Improve rendering performance by not relying on coroutines

[FSMaxB]

It's possible to get a 20-25% speedup in rendering performance, by not using coroutines in slt2.render. This can even be implemented without changing the API and without breaking slt2.render_co. See https://github.com/FSMaxB/liluat/pull/7 for reference.

The only thing that has to be changed is the code generation in slt2.loadstring.