lib/unictype/pr_byname.c

   1 /* Properties of Unicode characters.
   2    Copyright (C) 2007, 2011-2021 Free Software Foundation, Inc.
   3    Written by Bruno Haible <bruno@clisp.org>, 2007.
   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 "unictype.h"
  22
  23 #include <stdlib.h>
  24 #include <string.h>
  25
  26 /* Indices stored in the 'struct named_category' elements of the perfect hash
  27    table.  We don't use uc_general_category_t values or their addresses
  28    directly, because this would introduce load-time relocations.  */
  29 enum
  30 {
  31   /* General.  */
  32   UC_PROPERTY_INDEX_WHITE_SPACE,
  33   UC_PROPERTY_INDEX_ALPHABETIC,
  34   UC_PROPERTY_INDEX_OTHER_ALPHABETIC,
  35   UC_PROPERTY_INDEX_NOT_A_CHARACTER,
  36   UC_PROPERTY_INDEX_DEFAULT_IGNORABLE_CODE_POINT,
  37   UC_PROPERTY_INDEX_OTHER_DEFAULT_IGNORABLE_CODE_POINT,
  38   UC_PROPERTY_INDEX_DEPRECATED,
  39   UC_PROPERTY_INDEX_LOGICAL_ORDER_EXCEPTION,
  40   UC_PROPERTY_INDEX_VARIATION_SELECTOR,
  41   UC_PROPERTY_INDEX_PRIVATE_USE,
  42   UC_PROPERTY_INDEX_UNASSIGNED_CODE_VALUE,
  43   /* Case.  */
  44   UC_PROPERTY_INDEX_UPPERCASE,
  45   UC_PROPERTY_INDEX_OTHER_UPPERCASE,
  46   UC_PROPERTY_INDEX_LOWERCASE,
  47   UC_PROPERTY_INDEX_OTHER_LOWERCASE,
  48   UC_PROPERTY_INDEX_TITLECASE,
  49   UC_PROPERTY_INDEX_CASED,
  50   UC_PROPERTY_INDEX_CASE_IGNORABLE,
  51   UC_PROPERTY_INDEX_CHANGES_WHEN_LOWERCASED,
  52   UC_PROPERTY_INDEX_CHANGES_WHEN_UPPERCASED,
  53   UC_PROPERTY_INDEX_CHANGES_WHEN_TITLECASED,
  54   UC_PROPERTY_INDEX_CHANGES_WHEN_CASEFOLDED,
  55   UC_PROPERTY_INDEX_CHANGES_WHEN_CASEMAPPED,
  56   UC_PROPERTY_INDEX_SOFT_DOTTED,
  57   /* Identifiers.  */
  58   UC_PROPERTY_INDEX_ID_START,
  59   UC_PROPERTY_INDEX_OTHER_ID_START,
  60   UC_PROPERTY_INDEX_ID_CONTINUE,
  61   UC_PROPERTY_INDEX_OTHER_ID_CONTINUE,
  62   UC_PROPERTY_INDEX_XID_START,
  63   UC_PROPERTY_INDEX_XID_CONTINUE,
  64   UC_PROPERTY_INDEX_PATTERN_WHITE_SPACE,
  65   UC_PROPERTY_INDEX_PATTERN_SYNTAX,
  66   /* Shaping and rendering.  */
  67   UC_PROPERTY_INDEX_JOIN_CONTROL,
  68   UC_PROPERTY_INDEX_GRAPHEME_BASE,
  69   UC_PROPERTY_INDEX_GRAPHEME_EXTEND,
  70   UC_PROPERTY_INDEX_OTHER_GRAPHEME_EXTEND,
  71   UC_PROPERTY_INDEX_GRAPHEME_LINK,
  72   /* Bidi.  */
  73   UC_PROPERTY_INDEX_BIDI_CONTROL,
  74   UC_PROPERTY_INDEX_BIDI_LEFT_TO_RIGHT,
  75   UC_PROPERTY_INDEX_BIDI_HEBREW_RIGHT_TO_LEFT,
  76   UC_PROPERTY_INDEX_BIDI_ARABIC_RIGHT_TO_LEFT,
  77   UC_PROPERTY_INDEX_BIDI_EUROPEAN_DIGIT,
  78   UC_PROPERTY_INDEX_BIDI_EUR_NUM_SEPARATOR,
  79   UC_PROPERTY_INDEX_BIDI_EUR_NUM_TERMINATOR,
  80   UC_PROPERTY_INDEX_BIDI_ARABIC_DIGIT,
  81   UC_PROPERTY_INDEX_BIDI_COMMON_SEPARATOR,
  82   UC_PROPERTY_INDEX_BIDI_BLOCK_SEPARATOR,
  83   UC_PROPERTY_INDEX_BIDI_SEGMENT_SEPARATOR,
  84   UC_PROPERTY_INDEX_BIDI_WHITESPACE,
