1 /* |
|
2 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. |
|
3 * |
|
4 * This code is free software; you can redistribute it and/or modify it |
|
5 * under the terms of the GNU General Public License version 2 only, as |
|
6 * published by the Free Software Foundation. Oracle designates this |
|
7 * particular file as subject to the "Classpath" exception as provided |
|
8 * by Oracle in the LICENSE file that accompanied this code. |
|
9 * |
|
10 * This code is distributed in the hope that it will be useful, but WITHOUT |
|
11 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or |
|
12 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License |
|
13 * version 2 for more details (a copy is included in the LICENSE file that |
|
14 * accompanied this code). |
|
15 * |
|
16 * You should have received a copy of the GNU General Public License version |
|
17 * 2 along with this work; if not, write to the Free Software Foundation, |
|
18 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. |
|
19 * |
|
20 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA |
|
21 * or visit www.oracle.com if you need additional information or have any |
|
22 * questions. |
|
23 * |
|
24 */ |
|
25 |
|
26 /* |
|
27 * |
|
28 * (C) Copyright IBM Corp. 1998-2013 - All Rights Reserved |
|
29 * |
|
30 */ |
|
31 |
|
32 #include "LETypes.h" |
|
33 #include "LEScripts.h" |
|
34 #include "OpenTypeTables.h" |
|
35 #include "OpenTypeUtilities.h" |
|
36 #include "IndicReordering.h" |
|
37 |
|
38 U_NAMESPACE_BEGIN |
|
39 |
|
40 // Split matra table indices |
|
41 #define _x1 (1 << CF_INDEX_SHIFT) |
|
42 #define _x2 (2 << CF_INDEX_SHIFT) |
|
43 #define _x3 (3 << CF_INDEX_SHIFT) |
|
44 #define _x4 (4 << CF_INDEX_SHIFT) |
|
45 #define _x5 (5 << CF_INDEX_SHIFT) |
|
46 #define _x6 (6 << CF_INDEX_SHIFT) |
|
47 #define _x7 (7 << CF_INDEX_SHIFT) |
|
48 #define _x8 (8 << CF_INDEX_SHIFT) |
|
49 #define _x9 (9 << CF_INDEX_SHIFT) |
|
50 |
|
51 // simple classes |
|
52 #define _xx (CC_RESERVED) |
|
53 #define _ma (CC_VOWEL_MODIFIER | CF_POS_ABOVE) |
|
54 #define _mp (CC_VOWEL_MODIFIER | CF_POS_AFTER) |
|
55 #define _sa (CC_STRESS_MARK | CF_POS_ABOVE) |
|
56 #define _sb (CC_STRESS_MARK | CF_POS_BELOW) |
|
57 #define _iv (CC_INDEPENDENT_VOWEL) |
|
58 #define _i2 (CC_INDEPENDENT_VOWEL_2) |
|
59 #define _i3 (CC_INDEPENDENT_VOWEL_3) |
|
60 #define _ct (CC_CONSONANT | CF_CONSONANT) |
|
61 #define _cn (CC_CONSONANT_WITH_NUKTA | CF_CONSONANT) |
|
62 #define _nu (CC_NUKTA) |
|
63 #define _dv (CC_DEPENDENT_VOWEL) |
|
64 #define _dl (_dv | CF_POS_BEFORE) |
|
65 #define _db (_dv | CF_POS_BELOW) |
|
66 #define _da (_dv | CF_POS_ABOVE) |
|
67 #define _dr (_dv | CF_POS_AFTER) |
|
68 #define _lm (_dv | CF_LENGTH_MARK) |
|
69 #define _l1 (CC_SPLIT_VOWEL_PIECE_1 | CF_POS_BEFORE) |
|
70 #define _a1 (CC_SPLIT_VOWEL_PIECE_1 | CF_POS_ABOVE) |
|
71 #define _b2 (CC_SPLIT_VOWEL_PIECE_2 | CF_POS_BELOW) |
|
72 #define _r2 (CC_SPLIT_VOWEL_PIECE_2 | CF_POS_AFTER) |
|
73 #define _m2 (CC_SPLIT_VOWEL_PIECE_2 | CF_LENGTH_MARK) |
