dlls/usp10/breaking.c

   1 /*
   2  * Implementation of line breaking algorithm for the Uniscribe Script Processor
   3  *
   4  * Copyright 2011 CodeWeavers, Aric Stewart
   5  *
   6  * This library is free software; you can redistribute it and/or
   7  * modify it under the terms of the GNU Lesser General Public
   8  * License as published by the Free Software Foundation; either
   9  * version 2.1 of the License, or (at your option) any later version.
  10  *
  11  * This library is distributed in the hope that it will be useful,
  12  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  13  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14  * Lesser General Public License for more details.
  15  *
  16  * You should have received a copy of the GNU Lesser General Public
  17  * License along with this library; if not, write to the Free Software
  18  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
  19  *
  20  */
  21 #include "config.h"
  22 #include <stdarg.h>
  23 #include <stdio.h>
  24 #include <stdlib.h>
  25
  26 #include "windef.h"
  27 #include "winbase.h"
  28 #include "winuser.h"
  29 #include "wingdi.h"
  30 #include "winnls.h"
  31 #include "usp10.h"
  32 #include "winternl.h"
  33
  34 #include "wine/debug.h"
  35 #include "usp10_internal.h"
  36
  37 WINE_DEFAULT_DEBUG_CHANNEL(uniscribe);
  38
  39 extern const unsigned short wine_linebreak_table[];
  40
  41 enum breaking_types { b_BK=1, b_CR, b_LF, b_CM, b_SG, b_GL, b_CB, b_SP, b_ZW, b_NL, b_WJ, b_JL, b_JV, b_JT, b_H2, b_H3, b_XX, b_OP, b_CL, b_CP, b_QU, b_NS, b_EX, b_SY, b_IS, b_PR, b_PO, b_NU, b_AL, b_ID, b_IN, b_HY, b_BB, b_BA, b_SA, b_AI, b_B2};
  42
  43 enum breaking_class {b_r=1, b_s, b_x};
  44
  45 static void debug_output_breaks(const short* breaks, int count)
  46 {
  47     if (TRACE_ON(uniscribe))
  48     {
  49         int i;
  50         TRACE("[");
  51         for (i = 0; i < count && i < 200; i++)
  52         {
  53             switch (breaks[i])
  54             {
  55                 case b_x: TRACE("x"); break;
  56                 case b_r: TRACE("!"); break;
  57                 case b_s: TRACE("+"); break;
  58                 default: TRACE("*");
  59             }
  60         }
  61         if (i == 200)
  62             TRACE("...");
  63         TRACE("]\n");
  64     }
  65 }
  66
  67 static inline void else_break(short* before, short class)
  68 {
  69     if (*before == 0)  *before = class;
  70 }
  71
  72 void BREAK_line(const WCHAR *chars, int count, const SCRIPT_ANALYSIS *sa, SCRIPT_LOGATTR *la)
  73 {
  74     int i,j;
  75     short *break_class;
  76     short *break_before;
  77
  78     TRACE("In      %s\n",debugstr_wn(chars,count));
  79
  80     break_class = HeapAlloc(GetProcessHeap(),0, count * sizeof(short));
  81     break_before = HeapAlloc(GetProcessHeap(),0, count * sizeof(short));
  82
  83     for (i = 0; i < count; i++)
  84     {
  85         break_class[i] = get_table_entry( wine_linebreak_table, chars[i] );
  86         break_before[i] = 0;
  87
  88         memset(&la[i],0,sizeof(SCRIPT_LOGATTR));
  89
  90         la[i].fCharStop = TRUE;
  91         switch (break_class[i])
  92         {
  93             case b_BK:
  94             case b_ZW:
  95             case b_SP:
  96                 la[i].fWhiteSpace = TRUE;
  97                 break;
  98             case b_CM:
  99                 la[i].fCharStop = FALSE;
 100         }
 101     }
 102
 103     /* LB1 */
 104     /* TODO: Have outside algorithms for these scripts */
 105     for (i = 0; i < count; i++)
 106     {
 107         switch(break_class[i])
 108         {
 109             case b_AI:
 110             case b_SA:
 111             case b_SG:
 112             case b_XX:
 113                 break_class[i] = b_AL;
 114         }
 115     }
 116
 117     /* LB2 - LB3 */
 118     break_before[0] = b_x;
 119     for (i = 0; i < count; i++)
 120     {
 121         switch(break_class[i])
 122         {
 123             /* LB4 - LB6 */
 124             case b_CR:
 125                 if (i < count-1 && break_class[i+1] == b_LF)
 126                 {
 127                     else_break(&break_before[i],b_x);
 128                     else_break(&break_before[i+1],b_x);
 129                     break;
 130                 }
 131             case b_LF:
 132             case b_NL:
 133             case b_BK:
 134                 if (i < count-1) else_break(&break_before[i+1],b_r);
 135                     else_break(&break_before[i],b_x);
 136                 break;
 137             /* LB7 */
 138             case b_SP:
 139                 else_break(&break_before[i],b_x);
 140                 break;
 141             case b_ZW:
 142                 else_break(&break_before[i],b_x);
 143             /* LB8 */
 144                 while (i < count-1 && break_class[i+1] == b_SP)
 145                     i++;
 146                 else_break(&break_before[i],b_s);
 147                 break;
 148         }
 149     }
 150
