1// © 2016 and later: Unicode, Inc. and others.
2// License & terms of use: http://www.unicode.org/copyright.html
3/*
4******************************************************************************
5*
6* Copyright (C) 2000-2012, International Business Machines
7* Corporation and others. All Rights Reserved.
8*
9******************************************************************************
10* file name: ushape.h
11* encoding: UTF-8
12* tab size: 8 (not used)
13* indentation:4
14*
15* created on: 2000jun29
16* created by: Markus W. Scherer
17*/
18
19#ifndef __USHAPE_H__
20#define __USHAPE_H__
21
22#include "unicode/utypes.h"
23
24/**
25 * \file
26 * \brief C API: Arabic shaping
27 *
28 */
29
30/**
31 * Shape Arabic text on a character basis.
32 *
33 * <p>This function performs basic operations for "shaping" Arabic text. It is most
34 * useful for use with legacy data formats and legacy display technology
35 * (simple terminals). All operations are performed on Unicode characters.</p>
36 *
37 * <p>Text-based shaping means that some character code points in the text are
38 * replaced by others depending on the context. It transforms one kind of text
39 * into another. In comparison, modern displays for Arabic text select
40 * appropriate, context-dependent font glyphs for each text element, which means
41 * that they transform text into a glyph vector.</p>
42 *
43 * <p>Text transformations are necessary when modern display technology is not
44 * available or when text needs to be transformed to or from legacy formats that
45 * use "shaped" characters. Since the Arabic script is cursive, connecting
46 * adjacent letters to each other, computers select images for each letter based
47 * on the surrounding letters. This usually results in four images per Arabic
48 * letter: initial, middle, final, and isolated forms. In Unicode, on the other
49 * hand, letters are normally stored abstract, and a display system is expected
50 * to select the necessary glyphs. (This makes searching and other text
51 * processing easier because the same letter has only one code.) It is possible
52 * to mimic this with text transformations because there are characters in
53 * Unicode that are rendered as letters with a specific shape
54 * (or cursive connectivity). They were included for interoperability with
55 * legacy systems and codepages, and for unsophisticated display systems.</p>
56 *
57 * <p>A second kind of text transformations is supported for Arabic digits:
58 * For compatibility with legacy codepages that only include European digits,
59 * it is possible to replace one set of digits by another, changing the
60 * character code points. These operations can be performed for either
61 * Arabic-Indic Digits (U+0660...U+0669) or Eastern (Extended) Arabic-Indic
62 * digits (U+06f0...U+06f9).</p>
63 *
64 * <p>Some replacements may result in more or fewer characters (code points).
65 * By default, this means that the destination buffer may receive text with a
66 * length different from the source length. Some legacy systems rely on the
67 * length of the text to be constant. They expect extra spaces to be added
68 * or consumed either next to the affected character or at the end of the
69 * text.</p>
70 *
71 * <p>For details about the available operations, see the description of the
72 * <code>U_SHAPE_...</code> options.</p>
73 *
74 * @param source The input text.
75 *
76 * @param sourceLength The number of UChars in <code>source</code>.
77 *
78 * @param dest The destination buffer that will receive the results of the
79 * requested operations. It may be <code>NULL</code> only if
80 * <code>destSize</code> is 0. The source and destination must not
81 * overlap.
82 *
83 * @param destSize The size (capacity) of the destination buffer in UChars.
84 * If <code>destSize</code> is 0, then no output is produced,
85 * but the necessary buffer size is returned ("preflighting").
86 *
87 * @param options This is a 32-bit set of flags that specify the operations
88 * that are performed on the input text. If no error occurs,
89 * then the result will always be written to the destination
90 * buffer.
91 *
92 * @param pErrorCode must be a valid pointer to an error code value,
93 * which must not indicate a failure before the function call.
94 *
95 * @return The number of UChars written to the destination buffer.
96 * If an error occurred, then no output was written, or it may be
97 * incomplete. If <code>U_BUFFER_OVERFLOW_ERROR</code> is set, then
98 * the return value indicates the necessary destination buffer size.
