Files
sex/types.scm
alex-eg 1f445f0f9b implement type inference
Two things out of one mechanism. `_' as a type means "work it out from
the initializer", so (var n _ (strlen s)) stops needing size-t spelled
out; `type-of' hands a macro the type of an expression, so a macro can
dispatch on what it was handed rather than on what was declared. Both
read the same answers from two sides.

Algorithm W's core, intra-procedural, with the extensions C forces:

  - an unknown type, since (include stdio.h) brings in names we never
    parsed. Unification is consistency rather than equality, so
    anything touching an unparsed declaration stops constraining
    instead of rejecting a program that compiled yesterday;
  - the usual arithmetic conversions, since `+' is not a function of
    one type;
  - checking mode for initializers, since #(0 0) has no type of its own
    and takes one from its context. #(T : ...) is the way out of that.

What it wanted on the way:

  - what type a *name* has, which neither the typedef nor the tag
    database recorded. One table serves functions and variables, since
    a function type already has a surface spelling;
  - a scope chain, so a (var c int 9) inside a do ends with the block;
  - form-type, keyed by cons cell, so one form has one type;
  - macros expanded during the walk rather than before it, so type-of
    is answered in the scope the macro was written in.

Closures take the same machinery: a receiver whose type comes from a
call, captures written (name expr) and typed from the expression, and
conversion from a bare function wherever a closure is expected.

type-match grew `_' on the pattern side, since (closure ((int)) int)
and (closure ((float)) int) were separate clauses for one case.
2026-09-30 00:09:32 +03:00