|
74 #define _m3 (CC_SPLIT_VOWEL_PIECE_3 | CF_LENGTH_MARK) |
|
75 #define _vr (CC_VIRAMA) |
|
76 #define _al (CC_AL_LAKUNA) |
|
77 |
|
78 // split matras |
|
79 #define _s1 (_dv | _x1) |
|
80 #define _s2 (_dv | _x2) |
|
81 #define _s3 (_dv | _x3) |
|
82 #define _s4 (_dv | _x4) |
|
83 #define _s5 (_dv | _x5) |
|
84 #define _s6 (_dv | _x6) |
|
85 #define _s7 (_dv | _x7) |
|
86 #define _s8 (_dv | _x8) |
|
87 #define _s9 (_dv | _x9) |
|
88 |
|
89 // consonants with special forms |
|
90 // NOTE: this assumes that no consonants with nukta have |
|
91 // special forms... (Bengali RA?) |
|
92 #define _bb (_ct | CF_BELOW_BASE) |
|
93 #define _pb (_ct | CF_POST_BASE) |
|
94 #define _fb (_ct | CF_PRE_BASE) |
|
95 #define _vt (_bb | CF_VATTU) |
|
96 #define _rv (_vt | CF_REPH) |
|
97 #define _rp (_pb | CF_REPH) |
|
98 #define _rb (_bb | CF_REPH) |
|
99 |
|
100 // |
|
101 // Character class tables |
|
102 // |
|
103 static const IndicClassTable::CharClass devaCharClasses[] = |
|
104 { |
|
105 _xx, _ma, _ma, _mp, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, // 0900 - 090F |
|
106 _iv, _iv, _iv, _iv, _iv, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, // 0910 - 091F |
|
107 _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _cn, _ct, _ct, _ct, _ct, _ct, _ct, // 0920 - 092F |
|
108 _rv, _cn, _ct, _ct, _cn, _ct, _ct, _ct, _ct, _ct, _xx, _xx, _nu, _xx, _dr, _dl, // 0930 - 093F |
|
109 _dr, _db, _db, _db, _db, _da, _da, _da, _da, _dr, _dr, _dr, _dr, _vr, _xx, _xx, // 0940 - 094F |
|
110 _xx, _sa, _sb, _sa, _sa, _xx, _xx, _xx, _cn, _cn, _cn, _cn, _cn, _cn, _cn, _cn, // 0950 - 095F |
|
111 _iv, _iv, _db, _db, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0960 - 096F |
|
112 _xx // 0970 |
|
113 }; |
|
114 |
|
115 static const IndicClassTable::CharClass bengCharClasses[] = |
|
116 { |
|
117 _xx, _ma, _mp, _mp, _xx, _i2, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _xx, _xx, _i2, // 0980 - 098F |
|
118 _iv, _xx, _xx, _iv, _iv, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, // 0990 - 099F |
|
119 _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _xx, _ct, _ct, _bb, _ct, _ct, _pb, // 09A0 - 09AF |
|
120 _rv, _xx, _ct, _xx, _xx, _xx, _ct, _ct, _ct, _ct, _xx, _xx, _nu, _xx, _r2, _dl, // 09B0 - 09BF |
|
121 _dr, _db, _db, _db, _db, _xx, _xx, _l1, _dl, _xx, _xx, _s1, _s2, _vr, _xx, _xx, // 09C0 - 09CF |
|
122 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _m2, _xx, _xx, _xx, _xx, _cn, _cn, _xx, _cn, // 09D0 - 09DF |
|
123 _iv, _iv, _dv, _dv, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 09E0 - 09EF |
|
124 _rv, _ct, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx // 09F0 - 09FA |
|
125 }; |
|
126 |
|
127 static const IndicClassTable::CharClass punjCharClasses[] = |
|
128 { |
|
129 _xx, _ma, _ma, _mp, _xx, _iv, _iv, _iv, _iv, _iv, _iv, _xx, _xx, _xx, _xx, _iv, // 0A00 - 0A0F |
|
130 _iv, _xx, _xx, _i3, _iv, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, // 0A10 - 0A1F |
|
