Racket语言开发命令行程序十分简单，而且支持跨平台编译，我们可以用它来开发各种方便的工具程序。下面实现了一个完整的wc程序，只需50多行代码就完成了。
#lang racket
(require racket/cmdline)
(define show-bytes (make-parameter #t))
(define show-lines (make-parameter #t))
(define show-words (make-parameter #t))
(define files (make-parameter null))
(define re-chan (make-channel))
(define cmd-line
(command-line
#:program "r-wc"
#:once-any
[("-c" "--bytes" "--chars") "The number of bytes" (show-bytes #t) (show-lines #f) (show-words #f)]
[("-l" "--lines") "The number of lines" (show-lines #t) (show-bytes #f) (show-words #f)]
[("-w" "--words") "The muber of words" (show-words #t) (show-bytes #f) (show-lines #f)]
#:args file (files file)))
(define (do-work)
(if (null? (files)) (displayln "Please enter the file name")
(begin
(for-each scan-file (files))
(let loop ([finished 0])
(unless (= finished (length (files)))
(let ((re (channel-get re-chan)))
(match re
((list l w b f)
(displayln (append
(for/list ([d (list l w b)] [t (list (show-lines) (show-words) (show-bytes) )] #:when t) d) f)))))
(loop (add1 finished)))))))
(define (scan-file f)
(thread (lambda ()
(call-with-input-file f
(lambda (in)
(let loop
([a-line (read-line in)] [bytes 0] [lines 0] [words 0])
(if (eof-object? a-line) (channel-put re-chan (list lines words bytes f))
(let-values (((w b) (parse-line a-line)))
(loop (read-line in) (+ bytes b) (add1 lines) (+ words w))))))))))
(define (parse-line a-line)
(define s (open-input-string a-line))
(if (null? a-line)
(values 0 0)
(let loop ([w 0] [b 0] [pre-char #\a])
(define bl b)
(define c (read-char s))
(unless (eof-object? c) (set! bl (+ (char-utf-8-length c) b)))
(if (eof-object? c) (values (add1 w) (add1 b))
(case c
[(#\space) (loop (if (char-whitespace? pre-char) w (add1 w)) (add1 b) c)]
[else (loop w bl c)])))))
(do-work)
然后用raco生成一个可执行程序, 名字就叫r-wc，和wc稍微区分一下：
raco exe -o r-wc r-wc.rkt
就像使用wc命令一样，
➜ racket ./r-wc -h
usage: r-wc [ <option> ... ] [<file>] ...
<option> is one of
/ -c, --bytes, --chars
| The number of bytes
| -l, --lines
| The number of lines
| -w, --words
\ The muber of words
--help, -h
Show this help
--
Do not treat any remaining argument as a switch (at this level)
/|\ Brackets indicate mutually exclusive options.
Multiple single-letter switches can be combined after
one `-`. For example, `-h-` is the same as `-h --`.
上面的程序执行效率不是很好，主要表现在读取文件，和解析每一行时，占用的时间和内存太大。Racket提供了一个特性Sequences ，它可以与for 形式一样遍历数据结果，但是效率却高很多，因此可以使用Sequences来重构我们的程序，将上文的两个函数scan-file 和parse-line 修改为：
(define (scan-file f)
(thread (lambda ()
(define bytes 0)
(define lines 0)
(define words 0)
(call-with-input-file f
(lambda (in)
(for ([a-line (in-bytes-lines in)])
(let-values (((w b) (parse-line-seq a-line)))
(set! bytes (+ bytes b))
(set! lines (add1 lines))
(set! words (+ words w))))
(channel-put re-chan (list lines words bytes f)))))))
(define (parse-line a-line)
(define pre-char #\a)
(define w 0)
(define b 0)
(if (null? a-line) (values 0 0)
(for ([c (in-bytes a-line)])
(case c
[(#\space) (unless (char-whitespace? pre-char) (set! w (add1 w))) (set! b (add1 b))]
[else (set! b (add1 b))])
(set! pre-char c)))
(values (add1 w) (add1 b)))
两个版本对比一下(文件大小100MB)：
| Old version | Sequences version | wc | |
|---|---|---|---|
| 时间 | 4.6s | 0.8s | 0.4s |
| 峰值内存 | 132M | 115M | 1M |
经过Sequences重构之后，执行时间有了巨大的提升，并且最大内存使用量也有所减少，不过与wc还是有不少差距。希望有一天能够找到超越wc的方法。
欢迎加入Racket 隐修会 ：731859928(QQ)
