]> git.saurik.com Git - apple/icu.git/blame - icuSources/common/ushape.c
projects / apple / icu.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
ICU-8.11.tar.gz
[apple/icu.git] / icuSources / common / ushape.c
CommitLineData
b75a7d8f
A		 1/*
2******************************************************************************
3*
73c04bcf 4* Copyright (C) 2000-2006, International Business Machines
b75a7d8f
A		 5* Corporation and others. All Rights Reserved.
6*
7******************************************************************************
8* file name: ushape.c
9* encoding: US-ASCII
10* tab size: 8 (not used)
11* indentation:4
12*
13* created on: 2000jun29
14* created by: Markus W. Scherer
15*
16* Arabic letter shaping implemented by Ayman Roshdy
17*/
18
19#include "unicode/utypes.h"
20#include "unicode/uchar.h"
21#include "unicode/ustring.h"
b75a7d8f 22#include "unicode/ushape.h"
374ca955
A		 23#include "cmemory.h"
24#include "putilimp.h"
b75a7d8f 25#include "ustr_imp.h"
73c04bcf 26#include "ubidi_props.h"
b75a7d8f
A		 27
28#if UTF_SIZE<16
29 /*
30 * This implementation assumes that the internal encoding is UTF-16
31 * or UTF-32, not UTF-8.
32 * The main assumption is that the Arabic characters and their
33 * presentation forms each fit into a single UChar.
34 * With UTF-8, they occupy 2 or 3 bytes, and more than the ASCII
35 * characters.
36 */
37# error This implementation assumes UTF-16 or UTF-32 (check UTF_SIZE)
38#endif
39
40/*
41 * ### TODO in general for letter shaping:
42 * - the letter shaping code is UTF-16-unaware; needs update
43 * + especially invertBuffer()?!
44 * - needs to handle the "Arabic Tail" that is used in some legacy codepages
45 * as a glyph fragment of wide-glyph letters
46 * + IBM Unicode conversion tables map it to U+200B (ZWSP)
47 * + IBM Egypt has proposed to encode the tail in Unicode among Arabic Presentation Forms
48 */
49
50/* definitions for Arabic letter shaping ------------------------------------ */
51
52#define IRRELEVANT 4
53#define LAMTYPE 16
54#define ALEFTYPE 32
55#define LINKR 1
56#define LINKL 2
57
58static const UChar IrrelevantPos[] = {
59 0x0, 0x2, 0x4, 0x6,
60 0x8, 0xA, 0xC, 0xE,
61};
62
63static const UChar convertLamAlef[] =
64{
65/*FEF5*/ 0x0622,
66/*FEF6*/ 0x0622,
67/*FEF7*/ 0x0623,
68/*FEF8*/ 0x0623,
69/*FEF9*/ 0x0625,
70/*FEFA*/ 0x0625,
71/*FEFB*/ 0x0627,
72/*FEFC*/ 0x0627
73};
74
75static const UChar araLink[178]=
76{
77 1 + 32 + 256 * 0x11,/*0x0622*/
78 1 + 32 + 256 * 0x13,/*0x0623*/
79 1 + 256 * 0x15,/*0x0624*/
80 1 + 32 + 256 * 0x17,/*0x0625*/
81 1 + 2 + 256 * 0x19,/*0x0626*/
82 1 + 32 + 256 * 0x1D,/*0x0627*/
83 1 + 2 + 256 * 0x1F,/*0x0628*/
84 1 + 256 * 0x23,/*0x0629*/
85 1 + 2 + 256 * 0x25,/*0x062A*/
86 1 + 2 + 256 * 0x29,/*0x062B*/
87 1 + 2 + 256 * 0x2D,/*0x062C*/
88 1 + 2 + 256 * 0x31,/*0x062D*/
89 1 + 2 + 256 * 0x35,/*0x062E*/
90 1 + 256 * 0x39,/*0x062F*/
91 1 + 256 * 0x3B,/*0x0630*/
92 1 + 256 * 0x3D,/*0x0631*/
93 1 + 256 * 0x3F,/*0x0632*/
94 1 + 2 + 256 * 0x41,/*0x0633*/
95 1 + 2 + 256 * 0x45,/*0x0634*/
96 1 + 2 + 256 * 0x49,/*0x0635*/
97 1 + 2 + 256 * 0x4D,/*0x0636*/
98 1 + 2 + 256 * 0x51,/*0x0637*/
99 1 + 2 + 256 * 0x55,/*0x0638*/
100 1 + 2 + 256 * 0x59,/*0x0639*/
101 1 + 2 + 256 * 0x5D,/*0x063A*/
102 0, 0, 0, 0, 0, /*0x063B-0x063F*/
103 1 + 2, /*0x0640*/
104 1 + 2 + 256 * 0x61,/*0x0641*/
105 1 + 2 + 256 * 0x65,/*0x0642*/
106 1 + 2 + 256 * 0x69,/*0x0643*/
107 1 + 2 + 16 + 256 * 0x6D,/*0x0644*/
108 1 + 2 + 256 * 0x71,/*0x0645*/
109 1 + 2 + 256 * 0x75,/*0x0646*/
110 1 + 2 + 256 * 0x79,/*0x0647*/
111 1 + 256 * 0x7D,/*0x0648*/
112 1 + 256 * 0x7F,/*0x0649*/
113 1 + 2 + 256 * 0x81,/*0x064A*/
114 4, 4, 4, 4, /*0x064B-0x064E*/
115 4, 4, 4, 4, /*0x064F-0x0652*/
116 4, 4, 4, 0, 0, /*0x0653-0x0657*/
117 0, 0, 0, 0, /*0x0658-0x065B*/
118 1 + 256 * 0x85,/*0x065C*/
119 1 + 256 * 0x87,/*0x065D*/
120 1 + 256 * 0x89,/*0x065E*/
121 1 + 256 * 0x8B,/*0x065F*/
122 0, 0, 0, 0, 0, /*0x0660-0x0664*/
123 0, 0, 0, 0, 0, /*0x0665-0x0669*/
124 0, 0, 0, 0, 0, 0, /*0x066A-0x066F*/
125 4, /*0x0670*/
126 0, /*0x0671*/
127 1 + 32, /*0x0672*/
