Implementing Half Floats in D

Era Scarecrow rtcvb32 at yahoo.com
Tue Jan 29 01:43:01 PST 2013


On Tuesday, 29 January 2013 at 08:48:24 UTC, Simen Kjaeraas wrote:
> On 2013-49-29 08:01, Tove <tove at fransson.se> wrote:
>> Maybe you could also demonstrate that it's possible to 
>> implement another literal syntax?
>> HalfFloat h = 1.3.hf;
>>
>> some people will prefer that for sure.
>
> Except that's not a literal. 1.3.hf would be a function call, 
> while hf!1.3f is a template instantiation.

  Unless it's an enum or something that requires it to know it 
during compile-time.

  enum h = 1.3.hf; //potentially correct

  But I doubt it would be that useful. More likely during 
optimization static values would be equal to it.


More information about the Digitalmars-d mailing list