int[] a1, a2;
// ....

int idx = 0;
a1[idx] = a2[idx++];

In Java this has another result like in D.
Java will evaluate from left to right, D does the increment befor doing the assignment, so the copied values go into different fields.

Is this a bug or design decision?

"Frank Benoit" <keinfarbton@googlemail.com> wrote in message news:fqubu4$2d2e$1@digitalmars.com...
> int[] a1, a2;
> // ....
>
> int idx = 0;
> a1[idx] = a2[idx++];
>
> In Java this has another result like in D.
> Java will evaluate from left to right, D does the increment befor doing
> the assignment, so the copied values go into different fields.
>
> Is this a bug or design decision?

http://www.digitalmars.com/d/1.0/expression.html

Very first section.  It's illegal code but it's up to the compiler to detect it or not.  I guess you could file a bug about it as "accepts-invalid".

Frank Benoit wrote:
> int[] a1, a2;
> // ....
> 
> int idx = 0;
> a1[idx] = a2[idx++];
> 
> In Java this has another result like in D.
> Java will evaluate from left to right, D does the increment befor doing the assignment, so the copied values go into different fields.
> 
> Is this a bug or design decision?

Currently order of evaluation dependencies are illegal in D. Did you run across it in some Java code?

Walter Bright schrieb:
> Frank Benoit wrote:
>> Is this a bug or design decision?
> 
> Currently order of evaluation dependencies are illegal in D. Did you run across it in some Java code?

Yes, see 2. + 3. change
http://www.dsource.org/projects/dwt-linux/changeset/204%3A8ab2f71515ba

My eyes will be more sensible for this patterns now :)

Frank Benoit wrote:
> Walter Bright schrieb:
>> Currently order of evaluation dependencies are illegal in D. Did you run across it in some Java code?
> 
> Yes, see 2. + 3. change
> http://www.dsource.org/projects/dwt-linux/changeset/204%3A8ab2f71515ba
> 
> My eyes will be more sensible for this patterns now :)

My old C eyes are pretty sensitive to that, too. But it's also clear that D should move to a defined left-to-right order-of-evaluation semantic.