1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/common/unicode/ushape.h Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,474 @@
1.4 +/*
1.5 +******************************************************************************
1.6 +*
1.7 +* Copyright (C) 2000-2012, International Business Machines
1.8 +* Corporation and others. All Rights Reserved.
1.9 +*
1.10 +******************************************************************************
1.11 +* file name: ushape.h
1.12 +* encoding: US-ASCII
1.13 +* tab size: 8 (not used)
1.14 +* indentation:4
1.15 +*
1.16 +* created on: 2000jun29
1.17 +* created by: Markus W. Scherer
1.18 +*/
1.19 +
1.20 +#ifndef __USHAPE_H__
1.21 +#define __USHAPE_H__
1.22 +
1.23 +#include "unicode/utypes.h"
1.24 +
1.25 +/**
1.26 + * \file
1.27 + * \brief C API: Arabic shaping
1.28 + *
1.29 + */
1.30 +
1.31 +/**
1.32 + * Shape Arabic text on a character basis.
1.33 + *
1.34 + * <p>This function performs basic operations for "shaping" Arabic text. It is most
1.35 + * useful for use with legacy data formats and legacy display technology
1.36 + * (simple terminals). All operations are performed on Unicode characters.</p>
1.37 + *
1.38 + * <p>Text-based shaping means that some character code points in the text are
1.39 + * replaced by others depending on the context. It transforms one kind of text
1.40 + * into another. In comparison, modern displays for Arabic text select
1.41 + * appropriate, context-dependent font glyphs for each text element, which means
1.42 + * that they transform text into a glyph vector.</p>
1.43 + *
1.44 + * <p>Text transformations are necessary when modern display technology is not
1.45 + * available or when text needs to be transformed to or from legacy formats that
1.46 + * use "shaped" characters. Since the Arabic script is cursive, connecting
1.47 + * adjacent letters to each other, computers select images for each letter based
1.48 + * on the surrounding letters. This usually results in four images per Arabic
1.49 + * letter: initial, middle, final, and isolated forms. In Unicode, on the other
1.50 + * hand, letters are normally stored abstract, and a display system is expected
1.51 + * to select the necessary glyphs. (This makes searching and other text
1.52 + * processing easier because the same letter has only one code.) It is possible
1.53 + * to mimic this with text transformations because there are characters in
1.54 + * Unicode that are rendered as letters with a specific shape
1.55 + * (or cursive connectivity). They were included for interoperability with
1.56 + * legacy systems and codepages, and for unsophisticated display systems.</p>
1.57 + *
1.58 + * <p>A second kind of text transformations is supported for Arabic digits:
1.59 + * For compatibility with legacy codepages that only include European digits,
1.60 + * it is possible to replace one set of digits by another, changing the
1.61 + * character code points. These operations can be performed for either
1.62 + * Arabic-Indic Digits (U+0660...U+0669) or Eastern (Extended) Arabic-Indic
1.63 + * digits (U+06f0...U+06f9).</p>
1.64 + *
1.65 + * <p>Some replacements may result in more or fewer characters (code points).
1.66 + * By default, this means that the destination buffer may receive text with a
1.67 + * length different from the source length. Some legacy systems rely on the
1.68 + * length of the text to be constant. They expect extra spaces to be added
1.69 + * or consumed either next to the affected character or at the end of the
1.70 + * text.</p>
1.71 + *
1.72 + * <p>For details about the available operations, see the description of the
1.73 + * <code>U_SHAPE_...</code> options.</p>
1.74 + *
1.75 + * @param source The input text.
1.76 + *
1.77 + * @param sourceLength The number of UChars in <code>source</code>.
1.78 + *
1.79 + * @param dest The destination buffer that will receive the results of the
1.80 + * requested operations. It may be <code>NULL</code> only if
1.81 + * <code>destSize</code> is 0. The source and destination must not
1.82 + * overlap.
1.83 + *
1.84 + * @param destSize The size (capacity) of the destination buffer in UChars.
1.85 + * If <code>destSize</code> is 0, then no output is produced,
1.86 + * but the necessary buffer size is returned ("preflighting").