128 1 + 32, /*0x0673*/
129 0, /*0x0674*/
130 1 + 32, /*0x0675*/
131 1, 1, /*0x0676-0x0677*/
132 1+2, 1+2, 1+2, 1+2, 1+2, 1+2, /*0x0678-0x067D*/
133 1+2, 1+2, 1+2, 1+2, 1+2, 1+2, /*0x067E-0x0683*/
134 1+2, 1+2, 1+2, 1+2, /*0x0684-0x0687*/
135 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /*0x0688-0x0691*/
136 1, 1, 1, 1, 1, 1, 1, 1, /*0x0692-0x0699*/
137 1+2, 1+2, 1+2, 1+2, 1+2, 1+2, /*0x069A-0x06A3*/
138 1+2, 1+2, 1+2, 1+2, /*0x069A-0x06A3*/
139 1+2, 1+2, 1+2, 1+2, 1+2, 1+2, /*0x06A4-0x06AD*/
140 1+2, 1+2, 1+2, 1+2, /*0x06A4-0x06AD*/
141 1+2, 1+2, 1+2, 1+2, 1+2, 1+2, /*0x06AE-0x06B7*/
142 1+2, 1+2, 1+2, 1+2, /*0x06AE-0x06B7*/
143 1+2, 1+2, 1+2, 1+2, 1+2, 1+2, /*0x06B8-0x06BF*/
144 1+2, 1+2, /*0x06B8-0x06BF*/
145 1, /*0x06C0*/
146 1+2, /*0x06C1*/
147 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /*0x06C2-0x06CB*/
148 1+2, /*0x06CC*/
149 1, /*0x06CD*/
150 1+2, 1+2, 1+2, 1+2, /*0x06CE-0x06D1*/
151 1, 1 /*0x06D2-0x06D3*/
152};
153
154static const UChar presLink[141]=
155{
156 1 + 2, /*0xFE70*/
157 1 + 2, /*0xFE71*/
158 1 + 2, 0, 1+ 2, 0, 1+ 2, /*0xFE72-0xFE76*/
159 1 + 2, /*0xFE77*/
160 1+ 2, 1 + 2, 1+2, 1 + 2, /*0xFE78-0xFE81*/
161 1+ 2, 1 + 2, 1+2, 1 + 2, /*0xFE82-0xFE85*/
162 0, 0 + 32, 1 + 32, 0 + 32, /*0xFE86-0xFE89*/
163 1 + 32, 0, 1, 0 + 32, /*0xFE8A-0xFE8D*/
164 1 + 32, 0, 2, 1 + 2, /*0xFE8E-0xFE91*/
165 1, 0 + 32, 1 + 32, 0, /*0xFE92-0xFE95*/
166 2, 1 + 2, 1, 0, /*0xFE96-0xFE99*/
167 1, 0, 2, 1 + 2, /*0xFE9A-0xFE9D*/
168 1, 0, 2, 1 + 2, /*0xFE9E-0xFEA1*/
169 1, 0, 2, 1 + 2, /*0xFEA2-0xFEA5*/
170 1, 0, 2, 1 + 2, /*0xFEA6-0xFEA9*/
171 1, 0, 2, 1 + 2, /*0xFEAA-0xFEAD*/
172 1, 0, 1, 0, /*0xFEAE-0xFEB1*/
173 1, 0, 1, 0, /*0xFEB2-0xFEB5*/
174 1, 0, 2, 1+2, /*0xFEB6-0xFEB9*/
175 1, 0, 2, 1+2, /*0xFEBA-0xFEBD*/
176 1, 0, 2, 1+2, /*0xFEBE-0xFEC1*/
177 1, 0, 2, 1+2, /*0xFEC2-0xFEC5*/
178 1, 0, 2, 1+2, /*0xFEC6-0xFEC9*/
179 1, 0, 2, 1+2, /*0xFECA-0xFECD*/
180 1, 0, 2, 1+2, /*0xFECE-0xFED1*/
181 1, 0, 2, 1+2, /*0xFED2-0xFED5*/
182 1, 0, 2, 1+2, /*0xFED6-0xFED9*/
183 1, 0, 2, 1+2, /*0xFEDA-0xFEDD*/
184 1, 0, 2, 1+2, /*0xFEDE-0xFEE1*/
185 1, 0 + 16, 2 + 16, 1 + 2 +16, /*0xFEE2-0xFEE5*/
186 1 + 16, 0, 2, 1+2, /*0xFEE6-0xFEE9*/
187 1, 0, 2, 1+2, /*0xFEEA-0xFEED*/
188 1, 0, 2, 1+2, /*0xFEEE-0xFEF1*/
189 1, 0, 1, 0, /*0xFEF2-0xFEF5*/
190 1, 0, 2, 1+2, /*0xFEF6-0xFEF9*/
191 1, 0, 1, 0, /*0xFEFA-0xFEFD*/
192 1, 0, 1, 0,
193 1
194};
195
196static const UChar convertFEto06[] =
197{
198/***********0******1******2******3******4******5******6******7******8******9******A******B******C******D******E******F***/
199/*FE7*/ 0x64B, 0x64B, 0x64C, 0x64C, 0x64D, 0x64D, 0x64E, 0x64E, 0x64F, 0x64F, 0x650, 0x650, 0x651, 0x651, 0x652, 0x652,
200/*FE8*/ 0x621, 0x622, 0x622, 0x623, 0x623, 0x624, 0x624, 0x625, 0x625, 0x626, 0x626, 0x626, 0x626, 0x627, 0x627, 0x628,
201/*FE9*/ 0x628, 0x628, 0x628, 0x629, 0x629, 0x62A, 0x62A, 0x62A, 0x62A, 0x62B, 0x62B, 0x62B, 0x62B, 0x62C, 0x62C, 0x62C,
202/*FEA*/ 0x62C, 0x62D, 0x62D, 0x62D, 0x62D, 0x62E, 0x62E, 0x62E, 0x62E, 0x62F, 0x62F, 0x630, 0x630, 0x631, 0x631, 0x632,
203/*FEB*/ 0x632, 0x633, 0x633, 0x633, 0x633, 0x634, 0x634, 0x634, 0x634, 0x635, 0x635, 0x635, 0x635, 0x636, 0x636, 0x636,
204/*FEC*/ 0x636, 0x637, 0x637, 0x637, 0x637, 0x638, 0x638, 0x638, 0x638, 0x639, 0x639, 0x639, 0x639, 0x63A, 0x63A, 0x63A,
205/*FED*/ 0x63A, 0x641, 0x641, 0x641, 0x641, 0x642, 0x642, 0x642, 0x642, 0x643, 0x643, 0x643, 0x643, 0x644, 0x644, 0x644,
206/*FEE*/ 0x644, 0x645, 0x645, 0x645, 0x645, 0x646, 0x646, 0x646, 0x646, 0x647, 0x647, 0x647, 0x647, 0x648, 0x648, 0x649,
207/*FEF*/ 0x649, 0x64A, 0x64A, 0x64A, 0x64A, 0x65C, 0x65C, 0x65D, 0x65D, 0x65E, 0x65E, 0x65F, 0x65F
208};
209
210static const UChar shapeTable[4][4][4]=
211{
212 { {0,0,0,0}, {0,0,0,0}, {0,1,0,3}, {0,1,0,1} },
213 { {0,0,2,2}, {0,0,1,2}, {0,1,1,2}, {0,1,1,3} },
214 { {0,0,0,0}, {0,0,0,0}, {0,1,0,3}, {0,1,0,3} },
215 { {0,0,1,2}, {0,0,1,2}, {0,1,1,2}, {0,1,1,3} }
216};
217
218/*
219 * This function shapes European digits to Arabic-Indic digits
220 * in-place, writing over the input characters.