 151     debug_output_breaks(break_before,count);
 152
 153     /* LB9 - LB10 */
 154     for (i = 0; i < count; i++)
 155     {
 156         if (break_class[i] == b_CM)
 157         {
 158             if (i > 0)
 159             {
 160                 switch (break_class[i-1])
 161                 {
 162                     case b_SP:
 163                     case b_BK:
 164                     case b_CR:
 165                     case b_LF:
 166                     case b_NL:
 167                     case b_ZW:
 168                         break_class[i] = b_AL;
 169                         break;
 170                     default:
 171                         break_class[i] = break_class[i-1];
 172                 }
 173             }
 174             else break_class[i] = b_AL;
 175         }
 176     }
 177
 178     for (i = 0; i < count; i++)
 179     {
 180         switch(break_class[i])
 181         {
 182             /* LB11 */
 183             case b_WJ:
 184                 else_break(&break_before[i],b_x);
 185                 if (i < count-1)
 186                     else_break(&break_before[i+1],b_x);
 187                 break;
 188             /* LB12 */
 189             case b_GL:
 190                 if (i < count-1)
 191                     else_break(&break_before[i+1],b_x);
 192             /* LB12a */
 193                 if (i > 0)
 194                 {
 195                     if (break_class[i-1] != b_SP &&
 196                         break_class[i-1] != b_BA &&
 197                         break_class[i-1] != b_HY)
 198                         else_break(&break_before[i],b_x);
 199                 }
 200                 break;
 201             /* LB13 */
 202             case b_CL:
 203             case b_CP:
 204             case b_EX:
 205             case b_IS:
 206             case b_SY:
 207                 else_break(&break_before[i],b_x);
 208                 break;
 209             /* LB14 */
 210             case b_OP:
 211                 while (i < count-1 && break_class[i+1] == b_SP)
 212                 {
 213                     else_break(&break_before[i+1],b_x);
 214                     i++;
 215                 }
 216                 else_break(&break_before[i+1],b_x);
 217                 break;
 218             /* LB15 */
 219             case b_QU:
 220                 j = i+1;
 221                 while (j < count-1 && break_class[j] == b_SP)
 222                     j++;
 223                 if (break_class[j] == b_OP)
 224                 {
 225                     for (; j > i; j--)
 226                         else_break(&break_before[j],b_x);
 227                 }
 228                 break;
 229             /* LB16 */
 230             case b_NS:
 231                 j = i-1;
 232                 while(j > 0 && break_class[j] == b_SP)
 233                     j--;
 234                 if (break_class[j] == b_CL || break_class[j] == b_CP)
 235                 {
 236                     for (j++; j <= i; j++)
 237                         else_break(&break_before[j],b_x);
 238                 }
 239                 break;
 240             /* LB17 */
 241             case b_B2:
 242                 j = i+1;
 243                 while (j < count && break_class[j] == b_SP)
 244                     j++;
 245                 if (break_class[j] == b_B2)
 246                 {
 247                     for (; j > i; j--)
 248                         else_break(&break_before[j],b_x);
 249                 }
 250                 break;
 251         }
 252     }
 253
 254     debug_output_breaks(break_before,count);
 255
 256     for (i = 0; i < count; i++)
 257     {
 258         switch(break_class[i])
 259         {
 260             /* LB18 */
 261             case b_SP:
 262                 if (i < count-1)
 263                     else_break(&break_before[i+1],b_s);
 264                 break;
 265             /* LB19 */
 266             case b_QU:
 267                 else_break(&break_before[i],b_x);
 268                 if (i < count-1)
 269                     else_break(&break_before[i+1],b_x);
 270                 break;
 271             /* LB20 */
 272             case b_CB:
 273                 else_break(&break_before[i],b_s);
 274                 if (i < count-1)
 275                     else_break(&break_before[i+1],b_s);
 276             /* LB21 */
 277             case b_BA:
 278             case b_HY:
 279             case b_NS:
 280                 else_break(&break_before[i],b_x);
 281                 break;
 282             case b_BB:
 283                 if (i < count-1)
 284                     else_break(&break_before[i+1],b_x);
 285                 break;
 286             /* LB22 */
 287             case b_IN:
 288                 if (i > 0)
 289                 {
 290                     switch (break_class[i-1])
 291                     {
 292                         case b_AL:
 293                         case b_ID:
 294                         case b_IN:
