net/uri-codec: added `uri-path-segment-unreserved-encode'

collects/net/scribblings/uri-codec.scrbl

@@ -28,7 +28,11 @@ less than 128).
 The encoding, in line with RFC 2396's recommendation, represents a
 character as-is, if possible.  The decoding allows any characters
 to be represented by their hex values, and allows characters to be
-incorrectly represented as-is.
+incorrectly represented as-is. The library provides ``unreserved''
+encoders that encode @litchar{!}, @litchar{*}, @litchar{'},
+@litchar{(}, and @litchar{)} using their hex representation,
+which is not recommended by RFC 2396 but avoids problems with some
+contexts.
 
 The rules for the @tt{application/x-www-form-urlencoded} mimetype
 given in the HTML 4.0 spec are:
@@ -52,15 +56,17 @@ given in the HTML 4.0 spec are:
 
 ]
 
-These rules differs slightly from the straight encoding in RFC 2396 in
+These @tt{application/x-www-form-urlencoded} rules differs slightly from the straight encoding in RFC 2396 in
 that @litchar{+} is allowed, and it represents a space.  The
 @racketmodname[net/uri-codec] library follows this convention,
 encoding a space as @litchar{+} and decoding @litchar{+} as a space.
-In addtion, since there appear to be some brain-dead decoders on the
+In addition, since there appear to be some broken decoders on the
 web, the library also encodes @litchar{!}, @litchar{~}, @litchar{'},
 @litchar{(}, and @litchar{)} using their hex representation, which is
 the same choice as made by the Java's @tt{URLEncoder}.
 
+
+
 @; ----------------------------------------
 
 @section[#:tag "uri-codec-proc"]{Functions}
@@ -92,6 +98,14 @@ Encodes a string according to the rules in @cite["RFC3986"](section 2.3) for the
 @defproc[(uri-unreserved-decode [str string?]) string?]{
 Decodes a string according to the rules in @cite["RFC3986"](section 2.3) for the unreserved characters.
 }
+@defproc[(uri-path-segment-unreserved-encode [str string?]) string?]{
+Encodes a string according to the rules in @cite["RFC3986"] for path segments,
+but also encodes characters that @racket[uri-unreserved-encode] encodes
+and that @racket[uri-encode] does not.
+}
+@defproc[(uri-path-segment-unreserved-decode [str string?]) string?]{
+Decodes a string according to the rules in @cite["RFC3986"] for path segments.
+}
 
 
 @defproc[(form-urlencoded-encode [str string?]) string?]{
@@ -184,7 +198,9 @@ Imports nothing, exports @racket[uri-codec^].}
 
 @defsignature[uri-codec^ ()]{}
 
-Includes everything exported by the @racketmodname[net/uri-codec] module.
+Includes everything exported by the @racketmodname[net/uri-codec]
+module except @racket[uri-path-segment-unreserved-encode] and
+@racket[uri-path-segment-unreserved-decode].
 
 
 @close-eval[uri-codec-eval]

collects/net/uri-codec.rkt

@@ -93,6 +93,8 @@ See more in PR8831.
          uri-userinfo-decode
          uri-unreserved-encode
          uri-unreserved-decode
+         uri-path-segment-unreserved-encode
+         uri-path-segment-unreserved-decode
          form-urlencoded-encode
          form-urlencoded-decode
          alist->form-urlencoded
@@ -115,16 +117,21 @@ See more in PR8831.
 (define uri-mapping (append alphanumeric-mapping safe-mapping))
 
 ;; The uri path segment mapping from RFC 3986
+(define path-segment-extra-mapping (self-map-chars "@+,=$&:"))
 (define uri-path-segment-mapping
-  (append alphanumeric-mapping
+  (append uri-mapping
-          safe-mapping
+          path-segment-extra-mapping))
-          (self-map-chars "@+,=$&:")))
 
 ;; from RFC 3986
 (define unreserved-mapping
   (append alphanumeric-mapping
           (self-map-chars "-._~")))
 
+;; The uri path segment mapping from RFC 3986
+(define uri-path-segment-unreserved-mapping
+  (append unreserved-mapping
+          path-segment-extra-mapping))
+
 ;; from RFC 3986
 (define sub-delims-mapping
   (self-map-chars "!$&'()*+,;="))
@@ -176,6 +183,10 @@ See more in PR8831.
                 uri-unreserved-decoding-vector)
   (make-codec-tables unreserved-mapping))
 
+(define-values (uri-path-segment-unreserved-encoding-vector
+                uri-path-segment-unreserved-decoding-vector)
+  (make-codec-tables uri-path-segment-unreserved-mapping))
+
 (define-values (form-urlencoded-encoding-vector
                 form-urlencoded-decoding-vector)
   (make-codec-tables form-urlencoded-mapping))
@@ -249,6 +260,14 @@ See more in PR8831.
 (define (uri-unreserved-decode str)
   (decode uri-unreserved-decoding-vector str))
 
+;; string -> string
+(define (uri-path-segment-unreserved-encode str)
+  (encode uri-path-segment-unreserved-encoding-vector str))
+
+;; string -> string
+(define (uri-path-segment-unreserved-decode str)
+  (decode uri-path-segment-unreserved-decoding-vector str))
+
 ;; string -> string
 (define (form-urlencoded-encode str)
   (encode form-urlencoded-encoding-vector str))

collects/tests/net/uri-codec.rkt

@@ -75,11 +75,16 @@
         (uri-path-segment-encode "M~(@; ") =>  "M~(@%3B%20"
         (uri-userinfo-encode "M~(@; ")     =>  "M~(%40;%20"
         (uri-unreserved-encode "M~(@; ")   =>  "M~%28%40%3B%20"         
+        (uri-path-segment-unreserved-encode "M~(@; ") =>  "M~%28@%3B%20"
         ;; matching decodes:
         (uri-decode "M~(%40%3B%20")              =>  "M~(@; "
         (uri-path-segment-decode "M~(@%3B%20")   =>  "M~(@; "
         (uri-userinfo-decode "M~(%40;%20")       =>  "M~(@; "
         (uri-unreserved-decode "M~%28%40%3B%20") =>  "M~(@; "
+        (uri-path-segment-unreserved-decode "M~%28@%3B%20")   =>  "M~(@; "
+
+        (uri-path-segment-decode "M~%28@%3B%20")   =>  "M~(@; "
+        (uri-path-segment-unreserved-decode "M~(@%3B%20")   =>  "M~(@; "
         ))
 
 ;; tests adapted from Noel Welsh's original test suite

doc/release-notes/racket/HISTORY.txt

@@ -4,6 +4,7 @@ r5rs, r6rs: fixed case and cond to disallow internal definitions
  in clauses
 Add #fx() and #fl() reader forms for flvectors and fxvectors
 racket/serialize: fxvectors and flvectors are serializable
+racket/net: added uri-path-segment-unreserved-encode
 
 Version 5.3.1.8
 file/untar: added