Just downloaded DMD114 and found that the .init property of arrays sadly no longer initializes all the contents of the array.  I know, it's more consistent this way, but.. could there perhaps be a .clear property then?

Anyway I found the problem when I was looking for another way to quickly initialize the contents of an array.  I decided to see what exactly the .init property was returning, when I came across this.. bug?

import std.stdio;

void main()
{
 int[] a;
 a.length=2;
 a[0]=5;
 a[1]=10;
 int[] b=a.init;
 writefln(b[0],", ",b.length);
}

This causes an access violation on the line "int[] b=a.init".

However, if I change the first two lines like so:

import std.stdio;

void main()
{
 int[] a=new int[2];
 a[0]=5;
 a[1]=10;
 int[] b=a.init;
 writefln(b[0],", ",b.length);
}

It works.

Uhh.. is this not a problem?  I'd say a MAV for no reason is a problem.

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Jarrett Billingsley schrieb am Mon, 28 Feb 2005 17:02:56 -0500:
> Just downloaded DMD114 and found that the .init property of arrays sadly no longer initializes all the contents of the array.  I know, it's more consistent this way, but.. could there perhaps be a .clear property then?
>
> Anyway I found the problem when I was looking for another way to quickly initialize the contents of an array.  I decided to see what exactly the .init property was returning, when I came across this.. bug?
>
> import std.stdio;
>
> void main()
> {
>  int[] a;
>  a.length=2;
>  a[0]=5;
>  a[1]=10;
>  int[] b=a.init;
>  writefln(b[0],", ",b.length);
> }
>
> This causes an access violation on the line "int[] b=a.init".

I can't reproduce this on Linux. However your code has to segfault due
to b[0] (b.length is 0).

Added to DStress as http://dstress.kuehne.cn/run/array_initialization_05.d

Thomas

-----BEGIN PGP SIGNATURE-----

iD8DBQFCOo5+3w+/yD4P9tIRAoFrAJ48Rdq2NQm49fM3iIE7xusotYI0HACgqPDn
pRhdL+vgMEo/NOoXP2OoE+k=
=EKoq
-----END PGP SIGNATURE-----