Home | History | Annotate | Download | only in src
      1 /*
      2  * Copyright  2009  Red Hat, Inc.
      3  * Copyright  2011  Codethink Limited
      4  * Copyright  2010,2011,2012  Google, Inc.
      5  *
      6  *  This is part of HarfBuzz, a text shaping library.
      7  *
      8  * Permission is hereby granted, without written agreement and without
      9  * license or royalty fees, to use, copy, modify, and distribute this
     10  * software and its documentation for any purpose, provided that the
     11  * above copyright notice and the following two paragraphs appear in
     12  * all copies of this software.
     13  *
     14  * IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
     15  * DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
     16  * ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
     17  * IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
     18  * DAMAGE.
     19  *
     20  * THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
     21  * BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
     22  * FITNESS FOR A PARTICULAR PURPOSE.  THE SOFTWARE PROVIDED HEREUNDER IS
     23  * ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
     24  * PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
     25  *
     26  * Red Hat Author(s): Behdad Esfahbod
     27  * Codethink Author(s): Ryan Lortie
     28  * Google Author(s): Behdad Esfahbod
     29  */
     30 
     31 #ifndef HB_UNICODE_PRIVATE_HH
     32 #define HB_UNICODE_PRIVATE_HH
     33 
     34 #include "hb-private.hh"
     35 
     36 #include "hb-unicode.h"
     37 #include "hb-object-private.hh"
     38 
     39 
     40 extern HB_INTERNAL const uint8_t _hb_modified_combining_class[256];
     41 
     42 /*
     43  * hb_unicode_funcs_t
     44  */
     45 
     46 #define HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS \
     47   HB_UNICODE_FUNC_IMPLEMENT (combining_class) \
     48   HB_UNICODE_FUNC_IMPLEMENT (eastasian_width) \
     49   HB_UNICODE_FUNC_IMPLEMENT (general_category) \
     50   HB_UNICODE_FUNC_IMPLEMENT (mirroring) \
     51   HB_UNICODE_FUNC_IMPLEMENT (script) \
     52   HB_UNICODE_FUNC_IMPLEMENT (compose) \
     53   HB_UNICODE_FUNC_IMPLEMENT (decompose) \
     54   HB_UNICODE_FUNC_IMPLEMENT (decompose_compatibility) \
     55   /* ^--- Add new callbacks here */
     56 
     57 /* Simple callbacks are those taking a hb_codepoint_t and returning a hb_codepoint_t */
     58 #define HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS_SIMPLE \
     59   HB_UNICODE_FUNC_IMPLEMENT (hb_unicode_combining_class_t, combining_class) \
     60   HB_UNICODE_FUNC_IMPLEMENT (unsigned int, eastasian_width) \
     61   HB_UNICODE_FUNC_IMPLEMENT (hb_unicode_general_category_t, general_category) \
     62   HB_UNICODE_FUNC_IMPLEMENT (hb_codepoint_t, mirroring) \
     63   HB_UNICODE_FUNC_IMPLEMENT (hb_script_t, script) \
     64   /* ^--- Add new simple callbacks here */
     65 
     66 struct hb_unicode_funcs_t {
     67   hb_object_header_t header;
     68   ASSERT_POD ();
     69 
     70   hb_unicode_funcs_t *parent;
     71 
     72   bool immutable;
     73 
     74 #define HB_UNICODE_FUNC_IMPLEMENT(return_type, name) \
     75   inline return_type name (hb_codepoint_t unicode) { return func.name (this, unicode, user_data.name); }
     76 HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS_SIMPLE
     77 #undef HB_UNICODE_FUNC_IMPLEMENT
     78 