131 _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _xx, _ct, _ct, _ct, _ct, _ct, _bb, // 0A20 - 0A2F |
|
132 _vt, _xx, _ct, _cn, _xx, _bb, _cn, _xx, _ct, _bb, _xx, _xx, _nu, _xx, _dr, _dl, // 0A30 - 0A3F |
|
133 _dr, _b2, _db, _xx, _xx, _xx, _xx, _da, _da, _xx, _xx, _a1, _da, _vr, _xx, _xx, // 0A40 - 0A4F |
|
134 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _cn, _cn, _cn, _ct, _xx, _cn, _xx, // 0A50 - 0A5F |
|
135 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0A60 - 0A6F |
|
136 _ma, _ma, _xx, _xx, _xx // 0A70 - 0A74 |
|
137 }; |
|
138 |
|
139 static const IndicClassTable::CharClass gujrCharClasses[] = |
|
140 { |
|
141 _xx, _ma, _ma, _mp, _xx, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _xx, _iv, _xx, _iv, // 0A80 - 0A8F |
|
142 _iv, _iv, _xx, _iv, _iv, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, // 0A90 - 0A9F |
|
143 _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _xx, _ct, _ct, _ct, _ct, _ct, _ct, // 0AA0 - 0AAF |
|
144 _rv, _xx, _ct, _ct, _xx, _ct, _ct, _ct, _ct, _ct, _xx, _xx, _nu, _xx, _dr, _dl, // 0AB0 - 0ABF |
|
145 _dr, _db, _db, _db, _db, _da, _xx, _da, _da, _dr, _xx, _dr, _dr, _vr, _xx, _xx, // 0AC0 - 0ACF |
|
146 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0AD0 - 0ADF |
|
147 _iv, _iv, _db, _db, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx // 0AE0 - 0AEF |
|
148 }; |
|
149 |
|
150 #if 1 |
|
151 static const IndicClassTable::CharClass oryaCharClasses[] = |
|
152 { |
|
153 _xx, _ma, _mp, _mp, _xx, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _xx, _xx, _iv, /* 0B00 - 0B0F */ |
|
154 _iv, _xx, _xx, _iv, _iv, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _ct, _bb, /* 0B10 - 0B1F */ |
|
155 _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _xx, _bb, _bb, _bb, _bb, _bb, _pb, /* 0B20 - 0B2F */ |
|
156 _rb, _xx, _bb, _bb, _xx, _bb, _bb, _bb, _bb, _bb, _xx, _xx, _nu, _xx, _dr, _da, /* 0B30 - 0B3F */ |
|
157 _dr, _db, _db, _db, _xx, _xx, _xx, _dl, _s1, _xx, _xx, _s2, _s3, _vr, _xx, _xx, /* 0B40 - 0B4F */ |
|
158 _xx, _xx, _xx, _xx, _xx, _xx, _da, _dr, _xx, _xx, _xx, _xx, _cn, _cn, _xx, _pb, /* 0B50 - 0B5F */ |
|
159 _iv, _iv, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, /* 0B60 - 0B6F */ |
|
160 _xx, _bb /* 0B70 - 0B71 */ |
|
161 }; |
|
162 #else |
|
163 static const IndicClassTable::CharClass oryaCharClasses[] = |
|
164 { |
|
165 _xx, _ma, _mp, _mp, _xx, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _xx, _xx, _iv, // 0B00 - 0B0F |
|
166 _iv, _xx, _xx, _iv, _iv, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, // 0B10 - 0B1F |
|
167 _ct, _ct, _ct, _ct, _bb, _ct, _ct, _ct, _bb, _xx, _ct, _ct, _bb, _bb, _bb, _pb, // 0B20 - 0B2F |
|
168 _rb, _xx, _bb, _bb, _xx, _ct, _ct, _ct, _ct, _ct, _xx, _xx, _nu, _xx, _r2, _da, // 0B30 - 0B3F |
|
169 _dr, _db, _db, _db, _xx, _xx, _xx, _l1, _s1, _xx, _xx, _s2, _s3, _vr, _xx, _xx, // 0B40 - 0B4F |
|
170 _xx, _xx, _xx, _xx, _xx, _xx, _m2, _m2, _xx, _xx, _xx, _xx, _cn, _cn, _xx, _cn, // 0B50 - 0B5F |
|
