DMD 0.148 release

[Ivan Senji]

Georg Wrede wrote:
> Ivan Senji wrote:
>> And if I did missunderstand something I apologize (it's late).
> 
> 
> No problem. The way this whole thread is going, half the time everybody 
> is clueless. :-)

:)

>> Are you now saying you want real booleans that are not ints?
> 
> 
> On the inside they should be ints. Mostly for efficiency reasons, but 
> also for established coding conventions.

I agree completly that on the inside they should be ints for some 
reasons. But not as a must. On a different arhitecture something else 
might be more efficient.

> 
> But on the outside, they should not be compatible with ints.
> 
> In other words, doing
> 
> bool foo = 55 && 5000;
> 
> should be legal, 

You trying to scare me away from this thread? ;)

I believe programming language constructs should(mostly) be translatable 
to a human language.
but:

if(55){Georg.giveApple();}

means:

if 55 give Georg an apple.

Does Georg expect to get an apple? I really don't know. This doesn't 
have any meaning.

> and it should store the 'true value' (that'd be a 1) 
> into foo.

If by "store the 'true value' into foo" you mean "compile time error: 
cannot implicitly conver from int to bool" than I agree! :)

> 
> But doing:
> 
> bool a = 1;  // legal, but poor coding
> bool b = 0;  // likewise
> whateverType bar = a + b;  // should produce a compiler error
>                            // since addition of bools is illegal

Hm! Addition of bools is illegal but assigning random integers into them 
is OK.

>>
>> My example was supposed to be an example for:
>> a) real bools being needed
>> b) int bool being bad
> 
> 
> I'm still waiting.

No need to wait, the examples are there. :)