     79   inline hb_bool_t compose (hb_codepoint_t a, hb_codepoint_t b,
     80 			    hb_codepoint_t *ab)
     81   {
     82     *ab = 0;
     83     if (unlikely (!a || !b)) return false;
     84     return func.compose (this, a, b, ab, user_data.compose);
     85   }
     86 
     87   inline hb_bool_t decompose (hb_codepoint_t ab,
     88 			      hb_codepoint_t *a, hb_codepoint_t *b)
     89   {
     90     *a = ab; *b = 0;
     91     return func.decompose (this, ab, a, b, user_data.decompose);
     92   }
     93 
     94   inline unsigned int decompose_compatibility (hb_codepoint_t  u,
     95 					       hb_codepoint_t *decomposed)
     96   {
     97     unsigned int ret = func.decompose_compatibility (this, u, decomposed, user_data.decompose_compatibility);
     98     if (ret == 1 && u == decomposed[0]) {
     99       decomposed[0] = 0;
    100       return 0;
    101     }
    102     decomposed[ret] = 0;
    103     return ret;
    104   }
    105 
    106 
    107   unsigned int
    108   modified_combining_class (hb_codepoint_t unicode)
    109   {
    110     /* XXX This hack belongs to the Myanmar shaper. */
    111     if (unicode == 0x1037) unicode = 0x103A;
    112 
    113     return _hb_modified_combining_class[combining_class (unicode)];
    114   }
    115 
    116   inline hb_bool_t
    117   is_variation_selector (hb_codepoint_t unicode)
    118   {
    119     return unlikely (hb_in_ranges<hb_codepoint_t> (unicode,
    120 						   0x180B, 0x180D, /* MONGOLIAN FREE VARIATION SELECTOR ONE..THREE */
    121 						   0xFE00, 0xFE0F, /* VARIATION SELECTOR-1..16 */
    122 						   0xE0100, 0xE01EF));  /* VARIATION SELECTOR-17..256 */
    123   }
    124 
    125   /* Default_Ignorable codepoints:
    126    *
    127    * Note that as of Oct 2012 (Unicode 6.2), U+180E MONGOLIAN VOWEL SEPARATOR
    128    * is NOT Default_Ignorable, but it really behaves in a way that it should
    129    * be.  That has been reported to the Unicode Technical Committee for
    130    * consideration.  As such, we include it here, since Uniscribe removes it.
    131    * It *is* in Unicode 6.3 however.  U+061C ARABIC LETTER MARK from Unicode
    132    * 6.3 is also added manually.  The new Unicode 6.3 bidi formatting
    133    * characters are encoded in a block that was Default_Ignorable already.
    134    *
    135    * Note: While U+115F and U+1160 are Default_Ignorable, we do NOT want to
    136    * hide them, as the way Uniscribe has implemented them is with regular
    137    * spacing glyphs, and that's the way fonts are made to work.  As such,
    138    * we make exceptions for those two.
    139    *
    140    * Gathered from:
    141    * http://unicode.org/cldr/utility/list-unicodeset.jsp?a=[:DI:]&abb=on&ucd=on&esc=on
    142    *
    143    * Last updated to the page with the following versions:
    144    * Version 3.6; ICU version: 50.0.1.0; Unicode version: 6.1.0.0
    145    *
    146    * 4,167 Code Points
    147    *
    148    * [\u00AD\u034F\u115F\u1160\u17B4\u17B5\u180B-\u180D\u200B-\u200F\u202A-\u202E\u2060-\u206F\u3164\uFE00-\uFE0F\uFEFF\uFFA0\uFFF0-\uFFF8\U0001D173-\U0001D17A\U000E0000-\U000E0FFF]
    149    *
    150    * 00AD ;SOFT HYPHEN
    151    * 034F ;COMBINING GRAPHEME JOINER
    152    * #115F ;HANGUL CHOSEONG FILLER
    153    * #1160 ;HANGUL JUNGSEONG FILLER
    154    * 17B4 ;KHMER VOWEL INHERENT AQ
    155    * 17B5 ;KHMER VOWEL INHERENT AA
    156    * 180B..180D ;MONGOLIAN FREE VARIATION SELECTOR THREE