1.87 + *
1.88 + * @param options This is a 32-bit set of flags that specify the operations
1.89 + * that are performed on the input text. If no error occurs,
1.90 + * then the result will always be written to the destination
1.91 + * buffer.
1.92 + *
1.93 + * @param pErrorCode must be a valid pointer to an error code value,
1.94 + * which must not indicate a failure before the function call.
1.95 + *
1.96 + * @return The number of UChars written to the destination buffer.
1.97 + * If an error occured, then no output was written, or it may be
1.98 + * incomplete. If <code>U_BUFFER_OVERFLOW_ERROR</code> is set, then
1.99 + * the return value indicates the necessary destination buffer size.
1.100 + * @stable ICU 2.0
1.101 + */
1.102 +U_STABLE int32_t U_EXPORT2
1.103 +u_shapeArabic(const UChar *source, int32_t sourceLength,
1.104 + UChar *dest, int32_t destSize,
1.105 + uint32_t options,
1.106 + UErrorCode *pErrorCode);
1.107 +
1.108 +/**
1.109 + * Memory option: allow the result to have a different length than the source.
1.110 + * Affects: LamAlef options
1.111 + * @stable ICU 2.0
1.112 + */
1.113 +#define U_SHAPE_LENGTH_GROW_SHRINK 0
1.114 +
1.115 +/**
1.116 + * Memory option: allow the result to have a different length than the source.
1.117 + * Affects: LamAlef options
1.118 + * This option is an alias to U_SHAPE_LENGTH_GROW_SHRINK
1.119 + * @stable ICU 4.2
1.120 + */
1.121 +#define U_SHAPE_LAMALEF_RESIZE 0
1.122 +
1.123 +/**
1.124 + * Memory option: the result must have the same length as the source.
1.125 + * If more room is necessary, then try to consume spaces next to modified characters.
1.126 + * @stable ICU 2.0
1.127 + */
1.128 +#define U_SHAPE_LENGTH_FIXED_SPACES_NEAR 1
1.129 +
1.130 +/**
1.131 + * Memory option: the result must have the same length as the source.
1.132 + * If more room is necessary, then try to consume spaces next to modified characters.
1.133 + * Affects: LamAlef options
1.134 + * This option is an alias to U_SHAPE_LENGTH_FIXED_SPACES_NEAR
1.135 + * @stable ICU 4.2
1.136 + */
1.137 +#define U_SHAPE_LAMALEF_NEAR 1
1.138 +
1.139 +/**
1.140 + * Memory option: the result must have the same length as the source.
1.141 + * If more room is necessary, then try to consume spaces at the end of the text.
1.142 + * @stable ICU 2.0
1.143 + */
1.144 +#define U_SHAPE_LENGTH_FIXED_SPACES_AT_END 2
1.145 +
1.146 +/**
1.147 + * Memory option: the result must have the same length as the source.
1.148 + * If more room is necessary, then try to consume spaces at the end of the text.
1.149 + * Affects: LamAlef options
1.150 + * This option is an alias to U_SHAPE_LENGTH_FIXED_SPACES_AT_END
1.151 + * @stable ICU 4.2
1.152 + */
1.153 +#define U_SHAPE_LAMALEF_END 2
1.154 +
1.155 +/**
1.156 + * Memory option: the result must have the same length as the source.
1.157 + * If more room is necessary, then try to consume spaces at the beginning of the text.
1.158 + * @stable ICU 2.0
1.159 + */
1.160 +#define U_SHAPE_LENGTH_FIXED_SPACES_AT_BEGINNING 3
1.161 +
1.162 +/**
1.163 + * Memory option: the result must have the same length as the source.
1.164 + * If more room is necessary, then try to consume spaces at the beginning of the text.
1.165 + * Affects: LamAlef options
1.166 + * This option is an alias to U_SHAPE_LENGTH_FIXED_SPACES_AT_BEGINNING
1.167 + * @stable ICU 4.2
1.168 + */
1.169 +#define U_SHAPE_LAMALEF_BEGIN 3
1.170 +
1.171 +
1.172 +/**
1.173 + * Memory option: the result must have the same length as the source.