221 * Since we know that we are only looking for BMP code points,
222 * we can safely just work with code units (again, at least UTF-16).
223 */
224static void
225_shapeToArabicDigitsWithContext(UChar *s, int32_t length,
226 UChar digitBase,
227 UBool isLogical, UBool lastStrongWasAL) {
73c04bcf
A		 228 const UBiDiProps *bdp;
229 UErrorCode errorCode;
230
b75a7d8f
A		 231 int32_t i;
232 UChar c;
233
73c04bcf
A		 234 errorCode=U_ZERO_ERROR;
235 bdp=ubidi_getSingleton(&errorCode);
236 if(U_FAILURE(errorCode)) {
237 return;
238 }
239
b75a7d8f
A		 240 digitBase-=0x30;
241
242 /* the iteration direction depends on the type of input */
243 if(isLogical) {
244 for(i=0; i<length; ++i) {
245 c=s[i];
73c04bcf 246 switch(ubidi_getClass(bdp, c)) {
b75a7d8f
A		 247 case U_LEFT_TO_RIGHT: /* L */
248 case U_RIGHT_TO_LEFT: /* R */
249 lastStrongWasAL=FALSE;
250 break;
251 case U_RIGHT_TO_LEFT_ARABIC: /* AL */
252 lastStrongWasAL=TRUE;
253 break;
254 case U_EUROPEAN_NUMBER: /* EN */
255 if(lastStrongWasAL && (uint32_t)(c-0x30)<10) {
256 s[i]=(UChar)(digitBase+c); /* digitBase+(c-0x30) - digitBase was modified above */
257 }
258 break;
259 default :
260 break;
261 }
262 }
263 } else {
264 for(i=length; i>0; /* pre-decrement in the body */) {
265 c=s[--i];
73c04bcf 266 switch(ubidi_getClass(bdp, c)) {
b75a7d8f
A		 267 case U_LEFT_TO_RIGHT: /* L */
268 case U_RIGHT_TO_LEFT: /* R */
269 lastStrongWasAL=FALSE;
270 break;
271 case U_RIGHT_TO_LEFT_ARABIC: /* AL */
272 lastStrongWasAL=TRUE;
273 break;
274 case U_EUROPEAN_NUMBER: /* EN */
275 if(lastStrongWasAL && (uint32_t)(c-0x30)<10) {
276 s[i]=(UChar)(digitBase+c); /* digitBase+(c-0x30) - digitBase was modified above */
277 }
278 break;
279 default :
280 break;
281 }
282 }
283 }
284}
285
286/*
287 *Name : invertBuffer
288 *Function : This function inverts the buffer, it's used
289 * in case the user specifies the buffer to be
290 * U_SHAPE_TEXT_DIRECTION_LOGICAL
291 */
292static void
293invertBuffer(UChar *buffer,int32_t size,uint32_t options,int32_t *spacesCountl,int32_t *spacesCountr) {
294
295 UChar temp;
296 int32_t i=0,j=0;
297 int32_t lowlimit = 0, highlimit = 0;
298
299 lowlimit = *spacesCountl;
300 highlimit = *spacesCountr;
301
302 for(i=lowlimit,j=size-highlimit-1;i<j;i++,j--) {
303 temp = buffer[i];
304 buffer[i] = buffer[j];
305 buffer[j] = temp;
306 }
307}
308
309/*
310 *Name : changeLamAlef
311 *Function : Converts the Alef characters into an equivalent
312 * LamAlef location in the 0x06xx Range, this is an
313 * intermediate stage in the operation of the program
314 * later it'll be converted into the 0xFExx LamAlefs
315 * in the shaping function.
