[fluxus] build error: PDataArithmetic.cpp

[james tittle]

hey all,

On Nov 28, 2005, at 7:43 AM, Dave Griffiths wrote:

>>
>> The recent changes to pdata in cvs give the following build error  
>> on osX
>> (ie. gcc 4.0)
>>
>> PDataArithmetic.cpp:172: error: specialization of 'template<class  
>> S, class
>> T> static fluxus::PData*
>> fluxus::ClosestOperator::Operate(fluxus::TypedPData<S>*, T)' in  
>> different
>> namespace
>>
>> PDataArithmetic.h:129: error:   from definition of 'template<class S,
>> class T> static fluxus::PData*
>> fluxus::ClosestOperator::Operate(fluxus::TypedPData<S>*, T)'
>
> I don't have gcc 4 - I can't see anything immediately obviously wrong
> here, it's defined in the same way as all the other functions - do  
> they
> all error?

...yeh, came across this too with gcc 4.0.1, and after a little  
hacking (meaning I don't know if this is right or not...), I got gcc  
to accept it...seems that in PDataArithmetic.h, the following is  
defined:

template<>
PData *ClosestOperator::Operate(TypedPData<dVector> *a, dVector *b);

...but then in PDataArithmetic.cpp:176, the following is used:

template <>
PData *ClosestOperator::Operate(TypedPData<dVector> *a, dVector b)

...so I added a "*" before "b", and that required doing the same  
inside the function, such that the new one starts out looking like this:

template <>
PData *ClosestOperator::Operate(TypedPData<dVector> *a, dVector *b)
{
	float closestdist=999999;
	dVector closest;
	for (unsigned int i=0; i<a->Size(); i++)
	{	
		float dist = a->m_Data[i].dist(*b);

...and then it compiles without warnings!  If this is a sane change,  
I can commit it to cvs...

l8r,
james