sure, that's not a problem, but it would probably be better if we stopped juggling around with so many different branches, it's been confusing enough with Gijs' and Hyde's work as is
MapStructure.nas shouldn't be heavily affected by this, it should be just 2-3 lines IIRC - and the prototype in geo.nas can then probably go away, P. ?
And yes Tom:, thank you VERY much for implementing this (and kudos, very dense code!).
@P.: if we now get the canvas/caching scheme working in time, most reported "performance" issues should be resolved because of the way MapStructure is "structured"
But if you want me to help with porting existing "layers", you should either carefully comment the code and document what's going in your .scontroller/lcontroller framework, or turn the design into a simple OOP class/interface that even people without any Nasal metaprogramming skills can understand
At the moment, I am getting segfaults when adding my own trivial (and valid!) code inside the Nasal/VM, which are basically impossible to troubleshoot currently, because of all the indirection via compile(), call() and bind() going on there