  85   UC_PROPERTY_INDEX_BIDI_NON_SPACING_MARK,
  86   UC_PROPERTY_INDEX_BIDI_BOUNDARY_NEUTRAL,
  87   UC_PROPERTY_INDEX_BIDI_PDF,
  88   UC_PROPERTY_INDEX_BIDI_EMBEDDING_OR_OVERRIDE,
  89   UC_PROPERTY_INDEX_BIDI_OTHER_NEUTRAL,
  90   /* Numeric.  */
  91   UC_PROPERTY_INDEX_HEX_DIGIT,
  92   UC_PROPERTY_INDEX_ASCII_HEX_DIGIT,
  93   /* CJK.  */
  94   UC_PROPERTY_INDEX_IDEOGRAPHIC,
  95   UC_PROPERTY_INDEX_UNIFIED_IDEOGRAPH,
  96   UC_PROPERTY_INDEX_RADICAL,
  97   UC_PROPERTY_INDEX_IDS_BINARY_OPERATOR,
  98   UC_PROPERTY_INDEX_IDS_TRINARY_OPERATOR,
  99   /* Misc.  */
 100   UC_PROPERTY_INDEX_ZERO_WIDTH,
 101   UC_PROPERTY_INDEX_SPACE,
 102   UC_PROPERTY_INDEX_NON_BREAK,
 103   UC_PROPERTY_INDEX_ISO_CONTROL,
 104   UC_PROPERTY_INDEX_FORMAT_CONTROL,
 105   UC_PROPERTY_INDEX_DASH,
 106   UC_PROPERTY_INDEX_HYPHEN,
 107   UC_PROPERTY_INDEX_PUNCTUATION,
 108   UC_PROPERTY_INDEX_LINE_SEPARATOR,
 109   UC_PROPERTY_INDEX_PARAGRAPH_SEPARATOR,
 110   UC_PROPERTY_INDEX_QUOTATION_MARK,
 111   UC_PROPERTY_INDEX_SENTENCE_TERMINAL,
 112   UC_PROPERTY_INDEX_TERMINAL_PUNCTUATION,
 113   UC_PROPERTY_INDEX_CURRENCY_SYMBOL,
 114   UC_PROPERTY_INDEX_MATH,
 115   UC_PROPERTY_INDEX_OTHER_MATH,
 116   UC_PROPERTY_INDEX_PAIRED_PUNCTUATION,
 117   UC_PROPERTY_INDEX_LEFT_OF_PAIR,
 118   UC_PROPERTY_INDEX_COMBINING,
 119   UC_PROPERTY_INDEX_COMPOSITE,
 120   UC_PROPERTY_INDEX_DECIMAL_DIGIT,
 121   UC_PROPERTY_INDEX_NUMERIC,
 122   UC_PROPERTY_INDEX_DIACRITIC,
 123   UC_PROPERTY_INDEX_EXTENDER,
 124   UC_PROPERTY_INDEX_IGNORABLE_CONTROL
 125 };
 126
 127 /* Get gperf generated lookup function.  */
 128 #include "unictype/pr_byname.h"
 129
 130 static const uc_property_t UC_PROPERTY_NONE = { NULL };
 131
 132 uc_property_t
 133 uc_property_byname (const char *property_name)
 134 {
 135   char buf[MAX_WORD_LENGTH + 1];
 136   const char *cp;
 137   char *bp;
 138   unsigned int count;
 139   const struct named_property *found;
 140
 141   for (cp = property_name, bp = buf, count = MAX_WORD_LENGTH + 1; ; cp++, bp++)
 142     {
 143       unsigned char c = (unsigned char) *cp;
 144       if (c >= 0x80)
 145         goto invalid;
 146       if (c >= 'A' && c <= 'Z')
 147         c += 'a' - 'A';
 148       else if (c == ' ' || c == '-')
 149         c = '_';
 150       *bp = c;
 151       if (c == '\0')
 152         break;
 153       if (--count == 0)
 154         goto invalid;
 155     }
 156   found = uc_property_lookup (buf, bp - buf);
 157   if (found != NULL)
 158     /* Use a 'switch' statement here, because a table would introduce load-time
 159        relocations.  */
 160     switch (found->property_index)
 161       {
 162       case UC_PROPERTY_INDEX_WHITE_SPACE:
 163         return UC_PROPERTY_WHITE_SPACE;
 164       case UC_PROPERTY_INDEX_ALPHABETIC:
 165         return UC_PROPERTY_ALPHABETIC;
