Duy Nguyen

> Documenting it sounds good. Is there an easy way for the compiler to check whether the return value of begin has a receiver? If there is, it could return...

> A straightforward way is to generate loading #\ by (begin), then expect it to be removed in dead-code elimination pass. Does not seem to work. The change (I assume)...

Aha. Good reason to close this issue then?

Off topic, I suppose you did consider at some point to make the precompiler generate C code instead of VM instructions and decided not to do it. After all you...

Slightly off topic again, but ``` ;;=============================================================== ;; Pass 3. Optimization stage 2 ;; ;; Closure optimization can introduce superfluous $LET, which can ;; be optimized further. (In fact, pass2...

> Which means we could repeat passes 2 and 3 at least a couple more times in precompile mode (if experiments show that it produces better code)? Bad news. I...

Regarding native code compiler (through C), this link looks interesting because emacs is lispy and has a stack-based VM. It looks like they take stack-based byte code (Gauche's pass5), optimize...

> Actually I'm doing it. It's not committed to the master branch. The idea is that I have an alternative pass5 which first construct basic blocks for abstract register machine,...

Do we even want to report these invisible bindings? The apropos command's description says "Show names of global bindings", and as far as I can tell, these are not global...

Another observation (not necessarily wrong) with apropos. I'm adding srfi 171 as a r7rs module and srfi-171.scm as (define-module srfi-171 (extend srfi.171)) Apropos output looks a bit funny, although it's...