1.174 + * Shaping Mode: For each LAMALEF character found, expand LAMALEF using space at end.
1.175 + * If there is no space at end, use spaces at beginning of the buffer. If there
1.176 + * is no space at beginning of the buffer, use spaces at the near (i.e. the space
1.177 + * after the LAMALEF character).
1.178 + * If there are no spaces found, an error U_NO_SPACE_AVAILABLE (as defined in utypes.h)
1.179 + * will be set in pErrorCode
1.180 + *
1.181 + * Deshaping Mode: Perform the same function as the flag equals U_SHAPE_LAMALEF_END.
1.182 + * Affects: LamAlef options
1.183 + * @stable ICU 4.2
1.184 + */
1.185 +#define U_SHAPE_LAMALEF_AUTO 0x10000
1.186 +
1.187 +/** Bit mask for memory options. @stable ICU 2.0 */
1.188 +#define U_SHAPE_LENGTH_MASK 0x10003 /* Changed old value 3 */
1.189 +
1.190 +
1.191 +/**
1.192 + * Bit mask for LamAlef memory options.
1.193 + * @stable ICU 4.2
1.194 + */
1.195 +#define U_SHAPE_LAMALEF_MASK 0x10003 /* updated */
1.196 +
1.197 +/** Direction indicator: the source is in logical (keyboard) order. @stable ICU 2.0 */
1.198 +#define U_SHAPE_TEXT_DIRECTION_LOGICAL 0
1.199 +
1.200 +/**
1.201 + * Direction indicator:
1.202 + * the source is in visual RTL order,
1.203 + * the rightmost displayed character stored first.
1.204 + * This option is an alias to U_SHAPE_TEXT_DIRECTION_LOGICAL
1.205 + * @stable ICU 4.2
1.206 + */
1.207 +#define U_SHAPE_TEXT_DIRECTION_VISUAL_RTL 0
1.208 +
1.209 +/**
1.210 + * Direction indicator:
1.211 + * the source is in visual LTR order,
1.212 + * the leftmost displayed character stored first.
1.213 + * @stable ICU 2.0
1.214 + */
1.215 +#define U_SHAPE_TEXT_DIRECTION_VISUAL_LTR 4
1.216 +
1.217 +/** Bit mask for direction indicators. @stable ICU 2.0 */
1.218 +#define U_SHAPE_TEXT_DIRECTION_MASK 4
1.219 +
1.220 +
1.221 +/** Letter shaping option: do not perform letter shaping. @stable ICU 2.0 */
1.222 +#define U_SHAPE_LETTERS_NOOP 0
1.223 +
1.224 +/** Letter shaping option: replace abstract letter characters by "shaped" ones. @stable ICU 2.0 */
1.225 +#define U_SHAPE_LETTERS_SHAPE 8
1.226 +
1.227 +/** Letter shaping option: replace "shaped" letter characters by abstract ones. @stable ICU 2.0 */
1.228 +#define U_SHAPE_LETTERS_UNSHAPE 0x10
1.229 +
1.230 +/**
1.231 + * Letter shaping option: replace abstract letter characters by "shaped" ones.
1.232 + * The only difference with U_SHAPE_LETTERS_SHAPE is that Tashkeel letters
1.233 + * are always "shaped" into the isolated form instead of the medial form
1.234 + * (selecting code points from the Arabic Presentation Forms-B block).
1.235 + * @stable ICU 2.0
1.236 + */
1.237 +#define U_SHAPE_LETTERS_SHAPE_TASHKEEL_ISOLATED 0x18
1.238 +
1.239 +
1.240 +/** Bit mask for letter shaping options. @stable ICU 2.0 */
1.241 +#define U_SHAPE_LETTERS_MASK 0x18
1.242 +
1.243 +
1.244 +/** Digit shaping option: do not perform digit shaping. @stable ICU 2.0 */
1.245 +#define U_SHAPE_DIGITS_NOOP 0
1.246 +
1.247 +/**
1.248 + * Digit shaping option:
1.249 + * Replace European digits (U+0030...) by Arabic-Indic digits.
