[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: implementation of SRFI 96 for Larceny
Will Clinger has successfully ported SRFI-96 and SLIB to Larceny.
Some of his comments in
are relevant to SRFI-96.
| ; FIXME: This should really be ".sch", but then SLIB wouldn't work.
| (define (scheme-file-suffix) ".scm")
scheme-file-suffix would have been more appropriately named
slib-source-file-suffix. It is specific to SLIB files.
| ;;; FIXME:
| ;;; SLIB appears to need these things, even though SRFI 96 doesn't
| ;;; specify them or indicate that implementations of SRFI 96 must
| ;;; define them.
| (define (macro:eval . args) (apply slib:eval args))
| (define (macro:load . args) (apply load args))
That was true of the first version of SRFI-96, and has been rectified.
| ;;; FIXME: With Larceny's implementation of defmacro,
| ;;; macros that have been defined using defmacro are
| ;;; indistinguishable from macros that have been defined
| ;;; using Larceny's low-level explicit-renaming facility.
| ;;; For SLIB, however, it's probably good enough to pretend
| ;;; that all low-level macros were defined using defmacro.
| ;;; FIXME: This is terribly representation-dependent,
| ;;; and will break when (not if) the representation of
| ;;; macros changes.
| (define (defmacro? m)
| (let ((x (environment-get-macro (interaction-environment) m)))
| (and x (procedure? (cadr x)))))
I will remove defmacro? from SRFI-96.
| ;;; FIXME: The specification of defmacro:eval says it
| ;;; has to use slib:eval, but the definition of
| ;;; defmacro:load says no such thing.
| (define defmacro:load load)
I will change the wording for defmacro:eval to mirror slib:eval
instead of referring to it.
Function: defmacro:eval e
Returns the value of evaluating scheme expression e where all its
defmacros have been expanded.
| ;;; FIXME: There doesn't seem to be any analogue of
| ;;; macroexpand-1 in Larceny.
| (define (macroexpand-1 e) (macro-expand e))
macroexpand-1 is a Common-Lisp legacy which I think was useful for
debugging defmacros, but which doesn't need to be supported. I will
remove it from SRFI-96.
I have added an explanation of the relationship between macroexpand
I have uploaded the updated srfi-96.html to: