avoid all characters that have to be encoded in Scribble-generated HTML anchors

svn: r7994

collects/net/doc.txt

@@ -2198,7 +2198,7 @@ mimetype given the in HTML 4.0 specification.  There are minor
 differences between the two encodings.
 
 The URI encoding uses allows a few characters to be represented `as
-is': a-Z, A-Z, 0-9, -, _, ., !, ~, *, ', ( and ).  The remaining
+is': a-z, A-Z, 0-9, -, _, ., !, ~, *, ', ( and ).  The remaining
 characters are encoded as %xx, where xx is the hex representation
 of the integer value of the character (where the mapping
 character<->integer is determined by US-ASCII if the integer is

collects/scribble/html-render.ss

@@ -39,19 +39,28 @@
           (main-collects-relative->path p))))
 
   ;; HTML anchors are case-insenstive. To make them
-  ;;  distinct, add a "^" in front of capital letters.
+  ;;  distinct, add a "." in front of capital letters.
+  ;;  Also clean up characters that give browers trouble
+  ;;  (i.e., the ones that are not allowed as-in in URI
+  ;;  codecs) by using "~" followed by a hex encoding.
   (define (anchor-name v)
     (let loop ([s (format "~a" v)])
       (cond
-       [(regexp-match-positions #rx"[A-Z:]" s)
+       [(regexp-match-positions #rx"[A-Z.]" s)
+        => (lambda (m)
+             (string-append
+              (loop (substring s 0 (caar m)))
+              "."
+              (substring s (caar m) (cdar m))
+              (loop (substring s (cdar m)))))]
+       [(regexp-match-positions #rx"[^-a-zA-Z0-9_!*'().]" s)
         => (lambda (m)
              (string-append
               (substring s 0 (caar m))
-              ":"
+              "~"
-              (substring s (caar m) (cdar m))
+              (format "~x" (char->integer (string-ref s (caar m))))
               (loop (substring s (cdar m)))))]
        [else s])))
-                                
 
   ;; ----------------------------------------
   ;;  main mixin