I believe this is possible. For example you could have a Brightscript channel that calls an NDK video playback function for playing video from your special DRM solution that we aren't going to build into the platform. The caveat is that jumping from Brightscript > NDK and back is slow.
NDK channels all launch from Brightscript in the first place.
Yeah, i saw roNDK.start(...), hence suspected it is doable. So i gather there might be alternative (possibly manu-matic) way, to assign different executables for different platforms. Funny i did not even think of that, native being different on different players - so used i am with BrSc giving reasonably "write once, run everywhere".
I like to scout ahead as not to get myself into a swamp*. And currently i am scratching my head, since there are 3 major segments of Roku players in the wild:
firmware 3, model# < 2200
no OpenGL, models 2450 .. 27xx
OpenGLES, everything else
Do you guys have any stats that you can share with us - of what the % breakdown between these is in the wild? Preferably from players still in use - but even if based on total sales figures, that's fine too. Rough break down. Any bone you can throw. Anything is better than nothing. :idea:
(*) Last time i blindly trusted something was this June, when i saw that GetChildNodes() has been added to roXML in April. And for not sticking to Lenin/Reagan's favorite proverb "Trust but verify", i got burned in August when started testing my code on fw3. Yeah, i am still bitter melon over it, sorry. The power of documentation... it should be written by someone with attention to detail.