io.encodings.string: faster for ascii and utf8.

basis/io/encodings/string/string.factor

@@ -1,23 +1,32 @@
 ! Copyright (C) 2008 Daniel Ehrenberg.
 ! See http://factorcode.org/license.txt for BSD license.
-USING: byte-vectors io io.encodings io.streams.byte-array
-io.streams.string kernel locals sbufs sequences io.private
-io.encodings.binary ;
+USING: accessors byte-arrays byte-vectors io io.encodings
+io.streams.byte-array io.streams.string kernel locals
+sbufs sequences io.private io.encodings.ascii
+io.encodings.binary io.encodings.private io.encodings.utf8 ;
 IN: io.encodings.string
 
 :: decode ( byte-array encoding -- string )
     encoding binary eq? [ byte-array ] [
-        byte-array encoding <byte-reader> :> reader
-        byte-array length
-        encoding guess-decoded-length
-        reader stream-exemplar-growable new-resizable :> buf
-        [ reader stream-read1 dup ] [ buf push ] while drop
-        buf reader stream-exemplar like
+        byte-array byte-array? encoding ascii eq? and [
+            byte-array byte-array>string-fast
+        ] [
+            byte-array encoding <byte-reader> :> reader
+            byte-array length
+            encoding guess-decoded-length
+            reader stream-exemplar-growable new-resizable :> buf
+            [ reader stream-read1 dup ] [ buf push ] while drop
+            buf reader stream-exemplar like
+        ] if
     ] if ; inline
 
 :: encode ( string encoding -- byte-array )
     encoding binary eq? [ string ] [
-        string length encoding guess-encoded-length <byte-vector> :> vec
-        string vec encoding <encoder> stream-write
-        vec B{ } like
+        string aux>> not encoding { ascii utf8 } member-eq? and [
+            string string>byte-array-fast
+        ] [
+            string length encoding guess-encoded-length <byte-vector> :> vec
+            string vec encoding <encoder> stream-write
+            vec B{ } like
+        ] if
     ] if ; inline

core/io/encodings/encodings.factor

@@ -36,6 +36,8 @@ PRIVATE>
 
 M: object decode-until (decode-until) ;
 
+CONSTANT: replacement-char 0xfffd
+
 <PRIVATE
 
 : string>byte-array-fast ( string -- byte-array )
@@ -47,6 +49,20 @@ M: object decode-until (decode-until) ;
         ] 2curry each-integer
     ] keep ; inline
 
+: byte-array>string-fast ( byte-array -- string )
+    { byte-array } declare
+    [ length ] keep over 0 <string> [
+        [
+            [
+                [
+                    nth-unsafe dup 127 <=
+                    [ drop replacement-char ] unless
+                ] 2keep drop
+            ]
+            [ set-string-nth ] bi*
+        ] 2curry each-integer
+    ] keep dup reset-string-hashcode ;
+
 PRIVATE>
 
 GENERIC: encode-char ( char stream encoding -- )
@@ -57,8 +73,6 @@ M: object encode-string [ encode-char ] 2curry each ; inline
 
 GENERIC: <decoder> ( stream encoding -- newstream )
 
-CONSTANT: replacement-char 0xfffd
-
 TUPLE: decoder { stream read-only } { code read-only } { cr boolean } ;
 INSTANCE: decoder input-stream