This may have been overlooked in my other thread so I wanted to ask again:

This seems very inconsistent, does a += b not lower to a = a + b? I guess not based on the below:

    ushort a = ushort.max, b = ushort.max;


    a += b; // Compiles fine
    a = a + b; // Error: cannot implicitly convert expression (cast(int)a + cast(int)b) of type int to ushort

On 11/05/2015 05:20 AM, ixid wrote:

> This seems very inconsistent, does a += b not lower to a = a + b?

Apparently not:

  http://dlang.org/expression.html#AssignExpression

It says "The right operand is implicitly converted to the type of the left operand". So, the rules are different.

Ali

On Thursday, 5 November 2015 at 13:20:26 UTC, ixid wrote:
> This may have been overlooked in my other thread so I wanted to ask again:
>
> This seems very inconsistent, does a += b not lower to a = a + b? I guess not based on the below:
>
>     ushort a = ushort.max, b = ushort.max;
>
>
>     a += b; // Compiles fine
>     a = a + b; // Error: cannot implicitly convert expression (cast(int)a + cast(int)b) of type int to ushort

What's inconsistent is the integral promotion of the add expression result that stops from 4 bytes int:

---
int a, b;
a += b;
a = a + b;
---

is compiled but according to the specs, a + b result should be widened to long:

http://dlang.org/expression.html#AddExpression

(ubyte, byte) until (uint int) should be widened and (long , ulong) wrapped. This behavior would match the specs better.

On 11/06/2015 04:56 PM, BBaz wrote:
> On Thursday, 5 November 2015 at 13:20:26 UTC, ixid wrote:
>> This may have been overlooked in my other thread so I wanted to ask
>> again:
>>
>> This seems very inconsistent, does a += b not lower to a = a + b? I
>> guess not based on the below:
>>
>>     ushort a = ushort.max, b = ushort.max;
>>
>>
>>     a += b; // Compiles fine
>>     a = a + b; // Error: cannot implicitly convert expression
>> (cast(int)a + cast(int)b) of type int to ushort
>
> What's inconsistent is the integral promotion of the add expression
> result that stops from 4 bytes int:
>
> ---
> int a, b;
> a += b;
> a = a + b;
> ---
>
> is compiled but according to the specs, a + b result should be widened
> to long:
>
> http://dlang.org/expression.html#AddExpression
>
> (ubyte, byte) until (uint int) should be widened and (long , ulong)
> wrapped. This behavior would match the specs better.

You say 'long' but according to integer promotions, a and b should both be promoted to 'int' and the result of 'a + b' is int:

  http://dlang.org/type.html#integer-promotions

Ali

On Saturday, 7 November 2015 at 01:10:01 UTC, Ali Çehreli wrote:
> On 11/06/2015 04:56 PM, BBaz wrote:
>> On Thursday, 5 November 2015 at 13:20:26 UTC, ixid wrote:
>>> [...]
>>
>> What's inconsistent is the integral promotion of the add expression
>> result that stops from 4 bytes int:
>>
>> ---
>> int a, b;
>> a += b;
>> a = a + b;
>> ---
>>
>> is compiled but according to the specs, a + b result should be widened
>> to long:
>>
>> http://dlang.org/expression.html#AddExpression
>>
>> (ubyte, byte) until (uint int) should be widened and (long , ulong)
>> wrapped. This behavior would match the specs better.
>
> You say 'long' but according to integer promotions, a and b should both be promoted to 'int' and the result of 'a + b' is int:
>
>   http://dlang.org/type.html#integer-promotions
>
> Ali

oh...sorry I thought that the widening was done to the follwing type that's bigger, eg byte->short, short->int.

So no inconsistence at all.