[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: Konqueror problem -- this one is worse :-(
From:       Kuba Ober <kuba () mareimbrium ! org>
Date:       2002-11-19 13:25:24
[Download RAW message or body]

On sobota 16 listopad 2002 01:30 pm, Harri Porten wrote:
> On Sat, 16 Nov 2002, CP Hennessy wrote:
> > > #9  0x445cd291 in KJS::AssignNode::value(KJS::ExecState*) const (
> > >     this=0x8309e58, exec=0xbfffe170) at nodes.cpp:1535
> >
> > Actually here the code looks like:
> > v = expr->value(exec);
> > but is it possible for "expr" to be a NULL pointer?
>
> No. The parser shouldn't allow this.

Maybe the parser returns a NULL expr (does it make any sense? probably no) or 
it just has a bug.

In any event, I presume adding a few assert()s here and there may help 
eradicate lingering bugs there.

I know that a few times even things like this:

void foo(type & bar) {
	assert(&bar != 0);
}

could trigger (they did in my sloppy code), so it's not a bad idea to exclude 
possibility of a null address whenever it won't be safely handled. With my 
luck (lack thereof at times), I always put asserts whenever the code wouldn't 
be able to handle certain error condition, or whenever something wouldn't 
make sense that way. It usually helps a lot.

Cheers, Kuba Ober
 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
[prev in list] [next in list] [prev in thread] [next in thread]