Adding Unicode operators to D

[Andrei Alexandrescu]

Simen Kjaeraas wrote:
> On Sun, 26 Oct 2008 22:28:16 +0100, Bill Baxter <wbaxter at gmail.com> wrote:
> 
>> On Sun, Oct 26, 2008 at 11:02 PM, Simen Kjaeraas 
>> <simen.kjaras at gmail.com> wrote:
>>> On Sat, 25 Oct 2008 12:14:47 +0200, Spacen Jasset 
>>> <spacenjasset at yahoo.co.uk>
>>> wrote:
>>>
>>>> Why unicode anyway? In the same way that editor support is required to
>>>> actually type them in, why not let the editor render them. So 
>>>> instead of
>>>> symbol 'x' in the source code, say:
>>>>
>>>> m3 = m1 cross_product m2
>>>>
>>>> as an infix notatation in a similar way to the (uniary) sizeof 
>>>> operator.
>>>>
>>>>
>>>> While cross_product is a bit long and unwieldy any editor capable can
>>>> replace the rendition of that keyword with a symbol for it. But in 
>>>> editors
>>>> that don't it means that it still can be typed in and/or displayed 
>>>> easily.
>>>>
>>>> Another option includes providing cross_product as an 'alias' and 'X'
>>>> aswell.
>>>>
>>>> Which then leads on to the introduction of a facility to add arbitary
>>>> operators, which could be interesting becuase you can supply any 
>>>> operator
>>>> you see fit for the domains that you use that require it. -- This 
>>>> provide
>>>> exactly the right solution though as all the additions would be 'non
>>>> standard' and I can see books in the future recommending people not use
>>>> unicode operators, becuase editors don't have support for them.
>>>
>>> This made me think. What if we /could/ define arbitrary infix 
>>> operators in
>>> D? I'm thinking something along the lines of:
>>>
>>>
>>> operator cross_product(T, U)
>>> {
>>>  static if (T.opCross)
>>>  {
>>>    T.opCross(T)
>>>  }
>>>  else static if (U.opCross)
>>>  {
>>>    U.opCross_r(T);
>>>  }
>>>  else
>>>  {
>>>    static assert(false, "Operator not applicable to operands.");
>>>  }
>>> }
>>>
>>> alias cross_product ×;
>>>
>>>
>>> I'm not sure if this is possible, but it sure would please downs. :P
>>
>> What's the precedence of your user-defined in-fix operator?
>>
>> --bb
> 
> Yup, I realized this myself as well. Seemed like such a great idea when 
> I only thought of it for three seconds. :p

An operator could always be defined to have the same precedent as an 
existing operator, which it has to specify.

Andrei