From 4c2a7c4f0ce420de71fe11a16356552674b11130 Mon Sep 17 00:00:00 2001
From: Matthew Butterick <mb@mbtype.com>
Date: Sun, 23 Feb 2014 13:08:27 -0800
Subject: [PATCH] add #:omit-tags parameter

---
 hyphenate/main.rkt                    | 37 ++++++++++++++++-----------
 hyphenate/scribblings/hyphenate.scrbl | 20 ++++++++++-----
 hyphenate/tests.rkt                   | 10 ++++++++
 3 files changed, 46 insertions(+), 21 deletions(-)

diff --git a/hyphenate/main.rkt b/hyphenate/main.rkt
index 7ca689bb..168bca34 100644
--- a/hyphenate/main.rkt
+++ b/hyphenate/main.rkt
@@ -151,21 +151,24 @@
   (if (char? joiner) (format "~a" joiner) joiner))
 
 ;; helper macro that applies proc to all strings found in xexpr input
-(define-syntax (apply-xexpr-strings stx)
-  (syntax-case stx ()
-    [(_ proc val) #'(let loop ([x val])
-                      (cond
-                        [(string? x) (proc x)]
-                        [(txexpr? x) (map-elements loop x)]
-                        [else x]))]))
+(define (apply-xexpr-strings proc x [tags-to-omit '()])
+;  ((procedure? txexpr?) ((or/c null (listof txexpr-tag?))) . ->* . txexpr?)
+  (let loop ([x x])
+    (cond
+      [(string? x) (proc x)]
+      [(and (txexpr? x) (not (member (car x) tags-to-omit))) (cons (car x) (map loop (cdr x)))]
+      [else x])))
+
 
 ;; Hyphenate using a filter procedure.
 (define+provide+safe (hyphenatef x proc [joiner default-joiner] 
-                                     #:exceptions [extra-exceptions '()]  
-                                     #:min-length [min-length default-min-length])
+                                 #:exceptions [extra-exceptions '()]  
+                                 #:min-length [min-length default-min-length]
+                                 #:omit-tags [tags-to-omit '()])
   ((xexpr? procedure?) ((or/c char? string?) 
                         #:exceptions (listof exception-word?) 
-                        #:min-length (or/c integer? #f)) . ->* . xexpr/c)
+                        #:min-length (or/c integer? #f)
+                        #:omit-tags (or/c null (listof txexpr-tag?))) . ->* . xexpr/c)
   
   ;; set up module data
   ;; todo?: change set! to parameterize
@@ -178,17 +181,21 @@
   (define (insert-hyphens text)
     (regexp-replace* word-pattern text (λ(word) (if (proc word) (string-join (word->hyphenation-points word min-length) joiner-string) word))))
   
-  (apply-xexpr-strings insert-hyphens x))
+  (apply-xexpr-strings insert-hyphens x tags-to-omit))
 
 
 ;; Default hyphenate is a special case of hyphenatef.
 (define+provide+safe (hyphenate x [joiner default-joiner]  
-                                    #:exceptions [extra-exceptions '()] 
-                                    #:min-length [min-length default-min-length])
+                                #:exceptions [extra-exceptions '()] 
+                                #:min-length [min-length default-min-length]
+                                #:omit-tags [tags-to-omit '()])
   ((xexpr/c) ((or/c char? string?) 
               #:exceptions (listof exception-word?) 
-              #:min-length (or/c integer? #f)) . ->* . xexpr/c)
-  (hyphenatef x (λ(x) #t) joiner #:exceptions extra-exceptions #:min-length min-length))
+              #:min-length (or/c integer? #f)
+              #:omit-tags (or/c null (listof txexpr-tag?))) . ->* . xexpr/c)
+  (hyphenatef x (λ(x) #t) joiner #:exceptions extra-exceptions #:min-length min-length #:omit-tags tags-to-omit))
+
+
 
 
 ;; Remove hyphens.
diff --git a/hyphenate/scribblings/hyphenate.scrbl b/hyphenate/scribblings/hyphenate.scrbl
index ad0b63f3..36c3df3f 100644
--- a/hyphenate/scribblings/hyphenate.scrbl
+++ b/hyphenate/scribblings/hyphenate.scrbl
@@ -1,9 +1,9 @@
 #lang scribble/manual
 
