Home | History | Annotate | Download | only in unicode
      1 /*
      2 ******************************************************************************
      3 *
      4 *   Copyright (C) 1999-2010, International Business Machines
      5 *   Corporation and others.  All Rights Reserved.
      6 *
      7 ******************************************************************************
      8 *   file name:  ubidi.h
      9 *   encoding:   US-ASCII
     10 *   tab size:   8 (not used)
     11 *   indentation:4
     12 *
     13 *   created on: 1999jul27
     14 *   created by: Markus W. Scherer, updated by Matitiahu Allouche
     15 */
     16 
     17 #ifndef UBIDI_H
     18 #define UBIDI_H
     19 
     20 #include "unicode/utypes.h"
     21 #include "unicode/uchar.h"
     22 #include "unicode/localpointer.h"
     23 
     24 /**
     25  *\file
     26  * \brief C API: Bidi algorithm
     27  *
     28  * <h2>Bidi algorithm for ICU</h2>
     29  *
     30  * This is an implementation of the Unicode Bidirectional algorithm.
     31  * The algorithm is defined in the
     32  * <a href="http://www.unicode.org/unicode/reports/tr9/">Unicode Standard Annex #9</a>,
     33  * version 13, also described in The Unicode Standard, Version 4.0 .<p>
     34  *
     35  * Note: Libraries that perform a bidirectional algorithm and
     36  * reorder strings accordingly are sometimes called "Storage Layout Engines".
     37  * ICU's Bidi and shaping (u_shapeArabic()) APIs can be used at the core of such
     38  * "Storage Layout Engines".
     39  *
     40  * <h3>General remarks about the API:</h3>
     41  *
     42  * In functions with an error code parameter,
     43  * the <code>pErrorCode</code> pointer must be valid
     44  * and the value that it points to must not indicate a failure before
     45  * the function call. Otherwise, the function returns immediately.
     46  * After the function call, the value indicates success or failure.<p>
     47  *
     48  * The &quot;limit&quot; of a sequence of characters is the position just after their
     49  * last character, i.e., one more than that position.<p>
     50  *
     51  * Some of the API functions provide access to &quot;runs&quot;.
     52  * Such a &quot;run&quot; is defined as a sequence of characters
     53  * that are at the same embedding level
     54  * after performing the Bidi algorithm.<p>
     55  *
     56  * @author Markus W. Scherer
     57  * @version 1.0
     58  *
     59  *
     60  * <h4> Sample code for the ICU Bidi API </h4>
     61  *
     62  * <h5>Rendering a paragraph with the ICU Bidi API</h5>
     63  *
     64  * This is (hypothetical) sample code that illustrates
     65  * how the ICU Bidi API could be used to render a paragraph of text.
     66  * Rendering code depends highly on the graphics system,
     67  * therefore this sample code must make a lot of assumptions,
     68  * which may or may not match any existing graphics system's properties.
     69  *
     70  * <p>The basic assumptions are:</p>
     71  * <ul>
     72  * <li>Rendering is done from left to right on a horizontal line.</li>
     73  * <li>A run of single-style, unidirectional text can be rendered at once.</li>
     74  * <li>Such a run of text is passed to the graphics system with
     75  *     characters (code units) in logical order.</li>
     76  * <li>The line-breaking algorithm is very complicated
     77  *     and Locale-dependent -
     78  *     and therefore its implementation omitted from this sample code.</li>
     79  * </ul>
     80  *
     81  * <pre>
     82  * \code
     83  *#include "unicode/ubidi.h"
     84  *
     85  *typedef enum {
     86  *     styleNormal=0, styleSelected=1,
     87  *     styleBold=2, styleItalics=4,
     88  *     styleSuper=8, styleSub=16
     89  *} Style;
     90  *
     91  *typedef struct { int32_t limit; Style style; } StyleRun;
     92  *
     93  *int getTextWidth(const UChar *text, int32_t start, int32_t limit,
     94  *                  const StyleRun *styleRuns, int styleRunCount);
     95  *
     96  * // set *pLimit and *pStyleRunLimit for a line
     97  * // from text[start] and from styleRuns[styleRunStart]
     98  * // using ubidi_getLogicalRun(para, ...)
     99  *void getLineBreak(const UChar *text, int32_t start, int32_t *pLimit,
    100  *                  UBiDi *para,
    101  *                  const StyleRun *styleRuns, int styleRunStart, int *pStyleRunLimit,
    102  *                  int *pLineWidth);
    103  *
    104  * // render runs on a line sequentially, always from left to right
    105  *
    106  * // prepare rendering a new line
    107  * void startLine(UBiDiDirection textDirection, int lineWidth);
    108  *
    109  * // render a run of text and advance to the right by the run width
    110  * // the text[start..limit-1] is always in logical order
    111  * void renderRun(const UChar *text, int32_t start, int32_t limit,
    112  *               UBiDiDirection textDirection, Style style);
    113  *
    114  * // We could compute a cross-product
    115  * // from the style runs with the directional runs
    116  * // and then reorder it.
    117  * // Instead, here we iterate over each run type
    118  * // and render the intersections -
    119  * // with shortcuts in simple (and common) cases.
    120  * // renderParagraph() is the main function.
    121  *
    122  * // render a directional run with
    123  * // (possibly) multiple style runs intersecting with it
    124  * void renderDirectionalRun(const UChar *text,
    125  *                           int32_t start, int32_t limit,
    126  *                           UBiDiDirection direction,
    127  *                           const StyleRun *styleRuns, int styleRunCount) {
    128  *     int i;
    129  *
    130  *     // iterate over style runs
    131  *     if(direction==UBIDI_LTR) {
    132  *         int styleLimit;
    133  *
    134  *         for(i=0; i<styleRunCount; ++i) {
    135  *             styleLimit=styleRun[i].limit;
    136  *             if(start<styleLimit) {
    137  *                 if(styleLimit>limit) { styleLimit=limit; }
    138  *                 renderRun(text, start, styleLimit,
    139  *                           direction, styleRun[i].style);
    140  *                 if(styleLimit==limit) { break; }
    141  *                 start=styleLimit;
    142  *             }
    143  *         }
    144  *     } else {
    145  *         int styleStart;
    146  *
    147  *         for(i=styleRunCount-1; i>=0; --i) {
    148  *             if(i>0) {
    149  *                 styleStart=styleRun[i-1].limit;
    150  *             } else {
    151  *                 styleStart=0;
    152  *             }
    153  *             if(limit>=styleStart) {
    154  *                 if(styleStart<start) { styleStart=start; }
    155  *                 renderRun(text, styleStart, limit,
    156  *                           direction, styleRun[i].style);
    157  *                 if(styleStart==start) { break; }
    158  *                 limit=styleStart;
    159  *             }
    160  *         }
    161  *     }
    162  * }
    163  *
    164  * // the line object represents text[start..limit-1]
    165  * void renderLine(UBiDi *line, const UChar *text,
    166  *                 int32_t start, int32_t limit,
    167  *                 const StyleRun *styleRuns, int styleRunCount) {
    168  *     UBiDiDirection direction=ubidi_getDirection(line);
    169  *     if(direction!=UBIDI_MIXED) {
    170  *         // unidirectional
    171  *         if(styleRunCount<=1) {
    172  *             renderRun(text, start, limit, direction, styleRuns[0].style);
    173  *         } else {
    174  *             renderDirectionalRun(text, start, limit,
    175  *                                  direction, styleRuns, styleRunCount);
    176  *         }
    177  *     } else {
    178  *         // mixed-directional
    179  *         int32_t count, i, length;
    180  *         UBiDiLevel level;
    181  *
    182  *         count=ubidi_countRuns(para, pErrorCode);
    183  *         if(U_SUCCESS(*pErrorCode)) {
    184  *             if(styleRunCount<=1) {
    185  *                 Style style=styleRuns[0].style;
    186  *
    187  *                 // iterate over directional runs
    188  *                for(i=0; i<count; ++i) {
    189  *                    direction=ubidi_getVisualRun(para, i, &start, &length);
    190  *                     renderRun(text, start, start+length, direction, style);
    191  *                }
    192  *             } else {
    193  *                 int32_t j;
    194  *
    195  *                 // iterate over both directional and style runs
    196  *                 for(i=0; i<count; ++i) {
    197  *                     direction=ubidi_getVisualRun(line, i, &start, &length);
    198  *                     renderDirectionalRun(text, start, start+length,
    199  *                                          direction, styleRuns, styleRunCount);
    200  *                 }
    201  *             }
    202  *         }
    203  *     }
    204  * }
    205  *
    206  *void renderParagraph(const UChar *text, int32_t length,
    207  *                     UBiDiDirection textDirection,
    208  *                      const StyleRun *styleRuns, int styleRunCount,
    209  *                      int lineWidth,
    210  *                      UErrorCode *pErrorCode) {
    211  *     UBiDi *para;
    212  *
    213  *     if(pErrorCode==NULL || U_FAILURE(*pErrorCode) || length<=0) {
    214  *         return;
    215  *     }
    216  *
    217  *     para=ubidi_openSized(length, 0, pErrorCode);
    218  *     if(para==NULL) { return; }
    219  *
    220  *     ubidi_setPara(para, text, length,
    221  *                   textDirection ? UBIDI_DEFAULT_RTL : UBIDI_DEFAULT_LTR,
    222  *                   NULL, pErrorCode);
    223  *     if(U_SUCCESS(*pErrorCode)) {
    224  *         UBiDiLevel paraLevel=1&ubidi_getParaLevel(para);
    225  *         StyleRun styleRun={ length, styleNormal };
    226  *         int width;
    227  *
    228  *         if(styleRuns==NULL || styleRunCount<=0) {
    229  *            styleRunCount=1;
    230  *             styleRuns=&styleRun;
    231  *         }
    232  *
    233  *        // assume styleRuns[styleRunCount-1].limit>=length
    234  *
    235  *         width=getTextWidth(text, 0, length, styleRuns, styleRunCount);
    236  *         if(width<=lineWidth) {
    237  *             // everything fits onto one line
    238  *
    239  *            // prepare rendering a new line from either left or right
    240  *             startLine(paraLevel, width);
    241  *
    242  *             renderLine(para, text, 0, length,
    243  *                        styleRuns, styleRunCount);
    244  *         } else {
    245  *             UBiDi *line;
    246  *
    247  *             // we need to render several lines
    248  *             line=ubidi_openSized(length, 0, pErrorCode);
    249  *             if(line!=NULL) {
    250  *                 int32_t start=0, limit;
    251  *                 int styleRunStart=0, styleRunLimit;
    252  *
    253  *                 for(;;) {
    254  *                     limit=length;
    255  *                     styleRunLimit=styleRunCount;
    256  *                     getLineBreak(text, start, &limit, para,
    257  *                                  styleRuns, styleRunStart, &styleRunLimit,
    258  *                                 &width);
    259  *                     ubidi_setLine(para, start, limit, line, pErrorCode);
    260  *                     if(U_SUCCESS(*pErrorCode)) {
    261  *                         // prepare rendering a new line
    262  *                         // from either left or right
    263  *                         startLine(paraLevel, width);
    264  *
    265  *                         renderLine(line, text, start, limit,
    266  *                                    styleRuns+styleRunStart,
    267  *                                    styleRunLimit-styleRunStart);
    268  *                     }
    269  *                     if(limit==length) { break; }
    270  *                     start=limit;
    271  *                     styleRunStart=styleRunLimit-1;
    272  *                     if(start>=styleRuns[styleRunStart].limit) {
    273  *                         ++styleRunStart;
    274  *                     }
    275  *                 }
    276  *
    277  *                 ubidi_close(line);
    278  *             }
    279  *        }
    280  *    }
    281  *
    282  *     ubidi_close(para);
    283  *}
    284  *\endcode
    285  * </pre>
    286  */
    287 
    288 /*DOCXX_TAG*/
    289 /*@{*/
    290 
    291 /**
    292  * UBiDiLevel is the type of the level values in this
    293  * Bidi implementation.
