struct Thing {
int[5] vals = 3;
}

void main() {
Thing t, g;

printf("%d %d\n", t.vals[0], g.vals[0]);
}

I get this output: 3 622879781

Is the compiler missing the array assignment on the second struct initialization?

In article <ds14nl$16h7$1@digitaldaemon.com>, Kevin Watters says...
>
>struct Thing {
>int[5] vals = 3;
>}
>
>void main() {
>Thing t, g;
>
>printf("%d %d\n", t.vals[0], g.vals[0]);
>}
>
>I get this output: 3 622879781
>
>Is the compiler missing the array assignment on the second struct initialization?
>
>


my output:

3 3

perhaps you aren't using the newest version of dmd.exe?

Definitely using 0.145, with Ares though--but I'm sure that wouldn't be the problem. I ran into it while crafting a template vector class.

Can anyone else reproduce this?

In article <ds1d21$1bml$1@digitaldaemon.com>, Chris says...
>
>In article <ds14nl$16h7$1@digitaldaemon.com>, Kevin Watters says...
>>
>>struct Thing {
>>int[5] vals = 3;
>>}
>>
>>void main() {
>>Thing t, g;
>>
>>printf("%d %d\n", t.vals[0], g.vals[0]);
>>}
>>
>>I get this output: 3 622879781
>>
>>Is the compiler missing the array assignment on the second struct initialization?
>>
>>
>
>
>my output:
>
>3 3
>
>perhaps you aren't using the newest version of dmd.exe?
>
>

0.145 (phobos) on both Linux and Windows working Ok for me as well. Try  it with phobos once just for kicks.

In article <ds1ijd$1efp$1@digitaldaemon.com>, Kevin Watters says...
>
>Definitely using 0.145, with Ares though--but I'm sure that wouldn't be the problem. I ran into it while crafting a template vector class.
>
>Can anyone else reproduce this?
>
>In article <ds1d21$1bml$1@digitaldaemon.com>, Chris says...
>>
>>In article <ds14nl$16h7$1@digitaldaemon.com>, Kevin Watters says...
>>>
>>>struct Thing {
>>>int[5] vals = 3;
>>>}
>>>
>>>void main() {
>>>Thing t, g;
>>>
>>>printf("%d %d\n", t.vals[0], g.vals[0]);
>>>}
>>>
>>>I get this output: 3 622879781
>>>
>>>Is the compiler missing the array assignment on the second struct initialization?
>>>
>>>
>>
>>
>>my output:
>>
>>3 3
>>
>>perhaps you aren't using the newest version of dmd.exe?
>>
>>
>
>

OK it did work with Phobos.

So it's a definite bug in Ares. I'll post it in the dsource.org Ares form.

Thanks for the help.

In article <ds2dfh$23j1$1@digitaldaemon.com>, Dave says...
>
>
>0.145 (phobos) on both Linux and Windows working Ok for me as well. Try  it with phobos once just for kicks.
>
>In article <ds1ijd$1efp$1@digitaldaemon.com>, Kevin Watters says...
>>
>>Definitely using 0.145, with Ares though--but I'm sure that wouldn't be the problem. I ran into it while crafting a template vector class.
>>
>>Can anyone else reproduce this?
>>
>>In article <ds1d21$1bml$1@digitaldaemon.com>, Chris says...
>>>
>>>In article <ds14nl$16h7$1@digitaldaemon.com>, Kevin Watters says...
>>>>
>>>>struct Thing {
>>>>int[5] vals = 3;
>>>>}
>>>>
>>>>void main() {
>>>>Thing t, g;
>>>>
>>>>printf("%d %d\n", t.vals[0], g.vals[0]);
>>>>}
>>>>
>>>>I get this output: 3 622879781
>>>>
>>>>Is the compiler missing the array assignment on the second struct initialization?
>>>>
>>>>
>>>
>>>
>>>my output:
>>>
>>>3 3
>>>
>>>perhaps you aren't using the newest version of dmd.exe?
>>>
>>>
>>
>>
>
>

Crazy.  I'll look into it.

Sean

Kevin Watters wrote:
> OK it did work with Phobos.
> 
> So it's a definite bug in Ares. I'll post it in the dsource.org Ares form.
> 
> Thanks for the help.
> 
> In article <ds2dfh$23j1$1@digitaldaemon.com>, Dave says...
>>
>> 0.145 (phobos) on both Linux and Windows working Ok for me as well. Try  it with
>> phobos once just for kicks.
>>
>> In article <ds1ijd$1efp$1@digitaldaemon.com>, Kevin Watters says...
>>> Definitely using 0.145, with Ares though--but I'm sure that wouldn't be the
>>> problem. I ran into it while crafting a template vector class.
>>>
>>> Can anyone else reproduce this?
>>>
>>> In article <ds1d21$1bml$1@digitaldaemon.com>, Chris says...
>>>> In article <ds14nl$16h7$1@digitaldaemon.com>, Kevin Watters says...
>>>>> struct Thing {
>>>>> int[5] vals = 3;
>>>>> }
>>>>>
>>>>> void main() {
>>>>> Thing t, g;
>>>>>
>>>>> printf("%d %d\n", t.vals[0], g.vals[0]);
>>>>> }
>>>>>
>>>>> I get this output: 3 622879781
>>>>>
>>>>> Is the compiler missing the array assignment on the second struct
>>>>> initialization?
>>>>>
>>>>>
>>>>
>>>> my output:
>>>>
>>>> 3 3
>>>>
>>>> perhaps you aren't using the newest version of dmd.exe?
>>>>
>>>>
>>>
>>
> 
> 

For what it's worth, I'm not seeing that error in my local build.  So this should either be fixed in the latest Ares release or it will be fixed in the next--I've been working on the C headers a bit so it could be a change that hasn't been made public yet.


Sean

Sean Kelly wrote:
> Crazy.  I'll look into it.
> 
> Sean
> 
> Kevin Watters wrote:
>> OK it did work with Phobos.
>>
>> So it's a definite bug in Ares. I'll post it in the dsource.org Ares form.
>>
>> Thanks for the help.
>>
>> In article <ds2dfh$23j1$1@digitaldaemon.com>, Dave says...
>>>
>>> 0.145 (phobos) on both Linux and Windows working Ok for me as well. Try  it with
>>> phobos once just for kicks.
>>>
>>> In article <ds1ijd$1efp$1@digitaldaemon.com>, Kevin Watters says...
>>>> Definitely using 0.145, with Ares though--but I'm sure that wouldn't be the
>>>> problem. I ran into it while crafting a template vector class.
>>>>
>>>> Can anyone else reproduce this?
>>>>
>>>> In article <ds1d21$1bml$1@digitaldaemon.com>, Chris says...
>>>>> In article <ds14nl$16h7$1@digitaldaemon.com>, Kevin Watters says...
>>>>>> struct Thing {
>>>>>> int[5] vals = 3;
>>>>>> }
>>>>>>
>>>>>> void main() {
>>>>>> Thing t, g;
>>>>>>
>>>>>> printf("%d %d\n", t.vals[0], g.vals[0]);
>>>>>> }
>>>>>>
>>>>>> I get this output: 3 622879781
>>>>>>
>>>>>> Is the compiler missing the array assignment on the second struct
>>>>>> initialization?
>>>>>>
>>>>>>
>>>>>
>>>>> my output:
>>>>>
>>>>> 3 3
>>>>>
>>>>> perhaps you aren't using the newest version of dmd.exe?
>>>>>
>>>>>
>>>>
>>>
>>
>>