firmware/bidi.c

   1 /***************************************************************************
   2  *             __________               __   ___.
   3  *   Open      \______   \ ____   ____ |  | _\_ |__   _______  ___
   4  *   Source     |       _//  _ \_/ ___\|  |/ /| __ \ /  _ \  \/  /
   5  *   Jukebox    |    |   (  <_> )  \___|    < | \_\ (  <_> > <  <
   6  *   Firmware   |____|_  /\____/ \___  >__|_ \|___  /\____/__/\_ \
   7  *                     \/            \/     \/    \/            \/
   8  * $Id$
   9  *
  10  * Copyright (C) 2005 by Gadi Cohen
  11  *
  12  * Largely based on php_hebrev by Zeev Suraski <zeev@php.net>
  13  * Heavily modified by Gadi Cohen aka Kinslayer <dragon@wastelands.net>
  14  *
  15  * This program is free software; you can redistribute it and/or
  16  * modify it under the terms of the GNU General Public License
  17  * as published by the Free Software Foundation; either version 2
  18  * of the License, or (at your option) any later version.
  19  *
  20  * This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
  21  * KIND, either express or implied.
  22  *
  23  ****************************************************************************/
  24 #include <stdio.h>
  25 #include <string.h>
  26 #include <ctype.h>
  27 #include "file.h"
  28 #include "lcd.h"
  29 #include "rbunicode.h"
  30 #include "arabjoin.h"
  31 #include "scroll_engine.h"
  32 #include "bidi.h"
  33
  34 /* #define _HEB_BUFFER_LENGTH (MAX_PATH + LCD_WIDTH/2 + 3 + 2 + 2) * 2 */
  35 #define _HEB_BLOCK_TYPE_ENG 1
  36 #define _HEB_BLOCK_TYPE_HEB 0
  37 #define _HEB_ORIENTATION_LTR 1
  38 #define _HEB_ORIENTATION_RTL 0
  39
  40 #define ischar(c) ((c > 0x0589 && c < 0x0700) || \
  41                    (c >= 0xfb50 && c <= 0xfefc) ? 1 : 0)
  42 #define _isblank(c) ((c==' ' || c=='\t') ? 1 : 0)
  43 #define _isnewline(c) ((c=='\n' || c=='\r') ? 1 : 0)
  44 #define XOR(a,b) ((a||b) && !(a&&b))
  45
  46 static const arab_t * arab_lookup(unsigned short uchar)
  47 {
  48     if (uchar >= 0x621 && uchar <= 0x63a)
  49         return &(jointable[uchar - 0x621]);
  50     if (uchar >= 0x640 && uchar <= 0x64a)
  51         return &(jointable[uchar - 0x621 - 5]);
  52     if (uchar >= 0x671 && uchar <= 0x6d5)
  53         return &(jointable[uchar - 0x621 - 5 - 38]);
  54     if (uchar == 0x200D) /* Support for the zero-width joiner */
  55         return &zwj;
  56     return 0;
  57 }
  58
  59 static void arabjoin(unsigned short * stringprt, int length)
  60 {
  61     bool connected = false;
  62     unsigned short * writeprt = stringprt;
  63
  64     const arab_t * prev = 0;
  65     const arab_t * cur;
  66     const arab_t * ligature = 0;
  67     short uchar;
  68
  69     int i;
  70     for (i = 0; i <= length; i++) {
  71         cur = arab_lookup(uchar = *stringprt++);
  72
  73         /* Skip non-arabic chars */
  74         if (cur == 0) {
  75             if (prev) {
  76                 /* Finish the last char */
  77                 if (connected) {
  78                     *writeprt++ = prev->final;
  79                     connected = false;
  80                 } else
  81                     *writeprt++ = prev->isolated;
  82                 prev = 0;
  83                 *writeprt++ = uchar;
  84             } else {
  85                 *writeprt++ = uchar;
  86             }
  87             continue;
  88         }
  89
  90         /* nothing to do for arabic char if the previous was non-arabic */
  91         if (prev == 0) {
  92             prev = cur;
  93             continue;
  94         }
  95
  96         /* if it's LAM, check for LAM+ALEPH ligatures */
  97         if (prev->isolated == 0xfedd) {
  98             switch (cur->isolated) {
  99                 case 0xfe8d:
 100                     ligature = &(lamaleph[0]);
 101                     break;
 102                 case 0xfe87:
 103                     ligature = &(lamaleph[1]);
 104                     break;
 105                 case 0xfe83:
 106                     ligature = &(lamaleph[2]);
 107                     break;
 108                 case 0xfe81:
 109                     ligature = &(lamaleph[3]);
 110             }
 111         }
 112
 113         if (ligature) { /* replace the 2 glyphs by their ligature */
 114             prev = ligature;
 115             ligature = 0;
 116         } else {
 117             if (connected) { /* previous char has something connected to it */
 118                 if (prev->medial && cur->final) /* Can we connect to it? */
 119                     *writeprt++ = prev->medial;
 120                 else {
 121                     *writeprt++ = prev->final;
 122                     connected = false;
 123                 }
 124             } else {
 125                 if (prev->initial && cur->final) { /* Can we connect to it? */
 126                     *writeprt++ = prev->initial;
 127                     connected = true;
