--- gforth/utf-8.fs	2005/01/05 22:13:40	1.13
+++ gforth/utf-8.fs	2006/02/19 22:15:05	1.20
@@ -1,6 +1,6 @@
 \ UTF-8 handling                                       12dec04py
 
-\ Copyright (C) 2004 Free Software Foundation, Inc.
+\ Copyright (C) 2004,2005 Free Software Foundation, Inc.
 
 \ This file is part of Gforth.
 
@@ -27,11 +27,12 @@ $80 Value max-single-byte
 : u8len ( u8 -- n )
     dup      max-single-byte u< IF  drop 1  EXIT  THEN \ special case ASCII
     $800  2 >r
-    BEGIN  2dup u>=  WHILE  5 lshift r> 1+ >r  REPEAT
+    BEGIN  2dup u>=  WHILE  5 lshift r> 1+ >r  dup 0= UNTIL  THEN
     2drop r> ;
 
 : u8@+ ( u8addr -- u8addr' u )
     count  dup max-single-byte u< ?EXIT  \ special case ASCII
+    dup $C2 u< IF  UTF-8-err throw  THEN  \ malformed character
     $7F and  $40 >r
     BEGIN  dup r@ and  WHILE  r@ xor
 	    6 lshift r> 5 lshift >r >r count
@@ -72,6 +73,7 @@ $80 Value max-single-byte
 
 : u8key ( -- u )
     defers key dup max-single-byte u< ?EXIT  \ special case ASCII
+    dup $C2 u< IF  UTF-8-err throw  THEN  \ malformed character
     $7F and  $40 >r
     BEGIN  dup r@ and  WHILE  r@ xor
 	    6 lshift r> 5 lshift >r >r defers key
@@ -126,6 +128,11 @@ $80 Value max-single-byte
 	nip nip over -
     then ;
 
+: u8width ( xcaddr u -- n )
+    0 rot rot over + swap ?DO
+	I xc@+ swap >r wcwidth +
+    r> I - +LOOP ;
+
 : set-encoding-utf-8 ( -- )
     ['] u8emit is xemit
     ['] u8key is xkey
@@ -137,95 +144,20 @@ $80 Value max-single-byte
     ['] u8!+? is xc!+?
     ['] u8@+ is xc@+
     ['] u8len is xc-size
+[ [IFDEF] x-width ]
+    ['] u8width is x-width
+[ [THEN] ]
     ['] -u8trailing-garbage is -trailing-garbage
 ;
 
-\ input editor
-
-[IFUNDEF] #esc  27 Constant #esc  [THEN]
+: utf-8-cold ( -- )
+    s" LC_ALL" getenv 2dup d0= IF  2drop
+	s" LC_CTYPE" getenv 2dup d0= IF  2drop
+	    s" LANG" getenv 2dup d0= IF  2drop
+		s" C"  THEN THEN THEN
+    s" UTF-8" search nip nip
+    IF  set-encoding-utf-8  ELSE  set-encoding-fixed-width  THEN ;
 
-: save-cursor ( -- )  #esc emit '7 emit ;
-: restore-cursor ( -- )  #esc emit '8 emit ;
-: .rest ( addr pos1 -- addr pos1 )
-    restore-cursor 2dup type ;
-: .all ( span addr pos1 -- span addr pos1 )
-    restore-cursor >r 2dup swap type r> ;
-
-: <u8ins>  ( max span addr pos1 u8char -- max span addr pos2 )
-    >r  2over r@ u8len + u< IF  rdrop bell  EXIT  THEN
-    >string over r@ u8len + swap move 2dup chars + r@ swap u8!+ drop
-    r> u8len >r  rot r@ chars + -rot r> chars + ;
-: (u8ins)  ( max span addr pos1 u8char -- max span addr pos2 )
-    <u8ins> .all .rest ;
-: u8back  ( max span addr pos1 -- max span addr pos2 f )
-    dup  IF  over + u8<< over -  0 max .all .rest
-    ELSE  bell  THEN 0 ;
-: u8forw  ( max span addr pos1 -- max span addr pos2 f )
-    2 pick over <> IF  over + u8@+ u8emit over -  ELSE  bell  THEN 0 ;
-: (u8del)  ( max span addr pos1 -- max span addr pos2 )
-    over + dup u8<< tuck - >r over -
-    >string over r@ + -rot move
-    rot r> - -rot ;
-: ?u8del ( max span addr pos1 -- max span addr pos2 0 )
-  dup  IF  (u8del) .all 2 spaces .rest  THEN  0 ;
-: <u8del> ( max span addr pos1 -- max span addr pos2 0 )
-  2 pick over <>
-    IF  u8forw drop (u8del) .all 2 spaces .rest
-    ELSE  bell  THEN  0 ;
-: u8eof  2 pick over or 0=  IF  bye  ELSE  <u8del>  THEN ;
-
-: u8first-pos  ( max span addr pos1 -- max span addr 0 0 )
-  drop 0 .all .rest 0 ;
-: u8end-pos  ( max span addr pos1 -- max span addr span 0 )
-  drop over .all 0 ;
-
-
-: u8clear-line ( max span addr pos1 -- max addr )
-    drop restore-cursor swap spaces restore-cursor ;
-: u8clear-tib ( max span addr pos -- max 0 addr 0 false )
-    u8clear-line 0 tuck dup ;
-
-: (u8enter)  ( max span addr pos1 -- max span addr pos2 true )
-    >r end^ 2@ hist-setpos
-    2dup swap history write-line drop ( throw ) \ don't worry about errors
-    hist-pos 2dup backward^ 2! end^ 2!
-    r> .all space true ;
-
-: u8kill-expand ( max span addr pos1 -- max span addr pos2 )
-    prefix-found cell+ @ ?dup IF  >r
-	r@ - >string over r@ + -rot move
-	rot r@ - -rot .all r> spaces .rest THEN ;
-
-: insert   ( string length buffer size -- )
-    rot over min >r  r@ - ( left over )
-    over dup r@ +  rot move   r> move  ;
-
-: u8tab-expand ( max span addr pos1 -- max span addr pos2 0 )
-    key? IF  #tab (u8ins) 0  EXIT  THEN
-    u8kill-expand 2dup extract-word dup 0= IF  nip EXIT  THEN
-    search-prefix tib-full?
-    IF    7 emit  2drop  prefix-off
-    ELSE  dup >r
-	2>r >string r@ + 2r> 2swap insert
-	r@ + rot r> + -rot
-    THEN
-    prefix-found @ IF  bl (u8ins)  ELSE  .all .rest  THEN  0 ;
-
-: utf-8-io ( -- )
-    ['] u8forw       ctrl F bindkey
-    ['] u8back       ctrl B bindkey
-    ['] ?u8del       ctrl H bindkey
-    ['] u8eof        ctrl D bindkey
-    ['] <u8del>      ctrl X bindkey
-    ['] u8clear-tib  ctrl K bindkey
-    ['] u8first-pos  ctrl A bindkey
-    ['] u8end-pos    ctrl E bindkey
-    ['] (u8enter)    #lf    bindkey
-    ['] (u8enter)    #cr    bindkey
-    ['] u8tab-expand #tab   bindkey
-    ['] (u8ins)      IS insert-char
-    ['] kill-prefix  IS everychar
-    ['] save-cursor  IS everyline
-    ['] u8key        IS key
-    ['] u8emit       IS emit ;
+' utf-8-cold INIT8 chained
 
+utf-8-cold