 295                         case b_NU:
 296                             else_break(&break_before[i], b_x);
 297                     }
 298                 }
 299                 break;
 300         }
 301
 302         if (i < count-1)
 303         {
 304             /* LB23 */
 305             if ((break_class[i] == b_ID && break_class[i+1] == b_PO) ||
 306                 (break_class[i] == b_AL && break_class[i+1] == b_NU) ||
 307                 (break_class[i] == b_NU && break_class[i+1] == b_AL))
 308                     else_break(&break_before[i+1],b_x);
 309             /* LB24 */
 310             if ((break_class[i] == b_PR && break_class[i+1] == b_ID) ||
 311                 (break_class[i] == b_PR && break_class[i+1] == b_AL) ||
 312                 (break_class[i] == b_PO && break_class[i+1] == b_AL))
 313                     else_break(&break_before[i+1],b_x);
 314
 315             /* LB25 */
 316             if ((break_class[i] == b_CL && break_class[i+1] == b_PO) ||
 317                 (break_class[i] == b_CP && break_class[i+1] == b_PO) ||
 318                 (break_class[i] == b_CL && break_class[i+1] == b_PR) ||
 319                 (break_class[i] == b_CP && break_class[i+1] == b_PR) ||
 320                 (break_class[i] == b_NU && break_class[i+1] == b_PO) ||
 321                 (break_class[i] == b_NU && break_class[i+1] == b_PR) ||
 322                 (break_class[i] == b_PO && break_class[i+1] == b_OP) ||
 323                 (break_class[i] == b_PO && break_class[i+1] == b_NU) ||
 324                 (break_class[i] == b_PR && break_class[i+1] == b_OP) ||
 325                 (break_class[i] == b_PR && break_class[i+1] == b_NU) ||
 326                 (break_class[i] == b_HY && break_class[i+1] == b_NU) ||
 327                 (break_class[i] == b_IS && break_class[i+1] == b_NU) ||
 328                 (break_class[i] == b_NU && break_class[i+1] == b_NU) ||
 329                 (break_class[i] == b_SY && break_class[i+1] == b_NU))
 330                     else_break(&break_before[i+1],b_x);
 331
 332             /* LB26 */
 333             if (break_class[i] == b_JL)
 334             {
 335                 switch (break_class[i+1])
 336                 {
 337                     case b_JL:
 338                     case b_JV:
 339                     case b_H2:
 340                     case b_H3:
 341                         else_break(&break_before[i+1],b_x);
 342                 }
 343             }
 344             if ((break_class[i] == b_JV || break_class[i] == b_H2) &&
 345                 (break_class[i+1] == b_JV || break_class[i+1] == b_JT))
 346                     else_break(&break_before[i+1],b_x);
 347             if ((break_class[i] == b_JT || break_class[i] == b_H3) &&
 348                  break_class[i+1] == b_JT)
 349                     else_break(&break_before[i+1],b_x);
 350
 351             /* LB27 */
 352             switch (break_class[i])
 353             {
 354                 case b_JL:
 355                 case b_JV:
 356                 case b_JT:
 357                 case b_H2:
 358                 case b_H3:
 359                     if (break_class[i+1] == b_IN || break_class[i+1] == b_PO)
 360                         else_break(&break_before[i+1],b_x);
 361             }
 362             if (break_class[i] == b_PO)
 363             {
 364                 switch (break_class[i+1])
 365                 {
 366                     case b_JL:
 367                     case b_JV:
 368                     case b_JT:
 369                     case b_H2:
 370                     case b_H3:
 371                         else_break(&break_before[i+1],b_x);
 372                 }
 373             }
 374
 375             /* LB28 */
 376             if (break_class[i] == b_AL && break_class[i+1] == b_AL)
 377                 else_break(&break_before[i+1],b_x);
 378
 379             /* LB29 */
 380             if (break_class[i] == b_IS && break_class[i+1] == b_AL)
 381                 else_break(&break_before[i+1],b_x);
 382
 383             /* LB30 */
 384             if ((break_class[i] == b_AL || break_class[i] == b_NU) &&
 385                  break_class[i+1] == b_OP)
 386                 else_break(&break_before[i+1],b_x);
 387             if (break_class[i] == b_CP &&
 388                 (break_class[i+1] == b_AL || break_class[i] == b_NU))
 389                 else_break(&break_before[i+1],b_x);
 390         }
 391     }
 392     debug_output_breaks(break_before,count);
 393
 394     /* LB31 */
 395     for (i = 0; i < count-1; i++)
 396         else_break(&break_before[i+1],b_s);
 397
 398     debug_output_breaks(break_before,count);
 399     for (i = 0; i < count; i++)
 400     {
 401         if (break_before[i] != b_x)
 402         {
 403             la[i].fSoftBreak = TRUE;
 404             la[i].fWordStop = TRUE;
 405         }
 406     }
 407
 408     HeapFree(GetProcessHeap(), 0, break_before);
 409     HeapFree(GetProcessHeap(), 0, break_class);
 410 }