    294  * It holds an embedding level and indicates the visual direction
    295  * by its bit&nbsp;0 (even/odd value).<p>
    296  *
    297  * It can also hold non-level values for the
    298  * <code>paraLevel</code> and <code>embeddingLevels</code>
    299  * arguments of <code>ubidi_setPara()</code>; there:
    300  * <ul>
    301  * <li>bit&nbsp;7 of an <code>embeddingLevels[]</code>
    302  * value indicates whether the using application is
    303  * specifying the level of a character to <i>override</i> whatever the
    304  * Bidi implementation would resolve it to.</li>
    305  * <li><code>paraLevel</code> can be set to the
    306  * pseudo-level values <code>UBIDI_DEFAULT_LTR</code>
    307  * and <code>UBIDI_DEFAULT_RTL</code>.</li>
    308  * </ul>
    309  *
    310  * @see ubidi_setPara
    311  *
    312  * <p>The related constants are not real, valid level values.
    313  * <code>UBIDI_DEFAULT_XXX</code> can be used to specify
    314  * a default for the paragraph level for
    315  * when the <code>ubidi_setPara()</code> function
    316  * shall determine it but there is no
    317  * strongly typed character in the input.<p>
    318  *
    319  * Note that the value for <code>UBIDI_DEFAULT_LTR</code> is even
    320  * and the one for <code>UBIDI_DEFAULT_RTL</code> is odd,
    321  * just like with normal LTR and RTL level values -
    322  * these special values are designed that way. Also, the implementation
    323  * assumes that UBIDI_MAX_EXPLICIT_LEVEL is odd.
    324  *
    325  * @see UBIDI_DEFAULT_LTR
    326  * @see UBIDI_DEFAULT_RTL
    327  * @see UBIDI_LEVEL_OVERRIDE
    328  * @see UBIDI_MAX_EXPLICIT_LEVEL
    329  * @stable ICU 2.0
    330  */
    331 typedef uint8_t UBiDiLevel;
    332 
    333 /** Paragraph level setting.<p>
    334  *
    335  * Constant indicating that the base direction depends on the first strong
    336  * directional character in the text according to the Unicode Bidirectional
    337  * Algorithm. If no strong directional character is present,
    338  * then set the paragraph level to 0 (left-to-right).<p>
    339  *
    340  * If this value is used in conjunction with reordering modes
    341  * <code>UBIDI_REORDER_INVERSE_LIKE_DIRECT</code> or
    342  * <code>UBIDI_REORDER_INVERSE_FOR_NUMBERS_SPECIAL</code>, the text to reorder
    343  * is assumed to be visual LTR, and the text after reordering is required
    344  * to be the corresponding logical string with appropriate contextual
    345  * direction. The direction of the result string will be RTL if either
    346  * the righmost or leftmost strong character of the source text is RTL
    347  * or Arabic Letter, the direction will be LTR otherwise.<p>
    348  *
    349  * If reordering option <code>UBIDI_OPTION_INSERT_MARKS</code> is set, an RLM may
    350  * be added at the beginning of the result string to ensure round trip
    351  * (that the result string, when reordered back to visual, will produce
    352  * the original source text).
    353  * @see UBIDI_REORDER_INVERSE_LIKE_DIRECT
    354  * @see UBIDI_REORDER_INVERSE_FOR_NUMBERS_SPECIAL
    355  * @stable ICU 2.0
    356  */
    357 #define UBIDI_DEFAULT_LTR 0xfe
    358 
    359 /** Paragraph level setting.<p>
    360  *
    361  * Constant indicating that the base direction depends on the first strong
    362  * directional character in the text according to the Unicode Bidirectional
    363  * Algorithm. If no strong directional character is present,
    364  * then set the paragraph level to 1 (right-to-left).<p>
    365  *
    366  * If this value is used in conjunction with reordering modes
    367  * <code>UBIDI_REORDER_INVERSE_LIKE_DIRECT</code> or
    368  * <code>UBIDI_REORDER_INVERSE_FOR_NUMBERS_SPECIAL</code>, the text to reorder
    369  * is assumed to be visual LTR, and the text after reordering is required
    370  * to be the corresponding logical string with appropriate contextual
    371  * direction. The direction of the result string will be RTL if either
    372  * the righmost or leftmost strong character of the source text is RTL
    373  * or Arabic Letter, or if the text contains no strong character;
    374  * the direction will be LTR otherwise.<p>
    375  *
    376  * If reordering option <code>UBIDI_OPTION_INSERT_MARKS</code> is set, an RLM may
    377  * be added at the beginning of the result string to ensure round trip
    378  * (that the result string, when reordered back to visual, will produce
    379  * the original source text).
    380  * @see UBIDI_REORDER_INVERSE_LIKE_DIRECT
    381  * @see UBIDI_REORDER_INVERSE_FOR_NUMBERS_SPECIAL
    382  * @stable ICU 2.0
    383  */
    384 #define UBIDI_DEFAULT_RTL 0xff
    385 
    386 /**
    387  * Maximum explicit embedding level.
    388  * (The maximum resolved level can be up to <code>UBIDI_MAX_EXPLICIT_LEVEL+1</code>).
    389  * @stable ICU 2.0
    390  */
    391 #define UBIDI_MAX_EXPLICIT_LEVEL 61
    392 
    393 /** Bit flag for level input.
    394  *  Overrides directional properties.
    395  * @stable ICU 2.0
    396  */
    397 #define UBIDI_LEVEL_OVERRIDE 0x80
    398 
    399 /**
    400  * Special value which can be returned by the mapping functions when a logical
    401  * index has no corresponding visual index or vice-versa. This may happen
    402  * for the logical-to-visual mapping of a Bidi control when option
    403  * <code>#UBIDI_OPTION_REMOVE_CONTROLS</code> is specified. This can also happen
    404  * for the visual-to-logical mapping of a Bidi mark (LRM or RLM) inserted
    405  * by option <code>#UBIDI_OPTION_INSERT_MARKS</code>.
    406  * @see ubidi_getVisualIndex
    407  * @see ubidi_getVisualMap
    408  * @see ubidi_getLogicalIndex
    409  * @see ubidi_getLogicalMap
    410  * @stable ICU 3.6
    411  */
    412 #define UBIDI_MAP_NOWHERE   (-1)
    413 
    414 /**
    415  * <code>UBiDiDirection</code> values indicate the text direction.
    416  * @stable ICU 2.0
    417  */
    418 enum UBiDiDirection {
    419     /** All left-to-right text. This is a 0 value. @stable ICU 2.0 */
    420     UBIDI_LTR,
    421     /** All right-to-left text. This is a 1 value. @stable ICU 2.0 */
    422     UBIDI_RTL,
    423     /** Mixed-directional text. @stable ICU 2.0 */
    424     UBIDI_MIXED
    425 };
    426 
    427 /** @stable ICU 2.0 */
    428 typedef enum UBiDiDirection UBiDiDirection;
    429 
    430 /**
    431  * Forward declaration of the <code>UBiDi</code> structure for the declaration of
    432  * the API functions. Its fields are implementation-specific.<p>
    433  * This structure holds information about a paragraph (or multiple paragraphs)
    434  * of text with Bidi-algorithm-related details, or about one line of
    435  * such a paragraph.<p>
    436  * Reordering can be done on a line, or on one or more paragraphs which are
    437  * then interpreted each as one single line.
