www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.learn - main return value, struct padding

reply bearophile <bearophileHUGS lycos.com> writes:
I post this here, but if later I see it fit I may post something in the main D
group too.

D allows to have the main() function of type void too. In such case I'd like
the program return 0 by default. If people agrees, this can become a feature
request.

A question: when D automatically adds some padding to a struct, is such unused
space guaranteed to be initialized to a default constant value, like zero?

Bye,
bearophile
Jan 20 2009
next sibling parent reply Bill Baxter <wbaxter gmail.com> writes:
On Tue, Jan 20, 2009 at 11:15 PM, bearophile <bearophileHUGS lycos.com> wrote:

 I post this here, but if later I see it fit I may post something in the main D
group too.

 D allows to have the main() function of type void too. In such case I'd like
the program return 0 by default. If people agrees, this can become a feature
request.



What does it return now?  I always assumed returned zero for void main.

--bb
Jan 20 2009
parent bearophile <bearophileHUGS lycos.com> writes:
Bill Baxter:


 What does it return now?



A random int, I presume.



I always assumed returned zero for void main.<



Me too.

Bye,
bearophile
Jan 20 2009
prev sibling next sibling parent "Jarrett Billingsley" <jarrett.billingsley gmail.com> writes:
On Tue, Jan 20, 2009 at 10:25 AM, bearophile <bearophileHUGS lycos.com> wrote:

 Bill Baxter:


 What does it return now?



 A random int, I presume.



I always assumed returned zero for void main.<



 Me too.



I've only ever seen void mains return 0.  I think they used to return
random ints but that was fixed long ago.  Maybe it's different on
Linux?
Jan 20 2009
prev sibling parent reply "Steven Schveighoffer" <schveiguy yahoo.com> writes:
"bearophile" wrote

I post this here, but if later I see it fit I may post something in the 
main D group too.

 D allows to have the main() function of type void too. In such case I'd 
 like the program return 0 by default. If people agrees, this can become a 
 feature request.



A test:
[steves localhost testing]$ cat testmainreturn.d
void main()
{
    return 1;
}
[steves localhost testing]$ cat testmainreturn2.d
void main()
{
}
[steves localhost testing]$ dmd testmainreturn.d
[steves localhost testing]$ dmd testmainreturn2.d
[steves localhost testing]$ ./testmainreturn
[steves localhost testing]$ echo $?
1
[steves localhost testing]$ ./testmainreturn2
[steves localhost testing]$ echo $?
0
[steves localhost testing]$ dmd | grep "Digital Mars D"
Digital Mars D Compiler v1.038
[steves localhost testing]$

disassembly:
[steves localhost testing]$ obj2asm testmainreturn.o
...
_Dmain:
                push    EBP
                mov     EBP,ESP
                mov     EAX,1
                pop     EBP
                ret
.text._Dmain    ends
...
[steves localhost testing]$ obj2asm testmainreturn2.o
...
_Dmain:
                push    EBP
                mov     EBP,ESP
                xor     EAX,EAX
                pop     EBP
                ret
.text._Dmain    ends
...

So it appears it returns 0 unless you return something else (!)

-Steve
Jan 20 2009
parent reply "Jarrett Billingsley" <jarrett.billingsley gmail.com> writes:
On Tue, Jan 20, 2009 at 12:21 PM, Steven Schveighoffer
<schveiguy yahoo.com> wrote:

 So it appears it returns 0 unless you return something else (!)



Ha!
Jan 20 2009
parent Denis Koroskin <2korden+dmd gmail.com> writes:
Jarrett Billingsley Wrote:


 On Tue, Jan 20, 2009 at 12:21 PM, Steven Schveighoffer
 <schveiguy yahoo.com> wrote:

 So it appears it returns 0 unless you return something else (!)



 Ha!



Yes (but not always), even though specs clearly say that void functions
evaluate return values and discard them.

IIRC, this one returns 0:

void main() { 
    return new Object();
}

while this one returns an object address:

void foo() { return new Object(); }

void main() {
    return foo();
}
Jan 20 2009