1.250 + * @stable ICU 2.0
1.251 + */
1.252 +#define U_SHAPE_DIGITS_EN2AN 0x20
1.253 +
1.254 +/**
1.255 + * Digit shaping option:
1.256 + * Replace Arabic-Indic digits by European digits (U+0030...).
1.257 + * @stable ICU 2.0
1.258 + */
1.259 +#define U_SHAPE_DIGITS_AN2EN 0x40
1.260 +
1.261 +/**
1.262 + * Digit shaping option:
1.263 + * Replace European digits (U+0030...) by Arabic-Indic digits if the most recent
1.264 + * strongly directional character is an Arabic letter
1.265 + * (<code>u_charDirection()</code> result <code>U_RIGHT_TO_LEFT_ARABIC</code> [AL]).<br>
1.266 + * The direction of "preceding" depends on the direction indicator option.
1.267 + * For the first characters, the preceding strongly directional character
1.268 + * (initial state) is assumed to be not an Arabic letter
1.269 + * (it is <code>U_LEFT_TO_RIGHT</code> [L] or <code>U_RIGHT_TO_LEFT</code> [R]).
1.270 + * @stable ICU 2.0
1.271 + */
1.272 +#define U_SHAPE_DIGITS_ALEN2AN_INIT_LR 0x60
1.273 +
1.274 +/**
1.275 + * Digit shaping option:
1.276 + * Replace European digits (U+0030...) by Arabic-Indic digits if the most recent
1.277 + * strongly directional character is an Arabic letter
1.278 + * (<code>u_charDirection()</code> result <code>U_RIGHT_TO_LEFT_ARABIC</code> [AL]).<br>
1.279 + * The direction of "preceding" depends on the direction indicator option.
1.280 + * For the first characters, the preceding strongly directional character
1.281 + * (initial state) is assumed to be an Arabic letter.
1.282 + * @stable ICU 2.0
1.283 + */
1.284 +#define U_SHAPE_DIGITS_ALEN2AN_INIT_AL 0x80
1.285 +
1.286 +/** Not a valid option value. May be replaced by a new option. @stable ICU 2.0 */
1.287 +#define U_SHAPE_DIGITS_RESERVED 0xa0
1.288 +
1.289 +/** Bit mask for digit shaping options. @stable ICU 2.0 */
1.290 +#define U_SHAPE_DIGITS_MASK 0xe0
1.291 +
1.292 +
1.293 +/** Digit type option: Use Arabic-Indic digits (U+0660...U+0669). @stable ICU 2.0 */
1.294 +#define U_SHAPE_DIGIT_TYPE_AN 0
1.295 +
1.296 +/** Digit type option: Use Eastern (Extended) Arabic-Indic digits (U+06f0...U+06f9). @stable ICU 2.0 */
1.297 +#define U_SHAPE_DIGIT_TYPE_AN_EXTENDED 0x100
1.298 +
1.299 +/** Not a valid option value. May be replaced by a new option. @stable ICU 2.0 */
1.300 +#define U_SHAPE_DIGIT_TYPE_RESERVED 0x200
1.301 +
1.302 +/** Bit mask for digit type options. @stable ICU 2.0 */
1.303 +#define U_SHAPE_DIGIT_TYPE_MASK 0x300 /* I need to change this from 0x3f00 to 0x300 */
1.304 +
1.305 +/**
1.306 + * Tashkeel aggregation option:
1.307 + * Replaces any combination of U+0651 with one of
1.308 + * U+064C, U+064D, U+064E, U+064F, U+0650 with
1.309 + * U+FC5E, U+FC5F, U+FC60, U+FC61, U+FC62 consecutively.
1.310 + * @stable ICU 3.6
1.311 + */
1.312 +#define U_SHAPE_AGGREGATE_TASHKEEL 0x4000
1.313 +/** Tashkeel aggregation option: do not aggregate tashkeels. @stable ICU 3.6 */
1.314 +#define U_SHAPE_AGGREGATE_TASHKEEL_NOOP 0
1.315 +/** Bit mask for tashkeel aggregation. @stable ICU 3.6 */
1.316 +#define U_SHAPE_AGGREGATE_TASHKEEL_MASK 0x4000
1.317 +
1.318 +/**
1.319 + * Presentation form option:
1.320 + * Don't replace Arabic Presentation Forms-A and Arabic Presentation Forms-B
1.321 + * characters with 0+06xx characters, before shaping.
