2
******************************************************************************
4
* Copyright (C) 2000-2010, International Business Machines
5
* Corporation and others. All Rights Reserved.
7
******************************************************************************
10
* tab size: 8 (not used)
13
* created on: 2000jun29
14
* created by: Markus W. Scherer
20
#include "unicode/utypes.h"
24
* \brief C API: Arabic shaping
29
* Shape Arabic text on a character basis.
31
* <p>This function performs basic operations for "shaping" Arabic text. It is most
32
* useful for use with legacy data formats and legacy display technology
33
* (simple terminals). All operations are performed on Unicode characters.</p>
35
* <p>Text-based shaping means that some character code points in the text are
36
* replaced by others depending on the context. It transforms one kind of text
37
* into another. In comparison, modern displays for Arabic text select
38
* appropriate, context-dependent font glyphs for each text element, which means
39
* that they transform text into a glyph vector.</p>
41
* <p>Text transformations are necessary when modern display technology is not
42
* available or when text needs to be transformed to or from legacy formats that
43
* use "shaped" characters. Since the Arabic script is cursive, connecting
44
* adjacent letters to each other, computers select images for each letter based
45
* on the surrounding letters. This usually results in four images per Arabic
46
* letter: initial, middle, final, and isolated forms. In Unicode, on the other
47
* hand, letters are normally stored abstract, and a display system is expected
48
* to select the necessary glyphs. (This makes searching and other text
49
* processing easier because the same letter has only one code.) It is possible
50
* to mimic this with text transformations because there are characters in
51
* Unicode that are rendered as letters with a specific shape
52
* (or cursive connectivity). They were included for interoperability with
53
* legacy systems and codepages, and for unsophisticated display systems.</p>
55
* <p>A second kind of text transformations is supported for Arabic digits:
56
* For compatibility with legacy codepages that only include European digits,
57
* it is possible to replace one set of digits by another, changing the
58
* character code points. These operations can be performed for either
59
* Arabic-Indic Digits (U+0660...U+0669) or Eastern (Extended) Arabic-Indic
60
* digits (U+06f0...U+06f9).</p>
62
* <p>Some replacements may result in more or fewer characters (code points).
63
* By default, this means that the destination buffer may receive text with a
64
* length different from the source length. Some legacy systems rely on the
65
* length of the text to be constant. They expect extra spaces to be added
66
* or consumed either next to the affected character or at the end of the
69
* <p>For details about the available operations, see the description of the
70
* <code>U_SHAPE_...</code> options.</p>
72
* @param source The input text.
74
* @param sourceLength The number of UChars in <code>source</code>.
76
* @param dest The destination buffer that will receive the results of the
77
* requested operations. It may be <code>NULL</code> only if
78
* <code>destSize</code> is 0. The source and destination must not
81
* @param destSize The size (capacity) of the destination buffer in UChars.
82
* If <code>destSize</code> is 0, then no output is produced,
83
* but the necessary buffer size is returned ("preflighting").
85
* @param options This is a 32-bit set of flags that specify the operations
86
* that are performed on the input text. If no error occurs,
87
* then the result will always be written to the destination
90
* @param pErrorCode must be a valid pointer to an error code value,
91
* which must not indicate a failure before the function call.
93
* @return The number of UChars written to the destination buffer.
94
* If an error occured, then no output was written, or it may be
95
* incomplete. If <code>U_BUFFER_OVERFLOW_ERROR</code> is set, then
96
* the return value indicates the necessary destination buffer size.
99
U_STABLE int32_t U_EXPORT2
100
u_shapeArabic(const UChar *source, int32_t sourceLength,
101
UChar *dest, int32_t destSize,
103
UErrorCode *pErrorCode);
106
* Memory option: allow the result to have a different length than the source.
107
* Affects: LamAlef options
110
#define U_SHAPE_LENGTH_GROW_SHRINK 0
113
* Memory option: allow the result to have a different length than the source.
114
* Affects: LamAlef options
115
* This option is an alias to U_SHAPE_LENGTH_GROW_SHRINK
118
#define U_SHAPE_LAMALEF_RESIZE 0
121
* Memory option: the result must have the same length as the source.
122
* If more room is necessary, then try to consume spaces next to modified characters.
125
#define U_SHAPE_LENGTH_FIXED_SPACES_NEAR 1
128
* Memory option: the result must have the same length as the source.
129
* If more room is necessary, then try to consume spaces next to modified characters.
