Can anyone tell me the expected runtime behavior of this program?

class Foo
{ }

void main()
{
    Foo foo;
    assert(foo == null);
}

It's certainly not what I expected, and I want to know if I should file a bug report or not.

Dave

David B. Held wrote:
> Can anyone tell me the expected runtime behavior of this program?
> 
> class Foo
> { }
> 
> void main()
> {
>     Foo foo;
>     assert(foo == null);
> }
> 
> It's certainly not what I expected, and I want to know if I should file a bug report or not.
> 
> Dave

Try 'assert(foo is null);'

Dave schrieb:
> David B. Held wrote:
>> Can anyone tell me the expected runtime behavior of this program?
>>
>> class Foo
>> { }
>>
>> void main()
>> {
>>     Foo foo;
>>     assert(foo == null);
>> }
>>
>> It's certainly not what I expected, and I want to know if I should file a bug report or not.
>>
>> Dave
> 
> Try 'assert(foo is null);'

See http://www.digitalmars.com/d/operatoroverloading.html

> Note: Comparing a reference to a class object against null  should be > done as:
>> if (a is null)
> and not as:
>> if (a == null)
> The latter is converted to:
>> if (a.opEquals(null))

which will fail if a is null because it can't use it to reference to opEquals()

david

davidb wrote:
> Dave schrieb:
>> David B. Held wrote:
>>> Can anyone tell me the expected runtime behavior of this program?
>>>
>>> class Foo
>>> { }
>>>
>>> void main()
>>> {
>>>     Foo foo;
>>>     assert(foo == null);
>>> }
>>>
>>> It's certainly not what I expected, and I want to know if I should file a bug report or not.
>>>
>>> Dave
>>
>> Try 'assert(foo is null);'
> 
> See http://www.digitalmars.com/d/operatoroverloading.html
> 
>  > Note: Comparing a reference to a class object against null  should be  > done as:
>  >> if (a is null)
>  > and not as:
>  >> if (a == null)
>  > The latter is converted to:
>  >> if (a.opEquals(null))
> 
> which will fail if a is null because it can't use it to reference to opEquals()

Aha...thanks guys.  However, I am disturbed by two things: 1) that this looks more like VB than C++ ;), and 2) that everyone in this thread is named "David".  Hmm...

Dave

David B. Held wrote:
> davidb wrote:
>> Dave schrieb:
>>> David B. Held wrote:
>>>> Can anyone tell me the expected runtime behavior of this program?
>>>>
>>>> class Foo
>>>> { }
>>>>
>>>> void main()
>>>> {
>>>>     Foo foo;
>>>>     assert(foo == null);
>>>> }
>>>>
>>>> It's certainly not what I expected, and I want to know if I should file a bug report or not.
>>>>
>>>> Dave
>>>
>>> Try 'assert(foo is null);'
>>
>> See http://www.digitalmars.com/d/operatoroverloading.html
>>
>>  > Note: Comparing a reference to a class object against null  should be  > done as:
>>  >> if (a is null)
>>  > and not as:
>>  >> if (a == null)
>>  > The latter is converted to:
>>  >> if (a.opEquals(null))
>>
>> which will fail if a is null because it can't use it to reference to opEquals()
> 
> Aha...thanks guys.  However, I am disturbed by two things: 1) that this looks more like VB than C++ ;), and 2) that everyone in this thread is named "David".  Hmm...
> 
> Dave

o_O  It only makes sense though, for a language named "D".

We have our army of Davids, and the (Ch|K)ris's... at least there's just one Walter. Could you imagine...

-- Chris (!) Nicholson-Sauls

"David B. Held" <dheld@codelogicconsulting.com> wrote in message news:f2ptar$2022$1@digitalmars.com...
>
> this looks more like VB than C++

Well, maybe VB got some things right ;)

Jarrett Billingsley wrote:
> "David B. Held" <dheld@codelogicconsulting.com> wrote in message news:f2ptar$2022$1@digitalmars.com...
>> this looks more like VB than C++
> 
> Well, maybe VB got some things right ;)

It's still the only language I've seen that lets me pass temporaries and literals as reference arguments... :)

-- 
int getRandomNumber()
{
    return 4; // chosen by fair dice roll.
              // guaranteed to be random.
}

http://xkcd.com/

v2sw5+8Yhw5ln4+5pr6OFPma8u6+7Lw4Tm6+7l6+7D i28a2Xs3MSr2e4/6+7t4TNSMb6HTOp5en5g6RAHCP  http://hackerkey.com/

"Daniel Keep" <daniel.keep.lists@gmail.com> wrote in message news:f2qf8m$2v85$1@digitalmars.com...
>
>
> Jarrett Billingsley wrote:
>> "David B. Held" <dheld@codelogicconsulting.com> wrote in message news:f2ptar$2022$1@digitalmars.com...
>>> this looks more like VB than C++
>>
>> Well, maybe VB got some things right ;)
>
> It's still the only language I've seen that lets me pass temporaries and literals as reference arguments... :)

Doesn't C++ have those ... && reference arguments?  Like int&& ?  You can pass temps to those.  Well I guess those are in C++09, so they're not part of the language yet.

"Jarrett Billingsley" <kb3ctd2@yahoo.com> wrote in message news:f2qgao$3176$1@digitalmars.com...
>
> Doesn't C++ have those ... && reference arguments?  Like int&& ?  You can pass temps to those.  Well I guess those are in C++09, so they're not part of the language yet.

Though D can do this:

struct S
{
    int x;

    static S opCall(int x)
    {
        S s;
        s.x = x;
        return s;
    }
}

void foo(ref S s)
{
    writefln(s.x);
}

void main()
{
    foo(S(4));
}

Which I guess is pretty close.

Jarrett Billingsley wrote:
> "Jarrett Billingsley" <kb3ctd2@yahoo.com> wrote in message news:f2qgao$3176$1@digitalmars.com...
>> Doesn't C++ have those ... && reference arguments?  Like int&& ?  You can pass temps to those.  Well I guess those are in C++09, so they're not part of the language yet.
> 
> Though D can do this:
> 
> struct S
> {
>     int x;
> 
>     static S opCall(int x)
>     {
>         S s;
>         s.x = x;
>         return s;
>     }
> }
> 
> void foo(ref S s)
> {
>     writefln(s.x);
> }
> 
> void main()
> {
>     foo(S(4));
> }
> 
> Which I guess is pretty close.

That's interesting; could that be NVRO?  That said, this still doesn't work...

> import std.stdio;
>
> void foo(ref int a)
> {
>     writefln("a: %s", a);
> }
>
> void main()
> {
>     foo(42);
> }

H:\WINDOWS\system32\cmd.exe /c bud -clean -exec refargs refargs.d(11): Error: constant 42 is not an lvalue

:(

This is a pain because I discovered that using ref arguments for my vector library can net me a *big* speed increase, even more so once I put in SSE optimisations.

The problem is that if I can't pass literals or temporaries, then it'll render the library much harder to use.  Then again, maybe since the vectors are structs, the above will work... need to give that a shot when I'm not so busy.

-- 
int getRandomNumber()
{
    return 4; // chosen by fair dice roll.
              // guaranteed to be random.
}

http://xkcd.com/

v2sw5+8Yhw5ln4+5pr6OFPma8u6+7Lw4Tm6+7l6+7D i28a2Xs3MSr2e4/6+7t4TNSMb6HTOp5en5g6RAHCP  http://hackerkey.com/