void func( int c )
{
    ubyte a;

    a = cast(ubyte)c + cast(ubyte)c;
}

dmd v2.065 complains about:
test.d(5): Error: cannot implicitly convert expression (cast(int)cast(ubyte)c + cast(int)cast(ubyte)c) of type int to ubyte

As far as I can tell, adding two values of the same type should result in the same type.

Shachar

the result of ubyte + ubyte is int I believe.

Try;
void func( int c )
{
     ubyte a;

     a = cast(ubyte)(cast(ubyte)c + cast(ubyte)c);
}

On Wednesday, 17 September 2014 at 13:03:05 UTC, flamencofantasy wrote:
> the result of ubyte + ubyte is int I believe.
>
> Try;
> void func( int c )
> {
>      ubyte a;
>
>      a = cast(ubyte)(cast(ubyte)c + cast(ubyte)c);
> }

From http://dlang.org/type, under Usual Arithmetic Conversions:
4. Else the integer promotions are done on each operand, followed by:
    1. If both are the same type, no more conversions are done.

So, as far as I can see, the specs disagree with you.

Shachar

On Wed, 17 Sep 2014 13:20:13 +0000
Shachar via Digitalmars-d-learn <digitalmars-d-learn@puremagic.com>
wrote:

>  From http://dlang.org/type, under Usual Arithmetic Conversions:
> 4. Else the integer promotions are done on each operand, followed
> by:
>      1. If both are the same type, no more conversions are done.
it's bug in specs, i believe, 'cause compiler promotes smaller types to int/uint.

On 17/09/14 16:32, ketmar via Digitalmars-d-learn wrote:
> On Wed, 17 Sep 2014 13:20:13 +0000
> Shachar via Digitalmars-d-learn <digitalmars-d-learn@puremagic.com>
> wrote:
>
>>   From http://dlang.org/type, under Usual Arithmetic Conversions:
>> 4. Else the integer promotions are done on each operand, followed
>> by:
>>       1. If both are the same type, no more conversions are done.
> it's bug in specs, i believe, 'cause compiler promotes smaller types to
> int/uint.
>

I don't understand. Why is this behavior preferrable to the one outlined by the specs?

Thanks,
Shachar

Because of overflow.



On Wednesday, 17 September 2014 at 13:36:42 UTC, Shachar Shemesh wrote:
> On 17/09/14 16:32, ketmar via Digitalmars-d-learn wrote:
>> On Wed, 17 Sep 2014 13:20:13 +0000
>> Shachar via Digitalmars-d-learn <digitalmars-d-learn@puremagic.com>
>> wrote:
>>
>>>  From http://dlang.org/type, under Usual Arithmetic Conversions:
>>> 4. Else the integer promotions are done on each operand, followed
>>> by:
>>>      1. If both are the same type, no more conversions are done.
>> it's bug in specs, i believe, 'cause compiler promotes smaller types to
>> int/uint.
>>
>
> I don't understand. Why is this behavior preferrable to the one outlined by the specs?
>
> Thanks,
> Shachar

http://www.drdobbs.com/tools/value-range-propagation/229300211

On Wed, 17 Sep 2014 16:36:41 +0300
Shachar Shemesh via Digitalmars-d-learn
<digitalmars-d-learn@puremagic.com> wrote:

> I don't understand. Why is this behavior preferrable to the one outlined by the specs?
'cause C does exactly this. there is no reason to confuse people with C background by silently removing such feature.

On Wednesday, 17 September 2014 at 13:20:15 UTC, Shachar wrote:
> On Wednesday, 17 September 2014 at 13:03:05 UTC, flamencofantasy wrote:
>> the result of ubyte + ubyte is int I believe.
>>
>> Try;
>> void func( int c )
>> {
>>     ubyte a;
>>
>>     a = cast(ubyte)(cast(ubyte)c + cast(ubyte)c);
>> }
>
> From http://dlang.org/type, under Usual Arithmetic Conversions:
> 4. Else the integer promotions are done on each operand, followed by:
>     1. If both are the same type, no more conversions are done.
>
> So, as far as I can see, the specs disagree with you.

You missed that "integer promotions" are done first.

<http://dlang.org/type.html#Integer%20Promotions>:
> Integer Promotions are conversions of the following types:
[...]
> from	to
[...]
> ubyte	int
[...]

So, integer promotions turn ubyte + ubyte into int + int.