 128                 } else
 129                     *writeprt++ = prev->isolated;
 130             }
 131             prev = cur;
 132         }
 133     }
 134 }
 135
 136 unsigned short *bidi_l2v(const unsigned char *str, int orientation)
 137 {
 138     int length = utf8length(str);
 139     static unsigned short  utf16_buf[SCROLL_LINE_SIZE];
 140     static unsigned short  bidi_buf[SCROLL_LINE_SIZE];
 141     unsigned short *heb_str, *target, *tmp; /* *broken_str */
 142     int block_start, block_end, block_type, block_length, i;
 143     /*
 144     long max_chars=0;
 145     int begin, end, char_count, orig_begin;
 146
 147     tmp = str;
 148     */
 149     target = tmp = utf16_buf;
 150     while (*str)
 151         str = utf8decode(str, target++);
 152     *target = 0;
 153
 154     if (target == utf16_buf) /* empty string */
 155         return target;
 156
 157     /* properly join any arabic chars */
 158     arabjoin(utf16_buf, length);
 159
 160     block_start=block_end=block_length=0;
 161
 162     heb_str = bidi_buf;
 163     if (orientation) {
 164         target = heb_str;
 165     } else {
 166         target = heb_str + length;
 167         *target = 0;
 168         target--;
 169     }
 170
 171     if (ischar(*tmp))
 172         block_type = _HEB_BLOCK_TYPE_HEB;
 173     else
 174         block_type = _HEB_BLOCK_TYPE_ENG;
 175
 176     do {
 177         while((XOR(ischar(*(tmp+1)),block_type)
 178                || _isblank(*(tmp+1)) || ispunct((int)*(tmp+1))
 179                || *(tmp+1)=='\n')
 180               && block_end < length-1) {
 181                 tmp++;
 182                 block_end++;
 183                 block_length++;
 184         }
 185
 186         if (block_type != orientation) {
 187             while ((_isblank(*tmp) || ispunct((int)*tmp))
 188                    && *tmp!='/' && *tmp!='-' && block_end>block_start) {
 189                 tmp--;
 190                 block_end--;
 191             }
 192         }
 193
 194         for (i=block_start; i<=block_end; i++) {
 195             *target = (block_type == orientation) ?
 196                       *(utf16_buf+i) : *(utf16_buf+block_end-i+block_start);
 197             if (block_type!=orientation) {
 198                 switch (*target) {
 199                 case '(':
 200                     *target = ')';
 201                     break;
 202                 case ')':
 203                     *target = '(';
 204                     break;
 205                 default:
 206                     break;
 207                 }
 208             }
 209             target += orientation ? 1 : -1;
 210         }
 211         block_type = !block_type;
 212         block_start=block_end+1;
 213     } while(block_end<length-1);
 214
 215     *target = 0;
 216
 217 #if 0 /* Is this code really necessary? */
 218     broken_str = utf16_buf;
 219     begin=end=length-1;
 220     target = broken_str;
 221
 222     while (1) {
 223         char_count=0;
 224         while ((!max_chars || char_count<max_chars) && begin>0) {
 225             char_count++;
 226             begin--;
 227             if (begin<=0 || _isnewline(heb_str[begin])) {
 228                 while(begin>0 && _isnewline(heb_str[begin-1])) {
 229                     begin--;
 230                     char_count++;
 231                 }
 232                 break;
 233             }
 234         }
 235         if (char_count==max_chars) { /* try to avoid breaking words */
 236             int new_char_count = char_count;
 237             int new_begin = begin;
 238
 239             while (new_char_count>0) {
 240                 if (_isblank(heb_str[new_begin]) ||
 241                     _isnewline(heb_str[new_begin])) {
 242                     break;
 243                 }
 244                 new_begin++;
 245                 new_char_count--;
 246             }
 247             if (new_char_count>0) {
 248                 char_count=new_char_count;
 249                 begin=new_begin;
 250             }
 251         }
 252         orig_begin=begin;
 253
 254         /* if (_isblank(heb_str[begin])) {
 255             heb_str[begin]='\n';
 256         } */
 257
 258         /* skip leading newlines */
 259         while (begin<=end && _isnewline(heb_str[begin])) {
 260             begin++;
 261         }
 262
 263         /* copy content */
 264         for (i=begin; i<=end; i++) {
 265             *target = heb_str[i];
 266             target++;
 267         }
 268
 269         for (i=orig_begin; i<=end && _isnewline(heb_str[i]); i++) {
 270             *target = heb_str[i];
 271             target++;
 272         }
 273         begin=orig_begin;
 274
 275         if (begin<=0) {
 276             *target = 0;
 277             break;
 278         }
 279         begin--;
 280         end=begin;
 281     }
 282     return broken_str;
 283 #endif
 284     return heb_str;
 285 }
 286