language_fan wrote:
I read the comments and I think some of them are justified. You cannot
really expect the way you built dmd to be the only alternative. Efficient
closures can be implemented differently if you have a VM that supports
precise generational gc, region inference, and the language is a bit more
value oriented (= functional).
I don't understand how the article implies that it describes the only
way to do it. It only describes *a* way.
- Re: How Nested Functions Work, part 2 Walter Bright
