Annotation of gforth/utf-8.fs, revision 1.10

1.1       pazsan      1: \ UTF-8 handling                                       12dec04py
                      2: 
                      3: \ Copyright (C) 2004 Free Software Foundation, Inc.
                      4: 
                      5: \ This file is part of Gforth.
                      6: 
                      7: \ Gforth is free software; you can redistribute it and/or
                      8: \ modify it under the terms of the GNU General Public License
                      9: \ as published by the Free Software Foundation; either version 2
                     10: \ of the License, or (at your option) any later version.
                     11: 
                     12: \ This program is distributed in the hope that it will be useful,
                     13: \ but WITHOUT ANY WARRANTY; without even the implied warranty of
                     14: \ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
                     15: \ GNU General Public License for more details.
                     16: 
                     17: \ You should have received a copy of the GNU General Public License
                     18: \ along with this program; if not, write to the Free Software
                     19: \ Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
                     20: 
                     21: \ short: u8 means utf-8 encoded address
                     22: 
1.7       pazsan     23: s" malformed UTF-8 character" exception Constant UTF-8-err
                     24: 
1.10    ! pazsan     25: $80 Value maxascii
        !            26: 
1.4       pazsan     27: : u8len ( u8 -- n )
1.10    ! pazsan     28:     dup      maxascii u< IF  drop 1  EXIT  THEN \ special case ASCII
1.1       pazsan     29:     $800  2 >r
                     30:     BEGIN  2dup u>=  WHILE  5 lshift r> 1+ >r  REPEAT
                     31:     2drop r> ;
                     32: 
                     33: : u8@+ ( u8addr -- u8addr' u )
1.10    ! pazsan     34:     count  dup maxascii u< ?EXIT  \ special case ASCII
1.1       pazsan     35:     $7F and  $40 >r
                     36:     BEGIN  dup r@ and  WHILE  r@ xor
                     37:            6 lshift r> 5 lshift >r >r count
1.7       pazsan     38:            dup $C0 and $80 <> IF   UTF-8-err throw  THEN
1.1       pazsan     39:            $3F and r> or
                     40:     REPEAT  rdrop ;
                     41: 
                     42: : u8!+ ( u u8addr -- u8addr' )
1.10    ! pazsan     43:     over maxascii u< IF  tuck c! 1+  EXIT  THEN \ special case ASCII
1.1       pazsan     44:     >r 0 swap  $3F
                     45:     BEGIN  2dup u>  WHILE
                     46:            2/ >r  dup $3F and $80 or swap 6 rshift r>
                     47:     REPEAT  $7F xor 2* or  r>
                     48:     BEGIN   over $80 u>= WHILE  tuck c! 1+  REPEAT  nip ;
                     49: 
1.8       pazsan     50: \ plug-in so that char and '<char> work for UTF-8
                     51: 
1.9       anton      52: [ifundef] char@ \ !! bootstrapping help
                     53:     Defer char@ ( addr u -- char addr' u' )
                     54:     :noname  over c@ -rot 1 /string ; IS char@
                     55: [then]
                     56: 
1.8       pazsan     57: :noname  ( addr u -- char addr' u' )
1.9       anton      58:     \ !! the if here seems to work around some breakage, but not
                     59:     \ entirely; e.g., try 'ç' with LANG=C.
                     60:     dup 1 u<= IF defers char@ EXIT THEN
1.8       pazsan     61:     over + >r u8@+ swap r> over - ; IS char@
                     62: 
1.1       pazsan     63: \ scan to next/previous character
                     64: 
                     65: : u8>> ( u8addr -- u8addr' )
1.10    ! pazsan     66:     BEGIN  count $C0 and maxascii <>  UNTIL ;
1.1       pazsan     67: : u8<< ( u8addr -- u8addr' )
1.10    ! pazsan     68:     BEGIN  1- dup c@ $C0 and maxascii <>  UNTIL ;
1.1       pazsan     69: 
                     70: \ utf key and emit
                     71: 
                     72: : u8key ( -- u )
1.10    ! pazsan     73:     defers key dup maxascii u< ?EXIT  \ special case ASCII
1.1       pazsan     74:     $7F and  $40 >r
                     75:     BEGIN  dup r@ and  WHILE  r@ xor
                     76:            6 lshift r> 5 lshift >r >r defers key
1.7       pazsan     77:            dup $C0 and $80 <> IF  UTF-8-err throw  THEN
1.1       pazsan     78:            $3F and r> or
                     79:     REPEAT  rdrop ;
                     80: 
                     81: : u8emit ( u -- )
1.10    ! pazsan     82:     dup maxascii u< IF  defers emit  EXIT  THEN \ special case ASCII
1.1       pazsan     83:     0 swap  $3F
                     84:     BEGIN  2dup u>  WHILE
                     85:            2/ >r  dup $3F and $80 or swap 6 rshift r>
                     86:     REPEAT  $7F xor 2* or
                     87:     BEGIN   dup $80 u>= WHILE  defers emit  REPEAT  drop ;
                     88: 
                     89: \ input editor
                     90: 
1.7       pazsan     91: [IFUNDEF] #esc  27 Constant #esc  [THEN]
                     92: 