99 * @stable ICU 2.0
100 */
101U_CAPI int32_t U_EXPORT2
102u_shapeArabic(const UChar *source, int32_t sourceLength,
103 UChar *dest, int32_t destSize,
104 uint32_t options,
105 UErrorCode *pErrorCode);
106
107/**
108 * Memory option: allow the result to have a different length than the source.
109 * Affects: LamAlef options
110 * @stable ICU 2.0
111 */
112#define U_SHAPE_LENGTH_GROW_SHRINK 0
113
114/**
115 * Memory option: allow the result to have a different length than the source.
116 * Affects: LamAlef options
117 * This option is an alias to U_SHAPE_LENGTH_GROW_SHRINK
118 * @stable ICU 4.2
119 */
120#define U_SHAPE_LAMALEF_RESIZE 0
121
122/**
123 * Memory option: the result must have the same length as the source.
124 * If more room is necessary, then try to consume spaces next to modified characters.
125 * @stable ICU 2.0
126 */
127#define U_SHAPE_LENGTH_FIXED_SPACES_NEAR 1
128
129/**
130 * Memory option: the result must have the same length as the source.
131 * If more room is necessary, then try to consume spaces next to modified characters.
132 * Affects: LamAlef options
133 * This option is an alias to U_SHAPE_LENGTH_FIXED_SPACES_NEAR
134 * @stable ICU 4.2
135 */
136#define U_SHAPE_LAMALEF_NEAR 1
137
138/**
139 * Memory option: the result must have the same length as the source.
140 * If more room is necessary, then try to consume spaces at the end of the text.
141 * @stable ICU 2.0
142 */
143#define U_SHAPE_LENGTH_FIXED_SPACES_AT_END 2
144
145/**
146 * Memory option: the result must have the same length as the source.
147 * If more room is necessary, then try to consume spaces at the end of the text.
148 * Affects: LamAlef options
149 * This option is an alias to U_SHAPE_LENGTH_FIXED_SPACES_AT_END
150 * @stable ICU 4.2
151 */
152#define U_SHAPE_LAMALEF_END 2
153
154/**
155 * Memory option: the result must have the same length as the source.
156 * If more room is necessary, then try to consume spaces at the beginning of the text.
157 * @stable ICU 2.0
158 */
159#define U_SHAPE_LENGTH_FIXED_SPACES_AT_BEGINNING 3
160
161/**
162 * Memory option: the result must have the same length as the source.
163 * If more room is necessary, then try to consume spaces at the beginning of the text.
164 * Affects: LamAlef options
165 * This option is an alias to U_SHAPE_LENGTH_FIXED_SPACES_AT_BEGINNING
166 * @stable ICU 4.2
167 */
168#define U_SHAPE_LAMALEF_BEGIN 3
169
170
171/**
172 * Memory option: the result must have the same length as the source.
173 * Shaping Mode: For each LAMALEF character found, expand LAMALEF using space at end.
174 * If there is no space at end, use spaces at beginning of the buffer. If there
175 * is no space at beginning of the buffer, use spaces at the near (i.e. the space
176 * after the LAMALEF character).
177 * If there are no spaces found, an error U_NO_SPACE_AVAILABLE (as defined in utypes.h)
178 * will be set in pErrorCode
179 *
180 * Deshaping Mode: Perform the same function as the flag equals U_SHAPE_LAMALEF_END.
181 * Affects: LamAlef options
182 * @stable ICU 4.2
183 */
184#define U_SHAPE_LAMALEF_AUTO 0x10000
185
186/** Bit mask for memory options. @stable ICU 2.0 */
187#define U_SHAPE_LENGTH_MASK 0x10003 /* Changed old value 3 */
188
189
190/**
191 * Bit mask for LamAlef memory options.
192 * @stable ICU 4.2
193 */
194#define U_SHAPE_LAMALEF_MASK 0x10003 /* updated */
195
196/** Direction indicator: the source is in logical (keyboard) order. @stable ICU 2.0 */
197#define U_SHAPE_TEXT_DIRECTION_LOGICAL 0
198
199/**
200 * Direction indicator:
201 * the source is in visual RTL order,
202 * the rightmost displayed character stored first.