130
* Affects: LamAlef options
131
* This option is an alias to U_SHAPE_LENGTH_FIXED_SPACES_NEAR
134
#define U_SHAPE_LAMALEF_NEAR 1
137
* Memory option: the result must have the same length as the source.
138
* If more room is necessary, then try to consume spaces at the end of the text.
141
#define U_SHAPE_LENGTH_FIXED_SPACES_AT_END 2
144
* Memory option: the result must have the same length as the source.
145
* If more room is necessary, then try to consume spaces at the end of the text.
146
* Affects: LamAlef options
147
* This option is an alias to U_SHAPE_LENGTH_FIXED_SPACES_AT_END
150
#define U_SHAPE_LAMALEF_END 2
153
* Memory option: the result must have the same length as the source.
154
* If more room is necessary, then try to consume spaces at the beginning of the text.
157
#define U_SHAPE_LENGTH_FIXED_SPACES_AT_BEGINNING 3
160
* Memory option: the result must have the same length as the source.
161
* If more room is necessary, then try to consume spaces at the beginning of the text.
162
* Affects: LamAlef options
163
* This option is an alias to U_SHAPE_LENGTH_FIXED_SPACES_AT_BEGINNING
166
#define U_SHAPE_LAMALEF_BEGIN 3
170
* Memory option: the result must have the same length as the source.
171
* Shaping Mode: For each LAMALEF character found, expand LAMALEF using space at end.
172
* If there is no space at end, use spaces at beginning of the buffer. If there
173
* is no space at beginning of the buffer, use spaces at the near (i.e. the space
174
* after the LAMALEF character).
175
* If there are no spaces found, an error U_NO_SPACE_AVAILABLE (as defined in utypes.h)
176
* will be set in pErrorCode
178
* Deshaping Mode: Perform the same function as the flag equals U_SHAPE_LAMALEF_END.
179
* Affects: LamAlef options
182
#define U_SHAPE_LAMALEF_AUTO 0x10000
184
/** Bit mask for memory options. @stable ICU 2.0 */
185
#define U_SHAPE_LENGTH_MASK 0x10003 /* Changed old value 3 */
189
* Bit mask for LamAlef memory options.
192
#define U_SHAPE_LAMALEF_MASK 0x10003 /* updated */
194
/** Direction indicator: the source is in logical (keyboard) order. @stable ICU 2.0 */
195
#define U_SHAPE_TEXT_DIRECTION_LOGICAL 0
198
* Direction indicator:
199
* the source is in visual RTL order,
200
* the rightmost displayed character stored first.
201
* This option is an alias to U_SHAPE_TEXT_DIRECTION_LOGICAL
204
#define U_SHAPE_TEXT_DIRECTION_VISUAL_RTL 0
207
* Direction indicator:
208
* the source is in visual LTR order,
209
* the leftmost displayed character stored first.
212
#define U_SHAPE_TEXT_DIRECTION_VISUAL_LTR 4
214
/** Bit mask for direction indicators. @stable ICU 2.0 */
215
#define U_SHAPE_TEXT_DIRECTION_MASK 4
218
/** Letter shaping option: do not perform letter shaping. @stable ICU 2.0 */
219
#define U_SHAPE_LETTERS_NOOP 0
221
/** Letter shaping option: replace abstract letter characters by "shaped" ones. @stable ICU 2.0 */
222
#define U_SHAPE_LETTERS_SHAPE 8
224
/** Letter shaping option: replace "shaped" letter characters by abstract ones. @stable ICU 2.0 */
225
#define U_SHAPE_LETTERS_UNSHAPE 0x10
228
* Letter shaping option: replace abstract letter characters by "shaped" ones.
229
* The only difference with U_SHAPE_LETTERS_SHAPE is that Tashkeel letters
230
* are always "shaped" into the isolated form instead of the medial form
231
* (selecting code points from the Arabic Presentation Forms-B block).
234
#define U_SHAPE_LETTERS_SHAPE_TASHKEEL_ISOLATED 0x18
237
/** Bit mask for letter shaping options. @stable ICU 2.0 */
238
#define U_SHAPE_LETTERS_MASK 0x18
241
/** Digit shaping option: do not perform digit shaping. @stable ICU 2.0 */
242
#define U_SHAPE_DIGITS_NOOP 0
245
* Digit shaping option:
246
* Replace European digits (U+0030...) by Arabic-Indic digits.
249
#define U_SHAPE_DIGITS_EN2AN 0x20
252
* Digit shaping option:
253
* Replace Arabic-Indic digits by European digits (U+0030...).