316 */
73c04bcf 317static U_INLINE UChar
b75a7d8f
A		 318changeLamAlef(UChar ch) {
319
320 switch(ch) {
321 case 0x0622 :
73c04bcf 322 return 0x065C;
b75a7d8f 323 case 0x0623 :
73c04bcf 324 return 0x065D;
b75a7d8f 325 case 0x0625 :
73c04bcf 326 return 0x065E;
b75a7d8f 327 case 0x0627 :
73c04bcf 328 return 0x065F;
b75a7d8f 329 }
73c04bcf 330 return 0;
b75a7d8f
A		 331}
332
333/*
334 *Name : specialChar
335 *Function : Special Arabic characters need special handling in the shapeUnicode
336 * function, this function returns 1 or 2 for these special characters
337 */
338static int32_t
339specialChar(UChar ch) {
340
341 if( (ch>0x0621 && ch<0x0626)||(ch==0x0627)||(ch>0x062e && ch<0x0633)||
342 (ch>0x0647 && ch<0x064a)||(ch==0x0629) ) {
343 return (1);
344 }
73c04bcf 345 else if( ch>=0x064B && ch<= 0x0652 )
b75a7d8f 346 return (2);
73c04bcf 347 else if( (ch>=0x0653 && ch<= 0x0655) || ch == 0x0670 ||
b75a7d8f
A		 348 (ch>=0xFE70 && ch<= 0xFE7F) )
349 return (3);
350 else
351 return (0);
352}
353
354/*
355 *Name : getLink
356 *Function : Resolves the link between the characters as
357 * Arabic characters have four forms :
358 * Isolated, Initial, Middle and Final Form
359 */
360static UChar
361getLink(UChar ch) {
362
363 if(ch >= 0x0622 && ch <= 0x06D3) {
364 return(araLink[ch-0x0622]);
365 } else if(ch == 0x200D) {
366 return(3);
367 } else if(ch >= 0x206D && ch <= 0x206F) {
368 return(4);
369 } else if(ch >= 0xFE70 && ch <= 0xFEFC) {
370 return(presLink[ch-0xFE70]);
371 } else {
372 return(0);
373 }
374}
375
376/*
377 *Name : countSpaces
378 *Function : Counts the number of spaces
379 * at each end of the logical buffer
380 */
381static void
382countSpaces(UChar *dest,int32_t size,uint32_t options,int32_t *spacesCountl,int32_t *spacesCountr) {
383
384 int32_t i = 0;
385 int32_t countl = 0,countr = 0;
386
387 while(dest[i] == 0x0020) {
388 countl++;
389 i++;
390 }
391 while(dest[size-1] == 0x0020) {
392 countr++;
393 size--;
394 }
395 *spacesCountl = countl;
396 *spacesCountr = countr;
397}
398
399/*
400 *Name : isTashkeelChar
401 *Function : Returns 1 for Tashkeel characters else return 0
402 */
73c04bcf 403static U_INLINE int32_t
b75a7d8f 404isTashkeelChar(UChar ch) {
73c04bcf 405 return (int32_t)( ch>=0x064B && ch<= 0x0652 );
b75a7d8f
A		 406}
407
408/*
409 *Name : isAlefChar
410 *Function : Returns 1 for Alef characters else return 0
411 */
73c04bcf 412static U_INLINE int32_t
b75a7d8f 413isAlefChar(UChar ch) {
73c04bcf 414 return (int32_t)( (ch==0x0622)||(ch==0x0623)||(ch==0x0625)||(ch==0x0627) );
b75a7d8f
A		 415}
416
417/*
418 *Name : isLamAlefChar
419 *Function : Returns 1 for LamAlef characters else return 0
420 */
73c04bcf 421static U_INLINE int32_t
b75a7d8f 422isLamAlefChar(UChar ch) {
73c04bcf 423 return (int32_t)( (ch>=0xFEF5)&&(ch<=0xFEFC) );
b75a7d8f
A		 424}
425
426/*
427 *Name : calculateSize
428 *Function : This function calculates the destSize to be used in preflighting
429 * when the destSize is equal to 0
430 */
431static int32_t
432calculateSize(const UChar *source, int32_t sourceLength,
433 int32_t destSize,uint32_t options) {
434
435 int32_t i = 0;
436 destSize = sourceLength;
437
438 switch(options&U_SHAPE_LETTERS_MASK) {
439
440 case U_SHAPE_LETTERS_SHAPE :
441 if((options&U_SHAPE_TEXT_DIRECTION_MASK)==U_SHAPE_TEXT_DIRECTION_VISUAL_LTR) {
442 for(i=0;i<sourceLength;i++) {
443 if( (isAlefChar(source[i]))&&(source[i+1]==0x0644) ) {
444 destSize--;
445 }
446 }
447 }
448 else
449 if((options&U_SHAPE_TEXT_DIRECTION_MASK)==U_SHAPE_TEXT_DIRECTION_LOGICAL) {
450 for(i=0;i<sourceLength;i++) {
451 if( (isAlefChar(source[i+1]))&&(source[i]==0x0644) ) {
452 destSize--;
453 }
454 }
455 }
456 break;
457
458 case U_SHAPE_LETTERS_UNSHAPE :
459 for(i=0;i<sourceLength;i++) {
460 if( isLamAlefChar(source[i]) ) {
461 destSize++;
462 }
463 }
464 break;
465
466 default :
467 /* will never occur because of validity checks at the begin of u_shapeArabic */
468 break;
469 }
470
471 return destSize;
472}
473
474/*
475 *Name : removeLamAlefSpaces
476 *Function : The shapeUnicode function converts Lam + Alef into LamAlef + space,
477 * this function removes the spaces behind the LamAlefs according to
478 * the options the user specifies, the spaces are removed to the end
479 * of the buffer, or shrink the buffer ab=nd remove spaces for good
480 * or leave the buffer as it is LamAlef + space.