203 * This option is an alias to U_SHAPE_TEXT_DIRECTION_LOGICAL
204 * @stable ICU 4.2
205 */
206#define U_SHAPE_TEXT_DIRECTION_VISUAL_RTL 0
207
208/**
209 * Direction indicator:
210 * the source is in visual LTR order,
211 * the leftmost displayed character stored first.
212 * @stable ICU 2.0
213 */
214#define U_SHAPE_TEXT_DIRECTION_VISUAL_LTR 4
215
216/** Bit mask for direction indicators. @stable ICU 2.0 */
217#define U_SHAPE_TEXT_DIRECTION_MASK 4
218
219
220/** Letter shaping option: do not perform letter shaping. @stable ICU 2.0 */
221#define U_SHAPE_LETTERS_NOOP 0
222
223/** Letter shaping option: replace abstract letter characters by "shaped" ones. @stable ICU 2.0 */
224#define U_SHAPE_LETTERS_SHAPE 8
225
226/** Letter shaping option: replace "shaped" letter characters by abstract ones. @stable ICU 2.0 */
227#define U_SHAPE_LETTERS_UNSHAPE 0x10
228
229/**
230 * Letter shaping option: replace abstract letter characters by "shaped" ones.
231 * The only difference with U_SHAPE_LETTERS_SHAPE is that Tashkeel letters
232 * are always "shaped" into the isolated form instead of the medial form
233 * (selecting code points from the Arabic Presentation Forms-B block).
234 * @stable ICU 2.0
235 */
236#define U_SHAPE_LETTERS_SHAPE_TASHKEEL_ISOLATED 0x18
237
238
239/** Bit mask for letter shaping options. @stable ICU 2.0 */
240#define U_SHAPE_LETTERS_MASK 0x18
241
242
243/** Digit shaping option: do not perform digit shaping. @stable ICU 2.0 */
244#define U_SHAPE_DIGITS_NOOP 0
245
246/**
247 * Digit shaping option:
248 * Replace European digits (U+0030...) by Arabic-Indic digits.
249 * @stable ICU 2.0
250 */
251#define U_SHAPE_DIGITS_EN2AN 0x20
252
253/**
254 * Digit shaping option:
255 * Replace Arabic-Indic digits by European digits (U+0030...).
256 * @stable ICU 2.0
257 */
258#define U_SHAPE_DIGITS_AN2EN 0x40
259
260/**
261 * Digit shaping option:
262 * Replace European digits (U+0030...) by Arabic-Indic digits if the most recent
263 * strongly directional character is an Arabic letter
264 * (<code>u_charDirection()</code> result <code>U_RIGHT_TO_LEFT_ARABIC</code> [AL]).<br>
265 * The direction of "preceding" depends on the direction indicator option.
266 * For the first characters, the preceding strongly directional character
267 * (initial state) is assumed to be not an Arabic letter
268 * (it is <code>U_LEFT_TO_RIGHT</code> [L] or <code>U_RIGHT_TO_LEFT</code> [R]).
269 * @stable ICU 2.0
270 */
271#define U_SHAPE_DIGITS_ALEN2AN_INIT_LR 0x60
272
273/**
274 * Digit shaping option:
275 * Replace European digits (U+0030...) by Arabic-Indic digits if the most recent
276 * strongly directional character is an Arabic letter
277 * (<code>u_charDirection()</code> result <code>U_RIGHT_TO_LEFT_ARABIC</code> [AL]).<br>
278 * The direction of "preceding" depends on the direction indicator option.
279 * For the first characters, the preceding strongly directional character
280 * (initial state) is assumed to be an Arabic letter.
