1/*
2 * Copyright (C) 2012 Grigori Goronzy <greg@kinoho.net>
3 *
4 * Permission to use, copy, modify, and/or distribute this software for any
5 * purpose with or without fee is hereby granted, provided that the above
6 * copyright notice and this permission notice appear in all copies.
7 *
8 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
15 */
16
17#include "mupdf/fitz.h"
18#include "mupdf/ucdn.h"
19
20#include <stdio.h>
21#include <stdlib.h>
22
23typedef struct {
24 unsigned char category;
25 unsigned char combining;
26 unsigned char bidi_class;
27 unsigned char east_asian_width;
28 unsigned char script;
29 unsigned char linebreak_class;
30} UCDRecord;
31
32typedef struct {
33 unsigned short from, to;
34} MirrorPair;
35
36typedef struct {
37 unsigned short from, to;
38 unsigned char type;
39} BracketPair;
40
41typedef struct {
42 unsigned int start;
43 short count, index;
44} Reindex;
45
46#include "ucdn_db.h"
47
48/* constants required for Hangul (de)composition */
49#define SBASE 0xAC00
50#define LBASE 0x1100
51#define VBASE 0x1161
52#define TBASE 0x11A7
53#define SCOUNT 11172
54#define LCOUNT 19
55#define VCOUNT 21
56#define TCOUNT 28
57#define NCOUNT (VCOUNT * TCOUNT)
58
59static const UCDRecord *get_ucd_record(uint32_t code)
60{
61 int index, offset;
62
63 if (code >= 0x110000)
64 index = 0;
65 else {
66 index = index0[code >> (SHIFT1+SHIFT2)] << SHIFT1;
67 offset = (code >> SHIFT2) & ((1<<SHIFT1) - 1);
68 index = index1[index + offset] << SHIFT2;
69 offset = code & ((1<<SHIFT2) - 1);
70 index = index2[index + offset];
71 }
72
73 return &ucd_records[index];
74}
75
76static const unsigned short *get_decomp_record(uint32_t code)
77{
78 int index, offset;
79
80 if (code >= 0x110000)
81 index = 0;
82 else {
83 index = decomp_index0[code >> (DECOMP_SHIFT1+DECOMP_SHIFT2)]
84 << DECOMP_SHIFT1;
85 offset = (code >> DECOMP_SHIFT2) & ((1<<DECOMP_SHIFT1) - 1);
86 index = decomp_index1[index + offset] << DECOMP_SHIFT2;
87 offset = code & ((1<<DECOMP_SHIFT2) - 1);
88 index = decomp_index2[index + offset];
89 }
90
91 return &decomp_data[index];
92}
93
94static int compare_reindex(const void *a, const void *b)
95{
96 Reindex *ra = (Reindex *)a;
97 Reindex *rb = (Reindex *)b;
98
99 if (ra->start < rb->start)
100 return -1;
101 else if (ra->start > (rb->start + rb->count))
102 return 1;
103 else
104 return 0;
105}
106
107static int get_comp_index(uint32_t code, const Reindex *idx, size_t len)
108{
109 Reindex *res;
110 Reindex r = {0, 0, 0};
111 r.start = code;
112 res = (Reindex *) bsearch(&r, idx, len, sizeof(Reindex), compare_reindex);
113
114 if (res != NULL)
115 return res->index + (code - res->start);
116 else
117 return -1;
118}
119
120static int compare_mp(const void *a, const void *b)
121{
122 MirrorPair *mpa = (MirrorPair *)a;
123 MirrorPair *mpb = (MirrorPair *)b;
124 return mpa->from - mpb->from;
125}
126
127static int compare_bp(const void *a, const void *b)
128{
129 BracketPair *bpa = (BracketPair *)a;
130 BracketPair *bpb = (BracketPair *)b;
131 return bpa->from - bpb->from;
132}
133
134static BracketPair *search_bp(uint32_t code)
135{
136 BracketPair bp = {0,0,2};
137 BracketPair *res;
138
139 bp.from = code;
140 res = (BracketPair *) bsearch(&bp, bracket_pairs, BIDI_BRACKET_LEN,
141 sizeof(BracketPair), compare_bp);
142 return res;
143}
144
145static int hangul_pair_decompose(uint32_t code, uint32_t *a, uint32_t *b)
146{
147 int si = code - SBASE;
148
149 if (si < 0 || si >= SCOUNT)
150 return 0;
151
152 if (si % TCOUNT) {
153 /* LV,T */
154 *a = SBASE + (si / TCOUNT) * TCOUNT;
155 *b = TBASE + (si % TCOUNT);
156 return 3;
157 } else {
158 /* L,V */
159 *a = LBASE + (si / NCOUNT);
160 *b = VBASE + (si % NCOUNT) / TCOUNT;
161 return 2;
162 }
163}
164
165static int hangul_pair_compose(uint32_t *code, uint32_t a, uint32_t b)
166{
167 if (a >= SBASE && a < (SBASE + SCOUNT) && b >= TBASE && b < (TBASE + TCOUNT)) {
168 /* LV,T */
169 *code = a + (b - TBASE);
170 return 3;
171 } else if (a >= LBASE && a < (LBASE + LCOUNT) && b >= VBASE && b < (VBASE + VCOUNT)) {
172 /* L,V */
173 int li = a - LBASE;
174 int vi = b - VBASE;
175 *code = SBASE + li * NCOUNT + vi * TCOUNT;