    438  * @stable ICU 2.0
    439  */
    440 struct UBiDi;
    441 
    442 /** @stable ICU 2.0 */
    443 typedef struct UBiDi UBiDi;
    444 
    445 /**
    446  * Allocate a <code>UBiDi</code> structure.
    447  * Such an object is initially empty. It is assigned
    448  * the Bidi properties of a piece of text containing one or more paragraphs
    449  * by <code>ubidi_setPara()</code>
    450  * or the Bidi properties of a line within a paragraph by
    451  * <code>ubidi_setLine()</code>.<p>
    452  * This object can be reused for as long as it is not deallocated
    453  * by calling <code>ubidi_close()</code>.<p>
    454  * <code>ubidi_setPara()</code> and <code>ubidi_setLine()</code> will allocate
    455  * additional memory for internal structures as necessary.
    456  *
    457  * @return An empty <code>UBiDi</code> object.
    458  * @stable ICU 2.0
    459  */
    460 U_STABLE UBiDi * U_EXPORT2
    461 ubidi_open(void);
    462 
    463 /**
    464  * Allocate a <code>UBiDi</code> structure with preallocated memory
    465  * for internal structures.
    466  * This function provides a <code>UBiDi</code> object like <code>ubidi_open()</code>
    467  * with no arguments, but it also preallocates memory for internal structures
    468  * according to the sizings supplied by the caller.<p>
    469  * Subsequent functions will not allocate any more memory, and are thus
    470  * guaranteed not to fail because of lack of memory.<p>
    471  * The preallocation can be limited to some of the internal memory
    472  * by setting some values to 0 here. That means that if, e.g.,
    473  * <code>maxRunCount</code> cannot be reasonably predetermined and should not
    474  * be set to <code>maxLength</code> (the only failproof value) to avoid
    475  * wasting memory, then <code>maxRunCount</code> could be set to 0 here
    476  * and the internal structures that are associated with it will be allocated
    477  * on demand, just like with <code>ubidi_open()</code>.
    478  *
    479  * @param maxLength is the maximum text or line length that internal memory
    480  *        will be preallocated for. An attempt to associate this object with a
    481  *        longer text will fail, unless this value is 0, which leaves the allocation
    482  *        up to the implementation.
    483  *
    484  * @param maxRunCount is the maximum anticipated number of same-level runs
    485  *        that internal memory will be preallocated for. An attempt to access
    486  *        visual runs on an object that was not preallocated for as many runs
    487  *        as the text was actually resolved to will fail,
    488  *        unless this value is 0, which leaves the allocation up to the implementation.<br><br>
    489  *        The number of runs depends on the actual text and maybe anywhere between
    490  *        1 and <code>maxLength</code>. It is typically small.
    491  *
    492  * @param pErrorCode must be a valid pointer to an error code value.
    493  *
    494  * @return An empty <code>UBiDi</code> object with preallocated memory.
    495  * @stable ICU 2.0
    496  */
    497 U_STABLE UBiDi * U_EXPORT2
    498 ubidi_openSized(int32_t maxLength, int32_t maxRunCount, UErrorCode *pErrorCode);
    499 
    500 /**
    501  * <code>ubidi_close()</code> must be called to free the memory
    502  * associated with a UBiDi object.<p>
    503  *
    504  * <strong>Important: </strong>
    505  * A parent <code>UBiDi</code> object must not be destroyed or reused if
    506  * it still has children.
    507  * If a <code>UBiDi</code> object has become the <i>child</i>
    508  * of another one (its <i>parent</i>) by calling
    509  * <code>ubidi_setLine()</code>, then the child object must
    510  * be destroyed (closed) or reused (by calling
    511  * <code>ubidi_setPara()</code> or <code>ubidi_setLine()</code>)
    512  * before the parent object.
    513  *
    514  * @param pBiDi is a <code>UBiDi</code> object.
    515  *
    516  * @see ubidi_setPara
    517  * @see ubidi_setLine
    518  * @stable ICU 2.0
    519  */
    520 U_STABLE void U_EXPORT2
    521 ubidi_close(UBiDi *pBiDi);
    522 
    523 #if U_SHOW_CPLUSPLUS_API
    524 
    525 U_NAMESPACE_BEGIN
    526 
    527 /**
    528  * \class LocalUBiDiPointer
    529  * "Smart pointer" class, closes a UBiDi via ubidi_close().
    530  * For most methods see the LocalPointerBase base class.
    531  *
    532  * @see LocalPointerBase
    533  * @see LocalPointer
    534  * @draft ICU 4.4
    535  */
    536 U_DEFINE_LOCAL_OPEN_POINTER(LocalUBiDiPointer, UBiDi, ubidi_close);
    537 
    538 U_NAMESPACE_END
    539 
    540 #endif
    541 
    542 /**
    543  * Modify the operation of the Bidi algorithm such that it
    544  * approximates an "inverse Bidi" algorithm. This function
    545  * must be called before <code>ubidi_setPara()</code>.
    546  *
    547  * <p>The normal operation of the Bidi algorithm as described
    548  * in the Unicode Technical Report is to take text stored in logical
    549  * (keyboard, typing) order and to determine the reordering of it for visual
    550  * rendering.
    551  * Some legacy systems store text in visual order, and for operations
    552  * with standard, Unicode-based algorithms, the text needs to be transformed
    553  * to logical order. This is effectively the inverse algorithm of the
    554  * described Bidi algorithm. Note that there is no standard algorithm for
    555  * this "inverse Bidi" and that the current implementation provides only an
    556  * approximation of "inverse Bidi".</p>
    557  *
    558  * <p>With <code>isInverse</code> set to <code>TRUE</code>,
    559  * this function changes the behavior of some of the subsequent functions
    560  * in a way that they can be used for the inverse Bidi algorithm.
    561  * Specifically, runs of text with numeric characters will be treated in a
    562  * special way and may need to be surrounded with LRM characters when they are
    563  * written in reordered sequence.</p>
    564  *
    565  * <p>Output runs should be retrieved using <code>ubidi_getVisualRun()</code>.
    566  * Since the actual input for "inverse Bidi" is visually ordered text and
    567  * <code>ubidi_getVisualRun()</code> gets the reordered runs, these are actually
    568  * the runs of the logically ordered output.</p>
    569  *
    570  * <p>Calling this function with argument <code>isInverse</code> set to
    571  * <code>TRUE</code> is equivalent to calling
    572  * <code>ubidi_setReorderingMode</code> with argument
    573  * <code>reorderingMode</code>
    574  * set to <code>#UBIDI_REORDER_INVERSE_NUMBERS_AS_L</code>.<br>
    575  * Calling this function with argument <code>isInverse</code> set to
    576  * <code>FALSE</code> is equivalent to calling
    577  * <code>ubidi_setReorderingMode</code> with argument
    578  * <code>reorderingMode</code>
    579  * set to <code>#UBIDI_REORDER_DEFAULT</code>.
    580  *
    581  * @param pBiDi is a <code>UBiDi</code> object.
    582  *
    583  * @param isInverse specifies "forward" or "inverse" Bidi operation.
    584  *
    585  * @see ubidi_setPara
    586  * @see ubidi_writeReordered
    587  * @see ubidi_setReorderingMode
    588  * @stable ICU 2.0
    589  */
    590 U_STABLE void U_EXPORT2
    591 ubidi_setInverse(UBiDi *pBiDi, UBool isInverse);
    592 
    593 /**
    594  * Is this Bidi object set to perform the inverse Bidi algorithm?
    595  * <p>Note: calling this function after setting the reordering mode with
    596  * <code>ubidi_setReorderingMode</code> will return <code>TRUE</code> if the
    597  * reordering mode was set to <code>#UBIDI_REORDER_INVERSE_NUMBERS_AS_L</code>,
    598  * <code>FALSE</code> for all other values.</p>
    599  *
    600  * @param pBiDi is a <code>UBiDi</code> object.
    601  * @return TRUE if the Bidi object is set to perform the inverse Bidi algorithm
    602  * by handling numbers as L.
    603  *
    604  * @see ubidi_setInverse
    605  * @see ubidi_setReorderingMode
    606  * @stable ICU 2.0
    607  */
    608 
    609 U_STABLE UBool U_EXPORT2
    610 ubidi_isInverse(UBiDi *pBiDi);
    611 
    612 /**
    613  * Specify whether block separators must be allocated level zero,
    614  * so that successive paragraphs will progress from left to right.
    615  * This function must be called before <code>ubidi_setPara()</code>.
    616  * Paragraph separators (B) may appear in the text.  Setting them to level zero
    617  * means that all paragraph separators (including one possibly appearing
    618  * in the last text position) are kept in the reordered text after the text
    619  * that they follow in the source text.
    620  * When this feature is not enabled, a paragraph separator at the last
    621  * position of the text before reordering will go to the first position
    622  * of the reordered text when the paragraph level is odd.
    623  *
    624  * @param pBiDi is a <code>UBiDi</code> object.
    625  *
    626  * @param orderParagraphsLTR specifies whether paragraph separators (B) must
    627  * receive level 0, so that successive paragraphs progress from left to right.
