Using DMD 0.100, Windows 98SE.

When a slice is assigned by a scalar, the value of the assignment expression has the length of the whole array, not that of the slice.

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

void show(int[] s) {
    foreach (int i; s) {
        writef("%d ", i);
    }
    writefln();
}

void main() {
    int[] qwert = new int[6];
    int[] yuiop;
    yuiop = qwert[2..5] = 3;
    show(yuiop);
    show(qwert[2..5] = 4);
    show(qwert[2..5]);
    show(qwert);
    show(yuiop[2..5] = qwert[1..4]);

    yuiop = qwert[2..5];
    show(yuiop[1..3] = 6);
    writefln((yuiop[1..3] = 7).length);
}
----------

Output:

3 3 3 0 0 0
4 4 4 0 0 0
4 4 4
0 0 4 4 4 0
0 4 4
6 6 4
3

Expected output:

3 3 3
4 4 4
4 4 4
0 0 4 4 4 0
0 4 4
6 6
2

Stewart.

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

In article <ch44o1$211s$1@digitaldaemon.com>, Stewart Gordon says...
>
>Using DMD 0.100, Windows 98SE.
>
>When a slice is assigned by a scalar, the value of the assignment expression has the length of the whole array, not that of the slice.

I don't agree that this is a bug. a[2..5] = 5 is an assignment, not a slice. The result of the assignment is the old array a with elements 2,3,4 set to 5, which is what should be returned.

Nick

Nick wrote:

> In article <ch44o1$211s$1@digitaldaemon.com>, Stewart Gordon says...
> 
>> Using DMD 0.100, Windows 98SE.
>> 
>> When a slice is assigned by a scalar, the value of the assignment expression has the length of the whole array, not that of the slice.
> 
> I don't agree that this is a bug. a[2..5] = 5 is an assignment, not a slice.   The result of the assignment is the old array a with elements
> 2,3,4 set to 5, which is what should be returned.

By your logic, the same would apply to a[3] = 5 or an assignment to a class/struct/union member.

The whole point of AssignExpression being an rvalue is that

    qwert = yuiop = asdfg;

is identical to

    yuiop = asdfg;
    qwert = yuiop;

except that yuiop is evaluated only once.

Anyway, your description isn't what the bug is doing.  Look at the output from my testcase again.

Stewart.

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

In article <ch4h5a$280v$1@digitaldaemon.com>, Stewart Gordon says...
>
>The whole point of AssignExpression being an rvalue is that
>
>     qwert = yuiop = asdfg;
>
>is identical to
>
>     yuiop = asdfg;
>     qwert = yuiop;
>
>except that yuiop is evaluated only once.

Ok, I can agree with that. But the docs say exactly nothing about this, so I'm not really sure what I would call expected behaviour here.

>Anyway, your description isn't what the bug is doing.  Look at the output from my testcase again.

That looks like a bug, yep. It looks like it was intended to work the way you describe, but it doesn't.

Nick

Nick wrote:
<snip>
> Ok, I can agree with that. But the docs say exactly nothing about this, so I'm
> not really sure what I would call expected behaviour here.
<snip>

You mean this doesn't count?

http://www.digitalmars.com/d/expression.html#AssignExpression

Stewart.

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

In article <ch4nva$2alo$1@digitaldaemon.com>, Stewart Gordon says...
>
>Nick wrote:
><snip>
>> Ok, I can agree with that. But the docs say exactly nothing about this, so I'm not really sure what I would call expected behaviour here.
><snip>
>
>You mean this doesn't count?
>
>http://www.digitalmars.com/d/expression.html#AssignExpression

I was thinking something more specifically about arrays, but again I agree with you. The link you refer to should cover the array slice case as well.

Nick