481 */
482static int32_t
483removeLamAlefSpaces(UChar *dest, int32_t sourceLength,
484 int32_t destSize,
485 uint32_t options,
486 UErrorCode *pErrorCode) {
487
488 int32_t i = 0, j = 0;
489 int32_t count = 0;
490 UChar *tempbuffer=NULL;
491
492 switch(options&U_SHAPE_LENGTH_MASK) {
493 case U_SHAPE_LENGTH_GROW_SHRINK :
494 tempbuffer = (UChar *)uprv_malloc((sourceLength+1)*U_SIZEOF_UCHAR);
495 /* Test for NULL */
496 if(tempbuffer == NULL) {
497 *pErrorCode = U_MEMORY_ALLOCATION_ERROR;
498 return 0;
499 }
500
501 uprv_memset(tempbuffer, 0, (sourceLength+1)*U_SIZEOF_UCHAR);
502
503 i = j = 0;
504 while(i < sourceLength) {
505 if(dest[i] == 0xFFFF) {
506 j--;
507 count++;
508 }
509 else
510 tempbuffer[j] = dest[i];
511 i++;
512 j++;
513 }
514
515 while(count >= 0) {
516 tempbuffer[i] = 0x0000;
517 i--;
518 count--;
519 }
520
521 uprv_memcpy(dest, tempbuffer, sourceLength*U_SIZEOF_UCHAR);
522 destSize = u_strlen(dest);
523 break;
524
525 case U_SHAPE_LENGTH_FIXED_SPACES_NEAR :
526 /* Lam+Alef is already shaped into LamAlef + FFFF */
527 i = 0;
528 while(i < sourceLength) {
529 if(dest[i] == 0xFFFF)
530 dest[i] = 0x0020;
531 i++;
532 }
533 destSize = sourceLength;
534 break;
535
536 case U_SHAPE_LENGTH_FIXED_SPACES_AT_BEGINNING :
537 tempbuffer = (UChar *)uprv_malloc((sourceLength+1)*U_SIZEOF_UCHAR);
538
539 /* Test for NULL */
540 if(tempbuffer == NULL) {
541 *pErrorCode = U_MEMORY_ALLOCATION_ERROR;
542 return 0;
543 }
544
545 uprv_memset(tempbuffer, 0, (sourceLength+1)*U_SIZEOF_UCHAR);
546
547 i = j = sourceLength;
548 while(i >= 0) {
549 if(dest[i] == 0xFFFF) {
550 j++;
551 count++;
552 }
553 else
554 tempbuffer[j] = dest[i];
555 i--;
556 j--;
557 }
558 for(i=0;i<count;i++)
559 tempbuffer[i] = 0x0020;
560
561 uprv_memcpy(dest, tempbuffer, sourceLength*U_SIZEOF_UCHAR);
562 destSize = sourceLength;
563 break;
564
565 case U_SHAPE_LENGTH_FIXED_SPACES_AT_END :
566 tempbuffer = (UChar *)uprv_malloc((sourceLength+1)*U_SIZEOF_UCHAR);
567
568 /* Test for NULL */
569 if(tempbuffer == NULL) {
570 *pErrorCode = U_MEMORY_ALLOCATION_ERROR;
571 return 0;
572 }
573
574 uprv_memset(tempbuffer, 0, (sourceLength+1)*U_SIZEOF_UCHAR);
575
576 i = j = 0;
577 while(i < sourceLength) {
578 if(dest[i] == 0xFFFF) {
579 j--;
580 count++;
581 }
582 else
583 tempbuffer[j] = dest[i];
584 i++;
585 j++;
586 }
587
588 while(count >= 0) {
589 tempbuffer[i] = 0x0020;
590 i--;
591 count--;
592 }
593
594 uprv_memcpy(dest,tempbuffer, sourceLength*U_SIZEOF_UCHAR);
595 destSize = sourceLength;
596 break;
597
598 default :
599 /* will not occur */
600 break;
601 }
602
603 if(tempbuffer)
604 uprv_free(tempbuffer);
605
606 return destSize;
607}
608
609/*
610 *Name : expandLamAlef
611 *Function : LamAlef needs special handling as the LamAlef is
612 * one character while expanding it will give two
613 * characters Lam + Alef, so we need to expand the LamAlef
614 * in near or far spaces according to the options the user
615 * specifies or increase the buffer size.
616 * If there are no spaces to expand the LamAlef, an error
617 * will be set to U_NO_SPACE_AVAILABLE as defined in utypes.h
618 */
619static int32_t
620expandLamAlef(UChar *dest, int32_t sourceLength,
621 int32_t destSize,uint32_t options,
622 UErrorCode *pErrorCode) {
623
624 int32_t i = 0,j = 0;
625 int32_t countl = 0;
626 int32_t countr = 0;
627 int32_t inpsize = sourceLength;
628 UChar lamalefChar;
629 UChar *tempbuffer=NULL;
630
631 switch(options&U_SHAPE_LENGTH_MASK) {
632
633 case U_SHAPE_LENGTH_GROW_SHRINK :
634 destSize = calculateSize(dest,sourceLength,destSize,options);
635 tempbuffer = (UChar *)uprv_malloc((destSize+1)*U_SIZEOF_UCHAR);
636
637 /* Test for NULL */
638 if(tempbuffer == NULL) {
639 *pErrorCode = U_MEMORY_ALLOCATION_ERROR;
640 return 0;
641 }
642
643 uprv_memset(tempbuffer, 0, (destSize+1)*U_SIZEOF_UCHAR);
644
645 i = j = 0;
646 while(i < destSize && j < destSize) {
647 if( isLamAlefChar(dest[i]) ) {
648 tempbuffer[j] = convertLamAlef[ dest[i] - 0xFEF5 ];
649 tempbuffer[j+1] = 0x0644;
650 j++;
651 }
652 else
653 tempbuffer[j] = dest[i];
654 i++;
655 j++;
656 }
657
658 uprv_memcpy(dest, tempbuffer, destSize*U_SIZEOF_UCHAR);
659 break;
660
661 case U_SHAPE_LENGTH_FIXED_SPACES_NEAR :
662 for(i = 0;i<sourceLength;i++) {
663 if((dest[i] == 0x0020) && isLamAlefChar(dest[i+1])) {
664 lamalefChar = dest[i+1];
665 dest[i+1] = 0x0644;
666 dest[i] = convertLamAlef[ lamalefChar - 0xFEF5 ];
667 }
668 else
669 if((dest[i] != 0x0020) && isLamAlefChar(dest[i+1])) {
670 *pErrorCode=U_NO_SPACE_AVAILABLE;
671 }
672 }
673 destSize = sourceLength;
674 break;
675
676 case U_SHAPE_LENGTH_FIXED_SPACES_AT_BEGINNING :
677 tempbuffer = (UChar *)uprv_malloc((sourceLength+1)*U_SIZEOF_UCHAR);
678
679 /* Test for NULL */