171 _iv, _iv, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0B60 - 0B6F |
|
172 _xx, _ct // 0B70 - 0B71 |
|
173 }; |
|
174 #endif |
|
175 |
|
176 static const IndicClassTable::CharClass tamlCharClasses[] = |
|
177 { |
|
178 _xx, _xx, _ma, _xx, _xx, _iv, _iv, _iv, _iv, _iv, _iv, _xx, _xx, _xx, _iv, _iv, // 0B80 - 0B8F |
|
179 _iv, _xx, _iv, _iv, _iv, _ct, _xx, _xx, _xx, _ct, _ct, _xx, _ct, _xx, _ct, _ct, // 0B90 - 0B9F |
|
180 _xx, _xx, _xx, _ct, _ct, _xx, _xx, _xx, _ct, _ct, _ct, _xx, _xx, _xx, _ct, _ct, // 0BA0 - 0BAF |
|
181 _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _xx, _xx, _xx, _xx, _r2, _dr, // 0BB0 - 0BBF |
|
182 _da, _dr, _dr, _xx, _xx, _xx, _l1, _l1, _dl, _xx, _s1, _s2, _s3, _vr, _xx, _xx, // 0BC0 - 0BCF |
|
183 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _m2, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0BD0 - 0BDF |
|
184 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0BE0 - 0BEF |
|
185 _xx, _xx, _xx // 0BF0 - 0BF2 |
|
186 }; |
|
187 |
|
188 // FIXME: Should some of the bb's be pb's? (KA, NA, MA, YA, VA, etc. (approx 13)) |
|
189 // U+C43 and U+C44 are _lm here not _dr. Similar to the situation with U+CC3 and |
|
190 // U+CC4 in Kannada below. |
|
191 static const IndicClassTable::CharClass teluCharClasses[] = |
|
192 { |
|
193 _xx, _mp, _mp, _mp, _xx, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _xx, _iv, _iv, // 0C00 - 0C0F |
|
194 _iv, _xx, _iv, _iv, _iv, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, // 0C10 - 0C1F |
|
195 _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _xx, _bb, _bb, _bb, _bb, _bb, _bb, // 0C20 - 0C2F |
|
196 _bb, _bb, _bb, _bb, _xx, _bb, _bb, _bb, _bb, _bb, _xx, _xx, _xx, _xx, _da, _da, // 0C30 - 0C3F |
|
197 _da, _dr, _dr, _lm, _lm, _xx, _a1, _da, _s1, _xx, _da, _da, _da, _vr, _xx, _xx, // 0C40 - 0C4F |
|
198 _xx, _xx, _xx, _xx, _xx, _da, _m2, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0C50 - 0C5F |
|
199 _iv, _iv, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx // 0C60 - 0C6F |
|
200 }; |
|
201 |
|
202 // U+CC3 and U+CC4 are _lm here not _dr since the Kannada rendering |
|
203 // rules want them below and to the right of the entire cluster |
|
204 // |
|
205 // There's some information about this in: |
|
206 // |
|
207 // http://brahmi.sourceforge.net/docs/KannadaComputing.html |
|
208 static const IndicClassTable::CharClass kndaCharClasses[] = |
|
209 { |
|
210 _xx, _xx, _mp, _mp, _xx, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _xx, _iv, _iv, // 0C80 - 0C8F |
|
211 _iv, _xx, _iv, _iv, _iv, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, // 0C90 - 0C9F |
|
212 _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _bb, _xx, _bb, _bb, _bb, _bb, _bb, _bb, // 0CA0 - 0CAF |
|
213 _rb, _ct, _bb, _bb, _xx, _bb, _bb, _bb, _bb, _bb, _xx, _xx, _xx, _xx, _dr, _da, // 0CB0 - 0CBF |
|
214 _s1, _dr, _r2, _lm, _lm, _xx, _a1, _s2, _s3, _xx, _s4, _s5, _da, _vr, _xx, _xx, // 0CC0 - 0CCF |
|
215 _xx, _xx, _xx, _xx, _xx, _m3, _m2, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _ct, _xx, // 0CD0 - 0CDF |