240 lines
8.8 KiB
Scheme

;;; The type database.
;;;
;;; Every named aggregate, typedef and define the semantic engine
;;; walks past is recorded here, so that macros (or other forms) can
;;; ask what a type is made of. That is what lets a macro generate
;;; code from a struct's fields given nothing but its name.
;;;
;;; Entries are filled in as toplevel forms are processed, in order, so
;;; a type has to be declared before the macro that asks about it.
(import
scheme
(scheme base)
(chicken base)
srfi-1
srfi-69)
;;; Two namespaces: `struct point' and a `point' typedef are separate
;;; declarations. Tags -- struct, union and enum alike -- share the
;;; second table between them
(define +type-db+ (make-hash-table)) ; typedefs and defines
(define +tag-db+ (make-hash-table)) ; struct, union and enums
;;; What type a *name* has, which neither of the two above records:
;;;
;;; (fn sum ((a int) (b int)) int ...) -> (fn ((int) (int)) int)
;;; (var origin (struct point) ...) -> (struct point)
(define +name-db+ (make-hash-table))
(define (strip-pub form)
(if (eq? (car form) 'pub) (cdr form) form))
(define (comment-form? f)
(and (pair? f) (eq? (car f) 'comment)))
;;; Fields are written with the type last and one or more names before
;;; it, so ((x y int) (p (* char))) describes three fields. Flatten that
;;; into one (name type) per field, which is what a caller wants.
(define (normalize-fields fields)
(append-map
(lambda (field)
(if (comment-form? field)
(list)
(let ((type (last field))
(names (drop-right field 1)))
(map (lambda (name) (list name type)) names))))
(remove comment-form? fields)))
;;; ([pub] struct name (fields ...) . attrs)
(define (aggregate-fields form)
(let ((f (strip-pub form)))
(if (and (pair? (cddr f)) (list? (caddr f)))
(caddr f)
(list))))
(define (add-struct name form)
(hash-table-set! +tag-db+ name
(list 'struct name (normalize-fields (aggregate-fields form)))))
(define (add-union name form)
(hash-table-set! +tag-db+ name
(list 'union name (normalize-fields (aggregate-fields form)))))
;;; ([pub] enum name (value ...))
(define (add-enum name form)
(let ((f (strip-pub form)))
(hash-table-set! +tag-db+ name
(list 'enum name (if (and (pair? (cddr f)) (list? (caddr f)))
(caddr f)
(list))))))
;;; ([pub] typedef new-name target)
(define (add-typedef name form)
(hash-table-set! +type-db+ name
(list 'typedef name (last (strip-pub form)))))
;;; (define name value ...) -- a C #define, kept so a macro can read a
;;; compile-time constant rather than re-parse the source.
(define (add-define name form)
(hash-table-set! +type-db+ name
(list 'define name (cddr (strip-pub form)))))
;;; `(type-of x)' inside a macro body: the type of the expression the
;;; macro was handed, where `get-name-type' only answers for a name.
;;; The walker that can answer it lives in `semen', which is compiled
;;; after this, so it installs itself here for the length of one
;;; expansion. Outside one there is no scope to ask about, and the
;;; answer is #f.
(define current-type-of (make-parameter (lambda (form) #f)))
(define (type-of form) ((current-type-of) form))
(define (add-name-type! name type)
(hash-table-set! +name-db+ name type))
;;; #f for a name never declared, which is what `printf' looks like
;;; until something parses stdio.h. Not an error here; the caller
;;; decides.
(define (get-name-type name)
(hash-table-ref/default +name-db+ name #f))
;;; What `(make-adder 10)' has for a type: `make-adder's return type,
;;; or #f when NAME is not a function with a signature on record
(define (get-return-type name)
(let ((type (get-name-type name)))
(and (pair? type)
(eq? 'fn (car type))
(= 3 (length type))
(third type))))
(define (get-tag-info name)
(hash-table-ref/default +tag-db+ name #f))
;;; First look up the ordinary identifier, then tag of that id when
;;; no ordinary one was declared, as C does
(define (get-type-info name)
(or (hash-table-ref/default +type-db+ name #f)
(get-tag-info name)))
;;; ((name type) ...) for a struct or union, #f for anything else --
;;; including a name that was never declared. Callers give the better
;;; error, since they know what they wanted it for.
;;;
;;; A typedef is followed to what it stands for, so reflection over an
;;; alias works exactly as it does over the name it aliases.
(define (get-fields name)
(let ((info (resolve-type-info name)))
(and info
(memq (car info) '(struct union))
(caddr info))))
;;; Follow a typedef chain to the name it stands for. #f if NAME is
;;; not a typedef. A typedef that leads back to itself stops rather
;;; than spinning: nothing prevents one from being written.
(define (get-underlying-type name)
(let follow ((name name) (seen (list)))
(and (not (member name seen))
(let ((info (get-type-info name)))
(and info
(eq? (car info) 'typedef)
(let ((target (caddr info)))
(or (and (symbol? target) (follow target (cons name seen)))
target)))))))
;;; The declaration NAME ultimately names. For a typedef that is the
;;; entry of whatever it stands for, and for anything else it is
;;; NAME's own info
(define (target-tag target)
(cond
((symbol? target) target)
((and (pair? target)
(memq (car target) '(struct union enum))
(pair? (cdr target))
(symbol? (cadr target)))
(cadr target))
(else #f)))
(define (resolve-type-info name)
(let ((info (resolve-ordinary-type-info name)))
(if (and info (memq (car info) '(struct union enum)))
info
(or (get-tag-info name) info))))
(define (resolve-ordinary-type-info name)
(let ((info (get-type-info name)))
(and info
(if (eq? (car info) 'typedef)
(let ((tag (target-tag (get-underlying-type name))))
;; A typedef target is written in type position, where
;; `struct point' means the tag
(and tag (or (get-tag-info tag) (get-type-info tag))))
info))))
;;; Type matcher macro
;;; (type-match type
;;; (int ...)
;;; ((* const char) ...)
;;; ([int 10] ...)
;;; ((* _) ...) ; a pointer to anything
;;; ((¤ _ _) ...) ; an array of anything, any length
;;; (else ...))
;;;
;;; A type is a form, not an atom, so patterns are matched structurally
;;; rather than dispatched on like `case'. They are literal types and
;;; are not evaluated; `else' is optional and the whole thing is #f when
;;; nothing matches and there is no else.
;;;
;;; `_' in a pattern matches anything in that position, the same thing
;;; it means in a type. Without it every spelling has to be enumerated:
;;; `(closure ((int)) int)' and `(closure ((float)) int)' are separate
;;; clauses for what is one case.
;;;
;;; A `_' written last takes everything that remains, because a type's
;;; words are spread rather than nested -- `(* const char)' is three
;;; elements, so `(* _)' has to cover two of them to mean "a pointer to
;;; anything".
;;;
;;; Nothing destructures: a macro body is ordinary Scheme and a type is
;;; a list, so `(caddr (type-of x))' already reads the length out of
;;; `(¤ int 4)'.
(define-syntax type-match
(syntax-rules (else)
((_ type) #f)
((_ type (else body ...)) (begin body ...))
((_ type (pattern body ...) clause ...)
(if (type-pattern-matches? 'pattern type)
(begin body ...)
(type-match type clause ...)))))
(define (type-pattern-matches? pattern type)
(cond
((eq? pattern '_) #t)
((and (pair? pattern) (pair? type))
(if (and (eq? (car pattern) '_) (null? (cdr pattern)))
#t ; a trailing `_' takes the rest
(and (type-pattern-matches? (car pattern) (car type))
(type-pattern-matches? (cdr pattern) (cdr type)))))
(else (equal? pattern type))))
;;; Map function to each field/value of a structure/union/enum
;;; For enums, field-type is the type of the enum (since C 23)
;;; (map-fields type-name
;;; (lambda (field-name field-type) ...))
;;;
;;; Returns #f if nothing of that name was declared
(define (map-fields struct-union-enum fn)
(let ((info (resolve-type-info struct-union-enum)))
(and info
(case (car info)
((struct union)
(map (lambda (field) (fn (car field) (cadr field)))
(caddr info)))
;; An enumerator's type is the enum itself -- named as it was
;; declared, since `enum some-typedef' is not a C type.
((enum)
(let ((type (list 'enum (cadr info))))
(map (lambda (value) (fn value type))
(caddr info))))
(else #f)))))