281 * @stable ICU 2.0
282 */
283#define U_SHAPE_DIGITS_ALEN2AN_INIT_AL 0x80
284
285/** Not a valid option value. May be replaced by a new option. @stable ICU 2.0 */
286#define U_SHAPE_DIGITS_RESERVED 0xa0
287
288/** Bit mask for digit shaping options. @stable ICU 2.0 */
289#define U_SHAPE_DIGITS_MASK 0xe0
290
291
292/** Digit type option: Use Arabic-Indic digits (U+0660...U+0669). @stable ICU 2.0 */
293#define U_SHAPE_DIGIT_TYPE_AN 0
294
295/** Digit type option: Use Eastern (Extended) Arabic-Indic digits (U+06f0...U+06f9). @stable ICU 2.0 */
296#define U_SHAPE_DIGIT_TYPE_AN_EXTENDED 0x100
297
298/** Not a valid option value. May be replaced by a new option. @stable ICU 2.0 */
299#define U_SHAPE_DIGIT_TYPE_RESERVED 0x200
300
301/** Bit mask for digit type options. @stable ICU 2.0 */
302#define U_SHAPE_DIGIT_TYPE_MASK 0x300 /* I need to change this from 0x3f00 to 0x300 */
303
304/**
305 * Tashkeel aggregation option:
306 * Replaces any combination of U+0651 with one of
307 * U+064C, U+064D, U+064E, U+064F, U+0650 with
308 * U+FC5E, U+FC5F, U+FC60, U+FC61, U+FC62 consecutively.
309 * @stable ICU 3.6
310 */
311#define U_SHAPE_AGGREGATE_TASHKEEL 0x4000
312/** Tashkeel aggregation option: do not aggregate tashkeels. @stable ICU 3.6 */
313#define U_SHAPE_AGGREGATE_TASHKEEL_NOOP 0
314/** Bit mask for tashkeel aggregation. @stable ICU 3.6 */
315#define U_SHAPE_AGGREGATE_TASHKEEL_MASK 0x4000
316
317/**
318 * Presentation form option:
319 * Don't replace Arabic Presentation Forms-A and Arabic Presentation Forms-B
320 * characters with 0+06xx characters, before shaping.
321 * @stable ICU 3.6
322 */
323#define U_SHAPE_PRESERVE_PRESENTATION 0x8000
324/** Presentation form option:
325 * Replace Arabic Presentation Forms-A and Arabic Presentationo Forms-B with
326 * their unshaped correspondents in range 0+06xx, before shaping.
327 * @stable ICU 3.6
328 */
329#define U_SHAPE_PRESERVE_PRESENTATION_NOOP 0
330/** Bit mask for preserve presentation form. @stable ICU 3.6 */
331#define U_SHAPE_PRESERVE_PRESENTATION_MASK 0x8000
332
333/* Seen Tail option */
334/**
335 * Memory option: the result must have the same length as the source.
336 * Shaping mode: The SEEN family character will expand into two characters using space near
337 * the SEEN family character(i.e. the space after the character).
338 * If there are no spaces found, an error U_NO_SPACE_AVAILABLE (as defined in utypes.h)
339 * will be set in pErrorCode
340 *
341 * De-shaping mode: Any Seen character followed by Tail character will be
342 * replaced by one cell Seen and a space will replace the Tail.
343 * Affects: Seen options
344 * @stable ICU 4.2
345 */
346#define U_SHAPE_SEEN_TWOCELL_NEAR 0x200000
347
348/**
349 * Bit mask for Seen memory options.
350 * @stable ICU 4.2
351 */
352#define U_SHAPE_SEEN_MASK 0x700000
353
354/* YehHamza option */
355/**
356 * Memory option: the result must have the same length as the source.
357 * Shaping mode: The YEHHAMZA character will expand into two characters using space near it
358 * (i.e. the space after the character
359 * If there are no spaces found, an error U_NO_SPACE_AVAILABLE (as defined in utypes.h)
360 * will be set in pErrorCode
361 *
362 * De-shaping mode: Any Yeh (final or isolated) character followed by Hamza character will be
363 * replaced by one cell YehHamza and space will replace the Hamza.
364 * Affects: YehHamza options
365 * @stable ICU 4.2
366 */
367#define U_SHAPE_YEHHAMZA_TWOCELL_NEAR 0x1000000
368
369
370/**
371 * Bit mask for YehHamza memory options.
372 * @stable ICU 4.2
373 */
374#define U_SHAPE_YEHHAMZA_MASK 0x3800000
375
376/* New Tashkeel options */
377/**
378 * Memory option: the result must have the same length as the source.
379 * Shaping mode: Tashkeel characters will be replaced by spaces.