|
216 _iv, _iv, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx // 0CE0 - 0CEF |
|
217 }; |
|
218 |
|
219 // FIXME: this is correct for old-style Malayalam (MAL) but not for reformed Malayalam (MLR) |
|
220 // FIXME: should there be a REPH for old-style Malayalam? |
|
221 static const IndicClassTable::CharClass mlymCharClasses[] = |
|
222 { |
|
223 _xx, _xx, _mp, _mp, _xx, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _xx, _iv, _iv, // 0D00 - 0D0F |
|
224 _iv, _xx, _iv, _iv, _iv, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, // 0D10 - 0D1F |
|
225 _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _xx, _ct, _ct, _ct, _ct, _ct, _pb, // 0D20 - 0D2F |
|
226 _fb, _fb, _bb, _ct, _ct, _pb, _ct, _ct, _ct, _ct, _xx, _xx, _xx, _xx, _r2, _dr, // 0D30 - 0D3F |
|
227 _dr, _dr, _dr, _dr, _xx, _xx, _l1, _l1, _dl, _xx, _s1, _s2, _s3, _vr, _xx, _xx, // 0D40 - 0D4F |
|
228 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _m2, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0D50 - 0D5F |
|
229 _iv, _iv, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx // 0D60 - 0D6F |
|
230 }; |
|
231 |
|
232 static const IndicClassTable::CharClass sinhCharClasses[] = |
|
233 { |
|
234 _xx, _xx, _mp, _mp, _xx, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, _iv, // 0D80 - 0D8F |
|
235 _iv, _iv, _iv, _iv, _iv, _iv, _iv, _xx, _xx, _xx, _ct, _ct, _ct, _ct, _ct, _ct, // 0D90 - 0D9F |
|
236 _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, // 0DA0 - 0DAF |
|
237 _ct, _ct, _xx, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _ct, _xx, _ct, _xx, _xx, // 0DB0 - 0DBF |
|
238 _ct, _ct, _ct, _ct, _ct, _ct, _ct, _xx, _xx, _xx, _al, _xx, _xx, _xx, _xx, _dr, // 0DC0 - 0DCF |
|
239 _dr, _dr, _da, _da, _db, _xx, _db, _xx, _dr, _dl, _s1, _dl, _s2, _s3, _s4, _dr, // 0DD0 - 0DDF |
|
240 _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, _xx, // 0DE0 - 0DEF |
|
241 _xx, _xx, _dr, _dr, _xx // 0DF0 - 0DF4 |
|
242 }; |
|
243 |
|
244 // |
|
245 // Split matra tables |
|
246 // |
|
247 static const SplitMatra bengSplitTable[] = {{0x09C7, 0x09BE}, {0x09C7, 0x09D7}}; |
|
248 |
|
249 static const SplitMatra oryaSplitTable[] = {{0x0B47, 0x0B56}, {0x0B47, 0x0B3E}, {0x0B47, 0x0B57}}; |
|
250 |
|
251 static const SplitMatra tamlSplitTable[] = {{0x0BC6, 0x0BBE}, {0x0BC7, 0x0BBE}, {0x0BC6, 0x0BD7}}; |
|
252 |
|
253 static const SplitMatra teluSplitTable[] = {{0x0C46, 0x0C56}}; |
|
254 |
|
255 static const SplitMatra kndaSplitTable[] = {{0x0CBF, 0x0CD5}, {0x0CC6, 0x0CD5}, {0x0CC6, 0x0CD6}, {0x0CC6, 0x0CC2}, |
|
256 {0x0CC6, 0x0CC2, 0x0CD5}}; |
|
257 |
|
258 static const SplitMatra mlymSplitTable[] = {{0x0D46, 0x0D3E}, {0x0D47, 0x0D3E}, {0x0D46, 0x0D57}}; |
|
259 |
|
260 |
|
261 static const SplitMatra sinhSplitTable[] = {{0x0DD9, 0x0DCA}, {0x0DD9, 0x0DCF}, {0x0DD9, 0x0DCF,0x0DCA}, |
|
262 {0x0DD9, 0x0DDF}}; |
|
263 // |
|
264 // Script Flags |
|
265 // |
|
266 |
|
