lib/unilbrk/u16-possible-linebreaks.c

   1 /* Line breaking of UTF-16 strings.
   2    Copyright (C) 2001-2003, 2006-2017 Free Software Foundation, Inc.
   3    Written by Bruno Haible <bruno@clisp.org>, 2001.
   4
   5    This program is free software: you can redistribute it and/or modify it
   6    under the terms of the GNU Lesser General Public License as published
   7    by the Free Software Foundation; either version 3 of the License, or
   8    (at your option) any later version.
   9
  10    This program is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13    Lesser General Public License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public License
  16    along with this program.  If not, see <https://www.gnu.org/licenses/>.  */
  17
  18 #include <config.h>
  19
  20 /* Specification.  */
  21 #include "unilbrk.h"
  22
  23 #include <stdlib.h>
  24 #include <string.h>
  25
  26 #include "unilbrk/lbrktables.h"
  27 #include "uniwidth/cjk.h"
  28 #include "unistr.h"
  29
  30 void
  31 u16_possible_linebreaks (const uint16_t *s, size_t n, const char *encoding, char *p)
  32 {
  33   int LBP_AI_REPLACEMENT = (is_cjk_encoding (encoding) ? LBP_ID : LBP_AL);
  34   const uint16_t *s_end = s + n;
  35   int last_prop = LBP_BK; /* line break property of last non-space character */
  36   char *seen_space = NULL; /* Was a space seen after the last non-space character? */
  37   char *seen_space2 = NULL; /* At least two spaces after the last non-space? */
  38
  39   /* Don't break inside multibyte characters.  */
  40   memset (p, UC_BREAK_PROHIBITED, n);
  41
  42   while (s < s_end)
  43     {
  44       ucs4_t uc;
  45       int count = u16_mbtouc_unsafe (&uc, s, s_end - s);
  46       int prop = unilbrkprop_lookup (uc);
  47
  48       if (prop == LBP_BK)
  49         {
  50           /* Mandatory break.  */
  51           *p = UC_BREAK_MANDATORY;
  52           last_prop = LBP_BK;
  53           seen_space = NULL;
  54           seen_space2 = NULL;
  55         }
  56       else
  57         {
  58           char *q;
  59
  60           /* Resolve property values whose behaviour is not fixed.  */
  61           switch (prop)
  62             {
  63             case LBP_AI:
  64               /* Resolve ambiguous.  */
  65               prop = LBP_AI_REPLACEMENT;
  66               break;
  67             case LBP_CB:
  68               /* This is arbitrary.  */
  69               prop = LBP_ID;
  70               break;
  71             case LBP_SA:
  72               /* We don't handle complex scripts yet.
  73                  Treat LBP_SA like LBP_XX.  */
  74             case LBP_XX:
  75               /* This is arbitrary.  */
  76               prop = LBP_AL;
  77               break;
  78             }
  79
  80           /* Deal with spaces and combining characters.  */
  81           q = p;
  82           if (prop == LBP_SP)
  83             {
  84               /* Don't break just before a space.  */
  85               *p = UC_BREAK_PROHIBITED;
  86               seen_space2 = seen_space;
  87               seen_space = p;
  88             }
  89           else if (prop == LBP_ZW)
  90             {
  91               /* Don't break just before a zero-width space.  */
  92               *p = UC_BREAK_PROHIBITED;
  93               last_prop = LBP_ZW;
  94               seen_space = NULL;
  95               seen_space2 = NULL;
  96             }
  97           else if (prop == LBP_CM)
  98             {
  99               /* Don't break just before a combining character, except immediately after a
 100                  zero-width space.  */
 101               if (last_prop == LBP_ZW)
 102                 {
 103                   /* Break after zero-width space.  */
 104                   *p = UC_BREAK_POSSIBLE;
 105                   /* A combining character turns a preceding space into LBP_ID.  */
 106                   last_prop = LBP_ID;
 107                 }
 108               else
 109                 {
 110                   *p = UC_BREAK_PROHIBITED;
 111                   /* A combining character turns a preceding space into LBP_ID.  */
 112                   if (seen_space != NULL)
 113                     {
 114                       q = seen_space;
 115                       seen_space = seen_space2;
 116                       prop = LBP_ID;
 117                       goto lookup_via_table;
 118                     }
 119                 }
 120             }
 121           else
 122             {
 123              lookup_via_table:
 124               /* prop must be usable as an index for table 7.3 of UTR #14.  */
 125               if (!(prop >= 0 && prop < sizeof (unilbrk_table) / sizeof (unilbrk_table[0])))
 126                 abort ();
 127
 128               if (last_prop == LBP_BK)
 129                 {
 130                   /* Don't break at the beginning of a line.  */
 131                   *q = UC_BREAK_PROHIBITED;
 132                 }
 133               else if (last_prop == LBP_ZW)
 134                 {
 135                   /* Break after zero-width space.  */
 136                   *q = UC_BREAK_POSSIBLE;
 137                 }
 138               else
 139                 {
 140                   switch (unilbrk_table [last_prop] [prop])
 141                     {
 142                     case D:
 143                       *q = UC_BREAK_POSSIBLE;
 144                       break;
 145                     case I:
 146                       *q = (seen_space != NULL ? UC_BREAK_POSSIBLE : UC_BREAK_PROHIBITED);
 147                       break;
 148                     case P:
 149                       *q = UC_BREAK_PROHIBITED;
 150                       break;
 151                     default:
 152                       abort ();
 153                     }
 154                 }
 155               last_prop = prop;
 156               seen_space = NULL;
 157               seen_space2 = NULL;
 158             }
 159         }
 160
 161       s += count;
 162       p += count;
 163     }
 164 }