680 if(tempbuffer == NULL) {
681 *pErrorCode = U_MEMORY_ALLOCATION_ERROR;
682 return 0;
683 }
684
685 uprv_memset(tempbuffer, 0, (sourceLength+1)*U_SIZEOF_UCHAR);
686
687 i = 0;
688 while(dest[i] == 0x0020) {
689 countl++;
690 i++;
691 }
692
693 i = j = sourceLength-1;
694 while(i >= 0 && j >= 0) {
695 if( countl>0 && isLamAlefChar(dest[i]) ) {
696 tempbuffer[j] = 0x0644;
697 tempbuffer[j-1] = convertLamAlef[ dest[i] - 0xFEF5 ];
698 j--;
699 countl--;
700 }
701 else {
702 if( countl == 0 && isLamAlefChar(dest[i]) )
703 *pErrorCode=U_NO_SPACE_AVAILABLE;
704 tempbuffer[j] = dest[i];
705 }
706 i--;
707 j--;
708 }
709
710 uprv_memcpy(dest, tempbuffer, sourceLength*U_SIZEOF_UCHAR);
711 destSize = sourceLength;
712 break;
713
714 case U_SHAPE_LENGTH_FIXED_SPACES_AT_END :
715 /* LamAlef expansion below is done from right to left to make sure that we consume
716 * the spaces with the LamAlefs as they appear in the visual buffer from right to left
717 */
718 tempbuffer = (UChar *)uprv_malloc((sourceLength+1)*U_SIZEOF_UCHAR);
719
720 /* Test for NULL */
721 if(tempbuffer == NULL) {
722 *pErrorCode = U_MEMORY_ALLOCATION_ERROR;
723 return 0;
724 }
725
726 uprv_memset(tempbuffer, 0, (sourceLength+1)*U_SIZEOF_UCHAR);
727
728 while(dest[inpsize-1] == 0x0020) {
729 countr++;
730 inpsize--;
731 }
732
733 i = sourceLength - countr - 1;
734 j = sourceLength - 1;
735
736 while(i >= 0 && j >= 0) {
737 if( countr>0 && isLamAlefChar(dest[i]) ) {
738 tempbuffer[j] = 0x0644;
739 tempbuffer[j-1] = convertLamAlef[ dest[i] - 0xFEF5 ];
740 j--;
741 countr--;
742 }
743 else {
744 if( countr == 0 && isLamAlefChar(dest[i]) )
745 *pErrorCode=U_NO_SPACE_AVAILABLE;
746 tempbuffer[j] = dest[i];
747 }
748 i--;
749 j--;
750 }
751
752 if(countr > 0) {
753 uprv_memcpy(tempbuffer, tempbuffer+countr, sourceLength*U_SIZEOF_UCHAR);
754 if(u_strlen(tempbuffer) < sourceLength) {
755 for(i=sourceLength-1;i>=sourceLength-countr;i--)
756 tempbuffer[i] = 0x0020;
757 }
758 }
759
760 uprv_memcpy(dest, tempbuffer, sourceLength*U_SIZEOF_UCHAR);
761
762 destSize = sourceLength;
763 break;
764
765 default :
766 /* will never occur because of validity checks */
767 break;
768 }
769
770 if(tempbuffer)
771 uprv_free(tempbuffer);
772
773 return destSize;
774}
775
776/*
777 *Name : shapeUnicode
778 *Function : Converts an Arabic Unicode buffer in 06xx Range into a shaped
779 * arabic Unicode buffer in FExx Range
780 */
781static int32_t
782shapeUnicode(UChar *dest, int32_t sourceLength,
783 int32_t destSize,uint32_t options,
784 UErrorCode *pErrorCode,
785 int tashkeelFlag) {
786
787 int32_t i, iend;
788 int32_t step;
73c04bcf 789 int32_t lastPos,Nx, Nw;
b75a7d8f
A		 790 unsigned int Shape;
791 int32_t flag;
792 int32_t lamalef_found = 0;
793 UChar prevLink = 0, lastLink = 0, currLink, nextLink = 0;
794 UChar wLamalef;
795
796 /*
797 * Converts the input buffer from FExx Range into 06xx Range
798 * to make sure that all characters are in the 06xx range
799 * even the lamalef is converted to the special region in
800 * the 06xx range
801 */
802 for (i = 0; i < sourceLength; i++) {
803 UChar inputChar = dest[i];
804 if ( (inputChar >= 0xFE70) && (inputChar <= 0xFEFC)) {
805 dest[i] = convertFEto06 [ (inputChar - 0xFE70) ] ;
806 } else {
807 dest[i] = inputChar ;
808 }
809 }
810
811 /* sets the index to the end of the buffer, together with the step point to -1 */
812 i = sourceLength - 1;
813 iend = -1;
814 step = -1;
815
816 /*
817 * This function resolves the link between the characters .
818 * Arabic characters have four forms :
819 * Isolated Form, Initial Form, Middle Form and Final Form
820 */
821 currLink = getLink(dest[i]);
822
b75a7d8f
A		 823 lastPos = i;
824 Nx = -2, Nw = 0;
825
826 while (i != iend) {
827 /* If high byte of currLink > 0 then more than one shape */
828 if ((currLink & 0xFF00) > 0 || isTashkeelChar(dest[i])) {
829 Nw = i + step;
830 while (Nx < 0) { /* we need to know about next char */
831 if(Nw == iend) {
832 nextLink = 0;
833 Nx = 3000;
834 } else {
835 nextLink = getLink(dest[Nw]);
836 if((nextLink & IRRELEVANT) == 0) {
837 Nx = Nw;
838 } else {
839 Nw = Nw + step;
840 }
841 }
842 }
843
844 if ( ((currLink & ALEFTYPE) > 0) && ((lastLink & LAMTYPE) > 0) ) {
845 lamalef_found = 1;
846 wLamalef = changeLamAlef(dest[i]); /*get from 0x065C-0x065f */
847 if ( wLamalef != 0) {
848 dest[i] = 0xFFFF; /* The default case is to drop the Alef and replace */
849 dest[lastPos] =wLamalef; /* it by 0xFFFF which is the last character in the */
850 i=lastPos; /* unicode private use area, this is done to make */
851 } /* sure that removeLamAlefSpaces() handles only the */
852 lastLink = prevLink; /* spaces generated during lamalef generation. */