 166       case UC_PROPERTY_INDEX_OTHER_ALPHABETIC:
 167         return UC_PROPERTY_OTHER_ALPHABETIC;
 168       case UC_PROPERTY_INDEX_NOT_A_CHARACTER:
 169         return UC_PROPERTY_NOT_A_CHARACTER;
 170       case UC_PROPERTY_INDEX_DEFAULT_IGNORABLE_CODE_POINT:
 171         return UC_PROPERTY_DEFAULT_IGNORABLE_CODE_POINT;
 172       case UC_PROPERTY_INDEX_OTHER_DEFAULT_IGNORABLE_CODE_POINT:
 173         return UC_PROPERTY_OTHER_DEFAULT_IGNORABLE_CODE_POINT;
 174       case UC_PROPERTY_INDEX_DEPRECATED:
 175         return UC_PROPERTY_DEPRECATED;
 176       case UC_PROPERTY_INDEX_LOGICAL_ORDER_EXCEPTION:
 177         return UC_PROPERTY_LOGICAL_ORDER_EXCEPTION;
 178       case UC_PROPERTY_INDEX_VARIATION_SELECTOR:
 179         return UC_PROPERTY_VARIATION_SELECTOR;
 180       case UC_PROPERTY_INDEX_PRIVATE_USE:
 181         return UC_PROPERTY_PRIVATE_USE;
 182       case UC_PROPERTY_INDEX_UNASSIGNED_CODE_VALUE:
 183         return UC_PROPERTY_UNASSIGNED_CODE_VALUE;
 184       case UC_PROPERTY_INDEX_UPPERCASE:
 185         return UC_PROPERTY_UPPERCASE;
 186       case UC_PROPERTY_INDEX_OTHER_UPPERCASE:
 187         return UC_PROPERTY_OTHER_UPPERCASE;
 188       case UC_PROPERTY_INDEX_LOWERCASE:
 189         return UC_PROPERTY_LOWERCASE;
 190       case UC_PROPERTY_INDEX_OTHER_LOWERCASE:
 191         return UC_PROPERTY_OTHER_LOWERCASE;
 192       case UC_PROPERTY_INDEX_TITLECASE:
 193         return UC_PROPERTY_TITLECASE;
 194       case UC_PROPERTY_INDEX_CASED:
 195         return UC_PROPERTY_CASED;
 196       case UC_PROPERTY_INDEX_CASE_IGNORABLE:
 197         return UC_PROPERTY_CASE_IGNORABLE;
 198       case UC_PROPERTY_INDEX_CHANGES_WHEN_LOWERCASED:
 199         return UC_PROPERTY_CHANGES_WHEN_LOWERCASED;
 200       case UC_PROPERTY_INDEX_CHANGES_WHEN_UPPERCASED:
 201         return UC_PROPERTY_CHANGES_WHEN_UPPERCASED;
 202       case UC_PROPERTY_INDEX_CHANGES_WHEN_TITLECASED:
 203         return UC_PROPERTY_CHANGES_WHEN_TITLECASED;
 204       case UC_PROPERTY_INDEX_CHANGES_WHEN_CASEFOLDED:
 205         return UC_PROPERTY_CHANGES_WHEN_CASEFOLDED;
 206       case UC_PROPERTY_INDEX_CHANGES_WHEN_CASEMAPPED:
 207         return UC_PROPERTY_CHANGES_WHEN_CASEMAPPED;
 208       case UC_PROPERTY_INDEX_SOFT_DOTTED:
 209         return UC_PROPERTY_SOFT_DOTTED;
 210       case UC_PROPERTY_INDEX_ID_START:
 211         return UC_PROPERTY_ID_START;
 212       case UC_PROPERTY_INDEX_OTHER_ID_START:
 213         return UC_PROPERTY_OTHER_ID_START;
 214       case UC_PROPERTY_INDEX_ID_CONTINUE:
 215         return UC_PROPERTY_ID_CONTINUE;
 216       case UC_PROPERTY_INDEX_OTHER_ID_CONTINUE:
 217         return UC_PROPERTY_OTHER_ID_CONTINUE;
 218       case UC_PROPERTY_INDEX_XID_START:
 219         return UC_PROPERTY_XID_START;
 220       case UC_PROPERTY_INDEX_XID_CONTINUE:
 221         return UC_PROPERTY_XID_CONTINUE;