    628  *
    629  * @see ubidi_setPara
    630  * @stable ICU 3.4
    631  */
    632 U_STABLE void U_EXPORT2
    633 ubidi_orderParagraphsLTR(UBiDi *pBiDi, UBool orderParagraphsLTR);
    634 
    635 /**
    636  * Is this Bidi object set to allocate level 0 to block separators so that
    637  * successive paragraphs progress from left to right?
    638  *
    639  * @param pBiDi is a <code>UBiDi</code> object.
    640  * @return TRUE if the Bidi object is set to allocate level 0 to block
    641  *         separators.
    642  *
    643  * @see ubidi_orderParagraphsLTR
    644  * @stable ICU 3.4
    645  */
    646 U_STABLE UBool U_EXPORT2
    647 ubidi_isOrderParagraphsLTR(UBiDi *pBiDi);
    648 
    649 /**
    650  * <code>UBiDiReorderingMode</code> values indicate which variant of the Bidi
    651  * algorithm to use.
    652  *
    653  * @see ubidi_setReorderingMode
    654  * @stable ICU 3.6
    655  */
    656 typedef enum UBiDiReorderingMode {
    657     /** Regular Logical to Visual Bidi algorithm according to Unicode.
    658       * This is a 0 value.
    659       * @stable ICU 3.6 */
    660     UBIDI_REORDER_DEFAULT = 0,
    661     /** Logical to Visual algorithm which handles numbers in a way which
    662       * mimicks the behavior of Windows XP.
    663       * @stable ICU 3.6 */
    664     UBIDI_REORDER_NUMBERS_SPECIAL,
    665     /** Logical to Visual algorithm grouping numbers with adjacent R characters
    666       * (reversible algorithm).
    667       * @stable ICU 3.6 */
    668     UBIDI_REORDER_GROUP_NUMBERS_WITH_R,
    669     /** Reorder runs only to transform a Logical LTR string to the Logical RTL
    670       * string with the same display, or vice-versa.<br>
    671       * If this mode is set together with option
    672       * <code>#UBIDI_OPTION_INSERT_MARKS</code>, some Bidi controls in the source
    673       * text may be removed and other controls may be added to produce the
    674       * minimum combination which has the required display.
    675       * @stable ICU 3.6 */
    676     UBIDI_REORDER_RUNS_ONLY,
    677     /** Visual to Logical algorithm which handles numbers like L
    678       * (same algorithm as selected by <code>ubidi_setInverse(TRUE)</code>.
    679       * @see ubidi_setInverse
    680       * @stable ICU 3.6 */
    681     UBIDI_REORDER_INVERSE_NUMBERS_AS_L,
    682     /** Visual to Logical algorithm equivalent to the regular Logical to Visual
    683       * algorithm.
    684       * @stable ICU 3.6 */
    685     UBIDI_REORDER_INVERSE_LIKE_DIRECT,
    686     /** Inverse Bidi (Visual to Logical) algorithm for the
    687       * <code>UBIDI_REORDER_NUMBERS_SPECIAL</code> Bidi algorithm.
    688       * @stable ICU 3.6 */
    689     UBIDI_REORDER_INVERSE_FOR_NUMBERS_SPECIAL,
    690     /** Number of values for reordering mode.
    691       * @stable ICU 3.6 */
    692     UBIDI_REORDER_COUNT
    693 } UBiDiReorderingMode;
    694 
    695 /**
    696  * Modify the operation of the Bidi algorithm such that it implements some
    697  * variant to the basic Bidi algorithm or approximates an "inverse Bidi"
    698  * algorithm, depending on different values of the "reordering mode".
    699  * This function must be called before <code>ubidi_setPara()</code>, and stays
    700  * in effect until called again with a different argument.
    701  *
    702  * <p>The normal operation of the Bidi algorithm as described
    703  * in the Unicode Standard Annex #9 is to take text stored in logical
    704  * (keyboard, typing) order and to determine how to reorder it for visual
    705  * rendering.</p>
    706  *
    707  * <p>With the reordering mode set to a value other than
    708  * <code>#UBIDI_REORDER_DEFAULT</code>, this function changes the behavior of
    709  * some of the subsequent functions in a way such that they implement an
    710  * inverse Bidi algorithm or some other algorithm variants.</p>
    711  *
    712  * <p>Some legacy systems store text in visual order, and for operations
    713  * with standard, Unicode-based algorithms, the text needs to be transformed
    714  * into logical order. This is effectively the inverse algorithm of the
    715  * described Bidi algorithm. Note that there is no standard algorithm for
    716  * this "inverse Bidi", so a number of variants are implemented here.</p>
    717  *
    718  * <p>In other cases, it may be desirable to emulate some variant of the
    719  * Logical to Visual algorithm (e.g. one used in MS Windows), or perform a
    720  * Logical to Logical transformation.</p>
    721  *
    722  * <ul>
    723  * <li>When the reordering mode is set to <code>#UBIDI_REORDER_DEFAULT</code>,
    724  * the standard Bidi Logical to Visual algorithm is applied.</li>
    725  *
    726  * <li>When the reordering mode is set to
    727  * <code>#UBIDI_REORDER_NUMBERS_SPECIAL</code>,
    728  * the algorithm used to perform Bidi transformations when calling
    729  * <code>ubidi_setPara</code> should approximate the algorithm used in
    730  * Microsoft Windows XP rather than strictly conform to the Unicode Bidi
    731  * algorithm.
    732  * <br>
    733  * The differences between the basic algorithm and the algorithm addressed
    734  * by this option are as follows:
    735  * <ul>
    736  *   <li>Within text at an even embedding level, the sequence "123AB"
    737  *   (where AB represent R or AL letters) is transformed to "123BA" by the
    738  *   Unicode algorithm and to "BA123" by the Windows algorithm.</li>
    739  *   <li>Arabic-Indic numbers (AN) are handled by the Windows algorithm just
    740  *   like regular numbers (EN).</li>
    741  * </ul></li>
    742  *
    743  * <li>When the reordering mode is set to
    744  * <code>#UBIDI_REORDER_GROUP_NUMBERS_WITH_R</code>,
    745  * numbers located between LTR text and RTL text are associated with the RTL
    746  * text. For instance, an LTR paragraph with content "abc 123 DEF" (where
    747  * upper case letters represent RTL characters) will be transformed to
    748  * "abc FED 123" (and not "abc 123 FED"), "DEF 123 abc" will be transformed
    749  * to "123 FED abc" and "123 FED abc" will be transformed to "DEF 123 abc".
    750  * This makes the algorithm reversible and makes it useful when round trip
    751  * (from visual to logical and back to visual) must be achieved without
    752  * adding LRM characters. However, this is a variation from the standard
    753  * Unicode Bidi algorithm.<br>
    754  * The source text should not contain Bidi control characters other than LRM
    755  * or RLM.</li>
    756  *
    757  * <li>When the reordering mode is set to
    758  * <code>#UBIDI_REORDER_RUNS_ONLY</code>,
    759  * a "Logical to Logical" transformation must be performed:
    760  * <ul>
    761  * <li>If the default text level of the source text (argument <code>paraLevel</code>
    762  * in <code>ubidi_setPara</code>) is even, the source text will be handled as
    763  * LTR logical text and will be transformed to the RTL logical text which has
    764  * the same LTR visual display.</li>
    765  * <li>If the default level of the source text is odd, the source text
    766  * will be handled as RTL logical text and will be transformed to the
    767  * LTR logical text which has the same LTR visual display.</li>
    768  * </ul>
    769  * This mode may be needed when logical text which is basically Arabic or
    770  * Hebrew, with possible included numbers or phrases in English, has to be
    771  * displayed as if it had an even embedding level (this can happen if the
    772  * displaying application treats all text as if it was basically LTR).
    773  * <br>
    774  * This mode may also be needed in the reverse case, when logical text which is
    775  * basically English, with possible included phrases in Arabic or Hebrew, has to
    776  * be displayed as if it had an odd embedding level.
    777  * <br>
    778  * Both cases could be handled by adding LRE or RLE at the head of the text,
    779  * if the display subsystem supports these formatting controls. If it does not,
    780  * the problem may be handled by transforming the source text in this mode
    781  * before displaying it, so that it will be displayed properly.<br>
    782  * The source text should not contain Bidi control characters other than LRM
    783  * or RLM.</li>
    784  *
    785  * <li>When the reordering mode is set to
    786  * <code>#UBIDI_REORDER_INVERSE_NUMBERS_AS_L</code>, an "inverse Bidi" algorithm
    787  * is applied.
    788  * Runs of text with numeric characters will be treated like LTR letters and
    789  * may need to be surrounded with LRM characters when they are written in
    790  * reordered sequence (the option <code>#UBIDI_INSERT_LRM_FOR_NUMERIC</code> can
    791  * be used with function <code>ubidi_writeReordered</code> to this end. This
    792  * mode is equivalent to calling <code>ubidi_setInverse()</code> with
    793  * argument <code>isInverse</code> set to <code>TRUE</code>.</li>
    794  *
    795  * <li>When the reordering mode is set to
    796  * <code>#UBIDI_REORDER_INVERSE_LIKE_DIRECT</code>, the "direct" Logical to Visual
    797  * Bidi algorithm is used as an approximation of an "inverse Bidi" algorithm.
    798  * This mode is similar to mode <code>#UBIDI_REORDER_INVERSE_NUMBERS_AS_L</code>
    799  * but is closer to the regular Bidi algorithm.