1.322 + * @stable ICU 3.6
1.323 + */
1.324 +#define U_SHAPE_PRESERVE_PRESENTATION 0x8000
1.325 +/** Presentation form option:
1.326 + * Replace Arabic Presentation Forms-A and Arabic Presentationo Forms-B with
1.327 + * their unshaped correspondants in range 0+06xx, before shaping.
1.328 + * @stable ICU 3.6
1.329 + */
1.330 +#define U_SHAPE_PRESERVE_PRESENTATION_NOOP 0
1.331 +/** Bit mask for preserve presentation form. @stable ICU 3.6 */
1.332 +#define U_SHAPE_PRESERVE_PRESENTATION_MASK 0x8000
1.333 +
1.334 +/* Seen Tail option */
1.335 +/**
1.336 + * Memory option: the result must have the same length as the source.
1.337 + * Shaping mode: The SEEN family character will expand into two characters using space near
1.338 + * the SEEN family character(i.e. the space after the character).
1.339 + * If there are no spaces found, an error U_NO_SPACE_AVAILABLE (as defined in utypes.h)
1.340 + * will be set in pErrorCode
1.341 + *
1.342 + * De-shaping mode: Any Seen character followed by Tail character will be
1.343 + * replaced by one cell Seen and a space will replace the Tail.
1.344 + * Affects: Seen options
1.345 + * @stable ICU 4.2
1.346 + */
1.347 +#define U_SHAPE_SEEN_TWOCELL_NEAR 0x200000
1.348 +
1.349 +/**
1.350 + * Bit mask for Seen memory options.
1.351 + * @stable ICU 4.2
1.352 + */
1.353 +#define U_SHAPE_SEEN_MASK 0x700000
1.354 +
1.355 +/* YehHamza option */
1.356 +/**
1.357 + * Memory option: the result must have the same length as the source.
1.358 + * Shaping mode: The YEHHAMZA character will expand into two characters using space near it
1.359 + * (i.e. the space after the character
1.360 + * If there are no spaces found, an error U_NO_SPACE_AVAILABLE (as defined in utypes.h)
1.361 + * will be set in pErrorCode
1.362 + *
1.363 + * De-shaping mode: Any Yeh (final or isolated) character followed by Hamza character will be
1.364 + * replaced by one cell YehHamza and space will replace the Hamza.
1.365 + * Affects: YehHamza options
1.366 + * @stable ICU 4.2
1.367 + */
1.368 +#define U_SHAPE_YEHHAMZA_TWOCELL_NEAR 0x1000000
1.369 +
1.370 +
1.371 +/**
1.372 + * Bit mask for YehHamza memory options.
1.373 + * @stable ICU 4.2
1.374 + */
1.375 +#define U_SHAPE_YEHHAMZA_MASK 0x3800000
1.376 +
1.377 +/* New Tashkeel options */
1.378 +/**
1.379 + * Memory option: the result must have the same length as the source.
1.380 + * Shaping mode: Tashkeel characters will be replaced by spaces.
1.381 + * Spaces will be placed at beginning of the buffer
1.382 + *
1.383 + * De-shaping mode: N/A
1.384 + * Affects: Tashkeel options
1.385 + * @stable ICU 4.2
1.386 + */
1.387 +#define U_SHAPE_TASHKEEL_BEGIN 0x40000
1.388 +
1.389 +/**
1.390 + * Memory option: the result must have the same length as the source.
1.391 + * Shaping mode: Tashkeel characters will be replaced by spaces.
1.392 + * Spaces will be placed at end of the buffer
1.393 + *
1.394 + * De-shaping mode: N/A
1.395 + * Affects: Tashkeel options
1.396 + * @stable ICU 4.2
1.397 + */
1.398 +#define U_SHAPE_TASHKEEL_END 0x60000
1.399 +
1.400 +/**
1.401 + * Memory option: allow the result to have a different length than the source.
