatus and just return
> the error status from the '+CPBS?' query it still doesn't work.  The code
> in xgnokii_contacts.c treats an error return from the 5110 as indicating
> that there is no ME memory, but the error return from the Motorola just
> means that we don't have the required information, not that there is no
> ME memory.

OK. Understood.

> Should we distinguish between the different causes of an error return
> from GetMemoryStatus by returning different values?

Yeah, I think so. The 3810 case would be GN_ERR_INVALIDMEMORYTYPE, your
case whould be GN_ERR_NOTSUPPORTED. Would it make sense? Does it work?

pkot
-- 
mailto:address@hidden
http://www.gnokii.org/