256
#define U_SHAPE_DIGITS_AN2EN 0x40
259
* Digit shaping option:
260
* Replace European digits (U+0030...) by Arabic-Indic digits if the most recent
261
* strongly directional character is an Arabic letter
262
* (<code>u_charDirection()</code> result <code>U_RIGHT_TO_LEFT_ARABIC</code> [AL]).<br>
263
* The direction of "preceding" depends on the direction indicator option.
264
* For the first characters, the preceding strongly directional character
265
* (initial state) is assumed to be not an Arabic letter
266
* (it is <code>U_LEFT_TO_RIGHT</code> [L] or <code>U_RIGHT_TO_LEFT</code> [R]).
269
#define U_SHAPE_DIGITS_ALEN2AN_INIT_LR 0x60
272
* Digit shaping option:
273
* Replace European digits (U+0030...) by Arabic-Indic digits if the most recent
274
* strongly directional character is an Arabic letter
275
* (<code>u_charDirection()</code> result <code>U_RIGHT_TO_LEFT_ARABIC</code> [AL]).<br>
276
* The direction of "preceding" depends on the direction indicator option.
277
* For the first characters, the preceding strongly directional character
278
* (initial state) is assumed to be an Arabic letter.
281
#define U_SHAPE_DIGITS_ALEN2AN_INIT_AL 0x80
283
/** Not a valid option value. May be replaced by a new option. @stable ICU 2.0 */
284
#define U_SHAPE_DIGITS_RESERVED 0xa0
286
/** Bit mask for digit shaping options. @stable ICU 2.0 */
287
#define U_SHAPE_DIGITS_MASK 0xe0
290
/** Digit type option: Use Arabic-Indic digits (U+0660...U+0669). @stable ICU 2.0 */
291
#define U_SHAPE_DIGIT_TYPE_AN 0
293
/** Digit type option: Use Eastern (Extended) Arabic-Indic digits (U+06f0...U+06f9). @stable ICU 2.0 */
294
#define U_SHAPE_DIGIT_TYPE_AN_EXTENDED 0x100
296
/** Not a valid option value. May be replaced by a new option. @stable ICU 2.0 */
297
#define U_SHAPE_DIGIT_TYPE_RESERVED 0x200
299
/** Bit mask for digit type options. @stable ICU 2.0 */
300
#define U_SHAPE_DIGIT_TYPE_MASK 0x300 /* I need to change this from 0x3f00 to 0x300 */
303
* Tashkeel aggregation option:
304
* Replaces any combination of U+0651 with one of
305
* U+064C, U+064D, U+064E, U+064F, U+0650 with
306
* U+FC5E, U+FC5F, U+FC60, U+FC61, U+FC62 consecutively.
309
#define U_SHAPE_AGGREGATE_TASHKEEL 0x4000
310
/** Tashkeel aggregation option: do not aggregate tashkeels. @stable ICU 3.6 */
311
#define U_SHAPE_AGGREGATE_TASHKEEL_NOOP 0
312
/** Bit mask for tashkeel aggregation. @stable ICU 3.6 */
313
#define U_SHAPE_AGGREGATE_TASHKEEL_MASK 0x4000
316
* Presentation form option:
317
* Don't replace Arabic Presentation Forms-A and Arabic Presentation Forms-B
318
* characters with 0+06xx characters, before shaping.
321
#define U_SHAPE_PRESERVE_PRESENTATION 0x8000
322
/** Presentation form option:
323
* Replace Arabic Presentation Forms-A and Arabic Presentationo Forms-B with
324
* their unshaped correspondants in range 0+06xx, before shaping.
327
#define U_SHAPE_PRESERVE_PRESENTATION_NOOP 0
328
/** Bit mask for preserve presentation form. @stable ICU 3.6 */
329
#define U_SHAPE_PRESERVE_PRESENTATION_MASK 0x8000
331
/* Seen Tail option */
333
* Memory option: the result must have the same length as the source.
334
* Shaping mode: The SEEN family character will expand into two characters using space near
335
* the SEEN family character(i.e. the space after the character).
336
* If there are no spaces found, an error U_NO_SPACE_AVAILABLE (as defined in utypes.h)
337
* will be set in pErrorCode
339
* De-shaping mode: Any Seen character followed by Tail character will be
340
* replaced by one cell Seen and a space will replace the Tail.
341
* Affects: Seen options
344
#define U_SHAPE_SEEN_TWOCELL_NEAR 0x200000
347
* Bit mask for Seen memory options.