                     93: : save-cursor ( -- )  #esc emit '7 emit ;
                     94: : restore-cursor ( -- )  #esc emit '8 emit ;
1.1       pazsan     95: : .rest ( addr pos1 -- addr pos1 )
                     96:     restore-cursor 2dup type ;
                     97: : .all ( span addr pos1 -- span addr pos1 )
                     98:     restore-cursor >r 2dup swap type r> ;
                     99: 
1.3       pazsan    100: : <u8ins>  ( max span addr pos1 u8char -- max span addr pos2 )
1.6       pazsan    101:     >r  2over r@ u8len + u< IF  rdrop bell  EXIT  THEN
                    102:     >string over r@ u8len + swap move 2dup chars + r@ swap u8!+ drop
1.3       pazsan    103:     r> u8len >r  rot r@ chars + -rot r> chars + ;
1.1       pazsan    104: : (u8ins)  ( max span addr pos1 u8char -- max span addr pos2 )
1.3       pazsan    105:     <u8ins> .all .rest ;
1.1       pazsan    106: : u8back  ( max span addr pos1 -- max span addr pos2 f )
1.2       pazsan    107:     dup  IF  over + u8<< over -  0 max .all .rest
1.6       pazsan    108:     ELSE  bell  THEN 0 ;
1.1       pazsan    109: : u8forw  ( max span addr pos1 -- max span addr pos2 f )
1.6       pazsan    110:     2 pick over <> IF  over + u8@+ u8emit over -  ELSE  bell  THEN 0 ;
1.1       pazsan    111: : (u8del)  ( max span addr pos1 -- max span addr pos2 )
                    112:     over + dup u8<< tuck - >r over -
                    113:     >string over r@ + -rot move
1.3       pazsan    114:     rot r> - -rot ;
1.1       pazsan    115: : ?u8del ( max span addr pos1 -- max span addr pos2 0 )
1.3       pazsan    116:   dup  IF  (u8del) .all 2 spaces .rest  THEN  0 ;
1.1       pazsan    117: : <u8del> ( max span addr pos1 -- max span addr pos2 0 )
                    118:   2 pick over <>
1.3       pazsan    119:     IF  u8forw drop (u8del) .all 2 spaces .rest
1.6       pazsan    120:     ELSE  bell  THEN  0 ;
1.1       pazsan    121: : u8eof  2 pick over or 0=  IF  bye  ELSE  <u8del>  THEN ;
                    122: 
1.3       pazsan    123: : u8first-pos  ( max span addr pos1 -- max span addr 0 0 )
                    124:   drop 0 .all .rest 0 ;
                    125: : u8end-pos  ( max span addr pos1 -- max span addr span 0 )
                    126:   drop over .all 0 ;
                    127: 
                    128: 
                    129: : u8clear-line ( max span addr pos1 -- max addr )
                    130:     drop restore-cursor swap spaces restore-cursor ;
                    131: : u8clear-tib ( max span addr pos -- max 0 addr 0 false )
                    132:     u8clear-line 0 tuck dup ;
                    133: 
                    134: : (u8enter)  ( max span addr pos1 -- max span addr pos2 true )
                    135:     >r end^ 2@ hist-setpos
                    136:     2dup swap history write-line drop ( throw ) \ don't worry about errors
                    137:     hist-pos 2dup backward^ 2! end^ 2!
                    138:     r> .all space true ;
                    139: 
                    140: : u8kill-expand ( max span addr pos1 -- max span addr pos2 )
                    141:     prefix-found cell+ @ ?dup IF  >r
                    142:        r@ - >string over r@ + -rot move
                    143:        rot r@ - -rot .all r> spaces .rest THEN ;
                    144: 
                    145: : insert   ( string length buffer size -- )
                    146:     rot over min >r  r@ - ( left over )
                    147:     over dup r@ +  rot move   r> move  ;
                    148: 
                    149: : u8tab-expand ( max span addr pos1 -- max span addr pos2 0 )
                    150:     key? IF  #tab (u8ins) 0  EXIT  THEN
                    151:     u8kill-expand 2dup extract-word dup 0= IF  nip EXIT  THEN
1.5       pazsan    152:     search-prefix tib-full?
                    153:     IF    7 emit  2drop  prefix-off
1.3       pazsan    154:     ELSE  dup >r
                    155:        2>r >string r@ + 2r> 2swap insert
                    156:        r@ + rot r> + -rot
                    157:     THEN
1.5       pazsan    158:     prefix-found @ IF  bl (u8ins)  ELSE  .all .rest  THEN  0 ;
1.3       pazsan    159: 
1.4       pazsan    160: : utf-8-io ( -- )
                    161:     ['] u8forw       ctrl F bindkey
                    162:     ['] u8back       ctrl B bindkey
                    163:     ['] ?u8del       ctrl H bindkey
                    164:     ['] u8eof        ctrl D bindkey
                    165:     ['] <u8del>      ctrl X bindkey
                    166:     ['] u8clear-tib  ctrl K bindkey
                    167:     ['] u8first-pos  ctrl A bindkey
                    168:     ['] u8end-pos    ctrl E bindkey
                    169:     ['] (u8enter)    #lf    bindkey
                    170:     ['] (u8enter)    #cr    bindkey
                    171:     ['] u8tab-expand #tab   bindkey
                    172:     ['] (u8ins)      IS insert-char
1.5       pazsan    173:     ['] kill-prefix  IS everychar
1.4       pazsan    174:     ['] save-cursor  IS everyline
                    175:     ['] u8key        IS key
                    176:     ['] u8emit       IS emit ;
                    177: 

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>