Hi Christian,
On Wed, Jun 23, 2010 at 10:42 AM, Christian Henning wrote:
The only difference between what you suggested and what I'm posting is
the addition of some explicit construction calls:
I'm sorry I cannot make out your addition of some explicit constructor
calls. Did you send the correct patch snippet? Could we just post the
code without the patch syntax? I don't use patch.
Sorry, I could have been clearer. The original code used only
explicit construction calls. Your suggested code replaced them with
static_cast (which would work great if they were casting to the
correct type instead of just an integral type). I found that both
were necessary to suppress warnings. Here's the new code:
// Both source and destination are unsigned integral channels,
// the dst max value is less than (or equal to) the src max value,
// and the src max value is not divisible by the dst max value
template
struct channel_converter_unsigned_integral_nondivisible
{
DstChannelV operator()(SrcChannelV src) const {
typedef typename
unsigned_integral_max_value<SrcChannelV>::value_type src_integer_t;
typedef typename
unsigned_integral_max_value<DstChannelV>::value_type dst_integer_t;
static const double div =
unsigned_integral_max_value<SrcChannelV>::value /
double(unsigned_integral_max_value<DstChannelV>::value);
static const src_integer_t div2 = static_cast(div/2);
return DstChannelV(static_cast((src + div2) / div));
}
};
} // namespace detail
/////////////////////////////////////////////////////
/// bits32f conversion
/////////////////////////////////////////////////////
template <typename DstChannelV> struct
channel_converter_unsigned : public
std::unary_function {
DstChannelV operator()(bits32f x) const {
typedef typename
detail::unsigned_integral_max_value<DstChannelV>::value_type
dst_integer_t;
return DstChannelV(static_cast(x*channel_traits<DstChannelV>::max_value()+0.5f));
}
}
Thanks for your help on this, and for you work on the library. Good
luck on the IO review!
Nate
