Using DMD 0.129, Windows 98SE.

----------
void[10] data;
----------
d:\data\stewart\d\tests\bugs>dmd integral_void.d
integral constant must be scalar type, not void
----------

Both missing filename and line number, and completely meaningless in the context.  The code looks valid to me, judging by the fact that .length and slicing work on dynamic void[].

(DStress seems to be down at the mo - unless it's my Internet connection playing silly donkeys....)

Stewart.

-- 
My e-mail is valid but not my primary mailbox.  Please keep replies on on the 'group where everyone may benefit.

Stewart Gordon schrieb:
> (DStress seems to be down at the mo - unless it's my Internet connection
> playing silly donkeys....)

Seems to be your connection.

Please try the webinterface of the backup: http://svn.berlios.de/viewcvs/*checkout*/dstress/www/dstress.html

Thomas

Thomas Kühne wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Stewart Gordon schrieb:
> 
>> (DStress seems to be down at the mo - unless it's my Internet connection
>> playing silly donkeys....)
> 
> Seems to be your connection.
> 
> Please try the webinterface of the backup:
> http://svn.berlios.de/viewcvs/*checkout*/dstress/www/dstress.html

Thank you.  It seems to work at the moment....

So, is the testcase I just gave you in there?  I can't seem to find it and you haven't posted a link....

But I did notice a typo:

module dtsress.run.v.void_01;
        ^^

Stewart.

-- 
My e-mail is valid but not my primary mailbox.  Please keep replies on on the 'group where everyone may benefit.

Stewart Gordon schrieb:
>
> So, is the testcase I just gave you in there?  I can't seem to find it and you haven't posted a link....

> ----------
> void[10] data;
> ----------
> d:\data\stewart\d\tests\bugs>dmd integral_void.d
> integral constant must be scalar type, not void
> ----------
>
> Both missing filename and line number, and completely meaningless in the context.  The code looks valid to me, judging by the fact that .length and slicing work on dynamic void[].

The following code is illegal:
	void[10] a;
	void b;

Reason: void.init is undefined

You might use:
	void[10] c = void;

Added to DStress as http://dstress.kuehne.cn/nocompile/v/void_02_A.d http://dstress.kuehne.cn/nocompile/v/void_02_B.d http://dstress.kuehne.cn/run/v/void_02_C.d

> But I did notice a typo:
> 
> module dtsress.run.v.void_01;
>         ^^

Thanks, fixed

Thomas

Thomas Kühne wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Stewart Gordon schrieb:
> 
>> So, is the testcase I just gave you in there? I can't seem to find it and you haven't posted a link....
> 
>> ----------
>> void[10] data;
>> ----------
>> d:\data\stewart\d\tests\bugs>dmd integral_void.d
>> integral constant must be scalar type, not void
>> ----------
>> 
>> Both missing filename and line number, and completely meaningless in the context. The code looks valid to me, judging by the fact that .length and slicing work on dynamic void[].
> 
> The following code is illegal:
> 	void[10] a;
> 	void b;
> 
> Reason: void.init is undefined
<snip>

So why is the length of a void[] writable?

----------
import std.stdio;

void[] data;

void main() {
    data.length = 10;
    foreach (ubyte b; cast(ubyte[]) data) {
        writefln(b);
    }
}
----------

Stewart.

-- 
My e-mail is valid but not my primary mailbox.  Please keep replies on on the 'group where everyone may benefit.

Stewart Gordon schrieb:

> Thomas Kühne wrote:
> 
>> Stewart Gordon schrieb:
>>
>>> So, is the testcase I just gave you in there? I can't seem to find it and you haven't posted a link....
>>
>>> ----------
>>> void[10] data;
>>> ----------
>>> d:\data\stewart\d\tests\bugs>dmd integral_void.d
>>> integral constant must be scalar type, not void
>>> ----------
>>>
>>> Both missing filename and line number, and completely meaningless in the context. The code looks valid to me, judging by the fact that .length and slicing work on dynamic void[].
>>
>>
>> The following code is illegal:
>>     void[10] a;
>>     void b;
>>
>> Reason: void.init is undefined
> 
> <snip>
> 
> So why is the length of a void[] writable?
> 
> ----------
> import std.stdio;
> 
> void[] data;
> 
> void main() {
>     data.length = 10;
>     foreach (ubyte b; cast(ubyte[]) data) {
>         writefln(b);
>     }
> }
> ----------

http://www.digitalmars.com/d/arrays.html
# If the new array length is shorter, only enough are copied to fill the
# new array. If the new array length is longer, the remainder is filled
# out with the default initializer.

Decreasing the void[]'s length is legal, but increasing isn't.
Currently dynamic void's are useless as they can only contain zero elements.

solution A:
define void.init

solution B:
a combination of change length and VoidInitializer for the new array
elements

e.g.
	void[] data;
	data.length(10, void);


http://dstress.kuehne.cn/nocompile/a/array_initialization_19.d

Thomas