ID 235135 - Update the DD::Image::DeepPixel::getOrderedSample documentation

Follow

Problem summary
Update the DD::Image::DeepPixel::getOrderedSample   description to reflect how it actually works

Reproduced by support
This is found in:
https://www.thefoundry.co.uk/products/nuke/developers/100/ndkreference/Plugins/classDD_1_1Image_1_1DeepPixel.html

Expected behaviour
For the information to be accurate, as the developer describes it:

In deep pixel getOrderedSample we do one of three things based on ordering (and from what we can see the ordering is set to be unordered in every case):

if data is already in descending order, with largest Z first, we simply returning it by index - this should give the largest Z sample first.
if data is already in ascending order, with smallest Z first, we return it in the reversed order - this should also give the largest Z sample first.
in the unordered case we sort a std::vector< std::pair< float, size_t > > using std::sort where the float is depth, then we access the list in the reverse of this order - this should also give the largest Z sample first.

Actual behaviour
The comment for this function says the opposite:
      /**
      * Get a particle sample (by depth) and channel.  Samples will be ordered so
      * that if depthIdx == 0, it will fetch the sample with the lowest Z.
      */

    We're sorry to hear that

    Please tell us why