ker wrote:
Does anything speak against making that volume pointer argument
1. const
I think it should be const, but actually I just gave it a quick test and it doesn't compile. The problem is that the Volume::Sampler also needs to be fixed for const correctness (and then there may be further changes). This should be done, but means it's no longer a 5 minute job. I'll leave this bug open until this gets sorted.
ker wrote:
2. a reference
Not sure about this one... but it
probably should be changed. I'll need to think about it some more. One point of concern is that we now have the start of some image processing classes (VolumeResampler, LowPassFilter, etc). More will probably be added, and it is possible that it will be nice to chain them together and then execute the whole chain at once. Toolkits such as ITK use this approach. I don't know yet if it makes sense for PolyVox, but if it does then using pointers may make it easier to build such chains. We'll have to see how that develops.