176 return 2;
177 } else {
178 return 0;
179 }
180}
181
182static uint32_t decode_utf16(const unsigned short **code_ptr)
183{
184 const unsigned short *code = *code_ptr;
185
186 if (code[0] < 0xd800 || code[0] > 0xdc00) {
187 *code_ptr += 1;
188 return (uint32_t)code[0];
189 } else {
190 *code_ptr += 2;
191 return 0x10000 + ((uint32_t)code[1] - 0xdc00) +
192 (((uint32_t)code[0] - 0xd800) << 10);
193 }
194}
195
196const char *ucdn_get_unicode_version(void)
197{
198 return UNIDATA_VERSION;
199}
200
201int ucdn_get_combining_class(uint32_t code)
202{
203 return get_ucd_record(code)->combining;
204}
205
206int ucdn_get_east_asian_width(uint32_t code)
207{
208 return get_ucd_record(code)->east_asian_width;
209}
210
211int ucdn_get_general_category(uint32_t code)
212{
213 return get_ucd_record(code)->category;
214}
215
216int ucdn_get_bidi_class(uint32_t code)
217{
218 return get_ucd_record(code)->bidi_class;
219}
220
221int ucdn_get_mirrored(uint32_t code)
222{
223 return ucdn_mirror(code) != code;
224}
225
226int ucdn_get_script(uint32_t code)
227{
228 return get_ucd_record(code)->script;
229}
230
231int ucdn_get_linebreak_class(uint32_t code)
232{
233 return get_ucd_record(code)->linebreak_class;
234}
235
236int ucdn_get_resolved_linebreak_class(uint32_t code)
237{
238 const UCDRecord *record = get_ucd_record(code);
239
240 switch (record->linebreak_class)
241 {
242 case UCDN_LINEBREAK_CLASS_AI:
243 case UCDN_LINEBREAK_CLASS_SG:
244 case UCDN_LINEBREAK_CLASS_XX:
245 return UCDN_LINEBREAK_CLASS_AL;
246
247 case UCDN_LINEBREAK_CLASS_SA:
248 if (record->category == UCDN_GENERAL_CATEGORY_MC ||
249 record->category == UCDN_GENERAL_CATEGORY_MN)
250 return UCDN_LINEBREAK_CLASS_CM;
251 return UCDN_LINEBREAK_CLASS_AL;
252
253 case UCDN_LINEBREAK_CLASS_CJ:
254 return UCDN_LINEBREAK_CLASS_NS;
255
256 case UCDN_LINEBREAK_CLASS_CB:
257 return UCDN_LINEBREAK_CLASS_B2;
258
259 case UCDN_LINEBREAK_CLASS_NL:
260 return UCDN_LINEBREAK_CLASS_BK;
261
262 default:
263 return record->linebreak_class;
264 }
265}
266
267uint32_t ucdn_mirror(uint32_t code)
268{
269 MirrorPair mp = {0};
270 MirrorPair *res;
271
272 mp.from = code;
273 res = (MirrorPair *) bsearch(&mp, mirror_pairs, BIDI_MIRROR_LEN,
274 sizeof(MirrorPair), compare_mp);
275
276 if (res == NULL)
277 return code;
278 else
279 return res->to;
280}
281
282uint32_t ucdn_paired_bracket(uint32_t code)
283{
284 BracketPair *res = search_bp(code);
285 if (res == NULL)
286 return code;
287 else
288 return res->to;
289}
290
291int ucdn_paired_bracket_type(uint32_t code)
292{
293 BracketPair *res = search_bp(code);
294 if (res == NULL)
295 return UCDN_BIDI_PAIRED_BRACKET_TYPE_NONE;
296 else
297 return res->type;
298}
299
300int ucdn_decompose(uint32_t code, uint32_t *a, uint32_t *b)
301{
302 const unsigned short *rec;
303 int len;
304
305 if (hangul_pair_decompose(code, a, b))
306 return 1;
307
308 rec = get_decomp_record(code);
309 len = rec[0] >> 8;
310
311 if ((rec[0] & 0xff) != 0 || len == 0)
312 return 0;
313
314 rec++;
315 *a = decode_utf16(&rec);
316 if (len > 1)
317 *b = decode_utf16(&rec);
318 else
319 *b = 0;
320
321 return 1;
322}
323
324int ucdn_compose(uint32_t *code, uint32_t a, uint32_t b)
325{
326 int l, r, index, indexi, offset;
327
328 if (hangul_pair_compose(code, a, b))
329 return 1;
330
331 l = get_comp_index(a, nfc_first, sizeof(nfc_first) / sizeof(Reindex));
332 r = get_comp_index(b, nfc_last, sizeof(nfc_last) / sizeof(Reindex));
333
334 if (l < 0 || r < 0)
335 return 0;
336
337 indexi = l * TOTAL_LAST + r;
338 index = comp_index0[indexi >> (COMP_SHIFT1+COMP_SHIFT2)] << COMP_SHIFT1;
339 offset = (indexi >> COMP_SHIFT2) & ((1<<COMP_SHIFT1) - 1);
340 index = comp_index1[index + offset] << COMP_SHIFT2;
341 offset = indexi & ((1<<COMP_SHIFT2) - 1);
342 *code = comp_data[index + offset];
343
344 return *code != 0;
345}
346
347int ucdn_compat_decompose(uint32_t code, uint32_t *decomposed)
348{
349 int i, len;
350 const unsigned short *rec = get_decomp_record(code);
351 len = rec[0] >> 8;
352
353 if (len == 0)
354 return 0;
355
356 rec++;
357 for (i = 0; i < len; i++)
358 decomposed[i] = decode_utf16(&rec);
359
360 return len;
361}
362