    800  * <br>
    801  * For example, an LTR paragraph with the content "FED 123 456 CBA" (where
    802  * upper case represents RTL characters) will be transformed to
    803  * "ABC 456 123 DEF", as opposed to "DEF 123 456 ABC"
    804  * with mode <code>UBIDI_REORDER_INVERSE_NUMBERS_AS_L</code>.<br>
    805  * When used in conjunction with option
    806  * <code>#UBIDI_OPTION_INSERT_MARKS</code>, this mode generally
    807  * adds Bidi marks to the output significantly more sparingly than mode
    808  * <code>#UBIDI_REORDER_INVERSE_NUMBERS_AS_L</code> with option
    809  * <code>#UBIDI_INSERT_LRM_FOR_NUMERIC</code> in calls to
    810  * <code>ubidi_writeReordered</code>.</li>
    811  *
    812  * <li>When the reordering mode is set to
    813  * <code>#UBIDI_REORDER_INVERSE_FOR_NUMBERS_SPECIAL</code>, the Logical to Visual
    814  * Bidi algorithm used in Windows XP is used as an approximation of an "inverse Bidi" algorithm.
    815  * <br>
    816  * For example, an LTR paragraph with the content "abc FED123" (where
    817  * upper case represents RTL characters) will be transformed to "abc 123DEF."</li>
    818  * </ul>
    819  *
    820  * <p>In all the reordering modes specifying an "inverse Bidi" algorithm
    821  * (i.e. those with a name starting with <code>UBIDI_REORDER_INVERSE</code>),
    822  * output runs should be retrieved using
    823  * <code>ubidi_getVisualRun()</code>, and the output text with
    824  * <code>ubidi_writeReordered()</code>. The caller should keep in mind that in
    825  * "inverse Bidi" modes the input is actually visually ordered text and
    826  * reordered output returned by <code>ubidi_getVisualRun()</code> or
    827  * <code>ubidi_writeReordered()</code> are actually runs or character string
    828  * of logically ordered output.<br>
    829  * For all the "inverse Bidi" modes, the source text should not contain
    830  * Bidi control characters other than LRM or RLM.</p>
    831  *
    832  * <p>Note that option <code>#UBIDI_OUTPUT_REVERSE</code> of
    833  * <code>ubidi_writeReordered</code> has no useful meaning and should not be
    834  * used in conjunction with any value of the reordering mode specifying
    835  * "inverse Bidi" or with value <code>UBIDI_REORDER_RUNS_ONLY</code>.
    836  *
    837  * @param pBiDi is a <code>UBiDi</code> object.
    838  * @param reorderingMode specifies the required variant of the Bidi algorithm.
    839  *
    840  * @see UBiDiReorderingMode
    841  * @see ubidi_setInverse
    842  * @see ubidi_setPara
    843  * @see ubidi_writeReordered
    844  * @stable ICU 3.6
    845  */
    846 U_STABLE void U_EXPORT2
    847 ubidi_setReorderingMode(UBiDi *pBiDi, UBiDiReorderingMode reorderingMode);
    848 
    849 /**
    850  * What is the requested reordering mode for a given Bidi object?
    851  *
    852  * @param pBiDi is a <code>UBiDi</code> object.
    853  * @return the current reordering mode of the Bidi object
    854  * @see ubidi_setReorderingMode
    855  * @stable ICU 3.6
    856  */
    857 U_STABLE UBiDiReorderingMode U_EXPORT2
    858 ubidi_getReorderingMode(UBiDi *pBiDi);
    859 
    860 /**
    861  * <code>UBiDiReorderingOption</code> values indicate which options are
    862  * specified to affect the Bidi algorithm.
    863  *
    864  * @see ubidi_setReorderingOptions
    865  * @stable ICU 3.6
    866  */
    867 typedef enum UBiDiReorderingOption {
    868     /**
    869      * option value for <code>ubidi_setReorderingOptions</code>:
    870      * disable all the options which can be set with this function
    871      * @see ubidi_setReorderingOptions
    872      * @stable ICU 3.6
    873      */
    874     UBIDI_OPTION_DEFAULT = 0,
    875 
    876     /**
    877      * option bit for <code>ubidi_setReorderingOptions</code>:
    878      * insert Bidi marks (LRM or RLM) when needed to ensure correct result of
    879      * a reordering to a Logical order
    880      *
    881      * <p>This option must be set or reset before calling
    882      * <code>ubidi_setPara</code>.</p>
    883      *
    884      * <p>This option is significant only with reordering modes which generate
    885      * a result with Logical order, specifically:</p>
    886      * <ul>
    887      *   <li><code>#UBIDI_REORDER_RUNS_ONLY</code></li>
    888      *   <li><code>#UBIDI_REORDER_INVERSE_NUMBERS_AS_L</code></li>
    889      *   <li><code>#UBIDI_REORDER_INVERSE_LIKE_DIRECT</code></li>
    890      *   <li><code>#UBIDI_REORDER_INVERSE_FOR_NUMBERS_SPECIAL</code></li>
    891      * </ul>
    892      *
    893      * <p>If this option is set in conjunction with reordering mode
    894      * <code>#UBIDI_REORDER_INVERSE_NUMBERS_AS_L</code> or with calling
    895      * <code>ubidi_setInverse(TRUE)</code>, it implies
    896      * option <code>#UBIDI_INSERT_LRM_FOR_NUMERIC</code>
    897      * in calls to function <code>ubidi_writeReordered()</code>.</p>
    898      *
    899      * <p>For other reordering modes, a minimum number of LRM or RLM characters
    900      * will be added to the source text after reordering it so as to ensure
    901      * round trip, i.e. when applying the inverse reordering mode on the
    902      * resulting logical text with removal of Bidi marks
    903      * (option <code>#UBIDI_OPTION_REMOVE_CONTROLS</code> set before calling
    904      * <code>ubidi_setPara()</code> or option <code>#UBIDI_REMOVE_BIDI_CONTROLS</code>
    905      * in <code>ubidi_writeReordered</code>), the result will be identical to the
    906      * source text in the first transformation.
    907      *
    908      * <p>This option will be ignored if specified together with option
    909      * <code>#UBIDI_OPTION_REMOVE_CONTROLS</code>. It inhibits option
    910      * <code>UBIDI_REMOVE_BIDI_CONTROLS</code> in calls to function
    911      * <code>ubidi_writeReordered()</code> and it implies option
    912      * <code>#UBIDI_INSERT_LRM_FOR_NUMERIC</code> in calls to function
    913      * <code>ubidi_writeReordered()</code> if the reordering mode is
    914      * <code>#UBIDI_REORDER_INVERSE_NUMBERS_AS_L</code>.</p>
    915      *
    916      * @see ubidi_setReorderingMode
    917      * @see ubidi_setReorderingOptions
    918      * @stable ICU 3.6
    919      */
    920     UBIDI_OPTION_INSERT_MARKS = 1,
    921 
    922     /**
    923      * option bit for <code>ubidi_setReorderingOptions</code>:
    924      * remove Bidi control characters
    925      *
    926      * <p>This option must be set or reset before calling
    927      * <code>ubidi_setPara</code>.</p>
    928      *
    929      * <p>This option nullifies option <code>#UBIDI_OPTION_INSERT_MARKS</code>.
    930      * It inhibits option <code>#UBIDI_INSERT_LRM_FOR_NUMERIC</code> in calls
    931      * to function <code>ubidi_writeReordered()</code> and it implies option
    932      * <code>#UBIDI_REMOVE_BIDI_CONTROLS</code> in calls to that function.</p>
    933      *
    934      * @see ubidi_setReorderingMode
    935      * @see ubidi_setReorderingOptions
    936      * @stable ICU 3.6
    937      */
    938     UBIDI_OPTION_REMOVE_CONTROLS = 2,
    939 
    940     /**
    941      * option bit for <code>ubidi_setReorderingOptions</code>:
    942      * process the output as part of a stream to be continued
    943      *
    944      * <p>This option must be set or reset before calling
    945      * <code>ubidi_setPara</code>.</p>
    946      *
    947      * <p>This option specifies that the caller is interested in processing large
    948      * text object in parts.
    949      * The results of the successive calls are expected to be concatenated by the
    950      * caller. Only the call for the last part will have this option bit off.</p>
    951      *
    952      * <p>When this option bit is on, <code>ubidi_setPara()</code> may process
    953      * less than the full source text in order to truncate the text at a meaningful
    954      * boundary. The caller should call <code>ubidi_getProcessedLength()</code>
    955      * immediately after calling <code>ubidi_setPara()</code> in order to
    956      * determine how much of the source text has been processed.
    957      * Source text beyond that length should be resubmitted in following calls to
    958      * <code>ubidi_setPara</code>. The processed length may be less than
    959      * the length of the source text if a character preceding the last character of
    960      * the source text constitutes a reasonable boundary (like a block separator)
    961      * for text to be continued.<br>
    962      * If the last character of the source text constitutes a reasonable
    963      * boundary, the whole text will be processed at once.<br>
    964      * If nowhere in the source text there exists
    965      * such a reasonable boundary, the processed length will be zero.<br>
    966      * The caller should check for such an occurrence and do one of the following:
    967      * <ul><li>submit a larger amount of text with a better chance to include
    968      *         a reasonable boundary.</li>
    969      *     <li>resubmit the same text after turning off option
    970      *         <code>UBIDI_OPTION_STREAMING</code>.</li></ul>
    971      * In all cases, this option should be turned off before processing the last
    972      * part of the text.</p>
    973      *
    974      * <p>When the <code>UBIDI_OPTION_STREAMING</code> option is used,
    975      * it is recommended to call <code>ubidi_orderParagraphsLTR()</code> with
    976      * argument <code>orderParagraphsLTR</code> set to <code>TRUE</code> before
    977      * calling <code>ubidi_setPara</code> so that later paragraphs may be
    978      * concatenated to previous paragraphs on the right.</p>
    979      *
    980      * @see ubidi_setReorderingMode
    981      * @see ubidi_setReorderingOptions
    982      * @see ubidi_getProcessedLength
    983      * @see ubidi_orderParagraphsLTR
    984      * @stable ICU 3.6
    985      */
    986     UBIDI_OPTION_STREAMING = 4
    987 } UBiDiReorderingOption;
    988 
    989 /**
    990  * Specify which of the reordering options
    991  * should be applied during Bidi transformations.
