Added size limit check for buffer stride

[stefanatwork]

Preventing possible wrap to zero when itemCount * bufferStride exceed 64 bit uint limit.

[svenwoop]

This does still not do the entire job. There is some "buffer range out of bounds" check in buffer.h line 251. First that check is wrong IMHO, as it should test to buffer_in->numBytes only on the right hand side and not multiply by stride_in.
Second even if fixed, the calculation offset_in+stride_in*num_in can still overflow if a large offset_in is used.

[svenwoop]

That check in buffer.h should better be placed inside rtcore.cpp actually, thus all valid checks are at main API wrapper file.

[svenwoop]

The check offset + Nstride >= buffer_size can be written as Nstride >= buffer_size - offset. Also check first if offset <= buffer_size.

[stefanatwork]

@svenwoop can you take another look?