Alex Wood

Aubrey Jaffer wrote a brief article on this: https://arxiv.org/abs/1310.8121v1 The code in "Reading" is in Scheme, but should be easily adaptable. It's only a few more lines than what Eclector...

upon further examination, no adaptation of `round` is actually needed. So you can pretty much use the code directly, substituting `round` for `round-quotient`, and `float` for `exact->inexact` and such. Plus...

Sorry I'm only replying to this now - I have been too busy with cvm/bytecode to think about this library. > I understand your analogy of type definition to functions...

The `tagbody` page describes tags as "go tag"s with a link to the glossary in its Arguments and Values. I think that's sufficient. I mean, analogously the `defun` page does...

You have the rest type as bottom, which is probably not correct. Since the function takes keywords, any type of argument can be validly provided in the rest list (if...

a rest type of bottom indicates that that the function is not passed arguments other than required and optional. so if the rest type is bottom, keyword arguments, which are...

It's quite possible keyword parameter types are in fact broken. I have not used this library as thoroughly as I had hoped. I might be able to take a look...

okay, yeah, somewhat more fundamental problem is that I did not implement function type subtypep in the presence of keyword types: https://github.com/s-expressionists/ctype/blob/main/cfunction.lisp#L42-L44

with respect to `arithmetic-error-operands`, the suggestion to make them generic functions would probably involve the issue to make condition types into classes, which i put in a branch since it's...

cleavir-macroexpand-all has not been maintained.