267 // FIXME: post 'GSUB' reordering of MATRA_PRE's for Malayalam and Tamil |
|
268 // FIXME: reformed Malayalam needs to reorder VATTU to before base glyph... |
|
269 // FIXME: not sure passing ZWJ/ZWNJ is best way to render Malayalam Cillu... |
|
270 // FIXME: eyelash RA only for Devanagari?? |
|
271 #define DEVA_SCRIPT_FLAGS (SF_EYELASH_RA | SF_NO_POST_BASE_LIMIT | SF_FILTER_ZERO_WIDTH) |
|
272 #define BENG_SCRIPT_FLAGS (SF_REPH_AFTER_BELOW | SF_NO_POST_BASE_LIMIT | SF_FILTER_ZERO_WIDTH) |
|
273 #define PUNJ_SCRIPT_FLAGS (SF_NO_POST_BASE_LIMIT | SF_FILTER_ZERO_WIDTH) |
|
274 #define GUJR_SCRIPT_FLAGS (SF_NO_POST_BASE_LIMIT | SF_FILTER_ZERO_WIDTH) |
|
275 #define ORYA_SCRIPT_FLAGS (SF_REPH_AFTER_BELOW | SF_NO_POST_BASE_LIMIT | SF_FILTER_ZERO_WIDTH) |
|
276 #define TAML_SCRIPT_FLAGS (SF_MPRE_FIXUP | SF_NO_POST_BASE_LIMIT | SF_FILTER_ZERO_WIDTH) |
|
277 #define TELU_SCRIPT_FLAGS (SF_MATRAS_AFTER_BASE | SF_FILTER_ZERO_WIDTH | 3) |
|
278 #define KNDA_SCRIPT_FLAGS (SF_MATRAS_AFTER_BASE | SF_FILTER_ZERO_WIDTH | 3) |
|
279 #define MLYM_SCRIPT_FLAGS (SF_MPRE_FIXUP | SF_NO_POST_BASE_LIMIT /*| SF_FILTER_ZERO_WIDTH*/) |
|
280 #define SINH_SCRIPT_FLAGS (SF_NO_POST_BASE_LIMIT) |
|
281 |
|
282 // |
|
283 // Indic Class Tables |
|
284 // |
|
285 static const IndicClassTable devaClassTable = {0x0900, 0x0970, 2, DEVA_SCRIPT_FLAGS, devaCharClasses, NULL}; |
|
286 |
|
287 static const IndicClassTable bengClassTable = {0x0980, 0x09FA, 3, BENG_SCRIPT_FLAGS, bengCharClasses, bengSplitTable}; |
|
288 |
|
289 static const IndicClassTable punjClassTable = {0x0A00, 0x0A74, 2, PUNJ_SCRIPT_FLAGS, punjCharClasses, NULL}; |
|
290 |
|
291 static const IndicClassTable gujrClassTable = {0x0A80, 0x0AEF, 2, GUJR_SCRIPT_FLAGS, gujrCharClasses, NULL}; |
|
292 |
|
293 static const IndicClassTable oryaClassTable = {0x0B00, 0x0B71, 3, ORYA_SCRIPT_FLAGS, oryaCharClasses, oryaSplitTable}; |
|
294 |
|
295 static const IndicClassTable tamlClassTable = {0x0B80, 0x0BF2, 3, TAML_SCRIPT_FLAGS, tamlCharClasses, tamlSplitTable}; |
|
296 |
|
297 static const IndicClassTable teluClassTable = {0x0C00, 0x0C6F, 3, TELU_SCRIPT_FLAGS, teluCharClasses, teluSplitTable}; |
|
298 |
|
299 static const IndicClassTable kndaClassTable = {0x0C80, 0x0CEF, 4, KNDA_SCRIPT_FLAGS, kndaCharClasses, kndaSplitTable}; |
|
300 |
|
301 static const IndicClassTable mlymClassTable = {0x0D00, 0x0D6F, 3, MLYM_SCRIPT_FLAGS, mlymCharClasses, mlymSplitTable}; |
|
302 |
|
303 static const IndicClassTable sinhClassTable = {0x0D80, 0x0DF4, 4, SINH_SCRIPT_FLAGS, sinhCharClasses, sinhSplitTable}; |
|
304 |
|
305 // |
|
306 // IndicClassTable addresses |
|
307 // |
|
308 static const IndicClassTable * const indicClassTables[scriptCodeCount] = { |
|
309 NULL, /* 'zyyy' (COMMON) */ |
|
310 NULL, /* 'qaai' (INHERITED) */ |
|
311 NULL, /* 'arab' (ARABIC) */ |
|
312 NULL, /* 'armn' (ARMENIAN) */ |
|
313 &bengClassTable, /* 'beng' (BENGALI) */ |
|
314 NULL, /* 'bopo' (BOPOMOFO) */ |
|
315 NULL, /* 'cher' (CHEROKEE) */ |
|
316 NULL, /* 'copt' (COPTIC) */ |
|