853 currLink = getLink(wLamalef); /* 0xFFFF is added here and is replaced by spaces */
854 } /* in removeLamAlefSpaces() */
855 /*
856 * get the proper shape according to link ability of neighbors
857 * and of character; depends on the order of the shapes
858 * (isolated, initial, middle, final) in the compatibility area
859 */
860 flag = specialChar(dest[i]);
861
862 Shape = shapeTable[nextLink & (LINKR + LINKL)]
863 [lastLink & (LINKR + LINKL)]
864 [currLink & (LINKR + LINKL)];
865
866 if (flag == 1) {
867 Shape = (Shape == 1 || Shape == 3) ? 1 : 0;
868 }
73c04bcf 869 else if(flag == 2) {
b75a7d8f 870 if( (lastLink & LINKL) && (nextLink & LINKR) && (tashkeelFlag == 1) &&
73c04bcf
A		 871 dest[i] != 0x064C && dest[i] != 0x064D )
872 {
b75a7d8f 873 Shape = 1;
73c04bcf 874 if( (nextLink&ALEFTYPE) == ALEFTYPE && (lastLink&LAMTYPE) == LAMTYPE ) {
b75a7d8f 875 Shape = 0;
73c04bcf 876 }
b75a7d8f
A		 877 }
878 else {
879 Shape = 0;
880 }
881 }
882
883 if(flag == 2) {
884 dest[i] = 0xFE70 + IrrelevantPos[(dest[i] - 0x064B)] + Shape;
885 }
886 else
887 dest[i] = (UChar)(0xFE70 + (currLink >> 8) + Shape);
888 }
889
890 /* move one notch forward */
891 if ((currLink & IRRELEVANT) == 0) {
892 prevLink = lastLink;
893 lastLink = currLink;
b75a7d8f
A		 894 lastPos = i;
895 }
896
897 i = i + step;
898 if (i == Nx) {
899 currLink = nextLink;
900 Nx = -2;
901 }
902 else if(i != iend) {
903 currLink = getLink(dest[i]);
904 }
905 }
906
907 /* If there is lamalef in the buffer call expandLamAlef */
908 if(lamalef_found != 0)
909 destSize = removeLamAlefSpaces(dest,sourceLength,destSize,options,pErrorCode);
910 else
911 destSize = sourceLength;
912
913 return destSize;
914}
915
916/*
917 *Name : deShapeUnicode
918 *Function : Converts an Arabic Unicode buffer in FExx Range into unshaped
919 * arabic Unicode buffer in 06xx Range
920 */
921static int32_t
922deShapeUnicode(UChar *dest, int32_t sourceLength,
923 int32_t destSize,uint32_t options,
924 UErrorCode *pErrorCode) {
925 int32_t i = 0;
926 int32_t lamalef_found = 0;
927
928 /*
929 *This for loop changes the buffer from the Unicode FE range to
930 *the Unicode 06 range
931 */
932 for(i = 0; i < sourceLength; i++) {
933 UChar inputChar = dest[i];
934 if (( inputChar >= 0xFE70) && (inputChar <= 0xFEF4 )) { /* FExx Arabic range */
935 dest[i] = convertFEto06 [ (inputChar - 0xFE70) ] ;
936 } else {
937 dest[i] = inputChar ;
938 }
939 if( isLamAlefChar(dest[i]) )
940 lamalef_found = 1;
941 }
942
943 /* If there is lamalef in the buffer call expandLamAlef */
944 if(lamalef_found != 0)
945 destSize = expandLamAlef(dest,sourceLength,destSize,options,pErrorCode);
946 else
947 destSize = sourceLength;
948
949 return destSize;
950}
951
952U_CAPI int32_t U_EXPORT2
953u_shapeArabic(const UChar *source, int32_t sourceLength,
954 UChar *dest, int32_t destCapacity,
955 uint32_t options,
956 UErrorCode *pErrorCode) {
957
958 int32_t destLength;
959
960 /* usual error checking */
961 if(pErrorCode==NULL || U_FAILURE(*pErrorCode)) {
962 return 0;
963 }
964
965 /* make sure that no reserved options values are used; allow dest==NULL only for preflighting */
966 if( source==NULL || sourceLength<-1 ||
967 (dest==NULL && destCapacity!=0) || destCapacity<0 ||
968 options>=U_SHAPE_DIGIT_TYPE_RESERVED ||
969 (options&U_SHAPE_DIGITS_MASK)>=U_SHAPE_DIGITS_RESERVED
970 ) {
971 *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
972 return 0;
973 }
974
975 /* determine the source length */
976 if(sourceLength==-1) {
977 sourceLength=u_strlen(source);
978 }
73c04bcf 979 if(sourceLength<=0) {
b75a7d8f
A		 980 return u_terminateUChars(dest, destCapacity, 0, pErrorCode);
981 }
982
983 /* check that source and destination do not overlap */
984 if( dest!=NULL &&
985 ((source<=dest && dest<source+sourceLength) ||
986 (dest<=source && source<dest+destCapacity))
987 ) {
988 *pErrorCode=U_ILLEGAL_ARGUMENT_ERROR;
989 return 0;
990 }
991
992 if((options&U_SHAPE_LETTERS_MASK)!=U_SHAPE_LETTERS_NOOP) {
993 UChar buffer[300];
994 UChar *tempbuffer;
995 int32_t outputSize, spacesCountl=0, spacesCountr=0;
996
997 /* calculate destination size */
998 /* TODO: do we ever need to do this pure preflighting? */
999 if((options&U_SHAPE_LENGTH_MASK)==U_SHAPE_LENGTH_GROW_SHRINK) {
1000 outputSize=calculateSize(source,sourceLength,destCapacity,options);
1001 } else {
1002 outputSize=sourceLength;
1003 }
1004 if(outputSize>destCapacity) {
1005 *pErrorCode=U_BUFFER_OVERFLOW_ERROR;
1006 return outputSize;
1007 }
1008
1009 /*
1010 * need a temporary buffer of size max(outputSize, sourceLength)
1011 * because at first we copy source->temp
1012 */
1013 if(sourceLength>outputSize) {
1014 outputSize=sourceLength;
1015 }
1016
1017 /* Start of Arabic letter shaping part */
1018 if(outputSize<=sizeof(buffer)/U_SIZEOF_UCHAR) {
