Hi all, this sounds like a daft question, but I can't find in the docs where
it is addressed.
The following code on linux with dmd0.94 segfaults.
Object o;
assert (o != null);

Is there another way that I can assert that o is a valid reference before I use it?

Cheers
Brad

On Mon, 26 Jul 2004 19:41:19 +1200, Brad Beveridge wrote:

> Hi all, this sounds like a daft question, but I can't find in the docs where
> it is addressed.
> The following code on linux with dmd0.94 segfaults.
> Object o;
> assert (o != null);
> 
> Is there another way that I can assert that o is a valid reference before I use it?
> 
> Cheers
> Brad

Try
  assert (o !== null)

or

 assert (!(o === null));

or

 assert (!(o is null));


-- 
Derek
Melbourne, Australia
26/Jul/04 5:43:18 PM

Thank you, now that I see the === operator it rings a bell :)  It has been a long time since I read the D specs in fine detail.

Cheers
Brad

Derek Parnell wrote:

> On Mon, 26 Jul 2004 19:41:19 +1200, Brad Beveridge wrote:
> 
>> Hi all, this sounds like a daft question, but I can't find in the docs
>> where it is addressed.
>> The following code on linux with dmd0.94 segfaults.
>> Object o;
>> assert (o != null);
>> 
>> Is there another way that I can assert that o is a valid reference before I use it?
>> 
>> Cheers
>> Brad
> 
> Try
>   assert (o !== null)
> 
> or
> 
>  assert (!(o === null));
> 
> or
> 
>  assert (!(o is null));
> 
> 

Brad Beveridge schrieb:

> Is there another way that I can assert that o is a valid reference before I
> use it?

assert(o);

-eye

"Ilya Minkov" <minkov@cs.tum.edu> wrote in message news:ce2qo1$quf$3@digitaldaemon.com...
> Brad Beveridge schrieb:
>
> > Is there another way that I can assert that o is a valid reference
before I
> > use it?
>
> assert(o);
>
> -eye

That's a special case that calls the invariant on the object. If it's null, you'll get an access violation. I wish a simple o!==null would be automatically tested before the invariant, but oh well.

Vathix wrote:
> "Ilya Minkov" <minkov@cs.tum.edu> wrote in message
>>assert(o);
> 
> That's a special case that calls the invariant on the object. If it's null,
> you'll get an access violation. I wish a simple o!==null would be
> automatically tested before the invariant, but oh well.
> 

You could always do:

  assert(o !== null && o);

But that's kind of... well, ugly, to me.

-Chris S.
-Invironz