[regexp] string-only regexp matching

This commit is contained in:
ben 2015-12-06 21:50:51 -05:00
parent 9ae2731195
commit f2c3b38fbb
2 changed files with 73 additions and 1 deletions

View File

@ -47,6 +47,5 @@ This library's [Scribble docs](TODO) give full details on the static checks.
- `printf`
#### `dependent/regexp`
- `regexp-match?`
- `regexp-match`

73
regexp.rkt Normal file
View File

@ -0,0 +1,73 @@
#lang typed/racket/base
(provide
regexp-match!
;; (-> Pattern String Any * (U #f (List String *N+1)))
;; Match the regular expression pattern against a string.
;; If the pattern is determined statically, result will be either #f
;; or a list of N+1 strings, where N is the number of groups specified
;; the pattern.
;;
;; Will raise a compile-time exception if the pattern contains unmatched groups.
)
(require
(for-syntax racket/base syntax/parse racket/syntax))
;; =============================================================================
(define-syntax regexp-match!
(syntax-parser
[(f pat-stx arg* ...)
#:when (or (string? (syntax-e #'pat-stx))
(regexp? (syntax-e #'pat-stx)))
#:with num-groups (count-groups (format "~a" (syntax-e #'pat-stx)) #:src #'f)
#:with ((index* . group-id*) ...)
#`#,(for/list ([i (in-range (syntax-e #'num-groups))])
(cons i (format-id #'f "group-~a" i)))
;; Chaining list-ref?
#'(let ([m (regexp-match pat-stx arg* ...)])
(if m
(let ([group-id* (or (list-ref m index*) (error 'regexp-match! "Internal error, try Racket's `regexp-match`"))] ...)
(list (car m) group-id* ...))
m))]
[(f arg* ...)
(syntax/loc #'f (regexp-match arg* ...))]))
(define-for-syntax (count-groups v #:src stx)
(cond
[(string? v) (count-groups/string v #:src stx)]
[(regexp? v) (count-groups/regexp v #:src stx)]
[else (error 'regexp-match! "Internal error on input" v)]))
;; Count the number of matched parentheses in a regexp pattern.
;; Raise an exception if there are unmatched parens.
(define-for-syntax (count-groups/string str #:src stx)
(define last-index (- (string-length str) 1))
(let loop ([i 0] [in-paren #f] [num-groups 0])
(if (> i last-index)
(if in-paren
(group-error str (format "'(' at index ~a" in-paren))
num-groups)
(case (string-ref str i)
[(#\()
(loop (+ i 1) i num-groups)]
[(#\))
(unless in-paren
(group-error str (format "')' at index ~a" i)))
(loop (+ i 1) #f (+ 1 num-groups))]
[(#\\)
(if (and (< i last-index) (eq? #\\ (string-ref str (+ i 1))))
(loop (+ i 3) in-paren num-groups)
(loop (+ i 2) in-paren num-groups))]
[else
(loop (+ i 1) in-paren num-groups)]))))
(define-for-syntax (count-groups/regexp rxp #:src stx)
(error 'regexp-match! "Not implemented"))
(define-for-syntax (group-error str reason)
(raise-argument-error
'regexp-match!
(format "Valid regexp pattern (contains unmatched ~a)" reason)
str))