317 NULL, /* 'cyrl' (CYRILLIC) */ |
|
318 NULL, /* 'dsrt' (DESERET) */ |
|
319 &devaClassTable, /* 'deva' (DEVANAGARI) */ |
|
320 NULL, /* 'ethi' (ETHIOPIC) */ |
|
321 NULL, /* 'geor' (GEORGIAN) */ |
|
322 NULL, /* 'goth' (GOTHIC) */ |
|
323 NULL, /* 'grek' (GREEK) */ |
|
324 &gujrClassTable, /* 'gujr' (GUJARATI) */ |
|
325 &punjClassTable, /* 'guru' (GURMUKHI) */ |
|
326 NULL, /* 'hani' (HAN) */ |
|
327 NULL, /* 'hang' (HANGUL) */ |
|
328 NULL, /* 'hebr' (HEBREW) */ |
|
329 NULL, /* 'hira' (HIRAGANA) */ |
|
330 &kndaClassTable, /* 'knda' (KANNADA) */ |
|
331 NULL, /* 'kata' (KATAKANA) */ |
|
332 NULL, /* 'khmr' (KHMER) */ |
|
333 NULL, /* 'laoo' (LAO) */ |
|
334 NULL, /* 'latn' (LATIN) */ |
|
335 &mlymClassTable, /* 'mlym' (MALAYALAM) */ |
|
336 NULL, /* 'mong' (MONGOLIAN) */ |
|
337 NULL, /* 'mymr' (MYANMAR) */ |
|
338 NULL, /* 'ogam' (OGHAM) */ |
|
339 NULL, /* 'ital' (OLD-ITALIC) */ |
|
340 &oryaClassTable, /* 'orya' (ORIYA) */ |
|
341 NULL, /* 'runr' (RUNIC) */ |
|
342 &sinhClassTable, /* 'sinh' (SINHALA) */ |
|
343 NULL, /* 'syrc' (SYRIAC) */ |
|
344 &tamlClassTable, /* 'taml' (TAMIL) */ |
|
345 &teluClassTable, /* 'telu' (TELUGU) */ |
|
346 NULL, /* 'thaa' (THAANA) */ |
|
347 NULL, /* 'thai' (THAI) */ |
|
348 NULL, /* 'tibt' (TIBETAN) */ |
|
349 NULL, /* 'cans' (CANADIAN-ABORIGINAL) */ |
|
350 NULL, /* 'yiii' (YI) */ |
|
351 NULL, /* 'tglg' (TAGALOG) */ |
|
352 NULL, /* 'hano' (HANUNOO) */ |
|
353 NULL, /* 'buhd' (BUHID) */ |
|
354 NULL, /* 'tagb' (TAGBANWA) */ |
|
355 NULL, /* 'brai' (BRAILLE) */ |
|
356 NULL, /* 'cprt' (CYPRIOT) */ |
|
357 NULL, /* 'limb' (LIMBU) */ |
|
358 NULL, /* 'linb' (LINEAR_B) */ |
|
359 NULL, /* 'osma' (OSMANYA) */ |
|
360 NULL, /* 'shaw' (SHAVIAN) */ |
|
361 NULL, /* 'tale' (TAI_LE) */ |
|
362 NULL, /* 'ugar' (UGARITIC) */ |
|
363 NULL, /* 'hrkt' (KATAKANA_OR_HIRAGANA) */ |
|
364 NULL, /* 'bugi' (BUGINESE) */ |
|
365 NULL, /* 'glag' (GLAGOLITIC) */ |
|
366 NULL, /* 'khar' (KHAROSHTHI) */ |
|
367 NULL, /* 'sylo' (SYLOTI_NAGRI) */ |
|
368 NULL, /* 'talu' (NEW_TAI_LUE) */ |
|
369 NULL, /* 'tfng' (TIFINAGH) */ |
|
370 NULL, /* 'xpeo' (OLD_PERSIAN) */ |
|
371 NULL, /* 'bali' (BALINESE) */ |
|
372 NULL, /* 'batk' (BATK) */ |
|
373 NULL, /* 'blis' (BLIS) */ |
|
374 NULL, /* 'brah' (BRAH) */ |
|
375 NULL, /* 'cham' (CHAM) */ |
|
376 NULL, /* 'cirt' (CIRT) */ |
|
377 NULL, /* 'cyrs' (CYRS) */ |
|
378 NULL, /* 'egyd' (EGYD) */ |
|
379 NULL, /* 'egyh' (EGYH) */ |
|
380 NULL, /* 'egyp' (EGYP) */ |
|
381 NULL, /* 'geok' (GEOK) */ |
|
382 NULL, /* 'hans' (HANS) */ |
|
383 NULL, /* 'hant' (HANT) */ |
|
384 NULL, /* 'hmng' (HMNG) */ |
|
385 NULL, /* 'hung' (HUNG) */ |
|
386 NULL, /* 'inds' (INDS) */ |
|
387 NULL, /* 'java' (JAVA) */ |
|
388 NULL, /* 'kali' (KALI) */ |
|
389 NULL, /* 'latf' (LATF) */ |
|
390 NULL, /* 'latg' (LATG) */ |
|
391 NULL, /* 'lepc' (LEPC) */ |
|
392 NULL, /* 'lina' (LINA) */ |
|
393 NULL, /* 'mand' (MAND) */ |
|
394 NULL, /* 'maya' (MAYA) */ |
|
395 NULL, /* 'mero' (MERO) */ |
|
396 NULL, /* 'nko ' (NKO) */ |