 222       case UC_PROPERTY_INDEX_PATTERN_WHITE_SPACE:
 223         return UC_PROPERTY_PATTERN_WHITE_SPACE;
 224       case UC_PROPERTY_INDEX_PATTERN_SYNTAX:
 225         return UC_PROPERTY_PATTERN_SYNTAX;
 226       case UC_PROPERTY_INDEX_JOIN_CONTROL:
 227         return UC_PROPERTY_JOIN_CONTROL;
 228       case UC_PROPERTY_INDEX_GRAPHEME_BASE:
 229         return UC_PROPERTY_GRAPHEME_BASE;
 230       case UC_PROPERTY_INDEX_GRAPHEME_EXTEND:
 231         return UC_PROPERTY_GRAPHEME_EXTEND;
 232       case UC_PROPERTY_INDEX_OTHER_GRAPHEME_EXTEND:
 233         return UC_PROPERTY_OTHER_GRAPHEME_EXTEND;
 234       case UC_PROPERTY_INDEX_GRAPHEME_LINK:
 235         return UC_PROPERTY_GRAPHEME_LINK;
 236       case UC_PROPERTY_INDEX_BIDI_CONTROL:
 237         return UC_PROPERTY_BIDI_CONTROL;
 238       case UC_PROPERTY_INDEX_BIDI_LEFT_TO_RIGHT:
 239         return UC_PROPERTY_BIDI_LEFT_TO_RIGHT;
 240       case UC_PROPERTY_INDEX_BIDI_HEBREW_RIGHT_TO_LEFT:
 241         return UC_PROPERTY_BIDI_HEBREW_RIGHT_TO_LEFT;
 242       case UC_PROPERTY_INDEX_BIDI_ARABIC_RIGHT_TO_LEFT:
 243         return UC_PROPERTY_BIDI_ARABIC_RIGHT_TO_LEFT;
 244       case UC_PROPERTY_INDEX_BIDI_EUROPEAN_DIGIT:
 245         return UC_PROPERTY_BIDI_EUROPEAN_DIGIT;
 246       case UC_PROPERTY_INDEX_BIDI_EUR_NUM_SEPARATOR:
 247         return UC_PROPERTY_BIDI_EUR_NUM_SEPARATOR;
 248       case UC_PROPERTY_INDEX_BIDI_EUR_NUM_TERMINATOR:
 249         return UC_PROPERTY_BIDI_EUR_NUM_TERMINATOR;
 250       case UC_PROPERTY_INDEX_BIDI_ARABIC_DIGIT:
 251         return UC_PROPERTY_BIDI_ARABIC_DIGIT;
 252       case UC_PROPERTY_INDEX_BIDI_COMMON_SEPARATOR:
 253         return UC_PROPERTY_BIDI_COMMON_SEPARATOR;
 254       case UC_PROPERTY_INDEX_BIDI_BLOCK_SEPARATOR:
 255         return UC_PROPERTY_BIDI_BLOCK_SEPARATOR;
 256       case UC_PROPERTY_INDEX_BIDI_SEGMENT_SEPARATOR:
 257         return UC_PROPERTY_BIDI_SEGMENT_SEPARATOR;
 258       case UC_PROPERTY_INDEX_BIDI_WHITESPACE:
 259         return UC_PROPERTY_BIDI_WHITESPACE;
 260       case UC_PROPERTY_INDEX_BIDI_NON_SPACING_MARK:
 261         return UC_PROPERTY_BIDI_NON_SPACING_MARK;
 262       case UC_PROPERTY_INDEX_BIDI_BOUNDARY_NEUTRAL:
 263         return UC_PROPERTY_BIDI_BOUNDARY_NEUTRAL;
 264       case UC_PROPERTY_INDEX_BIDI_PDF:
 265         return UC_PROPERTY_BIDI_PDF;
 266       case UC_PROPERTY_INDEX_BIDI_EMBEDDING_OR_OVERRIDE:
 267         return UC_PROPERTY_BIDI_EMBEDDING_OR_OVERRIDE;
 268       case UC_PROPERTY_INDEX_BIDI_OTHER_NEUTRAL:
 269         return UC_PROPERTY_BIDI_OTHER_NEUTRAL;
 270       case UC_PROPERTY_INDEX_HEX_DIGIT:
 271         return UC_PROPERTY_HEX_DIGIT;
 272       case UC_PROPERTY_INDEX_ASCII_HEX_DIGIT:
 273         return UC_PROPERTY_ASCII_HEX_DIGIT;
 274       case UC_PROPERTY_INDEX_IDEOGRAPHIC:
 275         return UC_PROPERTY_IDEOGRAPHIC;