    992  *
    993  * @param pBiDi is a <code>UBiDi</code> object.
    994  * @param reorderingOptions is a combination of zero or more of the following
    995  * options:
    996  * <code>#UBIDI_OPTION_DEFAULT</code>, <code>#UBIDI_OPTION_INSERT_MARKS</code>,
    997  * <code>#UBIDI_OPTION_REMOVE_CONTROLS</code>, <code>#UBIDI_OPTION_STREAMING</code>.
    998  *
    999  * @see ubidi_getReorderingOptions
   1000  * @stable ICU 3.6
   1001  */
   1002 U_STABLE void U_EXPORT2
   1003 ubidi_setReorderingOptions(UBiDi *pBiDi, uint32_t reorderingOptions);
   1004 
   1005 /**
   1006  * What are the reordering options applied to a given Bidi object?
   1007  *
   1008  * @param pBiDi is a <code>UBiDi</code> object.
   1009  * @return the current reordering options of the Bidi object
   1010  * @see ubidi_setReorderingOptions
   1011  * @stable ICU 3.6
   1012  */
   1013 U_STABLE uint32_t U_EXPORT2
   1014 ubidi_getReorderingOptions(UBiDi *pBiDi);
   1015 
   1016 /**
   1017  * Perform the Unicode Bidi algorithm. It is defined in the
   1018  * <a href="http://www.unicode.org/unicode/reports/tr9/">Unicode Standard Anned #9</a>,
   1019  * version 13,
   1020  * also described in The Unicode Standard, Version 4.0 .<p>
   1021  *
   1022  * This function takes a piece of plain text containing one or more paragraphs,
   1023  * with or without externally specified embedding levels from <i>styled</i>
   1024  * text and computes the left-right-directionality of each character.<p>
   1025  *
   1026  * If the entire text is all of the same directionality, then
   1027  * the function may not perform all the steps described by the algorithm,
   1028  * i.e., some levels may not be the same as if all steps were performed.
   1029  * This is not relevant for unidirectional text.<br>
   1030  * For example, in pure LTR text with numbers the numbers would get
   1031  * a resolved level of 2 higher than the surrounding text according to
   1032  * the algorithm. This implementation may set all resolved levels to
   1033  * the same value in such a case.<p>
   1034  *
   1035  * The text can be composed of multiple paragraphs. Occurrence of a block
   1036  * separator in the text terminates a paragraph, and whatever comes next starts
   1037  * a new paragraph. The exception to this rule is when a Carriage Return (CR)
   1038  * is followed by a Line Feed (LF). Both CR and LF are block separators, but
   1039  * in that case, the pair of characters is considered as terminating the
   1040  * preceding paragraph, and a new paragraph will be started by a character
   1041  * coming after the LF.
   1042  *
   1043  * @param pBiDi A <code>UBiDi</code> object allocated with <code>ubidi_open()</code>
   1044  *        which will be set to contain the reordering information,
   1045  *        especially the resolved levels for all the characters in <code>text</code>.
   1046  *
   1047  * @param text is a pointer to the text that the Bidi algorithm will be performed on.
   1048  *        This pointer is stored in the UBiDi object and can be retrieved
   1049  *        with <code>ubidi_getText()</code>.<br>
   1050  *        <strong>Note:</strong> the text must be (at least) <code>length</code> long.
   1051  *
   1052  * @param length is the length of the text; if <code>length==-1</code> then
   1053  *        the text must be zero-terminated.
   1054  *
   1055  * @param paraLevel specifies the default level for the text;
   1056  *        it is typically 0 (LTR) or 1 (RTL).
   1057  *        If the function shall determine the paragraph level from the text,
   1058  *        then <code>paraLevel</code> can be set to
   1059  *        either <code>#UBIDI_DEFAULT_LTR</code>
   1060  *        or <code>#UBIDI_DEFAULT_RTL</code>; if the text contains multiple
   1061  *        paragraphs, the paragraph level shall be determined separately for
   1062  *        each paragraph; if a paragraph does not include any strongly typed
   1063  *        character, then the desired default is used (0 for LTR or 1 for RTL).
   1064  *        Any other value between 0 and <code>#UBIDI_MAX_EXPLICIT_LEVEL</code>
   1065  *        is also valid, with odd levels indicating RTL.
   1066  *
   1067  * @param embeddingLevels (in) may be used to preset the embedding and override levels,
   1068  *        ignoring characters like LRE and PDF in the text.
   1069  *        A level overrides the directional property of its corresponding
   1070  *        (same index) character if the level has the
   1071  *        <code>#UBIDI_LEVEL_OVERRIDE</code> bit set.<br><br>
   1072  *        Except for that bit, it must be
   1073  *        <code>paraLevel<=embeddingLevels[]<=UBIDI_MAX_EXPLICIT_LEVEL</code>,
   1074  *        with one exception: a level of zero may be specified for a paragraph
   1075  *        separator even if <code>paraLevel>0</code> when multiple paragraphs
   1076  *        are submitted in the same call to <code>ubidi_setPara()</code>.<br><br>
   1077  *        <strong>Caution: </strong>A copy of this pointer, not of the levels,
   1078  *        will be stored in the <code>UBiDi</code> object;
   1079  *        the <code>embeddingLevels</code> array must not be
   1080  *        deallocated before the <code>UBiDi</code> structure is destroyed or reused,
   1081  *        and the <code>embeddingLevels</code>
   1082  *        should not be modified to avoid unexpected results on subsequent Bidi operations.
   1083  *        However, the <code>ubidi_setPara()</code> and
   1084  *        <code>ubidi_setLine()</code> functions may modify some or all of the levels.<br><br>
   1085  *        After the <code>UBiDi</code> object is reused or destroyed, the caller
   1086  *        must take care of the deallocation of the <code>embeddingLevels</code> array.<br><br>
   1087  *        <strong>Note:</strong> the <code>embeddingLevels</code> array must be
   1088  *        at least <code>length</code> long.
   1089  *        This pointer can be <code>NULL</code> if this
   1090  *        value is not necessary.
   1091  *
   1092  * @param pErrorCode must be a valid pointer to an error code value.
   1093  * @stable ICU 2.0
   1094  */
   1095 U_STABLE void U_EXPORT2
   1096 ubidi_setPara(UBiDi *pBiDi, const UChar *text, int32_t length,
   1097               UBiDiLevel paraLevel, UBiDiLevel *embeddingLevels,
   1098               UErrorCode *pErrorCode);
   1099 
   1100 /**
   1101  * <code>ubidi_setLine()</code> sets a <code>UBiDi</code> to
   1102  * contain the reordering information, especially the resolved levels,
   1103  * for all the characters in a line of text. This line of text is
   1104  * specified by referring to a <code>UBiDi</code> object representing
   1105  * this information for a piece of text containing one or more paragraphs,
   1106  * and by specifying a range of indexes in this text.<p>
   1107  * In the new line object, the indexes will range from 0 to <code>limit-start-1</code>.<p>
   1108  *
   1109  * This is used after calling <code>ubidi_setPara()</code>
   1110  * for a piece of text, and after line-breaking on that text.
   1111  * It is not necessary if each paragraph is treated as a single line.<p>
   1112  *
   1113  * After line-breaking, rules (L1) and (L2) for the treatment of
   1114  * trailing WS and for reordering are performed on
   1115  * a <code>UBiDi</code> object that represents a line.<p>
   1116  *
   1117  * <strong>Important: </strong><code>pLineBiDi</code> shares data with
   1118  * <code>pParaBiDi</code>.
   1119  * You must destroy or reuse <code>pLineBiDi</code> before <code>pParaBiDi</code>.
   1120  * In other words, you must destroy or reuse the <code>UBiDi</code> object for a line
   1121  * before the object for its parent paragraph.<p>
   1122  *
   1123  * The text pointer that was stored in <code>pParaBiDi</code> is also copied,
   1124  * and <code>start</code> is added to it so that it points to the beginning of the
   1125  * line for this object.
   1126  *
   1127  * @param pParaBiDi is the parent paragraph object. It must have been set
   1128  * by a successful call to ubidi_setPara.
   1129  *
   1130  * @param start is the line's first index into the text.
   1131  *
   1132  * @param limit is just behind the line's last index into the text
   1133  *        (its last index +1).<br>
   1134  *        It must be <code>0<=start<limit<=</code>containing paragraph limit.
   1135  *        If the specified line crosses a paragraph boundary, the function
   1136  *        will terminate with error code U_ILLEGAL_ARGUMENT_ERROR.
   1137  *
   1138  * @param pLineBiDi is the object that will now represent a line of the text.
   1139  *
   1140  * @param pErrorCode must be a valid pointer to an error code value.
   1141  *
   1142  * @see ubidi_setPara
   1143  * @see ubidi_getProcessedLength
   1144  * @stable ICU 2.0
   1145  */
   1146 U_STABLE void U_EXPORT2
   1147 ubidi_setLine(const UBiDi *pParaBiDi,
   1148               int32_t start, int32_t limit,
   1149               UBiDi *pLineBiDi,
   1150               UErrorCode *pErrorCode);
   1151 
   1152 /**
   1153  * Get the directionality of the text.