|
397 NULL, /* 'orkh' (ORKH) */ |
|
398 NULL, /* 'perm' (PERM) */ |
|
399 NULL, /* 'phag' (PHAGS_PA) */ |
|
400 NULL, /* 'phnx' (PHOENICIAN) */ |
|
401 NULL, /* 'plrd' (PLRD) */ |
|
402 NULL, /* 'roro' (RORO) */ |
|
403 NULL, /* 'sara' (SARA) */ |
|
404 NULL, /* 'syre' (SYRE) */ |
|
405 NULL, /* 'syrj' (SYRJ) */ |
|
406 NULL, /* 'syrn' (SYRN) */ |
|
407 NULL, /* 'teng' (TENG) */ |
|
408 NULL, /* 'vai ' (VAII) */ |
|
409 NULL, /* 'visp' (VISP) */ |
|
410 NULL, /* 'xsux' (CUNEIFORM) */ |
|
411 NULL, /* 'zxxx' (ZXXX) */ |
|
412 NULL, /* 'zzzz' (UNKNOWN) */ |
|
413 NULL, /* 'cari' (CARI) */ |
|
414 NULL, /* 'jpan' (JPAN) */ |
|
415 NULL, /* 'lana' (LANA) */ |
|
416 NULL, /* 'lyci' (LYCI) */ |
|
417 NULL, /* 'lydi' (LYDI) */ |
|
418 NULL, /* 'olck' (OLCK) */ |
|
419 NULL, /* 'rjng' (RJNG) */ |
|
420 NULL, /* 'saur' (SAUR) */ |
|
421 NULL, /* 'sgnw' (SGNW) */ |
|
422 NULL, /* 'sund' (SUND) */ |
|
423 NULL, /* 'moon' (MOON) */ |
|
424 NULL, /* 'mtei' (MTEI) */ |
|
425 NULL, /* 'armi' (ARMI) */ |
|
426 NULL, /* 'avst' (AVST) */ |
|
427 NULL, /* 'cakm' (CAKM) */ |
|
428 NULL, /* 'kore' (KORE) */ |
|
429 NULL, /* 'kthi' (KTHI) */ |
|
430 NULL, /* 'mani' (MANI) */ |
|
431 NULL, /* 'phli' (PHLI) */ |
|
432 NULL, /* 'phlp' (PHLP) */ |
|
433 NULL, /* 'phlv' (PHLV) */ |
|
434 NULL, /* 'prti' (PRTI) */ |
|
435 NULL, /* 'samr' (SAMR) */ |
|
436 NULL, /* 'tavt' (TAVT) */ |
|
437 NULL, /* 'zmth' (ZMTH) */ |
|
438 NULL, /* 'zsym' (ZSYM) */ |
|
439 NULL, /* 'bamu' (BAMUM) */ |
|
440 NULL, /* 'lisu' (LISU) */ |
|
441 NULL, /* 'nkgb' (NKGB) */ |
|
442 NULL /* 'sarb' (OLD_SOUTH_ARABIAN) */ |
|
443 }; |
|
444 |
|
445 IndicClassTable::CharClass IndicClassTable::getCharClass(LEUnicode ch) const |
|
446 { |
|
447 if (ch == C_SIGN_ZWJ) { |
|
448 return CF_CONSONANT | CC_ZERO_WIDTH_MARK; |
|
449 } |
|
450 |
|
451 if (ch == C_SIGN_ZWNJ) { |
|
452 return CC_ZERO_WIDTH_MARK; |
|
453 } |
|
454 |
|
455 if (ch < firstChar || ch > lastChar) { |
|
456 return CC_RESERVED; |
|
457 } |
|
458 |
|
459 return classTable[ch - firstChar]; |
|
460 } |
|
461 |
|
462 const IndicClassTable *IndicClassTable::getScriptClassTable(le_int32 scriptCode) |
|
463 { |
|
464 if (scriptCode < 0 || scriptCode >= scriptCodeCount) { |
|
465 return NULL; |
|
466 } |
|
467 |
|
468 return indicClassTables[scriptCode]; |
|
469 } |
|
470 |
|
471 le_int32 IndicReordering::getWorstCaseExpansion(le_int32 scriptCode) |
|
472 { |
|
473 const IndicClassTable *classTable = IndicClassTable::getScriptClassTable(scriptCode); |
|
474 |
|
475 if (classTable == NULL) { |
|
476 return 1; |
|
477 } |
|
478 |
|
479 return classTable->getWorstCaseExpansion(); |
|
480 } |
|
481 |
|
482 le_bool IndicReordering::getFilterZeroWidth(le_int32 scriptCode) |
|
483 { |
|
484 const IndicClassTable *classTable = IndicClassTable::getScriptClassTable(scriptCode); |
|
485 |
|
486 if (classTable == NULL) { |
|
487 return TRUE; |
|
488 } |
|
489 |
|
490 return classTable->getFilterZeroWidth(); |
|
491 } |
|
492 |
|
493 U_NAMESPACE_END |
|