 276       case UC_PROPERTY_INDEX_UNIFIED_IDEOGRAPH:
 277         return UC_PROPERTY_UNIFIED_IDEOGRAPH;
 278       case UC_PROPERTY_INDEX_RADICAL:
 279         return UC_PROPERTY_RADICAL;
 280       case UC_PROPERTY_INDEX_IDS_BINARY_OPERATOR:
 281         return UC_PROPERTY_IDS_BINARY_OPERATOR;
 282       case UC_PROPERTY_INDEX_IDS_TRINARY_OPERATOR:
 283         return UC_PROPERTY_IDS_TRINARY_OPERATOR;
 284       case UC_PROPERTY_INDEX_ZERO_WIDTH:
 285         return UC_PROPERTY_ZERO_WIDTH;
 286       case UC_PROPERTY_INDEX_SPACE:
 287         return UC_PROPERTY_SPACE;
 288       case UC_PROPERTY_INDEX_NON_BREAK:
 289         return UC_PROPERTY_NON_BREAK;
 290       case UC_PROPERTY_INDEX_ISO_CONTROL:
 291         return UC_PROPERTY_ISO_CONTROL;
 292       case UC_PROPERTY_INDEX_FORMAT_CONTROL:
 293         return UC_PROPERTY_FORMAT_CONTROL;
 294       case UC_PROPERTY_INDEX_DASH:
 295         return UC_PROPERTY_DASH;
 296       case UC_PROPERTY_INDEX_HYPHEN:
 297         return UC_PROPERTY_HYPHEN;
 298       case UC_PROPERTY_INDEX_PUNCTUATION:
 299         return UC_PROPERTY_PUNCTUATION;
 300       case UC_PROPERTY_INDEX_LINE_SEPARATOR:
 301         return UC_PROPERTY_LINE_SEPARATOR;
 302       case UC_PROPERTY_INDEX_PARAGRAPH_SEPARATOR:
 303         return UC_PROPERTY_PARAGRAPH_SEPARATOR;
 304       case UC_PROPERTY_INDEX_QUOTATION_MARK:
 305         return UC_PROPERTY_QUOTATION_MARK;
 306       case UC_PROPERTY_INDEX_SENTENCE_TERMINAL:
 307         return UC_PROPERTY_SENTENCE_TERMINAL;
 308       case UC_PROPERTY_INDEX_TERMINAL_PUNCTUATION:
 309         return UC_PROPERTY_TERMINAL_PUNCTUATION;
 310       case UC_PROPERTY_INDEX_CURRENCY_SYMBOL:
 311         return UC_PROPERTY_CURRENCY_SYMBOL;
 312       case UC_PROPERTY_INDEX_MATH:
 313         return UC_PROPERTY_MATH;
 314       case UC_PROPERTY_INDEX_OTHER_MATH:
 315         return UC_PROPERTY_OTHER_MATH;
 316       case UC_PROPERTY_INDEX_PAIRED_PUNCTUATION:
 317         return UC_PROPERTY_PAIRED_PUNCTUATION;
 318       case UC_PROPERTY_INDEX_LEFT_OF_PAIR:
 319         return UC_PROPERTY_LEFT_OF_PAIR;
 320       case UC_PROPERTY_INDEX_COMBINING:
 321         return UC_PROPERTY_COMBINING;
 322       case UC_PROPERTY_INDEX_COMPOSITE:
 323         return UC_PROPERTY_COMPOSITE;
 324       case UC_PROPERTY_INDEX_DECIMAL_DIGIT:
 325         return UC_PROPERTY_DECIMAL_DIGIT;
 326       case UC_PROPERTY_INDEX_NUMERIC:
 327         return UC_PROPERTY_NUMERIC;
 328       case UC_PROPERTY_INDEX_DIACRITIC:
 329         return UC_PROPERTY_DIACRITIC;
 330       case UC_PROPERTY_INDEX_EXTENDER:
 331         return UC_PROPERTY_EXTENDER;
 332       case UC_PROPERTY_INDEX_IGNORABLE_CONTROL:
 333         return UC_PROPERTY_IGNORABLE_CONTROL;
 334       default:
 335         abort ();
 336       }
 337  invalid:
 338   return UC_PROPERTY_NONE;
 339 }