   1154  *
   1155  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1156  *
   1157  * @return a value of <code>UBIDI_LTR</code>, <code>UBIDI_RTL</code>
   1158  *         or <code>UBIDI_MIXED</code>
   1159  *         that indicates if the entire text
   1160  *         represented by this object is unidirectional,
   1161  *         and which direction, or if it is mixed-directional.
   1162  *
   1163  * @see UBiDiDirection
   1164  * @stable ICU 2.0
   1165  */
   1166 U_STABLE UBiDiDirection U_EXPORT2
   1167 ubidi_getDirection(const UBiDi *pBiDi);
   1168 
   1169 /**
   1170  * Get the pointer to the text.
   1171  *
   1172  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1173  *
   1174  * @return The pointer to the text that the UBiDi object was created for.
   1175  *
   1176  * @see ubidi_setPara
   1177  * @see ubidi_setLine
   1178  * @stable ICU 2.0
   1179  */
   1180 U_STABLE const UChar * U_EXPORT2
   1181 ubidi_getText(const UBiDi *pBiDi);
   1182 
   1183 /**
   1184  * Get the length of the text.
   1185  *
   1186  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1187  *
   1188  * @return The length of the text that the UBiDi object was created for.
   1189  * @stable ICU 2.0
   1190  */
   1191 U_STABLE int32_t U_EXPORT2
   1192 ubidi_getLength(const UBiDi *pBiDi);
   1193 
   1194 /**
   1195  * Get the paragraph level of the text.
   1196  *
   1197  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1198  *
   1199  * @return The paragraph level. If there are multiple paragraphs, their
   1200  *         level may vary if the required paraLevel is UBIDI_DEFAULT_LTR or
   1201  *         UBIDI_DEFAULT_RTL.  In that case, the level of the first paragraph
   1202  *         is returned.
   1203  *
   1204  * @see UBiDiLevel
   1205  * @see ubidi_getParagraph
   1206  * @see ubidi_getParagraphByIndex
   1207  * @stable ICU 2.0
   1208  */
   1209 U_STABLE UBiDiLevel U_EXPORT2
   1210 ubidi_getParaLevel(const UBiDi *pBiDi);
   1211 
   1212 /**
   1213  * Get the number of paragraphs.
   1214  *
   1215  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1216  *
   1217  * @return The number of paragraphs.
   1218  * @stable ICU 3.4
   1219  */
   1220 U_STABLE int32_t U_EXPORT2
   1221 ubidi_countParagraphs(UBiDi *pBiDi);
   1222 
   1223 /**
   1224  * Get a paragraph, given a position within the text.
   1225  * This function returns information about a paragraph.<br>
   1226  * Note: if the paragraph index is known, it is more efficient to
   1227  * retrieve the paragraph information using ubidi_getParagraphByIndex().<p>
   1228  *
   1229  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1230  *
   1231  * @param charIndex is the index of a character within the text, in the
   1232  *        range <code>[0..ubidi_getProcessedLength(pBiDi)-1]</code>.
   1233  *
   1234  * @param pParaStart will receive the index of the first character of the
   1235  *        paragraph in the text.
   1236  *        This pointer can be <code>NULL</code> if this
   1237  *        value is not necessary.
   1238  *
   1239  * @param pParaLimit will receive the limit of the paragraph.
   1240  *        The l-value that you point to here may be the
   1241  *        same expression (variable) as the one for
   1242  *        <code>charIndex</code>.
   1243  *        This pointer can be <code>NULL</code> if this
   1244  *        value is not necessary.
   1245  *
   1246  * @param pParaLevel will receive the level of the paragraph.
   1247  *        This pointer can be <code>NULL</code> if this
   1248  *        value is not necessary.
   1249  *
   1250  * @param pErrorCode must be a valid pointer to an error code value.
   1251  *
   1252  * @return The index of the paragraph containing the specified position.
   1253  *
   1254  * @see ubidi_getProcessedLength
   1255  * @stable ICU 3.4
   1256  */
   1257 U_STABLE int32_t U_EXPORT2
   1258 ubidi_getParagraph(const UBiDi *pBiDi, int32_t charIndex, int32_t *pParaStart,
   1259                    int32_t *pParaLimit, UBiDiLevel *pParaLevel,
   1260                    UErrorCode *pErrorCode);
   1261 
   1262 /**
   1263  * Get a paragraph, given the index of this paragraph.
   1264  *
   1265  * This function returns information about a paragraph.<p>
   1266  *
   1267  * @param pBiDi is the paragraph <code>UBiDi</code> object.
   1268  *
   1269  * @param paraIndex is the number of the paragraph, in the
   1270  *        range <code>[0..ubidi_countParagraphs(pBiDi)-1]</code>.
   1271  *
   1272  * @param pParaStart will receive the index of the first character of the
   1273  *        paragraph in the text.
   1274  *        This pointer can be <code>NULL</code> if this
   1275  *        value is not necessary.
   1276  *
   1277  * @param pParaLimit will receive the limit of the paragraph.
   1278  *        This pointer can be <code>NULL</code> if this
   1279  *        value is not necessary.
   1280  *
   1281  * @param pParaLevel will receive the level of the paragraph.
   1282  *        This pointer can be <code>NULL</code> if this
   1283  *        value is not necessary.
   1284  *
   1285  * @param pErrorCode must be a valid pointer to an error code value.
   1286  *
   1287  * @stable ICU 3.4
   1288  */
   1289 U_STABLE void U_EXPORT2
   1290 ubidi_getParagraphByIndex(const UBiDi *pBiDi, int32_t paraIndex,
   1291                           int32_t *pParaStart, int32_t *pParaLimit,
   1292                           UBiDiLevel *pParaLevel, UErrorCode *pErrorCode);
   1293 
   1294 /**
   1295  * Get the level for one character.
   1296  *
   1297  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1298  *
   1299  * @param charIndex the index of a character. It must be in the range
   1300  *         [0..ubidi_getProcessedLength(pBiDi)].
   1301  *
   1302  * @return The level for the character at charIndex (0 if charIndex is not
   1303  *         in the valid range).
   1304  *
   1305  * @see UBiDiLevel
   1306  * @see ubidi_getProcessedLength
   1307  * @stable ICU 2.0
   1308  */
   1309 U_STABLE UBiDiLevel U_EXPORT2
   1310 ubidi_getLevelAt(const UBiDi *pBiDi, int32_t charIndex);
   1311 
   1312 /**
   1313  * Get an array of levels for each character.<p>
   1314  *
   1315  * Note that this function may allocate memory under some
   1316  * circumstances, unlike <code>ubidi_getLevelAt()</code>.
   1317  *
   1318  * @param pBiDi is the paragraph or line <code>UBiDi</code> object, whose
   1319  *        text length must be strictly positive.
   1320  *
   1321  * @param pErrorCode must be a valid pointer to an error code value.
   1322  *
   1323  * @return The levels array for the text,
   1324  *         or <code>NULL</code> if an error occurs.
   1325  *
   1326  * @see UBiDiLevel
   1327  * @see ubidi_getProcessedLength
   1328  * @stable ICU 2.0
   1329  */
   1330 U_STABLE const UBiDiLevel * U_EXPORT2
   1331 ubidi_getLevels(UBiDi *pBiDi, UErrorCode *pErrorCode);
   1332 
   1333 /**
   1334  * Get a logical run.
   1335  * This function returns information about a run and is used
   1336  * to retrieve runs in logical order.<p>
   1337  * This is especially useful for line-breaking on a paragraph.
   1338  *
   1339  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1340  *
   1341  * @param logicalPosition is a logical position within the source text.
   1342  *
   1343  * @param pLogicalLimit will receive the limit of the corresponding run.
   1344  *        The l-value that you point to here may be the
   1345  *        same expression (variable) as the one for
   1346  *        <code>logicalPosition</code>.
   1347  *        This pointer can be <code>NULL</code> if this
   1348  *        value is not necessary.
   1349  *
   1350  * @param pLevel will receive the level of the corresponding run.
   1351  *        This pointer can be <code>NULL</code> if this
   1352  *        value is not necessary.
   1353  *
   1354  * @see ubidi_getProcessedLength
   1355  * @stable ICU 2.0
   1356  */
   1357 U_STABLE void U_EXPORT2
   1358 ubidi_getLogicalRun(const UBiDi *pBiDi, int32_t logicalPosition,
   1359                     int32_t *pLogicalLimit, UBiDiLevel *pLevel);
   1360 
   1361 /**
   1362  * Get the number of runs.
   1363  * This function may invoke the actual reordering on the
   1364  * <code>UBiDi</code> object, after <code>ubidi_setPara()</code>
   1365  * may have resolved only the levels of the text. Therefore,
   1366  * <code>ubidi_countRuns()</code> may have to allocate memory,
   1367  * and may fail doing so.
   1368  *
   1369  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1370  *
   1371  * @param pErrorCode must be a valid pointer to an error code value.
   1372  *
   1373  * @return The number of runs.
   1374  * @stable ICU 2.0
   1375  */
   1376 U_STABLE int32_t U_EXPORT2
   1377 ubidi_countRuns(UBiDi *pBiDi, UErrorCode *pErrorCode);
   1378 
   1379 /**
   1380  * Get one run's logical start, length, and directionality,
   1381  * which can be 0 for LTR or 1 for RTL.
   1382  * In an RTL run, the character at the logical start is
   1383  * visually on the right of the displayed run.
