Change iconv procedures to take optional instead of keyword arg

* module/ice-9/iconv.scm (call-with-encoded-output-string):
  (string->bytevector, bytevector->string): Take an optional instead of
  a keyword argument.

* doc/ref/api-data.texi (Representing Strings as Bytes): Adapt docs to
  change, and fix a number of errors.  Thanks to Ludovic Courtès for the
  pointers.

* test-suite/tests/iconv.test ("wide non-ascii string"): Add a test for
  the 'substitute path.

diff --git a/doc/ref/api-data.texi b/doc/ref/api-data.texi
index 3bd38d2..21398f4 100644 (file)
--- a/doc/ref/api-data.texi
+++ b/doc/ref/api-data.texi
@@ -4190,6 +4190,11 @@ sequences of bytes.  @xref{Bytevectors}, for more on how Guile
 represents raw byte sequences.  This module gets its name from the
 common @sc{unix} command of the same name.
 
+Note that often it is sufficient to just read and write strings from
+ports instead of using these functions.  To do this, specify the port
+encoding using @code{set-port-encoding!}.  @xref{Ports}, for more on
+ports and character encodings.
+
 Unlike the rest of the procedures in this section, you have to load the
 @code{iconv} module before having access to these procedures:
 
@@ -4197,31 +4202,32 @@ Unlike the rest of the procedures in this section, you have to load the
 (use-modules (ice-9 iconv))
 @end example
 
-@deffn string->bytevector string encoding [#:conversion-strategy='error]
+@deffn string->bytevector string encoding [conversion-strategy]
 Encode @var{string} as a sequence of bytes.
 
 The string will be encoded in the character set specified by the
 @var{encoding} string.  If the string has characters that cannot be
 represented in the encoding, by default this procedure raises an
-@code{encoding-error}, though the @code{#:conversion-strategy} keyword
-can specify other behaviors.
+@code{encoding-error}.  Pass a @var{conversion-strategy} argument to
+specify other behaviors.
 
 The return value is a bytevector.  @xref{Bytevectors}, for more on
 bytevectors.  @xref{Ports}, for more on character encodings and
 conversion strategies.
 @end deffn
 
-@deffn bytevector->string bytevector encoding 
+@deffn bytevector->string bytevector encoding [conversion-strategy]
 Decode @var{bytevector} into a string.
 
 The bytes will be decoded from the character set by the @var{encoding}
 string.  If the bytes do not form a valid encoding, by default this
-procedure raises an @code{decoding-error}, though that may be overridden
-with the @code{#:conversion-strategy} keyword.  @xref{Ports}, for more
-on character encodings and conversion strategies.
+procedure raises an @code{decoding-error}.  As with
+@code{string->bytevector}, pass the optional @var{conversion-strategy}
+argument to modify this behavior.  @xref{Ports}, for more on character
+encodings and conversion strategies.
 @end deffn
 
-@deffn call-with-output-encoded-string encoding proc [#:conversion-strategy='error]
+@deffn call-with-output-encoded-string encoding proc [conversion-strategy]
 Like @code{call-with-output-string}, but instead of returning a string,
 returns a encoding of the string according to @var{encoding}, as a
 bytevector.  This procedure can be more efficient than collecting a
@@ -4371,7 +4377,7 @@ If @var{lenp} is @code{NULL}, this function will return a null-terminated C
 string.  It will throw an error if the string contains a null
 character.
 
-The Scheme interface to this function is @code{encode-string}, from the
+The Scheme interface to this function is @code{string->bytevector}, from the
 @code{ice-9 iconv} module.  @xref{Representing Strings as Bytes}.
 @end deftypefn
 
@@ -4382,7 +4388,7 @@ string is passed as the ASCII, null-terminated C string @code{encoding}.
 The @var{handler} parameters suggests a strategy for dealing with
 unconvertable characters.
 
-The Scheme interface to this function is @code{decode-string}.
+The Scheme interface to this function is @code{bytevector->string}.
 @xref{Representing Strings as Bytes}.
 @end deftypefn
 

diff --git a/module/ice-9/iconv.scm b/module/ice-9/iconv.scm
index 40d5954..0f0c1a3 100644 (file)
--- a/module/ice-9/iconv.scm
+++ b/module/ice-9/iconv.scm
@@ -43,7 +43,8 @@
         bv))))
 
 (define* (call-with-encoded-output-string encoding proc
-                                          #:key (conversion-strategy 'error))
+                                          #:optional
+                                          (conversion-strategy 'error))
   (if (string-ci=? encoding "utf-8")
       ;; I don't know why, but this appears to be faster; at least for
       ;; serving examples/debug-sxml.scm (1464 reqs/s versus 850
@@ -59,16 +60,18 @@
 ;; TODO: Provide C implementations that call scm_from_stringn and
 ;; friends?
 
-(define* (string->bytevector str encoding #:key (conversion-strategy 'error))
+(define* (string->bytevector str encoding
+                             #:optional (conversion-strategy 'error))
   (if (string-ci=? encoding "utf-8")
       (string->utf8 str)
       (call-with-encoded-output-string
        encoding
        (lambda (port)
          (display str port))
-       #:conversion-strategy conversion-strategy)))
+       conversion-strategy)))
 
-(define* (bytevector->string bv encoding #:key (conversion-strategy 'error))
+(define* (bytevector->string bv encoding
+                             #:optional (conversion-strategy 'error))
   (if (string-ci=? encoding "utf-8")
       (utf8->string bv)
       (let ((p (open-bytevector-input-port bv)))

diff --git a/test-suite/tests/iconv.test b/test-suite/tests/iconv.test
index e6ee90d..9083cd2 100644 (file)
--- a/test-suite/tests/iconv.test
+++ b/test-suite/tests/iconv.test
@@ -112,4 +112,9 @@
       (string->bytevector s "ascii"))
 
     (pass-if-exception "encode as latin1" exception:encoding-error
-      (string->bytevector s "latin1"))))
+      (string->bytevector s "latin1"))
+
+    (pass-if "encode as ascii with substitutions"
+      (equal? (make-string (string-length s) #\?)
+              (bytevector->string (string->bytevector s "ascii" 'substitute)
+                                  "ascii")))))