380 * Spaces will be placed at beginning of the buffer
381 *
382 * De-shaping mode: N/A
383 * Affects: Tashkeel options
384 * @stable ICU 4.2
385 */
386#define U_SHAPE_TASHKEEL_BEGIN 0x40000
387
388/**
389 * Memory option: the result must have the same length as the source.
390 * Shaping mode: Tashkeel characters will be replaced by spaces.
391 * Spaces will be placed at end of the buffer
392 *
393 * De-shaping mode: N/A
394 * Affects: Tashkeel options
395 * @stable ICU 4.2
396 */
397#define U_SHAPE_TASHKEEL_END 0x60000
398
399/**
400 * Memory option: allow the result to have a different length than the source.
401 * Shaping mode: Tashkeel characters will be removed, buffer length will shrink.
402 * De-shaping mode: N/A
403 *
404 * Affect: Tashkeel options
405 * @stable ICU 4.2
406 */
407#define U_SHAPE_TASHKEEL_RESIZE 0x80000
408
409/**
410 * Memory option: the result must have the same length as the source.
411 * Shaping mode: Tashkeel characters will be replaced by Tatweel if it is connected to adjacent
412 * characters (i.e. shaped on Tatweel) or replaced by space if it is not connected.
413 *
414 * De-shaping mode: N/A
415 * Affects: YehHamza options
416 * @stable ICU 4.2
417 */
418#define U_SHAPE_TASHKEEL_REPLACE_BY_TATWEEL 0xC0000
419
420/**
421 * Bit mask for Tashkeel replacement with Space or Tatweel memory options.
422 * @stable ICU 4.2
423 */
424#define U_SHAPE_TASHKEEL_MASK 0xE0000
425
426
427/* Space location Control options */
428/**
429 * This option affect the meaning of BEGIN and END options. if this option is not used the default
430 * for BEGIN and END will be as following:
431 * The Default (for both Visual LTR, Visual RTL and Logical Text)
432 * 1. BEGIN always refers to the start address of physical memory.
433 * 2. END always refers to the end address of physical memory.
434 *
435 * If this option is used it will swap the meaning of BEGIN and END only for Visual LTR text.
436 *
437 * The effect on BEGIN and END Memory Options will be as following:
438 * A. BEGIN For Visual LTR text: This will be the beginning (right side) of the visual text(
439 * corresponding to the physical memory address end for Visual LTR text, Same as END in
440 * default behavior)
441 * B. BEGIN For Logical text: Same as BEGIN in default behavior.
442 * C. END For Visual LTR text: This will be the end (left side) of the visual text (corresponding
443 * to the physical memory address beginning for Visual LTR text, Same as BEGIN in default behavior.
444 * D. END For Logical text: Same as END in default behavior).
445 * Affects: All LamAlef BEGIN, END and AUTO options.
446 * @stable ICU 4.2
447 */
448#define U_SHAPE_SPACES_RELATIVE_TO_TEXT_BEGIN_END 0x4000000
449
450/**
451 * Bit mask for swapping BEGIN and END for Visual LTR text
452 * @stable ICU 4.2
453 */
454#define U_SHAPE_SPACES_RELATIVE_TO_TEXT_MASK 0x4000000
455
456/**
457 * If this option is used, shaping will use the new Unicode code point for TAIL (i.e. 0xFE73).
458 * If this option is not specified (Default), old unofficial Unicode TAIL code point is used (i.e. 0x200B)
459 * De-shaping will not use this option as it will always search for both the new Unicode code point for the
460 * TAIL (i.e. 0xFE73) or the old unofficial Unicode TAIL code point (i.e. 0x200B) and de-shape the
461 * Seen-Family letter accordingly.
462 *
463 * Shaping Mode: Only shaping.
464 * De-shaping Mode: N/A.
465 * Affects: All Seen options
466 * @stable ICU 4.8
467 */
468#define U_SHAPE_TAIL_NEW_UNICODE 0x8000000
469
470/**
471 * Bit mask for new Unicode Tail option
472 * @stable ICU 4.8
473 */
474#define U_SHAPE_TAIL_TYPE_MASK 0x8000000
475
476#endif
477