    157    * 200B..200F ;RIGHT-TO-LEFT MARK
    158    * 202A..202E ;RIGHT-TO-LEFT OVERRIDE
    159    * 2060..206F ;NOMINAL DIGIT SHAPES
    160    * 3164 ;HANGUL FILLER
    161    * FE00..FE0F ;VARIATION SELECTOR-16
    162    * FEFF ;ZERO WIDTH NO-BREAK SPACE
    163    * FFA0 ;HALFWIDTH HANGUL FILLER
    164    * FFF0..FFF8 ;<unassigned-FFF8>
    165    * 1D173..1D17A ;MUSICAL SYMBOL END PHRASE
    166    * E0000..E0FFF ;<unassigned-E0FFF>
    167    */
    168   inline hb_bool_t
    169   is_default_ignorable (hb_codepoint_t ch)
    170   {
    171     hb_codepoint_t plane = ch >> 16;
    172     if (likely (plane == 0))
    173     {
    174       /* BMP */
    175       hb_codepoint_t page = ch >> 8;
    176       switch (page) {
    177 	case 0x00: return unlikely (ch == 0x00AD);
    178 	case 0x03: return unlikely (ch == 0x034F);
    179 	case 0x06: return unlikely (ch == 0x061C);
    180 	case 0x17: return hb_in_range<hb_codepoint_t> (ch, 0x17B4, 0x17B5);
    181 	case 0x18: return hb_in_range<hb_codepoint_t> (ch, 0x180B, 0x180E);
    182 	case 0x20: return hb_in_ranges<hb_codepoint_t> (ch, 0x200B, 0x200F,
    183 							    0x202A, 0x202E,
    184 							    0x2060, 0x206F);
    185 	case 0x31: return unlikely (ch == 0x3164);
    186 	case 0xFE: return hb_in_range<hb_codepoint_t> (ch, 0xFE00, 0xFE0F) || ch == 0xFEFF;
    187 	case 0xFF: return hb_in_range<hb_codepoint_t> (ch, 0xFFF0, 0xFFF8) || ch == 0xFFA0;
    188 	default: return false;
    189       }
    190     }
    191     else
    192     {
    193       /* Other planes */
    194       switch (plane) {
    195 	case 0x01: return hb_in_range<hb_codepoint_t> (ch, 0x0001D173, 0x0001D17A);
    196 	case 0x0E: return hb_in_range<hb_codepoint_t> (ch, 0x000E0000, 0x000E0FFF);
    197 	default: return false;
    198       }
    199     }
    200   }
    201 
    202 
    203   struct {
    204 #define HB_UNICODE_FUNC_IMPLEMENT(name) hb_unicode_##name##_func_t name;
    205     HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS
    206 #undef HB_UNICODE_FUNC_IMPLEMENT
    207   } func;
    208 
    209   struct {
    210 #define HB_UNICODE_FUNC_IMPLEMENT(name) void *name;
    211     HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS
    212 #undef HB_UNICODE_FUNC_IMPLEMENT
    213   } user_data;
    214 
    215   struct {
    216 #define HB_UNICODE_FUNC_IMPLEMENT(name) hb_destroy_func_t name;
    217     HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS
    218 #undef HB_UNICODE_FUNC_IMPLEMENT
    219   } destroy;
    220 };
    221 
    222 
    223 extern HB_INTERNAL const hb_unicode_funcs_t _hb_unicode_funcs_nil;
    224 
    225 
    226 /* Modified combining marks */
    227 
    228 /* Hebrew
    229  *
    230  * We permute the "fixed-position" classes 10-26 into the order
    231  * described in the SBL Hebrew manual:
    232  *
    233  * http://www.sbl-site.org/Fonts/SBLHebrewUserManual1.5x.pdf
    234  *
    235  * (as recommended by:
    236  *  http://forum.fontlab.com/archive-old-microsoft-volt-group/vista-and-diacritic-ordering-t6751.0.html)
    237  *
    238  * More details here:
    239  * https://bugzilla.mozilla.org/show_bug.cgi?id=662055
    240  */
    241 #define HB_MODIFIED_COMBINING_CLASS_CCC10 22 /* sheva */
    242 #define HB_MODIFIED_COMBINING_CLASS_CCC11 15 /* hataf segol */
    243 #define HB_MODIFIED_COMBINING_CLASS_CCC12 16 /* hataf patah */
    244 #define HB_MODIFIED_COMBINING_CLASS_CCC13 17 /* hataf qamats */
