Passing arrays by value?

Lionello Lunesu lionello at lunesu.remove.com
Mon Feb 5 10:28:28 PST 2007


"Jarrett Billingsley" <kb3ctd2 at yahoo.com> wrote in message 
news:eq7c0j$115c$1 at digitaldaemon.com...
> "Lionello Lunesu" <lio at lunesu.remove.com> wrote in message 
> news:eq78nh$30b1$1 at digitaldaemon.com...
>> The spec doesn't specify how fixed-sized arrays are passed to functions 
>> (or, I did not find it):
>>
>> void foo( in float[2] array, float x, float y )
>> {
>>     array[0] = x;
>>     array[1] = y;
>> }
>>
>> It seems to me that an fixed-sized array like that could well be passed 
>> by value:
>>
>> void by_ref( in float[] a, ... );
>> void by_value( in float[2] a, ... );
>> void by_ref( inout float[2] a, ... );
>>
> Fixed-size arrays are passed by reference.  But I suppose passing smaller 
> arrays by value could have its uses..

Yeah. I just thought that it would be useful for that vector proposal in 
digitalmars.D. Walter mentioned that he'd like float[3] to behave like a 3D 
vector, but with vectors you'd want to have a choice to pass them either by 
value or by ref. Seemed logical to use in/inout for that, but at the moment, 
it doesn't really do anything for state/fixed-sized arrays.

L. 




More information about the Digitalmars-d-learn mailing list