1.402 + * Shaping mode: Tashkeel characters will be removed, buffer length will shrink.
1.403 + * De-shaping mode: N/A
1.404 + *
1.405 + * Affect: Tashkeel options
1.406 + * @stable ICU 4.2
1.407 + */
1.408 +#define U_SHAPE_TASHKEEL_RESIZE 0x80000
1.409 +
1.410 +/**
1.411 + * Memory option: the result must have the same length as the source.
1.412 + * Shaping mode: Tashkeel characters will be replaced by Tatweel if it is connected to adjacent
1.413 + * characters (i.e. shaped on Tatweel) or replaced by space if it is not connected.
1.414 + *
1.415 + * De-shaping mode: N/A
1.416 + * Affects: YehHamza options
1.417 + * @stable ICU 4.2
1.418 + */
1.419 +#define U_SHAPE_TASHKEEL_REPLACE_BY_TATWEEL 0xC0000
1.420 +
1.421 +/**
1.422 + * Bit mask for Tashkeel replacement with Space or Tatweel memory options.
1.423 + * @stable ICU 4.2
1.424 + */
1.425 +#define U_SHAPE_TASHKEEL_MASK 0xE0000
1.426 +
1.427 +
1.428 +/* Space location Control options */
1.429 +/**
1.430 + * This option affect the meaning of BEGIN and END options. if this option is not used the default
1.431 + * for BEGIN and END will be as following:
1.432 + * The Default (for both Visual LTR, Visual RTL and Logical Text)
1.433 + * 1. BEGIN always refers to the start address of physical memory.
1.434 + * 2. END always refers to the end address of physical memory.
1.435 + *
1.436 + * If this option is used it will swap the meaning of BEGIN and END only for Visual LTR text.
1.437 + *
1.438 + * The effect on BEGIN and END Memory Options will be as following:
1.439 + * A. BEGIN For Visual LTR text: This will be the beginning (right side) of the visual text(
1.440 + * corresponding to the physical memory address end for Visual LTR text, Same as END in
1.441 + * default behavior)
1.442 + * B. BEGIN For Logical text: Same as BEGIN in default behavior.
1.443 + * C. END For Visual LTR text: This will be the end (left side) of the visual text (corresponding
1.444 + * to the physical memory address beginning for Visual LTR text, Same as BEGIN in default behavior.
1.445 + * D. END For Logical text: Same as END in default behavior).
1.446 + * Affects: All LamAlef BEGIN, END and AUTO options.
1.447 + * @stable ICU 4.2
1.448 + */
1.449 +#define U_SHAPE_SPACES_RELATIVE_TO_TEXT_BEGIN_END 0x4000000
1.450 +
1.451 +/**
1.452 + * Bit mask for swapping BEGIN and END for Visual LTR text
1.453 + * @stable ICU 4.2
1.454 + */
1.455 +#define U_SHAPE_SPACES_RELATIVE_TO_TEXT_MASK 0x4000000
1.456 +
1.457 +/**
1.458 + * If this option is used, shaping will use the new Unicode code point for TAIL (i.e. 0xFE73).
1.459 + * If this option is not specified (Default), old unofficial Unicode TAIL code point is used (i.e. 0x200B)
1.460 + * De-shaping will not use this option as it will always search for both the new Unicode code point for the
1.461 + * TAIL (i.e. 0xFE73) or the old unofficial Unicode TAIL code point (i.e. 0x200B) and de-shape the
1.462 + * Seen-Family letter accordingly.
1.463 + *
1.464 + * Shaping Mode: Only shaping.
1.465 + * De-shaping Mode: N/A.
1.466 + * Affects: All Seen options
1.467 + * @stable ICU 4.8
1.468 + */
1.469 +#define U_SHAPE_TAIL_NEW_UNICODE 0x8000000
1.470 +
1.471 +/**
1.472 + * Bit mask for new Unicode Tail option
1.473 + * @stable ICU 4.8
1.474 + */
1.475 +#define U_SHAPE_TAIL_TYPE_MASK 0x8000000
1.476 +
1.477 +#endif