    245 #define HB_MODIFIED_COMBINING_CLASS_CCC14 23 /* hiriq */
    246 #define HB_MODIFIED_COMBINING_CLASS_CCC15 18 /* tsere */
    247 #define HB_MODIFIED_COMBINING_CLASS_CCC16 19 /* segol */
    248 #define HB_MODIFIED_COMBINING_CLASS_CCC17 20 /* patah */
    249 #define HB_MODIFIED_COMBINING_CLASS_CCC18 21 /* qamats */
    250 #define HB_MODIFIED_COMBINING_CLASS_CCC19 14 /* holam */
    251 #define HB_MODIFIED_COMBINING_CLASS_CCC20 24 /* qubuts */
    252 #define HB_MODIFIED_COMBINING_CLASS_CCC21 12 /* dagesh */
    253 #define HB_MODIFIED_COMBINING_CLASS_CCC22 25 /* meteg */
    254 #define HB_MODIFIED_COMBINING_CLASS_CCC23 13 /* rafe */
    255 #define HB_MODIFIED_COMBINING_CLASS_CCC24 10 /* shin dot */
    256 #define HB_MODIFIED_COMBINING_CLASS_CCC25 11 /* sin dot */
    257 #define HB_MODIFIED_COMBINING_CLASS_CCC26 26 /* point varika */
    258 
    259 /*
    260  * Arabic
    261  *
    262  * Modify to move Shadda (ccc=33) before other marks.  See:
    263  * http://unicode.org/faq/normalization.html#8
    264  * http://unicode.org/faq/normalization.html#9
    265  */
    266 #define HB_MODIFIED_COMBINING_CLASS_CCC27 28 /* fathatan */
    267 #define HB_MODIFIED_COMBINING_CLASS_CCC28 29 /* dammatan */
    268 #define HB_MODIFIED_COMBINING_CLASS_CCC29 30 /* kasratan */
    269 #define HB_MODIFIED_COMBINING_CLASS_CCC30 31 /* fatha */
    270 #define HB_MODIFIED_COMBINING_CLASS_CCC31 32 /* damma */
    271 #define HB_MODIFIED_COMBINING_CLASS_CCC32 33 /* kasra */
    272 #define HB_MODIFIED_COMBINING_CLASS_CCC33 27 /* shadda */
    273 #define HB_MODIFIED_COMBINING_CLASS_CCC34 34 /* sukun */
    274 #define HB_MODIFIED_COMBINING_CLASS_CCC35 35 /* superscript alef */
    275 
    276 /* Syriac */
    277 #define HB_MODIFIED_COMBINING_CLASS_CCC36 36 /* superscript alaph */
    278 
    279 /* Telugu
    280  *
    281  * Modify Telugu length marks (ccc=84, ccc=91).
    282  * These are the only matras in the main Indic scripts range that have
    283  * a non-zero ccc.  That makes them reorder with the Halant that is
    284  * ccc=9.  Just zero them, we don't need them in our Indic shaper.
    285  */
    286 #define HB_MODIFIED_COMBINING_CLASS_CCC84 0 /* length mark */
    287 #define HB_MODIFIED_COMBINING_CLASS_CCC91 0 /* ai length mark */
    288 
    289 /* Thai
    290  *
    291  * Modify U+0E38 and U+0E39 (ccc=103) to be reordered before U+0E3A (ccc=9).
    292  * Assign 3, which is unassigned otherwise.
    293  * Uniscribe does this reordering too.
    294  */
    295 #define HB_MODIFIED_COMBINING_CLASS_CCC103 3 /* sara u / sara uu */
    296 #define HB_MODIFIED_COMBINING_CLASS_CCC107 107 /* mai * */
    297 
    298 /* Lao */
    299 #define HB_MODIFIED_COMBINING_CLASS_CCC118 118 /* sign u / sign uu */
    300 #define HB_MODIFIED_COMBINING_CLASS_CCC122 122 /* mai * */
    301 
    302 /* Tibetan */
    303 #define HB_MODIFIED_COMBINING_CLASS_CCC129 129 /* sign aa */
    304 #define HB_MODIFIED_COMBINING_CLASS_CCC130 130 /* sign i */
    305 #define HB_MODIFIED_COMBINING_CLASS_CCC132 132 /* sign u */
    306 
    307 
    308 /* Misc */
    309 
    310 #define HB_UNICODE_GENERAL_CATEGORY_IS_MARK(gen_cat) \
    311 	(FLAG (gen_cat) & \
    312 	 (FLAG (HB_UNICODE_GENERAL_CATEGORY_SPACING_MARK) | \
    313 	  FLAG (HB_UNICODE_GENERAL_CATEGORY_ENCLOSING_MARK) | \
    314 	  FLAG (HB_UNICODE_GENERAL_CATEGORY_NON_SPACING_MARK)))
    315 
    316 
    317 #endif /* HB_UNICODE_PRIVATE_HH */
    318