Re: Problems with newest smsd

[Pawel Kot]

On Tue, 6 Aug 2002, Jan Derfinak wrote:

> On Tue, 6 Aug 2002, Markus Plail wrote:
>
> > Hi Jan!
> >
> > * Jan Derfinak writes:
> > > On Tue, 6 Aug 2002, Markus Plail wrote:
> > >> Ok. I realized what the problem is. Actually there are two:
> > >> 1. phone driver shouldn't segfault but exit with appropriate error code
> > >> 2. Jano changed my patch a bit, which lead to the segfault, but it
> > >> would still not work without the error in the phone driver.
> > >>
> > >> I realize that this change isn't obvious at all, and that actually
> > >> there is no real reason for it, not to work Jano's way, but only
> > >> mine. Still I think that, especially when one can't test the code, it
> > >> should have been better to ask me why I did this. Especially as it was
> >
> > > Ok. Please tell me why I must do the same initialization every iteration.
> >
> > Because it works? As I have written it's not obvious.
>
> But this is dirty. You made only workaround. You don't eliminate
> problem. It is dirty when you unexpected modify variable which should
> not be modified in logical sight.

Correct and agree.

> iteration. I want to draft that API functions must not change parts which
> are not pertinent to action which they are doing. If not this will route
> into untransparent API.

Agree. This is a bug that needs to be fixed.

> I think moving constant initialization before loop is right thing

Yes, it is.

> and GetSMS is writen unclear and should be fixed

You are talking about gsm-sms.c::GetSMS()? What is unclear there? Or you
are talking n general about getting SMS? If so, it's not unclear, it is
buggy.

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