I tested this code:

import std.stdio;

void main(){
	double[] a=[1,2];
	double[10] c;
	c[]=a[]+0.;
	writeln(c);
	double[] b=[1,2,3,4,5,6,7,8,9,10];
	double[10] d;
	double[] e=b[0..2];
	d[]=e[]+0.;	
	writeln(e);
	writeln(d);
	d[]=e[];	
}

with the following result:

[1, 2, 0, 1.28823e-231, 1.11255e-308, -1.43745e+306, 0, 1.28823e-231, 1.56648e-307, 0]
[1, 2]
[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
object.Exception@src\rt\arraycat.d(31): lengths don't match for array copy
----------------
415330
4151BB
406DC4
406DFE
406A1F
41C425
----------------

Unless I have missed something, it seems that:

= tests the length of the array and implies an exception

BUT:

+ does not test the length, and it is thus possible to read information which are not allowed...


For information, I was testing how []+ were handling smaller array (e.g. Error? completed with nan? completed with 0.? cyclic computation?)

I used:

DMD32 D Compiler v2.059

on windows 7.

On Monday, 16 July 2012 at 18:24:36 UTC, fded wrote:
> I tested this code:
>
[...]
>
> For information, I was testing how []+ were handling smaller array (e.g. Error? completed with nan? completed with 0.? cyclic computation?)

"fded" <fded.cmlxxiv@gmail.com> wrote in message news:jjxlvtkcohzeseqbixms@forum.dlang.org...
>I used:
>
> DMD32 D Compiler v2.059
>
> on windows 7.
>
> On Monday, 16 July 2012 at 18:24:36 UTC, fded wrote:
>> I tested this code:
>>
> [...]
>>
>> For information, I was testing how []+ were handling smaller array (e.g. Error? completed with nan? completed with 0.? cyclic computation?)
>
>

Please use the bug tracker instead of posting to this newsgroup directly. http://d.puremagic.com/issues/

At a guess, I would say this is bug 2547. http://d.puremagic.com/issues/show_bug.cgi?id=2547