5 replies to this topic

[Misery]

Hi there,

I do have a bitses matrix class defined as follows:

template <class INT_TYPE,class CONTAINER_INT=uINT,
						  INT_TYPE ContainerSize=sizeof(CONTAINER_INT)*8,
						  class DATA_CONTAINER=BitsetDataContainer<INT_TYPE,CONTAINER_INT,ContainerSize> >
   class BitsetMatrix
   {
	///code
   };

And I have a function:

template <class INT_TYPE>
   INT_TYPE Sum(const BitsetMatrix<INT_TYPE> &M)
   {
	INT_TYPE Rt;
	if (!M.IsEmpty())
	{
		//count sum using another function
	}
	return Rt;
   }

How to define this function that it deduces INT_TYPE argument automatically? In VC++ it works fine, in GCC I need to write:

BitsetMatrix<int> BM;
Sum<int>(BM);

How to define function Sum to avoid <int> in its call.

Regards

[Álvaro]

I only have a passing understanding of templates in C++, but the way I would do it is by making Sum be a template of the matrix type, and the matrix can provide a typedef to access the underlying INT_TYPE.

[wqking]

Quick code, I'm not sure if it's compilable or work, don't have time to test for now.

template <class INT_TYPE,class CONTAINER_INT=uINT,

												  INT_TYPE ContainerSize=sizeof(CONTAINER_INT)*8,

												  class DATA_CONTAINER=BitsetDataContainer<INT_TYPE,CONTAINER_INT,ContainerSize> >

   class BitsetMatrix

   {

   public:

	 typedef INT_TYPE IntType;

		///code

   };





template <class T>

   typename T::IntType Sum(const T &M)

   {

		typename T::IntType Rt;

		if (!M.IsEmpty())

		{

				//count sum using another function

		}

		return Rt;

   }

Edited by wqking, 15 October 2012 - 04:26 AM.

[Misery]

@wqking: that looks promising :]

[wqking]

Oh, after reading alvaro's reply again, I just found his opinion is exactly same as my code.

[Álvaro]

Oh, after reading alvaro's reply again, I just found his opinion is exactly same as my code.

Of course! I just thought you were implementing my suggestion.