Fix some special characters in titles like ? and ;

src/page-category.rkt

@@ -52,7 +52,7 @@
               ,@(map
                  (λ (result)
                    (define title (jp "/title" result))
-                   (define page-path (regexp-replace* #rx" " title "_"))
+                   (define page-path (page-title->path title))
                    `(li
                      (a (@ (href ,(format "/~a/wiki/~a" wikiname page-path)))
                         ,title)))

src/page-home.rkt

@@ -18,9 +18,9 @@
 
 (define examples
   '(("minecraft" "Bricks")
-    ("crosscode" "CrossCode_Wiki")
+    ("crosscode" "CrossCode Wiki")
-    ("undertale" "Hot_Dog...%3F")
+    ("undertale" "Hot Dog...?")
-    ("tardis" "Eleanor_Blake")
+    ("tardis" "Eleanor Blake")
     ("zelda" "Boomerang")))
 
 (define content
@@ -45,7 +45,7 @@
     (h2 "Example pages")
     (ul
      ,@(map (λ (x)
-              `(li (a (@ (href ,(apply format "/~a/wiki/~a" x)))
+              `(li (a (@ (href ,(format "/~a/wiki/~a" (car x) (page-title->path (cadr x)))))
                       ,(apply format "~a: ~a" x))))
             examples))
     (h2 "Testimonials")

src/page-search.rkt

@@ -38,7 +38,7 @@
          (ul ,@(map
                 (λ (result)
                   (let* ([title (jp "/title" result)]
-                         [page-path (regexp-replace* #rx" " title "_")]
+                         [page-path (page-title->path title)]
                          [timestamp (jp "/timestamp" result)]
                          [wordcount (jp "/wordcount" result)]
                          [size (jp "/size" result)])

src/url-utils.rkt

@@ -17,7 +17,9 @@
   ; prints "out: <url>"
  log-outgoing
  ; pass in a header, headers, or something useless. they'll all combine into a list
- build-headers)
+ build-headers
+ ; try to follow wikimedia's format for which characters should be encoded/replaced in page titles for the url
+ page-title->path)
 
 (module+ test
   (require "typed-rackunit.rkt"))
@@ -26,14 +28,18 @@
 
 ;; https://url.spec.whatwg.org/#urlencoded-serializing
 
-(define urlencoded-set '(#\! #\' #\( #\) #\~ ; urlencoded set
+(define path-set '(#\; ; semicolon is part of the userinfo set in the URL standard, but I'm putting it here
-                         #\$ #\% #\& #\+ #\, ; component set
+                   #\? #\` #\{ #\} ; path set
-                         #\/ #\: #\; #\= #\@ #\[ #\\ #\] #\^ #\| ; userinfo set
+                   #\  #\" #\# #\< #\> ; query set
-                         #\? #\` #\{ #\} ; path set
+                   ; c0 controls included elsewhere
-                         #\  #\" #\# #\< #\> ; query set
+                   ; higher ranges included elsewhere
-                         ; c0 controls included elsewhere
+                   ))
-                         ; higher ranges included elsewhere
+(define urlencoded-set (append
-                         ))
+                        '(#\! #\' #\( #\) #\~ ; urlencoded set
+                          #\$ #\% #\& #\+ #\, ; component set
+                          #\/ #\: #\= #\@ #\[ #\\ #\] #\^ #\| ; userinfo set
+                          )
+                        path-set))
 
 (: percent-encode (String (Listof Char) Boolean -> Bytes))
 (define (percent-encode value set space-as-plus)
@@ -98,3 +104,7 @@
            [(header? f) (list f)]
            [(pair? f) f]))
        fs)))
+
+(: page-title->path (String -> Bytes))
+(define (page-title->path title)
+  (percent-encode (regexp-replace* " " title "_") path-set #f))