I’ve heard that trampolining is an ineffective way of implementing TCO. How does DrScheme (PLAI Scheme, technically) do it? Does it do it the ‘right’ way (that is, produce assembly code which directly branches to the tail call, instead of going through the stack and trampolining)?
I’ve heard that trampolining is an ineffective way of implementing TCO. How does DrScheme
Share
The implementors of PLT Scheme are quite active in their Google group, where you can get a quick answer from the people who write the code.
I’m not sure they read SO, though, so your best bet would probably be asking there.