1019 outputSize=sizeof(buffer)/U_SIZEOF_UCHAR;
1020 tempbuffer=buffer;
1021 } else {
1022 tempbuffer = (UChar *)uprv_malloc(outputSize*U_SIZEOF_UCHAR);
1023
1024 /*Test for NULL*/
1025 if(tempbuffer == NULL) {
1026 *pErrorCode = U_MEMORY_ALLOCATION_ERROR;
1027 return 0;
1028 }
1029 }
1030 uprv_memcpy(tempbuffer, source, sourceLength*U_SIZEOF_UCHAR);
1031 if(sourceLength<outputSize) {
1032 uprv_memset(tempbuffer+sourceLength, 0, (outputSize-sourceLength)*U_SIZEOF_UCHAR);
1033 }
1034
1035 if((options&U_SHAPE_TEXT_DIRECTION_MASK) == U_SHAPE_TEXT_DIRECTION_LOGICAL) {
1036 countSpaces(tempbuffer,sourceLength,options,&spacesCountl,&spacesCountr);
1037 invertBuffer(tempbuffer,sourceLength,options,&spacesCountl,&spacesCountr);
1038 }
1039
1040 switch(options&U_SHAPE_LETTERS_MASK) {
1041 case U_SHAPE_LETTERS_SHAPE :
1042 /* Call the shaping function with tashkeel flag == 1 */
1043 destLength = shapeUnicode(tempbuffer,sourceLength,destCapacity,options,pErrorCode,1);
1044 break;
1045 case U_SHAPE_LETTERS_SHAPE_TASHKEEL_ISOLATED :
1046 /* Call the shaping function with tashkeel flag == 0 */
1047 destLength = shapeUnicode(tempbuffer,sourceLength,destCapacity,options,pErrorCode,0);
1048 break;
1049 case U_SHAPE_LETTERS_UNSHAPE :
1050 /* Call the deshaping function */
1051 destLength = deShapeUnicode(tempbuffer,sourceLength,destCapacity,options,pErrorCode);
1052 break;
1053 default :
1054 /* will never occur because of validity checks above */
1055 destLength = 0;
1056 break;
1057 }
1058
1059 /*
1060 * TODO: (markus 2002aug01)
1061 * For as long as we always preflight the outputSize above
1062 * we should U_ASSERT(outputSize==destLength)
1063 * except for the adjustment above before the tempbuffer allocation
1064 */
1065
1066 if((options&U_SHAPE_TEXT_DIRECTION_MASK) == U_SHAPE_TEXT_DIRECTION_LOGICAL) {
1067 countSpaces(tempbuffer,destLength,options,&spacesCountl,&spacesCountr);
1068 invertBuffer(tempbuffer,destLength,options,&spacesCountl,&spacesCountr);
1069 }
1070 uprv_memcpy(dest, tempbuffer, uprv_min(destLength, destCapacity)*U_SIZEOF_UCHAR);
1071
1072 if(tempbuffer!=buffer) {
1073 uprv_free(tempbuffer);
1074 }
1075
1076 if(destLength>destCapacity) {
1077 *pErrorCode=U_BUFFER_OVERFLOW_ERROR;
1078 return destLength;
1079 }
1080
1081 /* End of Arabic letter shaping part */
1082 } else {
1083 /*
1084 * No letter shaping:
1085 * just make sure the destination is large enough and copy the string.
1086 */
1087 if(destCapacity<sourceLength) {
1088 /* this catches preflighting, too */
1089 *pErrorCode=U_BUFFER_OVERFLOW_ERROR;
1090 return sourceLength;
1091 }
1092 uprv_memcpy(dest, source, sourceLength*U_SIZEOF_UCHAR);
1093 destLength=sourceLength;
1094 }
1095
1096 /*
1097 * Perform number shaping.
1098 * With UTF-16 or UTF-32, the length of the string is constant.
1099 * The easiest way to do this is to operate on the destination and
1100 * "shape" the digits in-place.
1101 */
1102 if((options&U_SHAPE_DIGITS_MASK)!=U_SHAPE_DIGITS_NOOP) {
1103 UChar digitBase;
1104 int32_t i;
1105
1106 /* select the requested digit group */
1107 switch(options&U_SHAPE_DIGIT_TYPE_MASK) {
1108 case U_SHAPE_DIGIT_TYPE_AN:
1109 digitBase=0x660; /* Unicode: "Arabic-Indic digits" */
1110 break;
1111 case U_SHAPE_DIGIT_TYPE_AN_EXTENDED:
1112 digitBase=0x6f0; /* Unicode: "Eastern Arabic-Indic digits (Persian and Urdu)" */
1113 break;
1114 default:
1115 /* will never occur because of validity checks above */
1116 digitBase=0;
1117 break;
1118 }
1119
1120 /* perform the requested operation */
1121 switch(options&U_SHAPE_DIGITS_MASK) {
1122 case U_SHAPE_DIGITS_EN2AN:
1123 /* add (digitBase-'0') to each European (ASCII) digit code point */
1124 digitBase-=0x30;
1125 for(i=0; i<destLength; ++i) {
1126 if(((uint32_t)dest[i]-0x30)<10) {
1127 dest[i]+=digitBase;
1128 }
1129 }
1130 break;
1131 case U_SHAPE_DIGITS_AN2EN:
1132 /* subtract (digitBase-'0') from each Arabic digit code point */
1133 for(i=0; i<destLength; ++i) {
1134 if(((uint32_t)dest[i]-(uint32_t)digitBase)<10) {
1135 dest[i]-=digitBase-0x30;
1136 }
1137 }
1138 break;
1139 case U_SHAPE_DIGITS_ALEN2AN_INIT_LR:
1140 _shapeToArabicDigitsWithContext(dest, destLength,
1141 digitBase,
1142 (UBool)((options&U_SHAPE_TEXT_DIRECTION_MASK)==U_SHAPE_TEXT_DIRECTION_LOGICAL),
1143 FALSE);
1144 break;
1145 case U_SHAPE_DIGITS_ALEN2AN_INIT_AL:
1146 _shapeToArabicDigitsWithContext(dest, destLength,
1147 digitBase,
1148 (UBool)((options&U_SHAPE_TEXT_DIRECTION_MASK)==U_SHAPE_TEXT_DIRECTION_LOGICAL),
1149 TRUE);
1150 break;
1151 default:
1152 /* will never occur because of validity checks above */
1153 break;
1154 }
1155 }
1156
1157 return u_terminateUChars(dest, destCapacity, destLength, pErrorCode);
1158}
mirror of ICU