Hi @YafimSimanovsky, I am one of the co-developers of Honeybee[+].
The short, and simple, reply to your question would be that these libraries have not been formally (if at all) tested on Rhino-Grasshopper installations on Mac OS-X systems. So, it is hard to tell which features might work and which might not.
For a long and “it’s complicated” reply read on…
The core libraries (i.e. the ones that do most of the numerical and computational work) within both LB[+] and HB[+] have been cross-platform-compatible for nearly two years.
Perhaps the most definitive testament to that is the fact that @mostapha, @MingboPeng and a few other programmers were able to create a prototypical cloud-based server through less than 36 hours of non-stop coding for performing HB[+] simulations on the cloud during the AEC Hackathon in New York a few months ago.More recently, @AntoineDao has made some really exciting contributions by creating docker-based implementations of the HB[+] library whereby, among other things, it is possible to scale HB[+] simulations to run on 100s of cores on the cloud.
There simply hasn’t been enough of a groundswell of demand ,and more critically, support, from Mac-based Rhino users for the development to gain significant traction on that platform. Besides the occasional how-to-make-this-work type questions from people, there haven’t been any serious users who’ve been interested in investing time and effort to test/deploy the code on OS-X platforms. As far as I know, none of the core developers of Ladybug Tools are using Mac systems as their primary computers. So, we need some active engagement from the community to make this happen. One doesn’t have to know coding to support this, however, we’ll require someone to run some scripts/components and report back about the crashes/results. If you are interested, I would suggest opening an issue in the Honeybee or Ladybug repository. It might be at least a couple of weeks to a month before any tangible progress is seen.
Once deployed, HB[+] users will be able to perform annual daylighting simulations at least 4 to 8 times faster than Windows machines with similar hardware configurations.