Consider :

class A { int x;}
A[10] t;
foreach(A a; t) // should be "A inout a"
    a = new A;

Later, an access to any element of 't' is likely to segfault (happened to me
:).
The compiler should raise an error.

In article <co0cg2$24hq$1@digitaldaemon.com>, Valéry Croizier says...
>
>Consider :
>
>class A { int x;}
>A[10] t;
>foreach(A a; t) // should be "A inout a"
>    a = new A;
>
>Later, an access to any element of 't' is likely to segfault (happened to me
>:).
>The compiler should raise an error.

This is completely legal from a syntactical perspective--D doesn't have any concept of logical const-ness.  All you're doing is assigning a new value to a variable.


Sean

On Tue, 23 Nov 2004 22:49:06 +0000 (UTC), Sean Kelly <sean@f4.ca> wrote:

> In article <co0cg2$24hq$1@digitaldaemon.com>, Valéry Croizier says...
>>
>> Consider :
>>
>> class A { int x;}
>> A[10] t;
>> foreach(A a; t) // should be "A inout a"
>>    a = new A;
>>
>> Later, an access to any element of 't' is likely to segfault (happened to me
>> :).
>> The compiler should raise an error.
>
> This is completely legal from a syntactical perspective--D doesn't have any
> concept of logical const-ness.  All you're doing is assigning a new value to a
> variable.

Correct, if however my earlier suggestion that modifying an 'in' variable should be illegal was used this would have given an error. This is the 3rd or 4th instance where the idea would have prevented a bug.

Regan

-- 
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/