350
#define U_SHAPE_SEEN_MASK 0x700000
352
/* YehHamza option */
354
* Memory option: the result must have the same length as the source.
355
* Shaping mode: The YEHHAMZA character will expand into two characters using space near it
356
* (i.e. the space after the character
357
* If there are no spaces found, an error U_NO_SPACE_AVAILABLE (as defined in utypes.h)
358
* will be set in pErrorCode
360
* De-shaping mode: Any Yeh (final or isolated) character followed by Hamza character will be
361
* replaced by one cell YehHamza and space will replace the Hamza.
362
* Affects: YehHamza options
365
#define U_SHAPE_YEHHAMZA_TWOCELL_NEAR 0x1000000
369
* Bit mask for YehHamza memory options.
372
#define U_SHAPE_YEHHAMZA_MASK 0x3800000
374
/* New Tashkeel options */
376
* Memory option: the result must have the same length as the source.
377
* Shaping mode: Tashkeel characters will be replaced by spaces.
378
* Spaces will be placed at beginning of the buffer
380
* De-shaping mode: N/A
381
* Affects: Tashkeel options
384
#define U_SHAPE_TASHKEEL_BEGIN 0x40000
387
* Memory option: the result must have the same length as the source.
388
* Shaping mode: Tashkeel characters will be replaced by spaces.
389
* Spaces will be placed at end of the buffer
391
* De-shaping mode: N/A
392
* Affects: Tashkeel options
395
#define U_SHAPE_TASHKEEL_END 0x60000
398
* Memory option: allow the result to have a different length than the source.
399
* Shaping mode: Tashkeel characters will be removed, buffer length will shrink.
400
* De-shaping mode: N/A
402
* Affect: Tashkeel options
405
#define U_SHAPE_TASHKEEL_RESIZE 0x80000
408
* Memory option: the result must have the same length as the source.
409
* Shaping mode: Tashkeel characters will be replaced by Tatweel if it is connected to adjacent
410
* characters (i.e. shaped on Tatweel) or replaced by space if it is not connected.
412
* De-shaping mode: N/A
413
* Affects: YehHamza options
416
#define U_SHAPE_TASHKEEL_REPLACE_BY_TATWEEL 0xC0000
419
* Bit mask for Tashkeel replacement with Space or Tatweel memory options.
422
#define U_SHAPE_TASHKEEL_MASK 0xE0000
425
/* Space location Control options */
427
* This option affect the meaning of BEGIN and END options. if this option is not used the default
428
* for BEGIN and END will be as following:
429
* The Default (for both Visual LTR, Visual RTL and Logical Text)
430
* 1. BEGIN always refers to the start address of physical memory.
431
* 2. END always refers to the end address of physical memory.
433
* If this option is used it will swap the meaning of BEGIN and END only for Visual LTR text.
435
* The effect on BEGIN and END Memory Options will be as following:
436
* A. BEGIN For Visual LTR text: This will be the beginning (right side) of the visual text(
437
* corresponding to the physical memory address end for Visual LTR text, Same as END in
439
* B. BEGIN For Logical text: Same as BEGIN in default behavior.
440
* C. END For Visual LTR text: This will be the end (left side) of the visual text (corresponding
441
* to the physical memory address beginning for Visual LTR text, Same as BEGIN in default behavior.
442
* D. END For Logical text: Same as END in default behavior).
443
* Affects: All LamAlef BEGIN, END and AUTO options.
446
#define U_SHAPE_SPACES_RELATIVE_TO_TEXT_BEGIN_END 0x4000000
449
* Bit mask for swapping BEGIN and END for Visual LTR text
452
#define U_SHAPE_SPACES_RELATIVE_TO_TEXT_MASK 0x4000000
455
* If this option is used, shaping will use the new Unicode code point for TAIL (i.e. 0xFE73).
456
* If this option is not specified (Default), old unofficial Unicode TAIL code point is used (i.e. 0x200B)
457
* De-shaping will not use this option as it will always search for both the new Unicode code point for the
458
* TAIL (i.e. 0xFE73) or the old unofficial Unicode TAIL code point (i.e. 0x200B) and de-shape the
459
* Seen-Family letter accordingly.
461
* Shaping Mode: Only shaping.
462
* De-shaping Mode: N/A.
463
* Affects: All Seen options
466
#define SHAPE_TAIL_NEW_UNICODE 0x8000000
469
* Bit mask for new Unicode Tail option
472
#define SHAPE_TAIL_TYPE_MASK 0x8000000