Eitaro Fukamachi
Eitaro Fukamachi
Try `40ants/reblocks` since every project of Ultralisp has user names as naming conventions.
@macrologist Good point. I agree that it's better to allow the same names for each constructor. But, unlike struct, `define-type` only defines a single type and can't tell the subtype...
Well, adding accessor dispatching at run-time may raise a run-time error if the object doesn't have the field. I don't think it is acceptable. ```common-lisp (define-type (Point :num) (P2D (.x...
How is it about `.z`?
I made some changes to the codegen. * Generate a dispatching function for accessors that share the same name. * Generate a direct function call when the name is unique....
> I think runtime errors in accessors is acceptable if it has a different syntax I'm not familiar with Coalton's naming convention, but it could be `unsafe-` for example? ```common-lisp...
Thank you for reporting! I believe it has been fixed in https://github.com/fukamachi/rove/commit/9868028edf511da5fe61355881af9f4c35f051df.
Thank you for your suggestion. It looks quite natural to do it in "the CLOS way", but as @daninus14 said, it's not how RDB works. I've been struggling to provide...