We can probably drop canvas2d support in the near future, but right now it means dropping support for IE9/IE10 and more importantly a percentage of users with blacklisted graphics drivers, which is probably still around 5% of Chrome users and possibly higher for other browsers.
I would rather wait until then simply so we don't need to reimplement the same features in a soon-to-be-dead technology with some really annoying aspects (e.g. the way tiling works in canvas2d is totally braindead), and so we don't cause a whole string of "why does my game randomly break on a small number of systems" issues, which will be a massive pain for everybody. Just because you don't think it's a big deal doesn't mean it won't be a big problem for others, especially if there's no clear reason a particular feature wouldn't work sometimes. It would be better to pull the plug in one go and simply require WebGL at some point, so the game either works correctly, or doesn't start. Then at the same time we can add some other WebGL-only features that we've been putting off.
Hello, this post is 2 years old but ... OpenGL has already been adopted and the tilemap object can not vary its angle, could you implement this ?.