-@(require scribble/eval (for-label racket "../main.rkt" xml))
+@(require scribble/eval (for-label txexpr racket "../main.rkt" xml))
 
 @(define my-eval (make-base-eval))
-@(my-eval `(require hyphenate xml))
+@(my-eval `(require (submod txexpr safe) (submod hyphenate safe) xml))
 
 
 @title{Hyphenate}
@@ -39,7 +39,8 @@ Safe mode enables the function contracts documented below. Use safe mode by impo
 [xexpr xexpr/c] 
 [joiner (or/c char? string?) (integer->char #x00AD)]
 [#:exceptions exceptions (listof string?) empty]
-[#:min-length length (or/c integer? false?) 5])
+[#:min-length length (or/c integer? false?) 5]
+[#:omit-tags tags (or/c null (listof txexpr-tag?)) null])
 xexpr/c]
 Hyphenate @racket[_xexpr] by calculating hyphenation points and inserting @racket[_joiner] at those points. By default, @racket[_joiner] is the soft hyphen (Unicode 00AD = decimal 173). Words shorter than @racket[#:min-length] @racket[_length] will not be hyphenated. To hyphenate words of any length, use @racket[#:min-length] @racket[#f].
 
@@ -96,14 +97,20 @@ You can send HTML-style X-expressions through @racket[hyphenate]. It will recurs
 
 Don't send raw HTML or XML through @racket[hyphenate]. It can't distinguish tags and attributes from textual content, so everything will be hyphenated, thus goofing up your file. But you can easily convert your HTML or XML to an X-expression, hyphenate it, and then convert back.
 
-@margin-note{In HTML, be careful not to include any @code{<script>} or @code{<style>} blocks, which contain non-hyphenatable data. The easiest way to protect that data — and arguably the right way — is to wrap it in a @code{<![CDATA[]]>} tag.}
-
 @examples[#:eval my-eval
     (define html "<body style=\"background: yellow\">Hello</body>")
     (hyphenate html #\-)
     (xexpr->string (hyphenate (string->xexpr html) #\-)) 
    ]   
 
+If you're working with HTML, be careful not to include any @code{<script>} or @code{<style>} blocks, which contain non-hyphenatable data. You can protect that data by using the #:omit-tags parameter, which will skip over the listed tags.
+
+@examples[#:eval my-eval
+     (hyphenate '(body "hyphenate" (script "don't hyphenate")) #\-)
+     (hyphenate '(body "hyphenate" (script "don't hyphenate")) #\- 
+     #:omit-tags '(script))
+   ]
+
 
 @defproc[
 (hyphenatef
@@ -111,7 +118,8 @@ Don't send raw HTML or XML through @racket[hyphenate]. It can't distinguish tags
 [pred procedure?]
 [joiner (or/c char? string?) (integer->char \#x00AD)]
 [#:exceptions exceptions (listof string?) empty]
-[#:min-length length (or/c integer? false?) 5])
+[#:min-length length (or/c integer? false?) 5]
+[#:omit-tags tags (or/c null (listof txexpr-tag?)) null])
 xexpr/c]
 Like @racket[hyphenate], but only words matching @racket[_pred] are hyphenated. Convenient if you want to prevent hyphenation of certain sets of words, like proper names:
 
diff --git a/hyphenate/tests.rkt b/hyphenate/tests.rkt
index f45995f8..8ed7bdb9 100644
--- a/hyphenate/tests.rkt
+++ b/hyphenate/tests.rkt
@@ -24,3 +24,13 @@
 (check-false (exception-word? "foobar!"))
 (check-true (exception-word? "foo-bar"))
 (check-false (exception-word? "foo bar"))
+
+
+(check-equal? (hyphenate '(p "circular polymorphism" amp (em "squandering")) #:omit-tags '(em)) 
+              '(p "cir\u00ADcu\u00ADlar poly\u00ADmor\u00ADphism" amp (em "squandering")))
+
+(check-equal? (hyphenate '(p "circular polymorphism" amp (em "squandering")) #:omit-tags '(p)) 
+              '(p "circular polymorphism" amp (em "squandering")))
+
+(check-equal? (hyphenate '(p  (foo "circular") (bar "circular") (zam "circular")) #:omit-tags '(foo zam)) 
+              '(p  (foo "circular") (bar "cir\u00ADcu\u00ADlar") (zam "circular")))
\ No newline at end of file