   1384  * The length is the number of characters in the run.<p>
   1385  * <code>ubidi_countRuns()</code> should be called
   1386  * before the runs are retrieved.
   1387  *
   1388  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1389  *
   1390  * @param runIndex is the number of the run in visual order, in the
   1391  *        range <code>[0..ubidi_countRuns(pBiDi)-1]</code>.
   1392  *
   1393  * @param pLogicalStart is the first logical character index in the text.
   1394  *        The pointer may be <code>NULL</code> if this index is not needed.
   1395  *
   1396  * @param pLength is the number of characters (at least one) in the run.
   1397  *        The pointer may be <code>NULL</code> if this is not needed.
   1398  *
   1399  * @return the directionality of the run,
   1400  *         <code>UBIDI_LTR==0</code> or <code>UBIDI_RTL==1</code>,
   1401  *         never <code>UBIDI_MIXED</code>.
   1402  *
   1403  * @see ubidi_countRuns
   1404  *
   1405  * Example:
   1406  * <pre>
   1407  * \code
   1408  * int32_t i, count=ubidi_countRuns(pBiDi),
   1409  *         logicalStart, visualIndex=0, length;
   1410  * for(i=0; i<count; ++i) {
   1411  *    if(UBIDI_LTR==ubidi_getVisualRun(pBiDi, i, &logicalStart, &length)) {
   1412  *         do { // LTR
   1413  *             show_char(text[logicalStart++], visualIndex++);
   1414  *         } while(--length>0);
   1415  *     } else {
   1416  *         logicalStart+=length;  // logicalLimit
   1417  *         do { // RTL
   1418  *             show_char(text[--logicalStart], visualIndex++);
   1419  *         } while(--length>0);
   1420  *     }
   1421  * }
   1422  *\endcode
   1423  * </pre>
   1424  *
   1425  * Note that in right-to-left runs, code like this places
   1426  * second surrogates before first ones (which is generally a bad idea)
   1427  * and combining characters before base characters.
   1428  * <p>
   1429  * Use of <code>ubidi_writeReordered()</code>, optionally with the
   1430  * <code>#UBIDI_KEEP_BASE_COMBINING</code> option, can be considered in order
   1431  * to avoid these issues.
   1432  * @stable ICU 2.0
   1433  */
   1434 U_STABLE UBiDiDirection U_EXPORT2
   1435 ubidi_getVisualRun(UBiDi *pBiDi, int32_t runIndex,
   1436                    int32_t *pLogicalStart, int32_t *pLength);
   1437 
   1438 /**
   1439  * Get the visual position from a logical text position.
   1440  * If such a mapping is used many times on the same
   1441  * <code>UBiDi</code> object, then calling
   1442  * <code>ubidi_getLogicalMap()</code> is more efficient.<p>
   1443  *
   1444  * The value returned may be <code>#UBIDI_MAP_NOWHERE</code> if there is no
   1445  * visual position because the corresponding text character is a Bidi control
   1446  * removed from output by the option <code>#UBIDI_OPTION_REMOVE_CONTROLS</code>.
   1447  * <p>
   1448  * When the visual output is altered by using options of
   1449  * <code>ubidi_writeReordered()</code> such as <code>UBIDI_INSERT_LRM_FOR_NUMERIC</code>,
   1450  * <code>UBIDI_KEEP_BASE_COMBINING</code>, <code>UBIDI_OUTPUT_REVERSE</code>,
   1451  * <code>UBIDI_REMOVE_BIDI_CONTROLS</code>, the visual position returned may not
   1452  * be correct. It is advised to use, when possible, reordering options
   1453  * such as <code>UBIDI_OPTION_INSERT_MARKS</code> and <code>UBIDI_OPTION_REMOVE_CONTROLS</code>.
   1454  * <p>
   1455  * Note that in right-to-left runs, this mapping places
   1456  * second surrogates before first ones (which is generally a bad idea)
   1457  * and combining characters before base characters.
   1458  * Use of <code>ubidi_writeReordered()</code>, optionally with the
   1459  * <code>#UBIDI_KEEP_BASE_COMBINING</code> option can be considered instead
   1460  * of using the mapping, in order to avoid these issues.
   1461  *
   1462  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1463  *
   1464  * @param logicalIndex is the index of a character in the text.
   1465  *
   1466  * @param pErrorCode must be a valid pointer to an error code value.
   1467  *
   1468  * @return The visual position of this character.
   1469  *
   1470  * @see ubidi_getLogicalMap
   1471  * @see ubidi_getLogicalIndex
   1472  * @see ubidi_getProcessedLength
   1473  * @stable ICU 2.0
   1474  */
   1475 U_STABLE int32_t U_EXPORT2
   1476 ubidi_getVisualIndex(UBiDi *pBiDi, int32_t logicalIndex, UErrorCode *pErrorCode);
   1477 
   1478 /**
   1479  * Get the logical text position from a visual position.
   1480  * If such a mapping is used many times on the same
   1481  * <code>UBiDi</code> object, then calling
   1482  * <code>ubidi_getVisualMap()</code> is more efficient.<p>
   1483  *
   1484  * The value returned may be <code>#UBIDI_MAP_NOWHERE</code> if there is no
   1485  * logical position because the corresponding text character is a Bidi mark
   1486  * inserted in the output by option <code>#UBIDI_OPTION_INSERT_MARKS</code>.
   1487  * <p>
   1488  * This is the inverse function to <code>ubidi_getVisualIndex()</code>.
   1489  * <p>
   1490  * When the visual output is altered by using options of
   1491  * <code>ubidi_writeReordered()</code> such as <code>UBIDI_INSERT_LRM_FOR_NUMERIC</code>,
   1492  * <code>UBIDI_KEEP_BASE_COMBINING</code>, <code>UBIDI_OUTPUT_REVERSE</code>,
   1493  * <code>UBIDI_REMOVE_BIDI_CONTROLS</code>, the logical position returned may not
   1494  * be correct. It is advised to use, when possible, reordering options
   1495  * such as <code>UBIDI_OPTION_INSERT_MARKS</code> and <code>UBIDI_OPTION_REMOVE_CONTROLS</code>.
   1496  *
   1497  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1498  *
   1499  * @param visualIndex is the visual position of a character.
   1500  *
   1501  * @param pErrorCode must be a valid pointer to an error code value.
   1502  *
   1503  * @return The index of this character in the text.
   1504  *
   1505  * @see ubidi_getVisualMap
   1506  * @see ubidi_getVisualIndex
   1507  * @see ubidi_getResultLength
   1508  * @stable ICU 2.0
   1509  */
   1510 U_STABLE int32_t U_EXPORT2
   1511 ubidi_getLogicalIndex(UBiDi *pBiDi, int32_t visualIndex, UErrorCode *pErrorCode);
   1512 
   1513 /**
   1514  * Get a logical-to-visual index map (array) for the characters in the UBiDi
   1515  * (paragraph or line) object.
   1516  * <p>
   1517  * Some values in the map may be <code>#UBIDI_MAP_NOWHERE</code> if the
   1518  * corresponding text characters are Bidi controls removed from the visual
   1519  * output by the option <code>#UBIDI_OPTION_REMOVE_CONTROLS</code>.
   1520  * <p>
   1521  * When the visual output is altered by using options of
   1522  * <code>ubidi_writeReordered()</code> such as <code>UBIDI_INSERT_LRM_FOR_NUMERIC</code>,
   1523  * <code>UBIDI_KEEP_BASE_COMBINING</code>, <code>UBIDI_OUTPUT_REVERSE</code>,
   1524  * <code>UBIDI_REMOVE_BIDI_CONTROLS</code>, the visual positions returned may not
   1525  * be correct. It is advised to use, when possible, reordering options
   1526  * such as <code>UBIDI_OPTION_INSERT_MARKS</code> and <code>UBIDI_OPTION_REMOVE_CONTROLS</code>.
   1527  * <p>
   1528  * Note that in right-to-left runs, this mapping places
   1529  * second surrogates before first ones (which is generally a bad idea)
   1530  * and combining characters before base characters.
   1531  * Use of <code>ubidi_writeReordered()</code>, optionally with the
   1532  * <code>#UBIDI_KEEP_BASE_COMBINING</code> option can be considered instead
   1533  * of using the mapping, in order to avoid these issues.
   1534  *
   1535  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1536  *
   1537  * @param indexMap is a pointer to an array of <code>ubidi_getProcessedLength()</code>
   1538  *        indexes which will reflect the reordering of the characters.
   1539  *        If option <code>#UBIDI_OPTION_INSERT_MARKS</code> is set, the number
   1540  *        of elements allocated in <code>indexMap</code> must be no less than
   1541  *        <code>ubidi_getResultLength()</code>.
   1542  *        The array does not need to be initialized.<br><br>
   1543  *        The index map will result in <code>indexMap[logicalIndex]==visualIndex</code>.
   1544  *
   1545  * @param pErrorCode must be a valid pointer to an error code value.
   1546  *
   1547  * @see ubidi_getVisualMap
   1548  * @see ubidi_getVisualIndex
   1549  * @see ubidi_getProcessedLength
   1550  * @see ubidi_getResultLength
   1551  * @stable ICU 2.0
   1552  */
   1553 U_STABLE void U_EXPORT2
   1554 ubidi_getLogicalMap(UBiDi *pBiDi, int32_t *indexMap, UErrorCode *pErrorCode);
   1555 
   1556 /**
   1557  * Get a visual-to-logical index map (array) for the characters in the UBiDi
   1558  * (paragraph or line) object.
   1559  * <p>
   1560  * Some values in the map may be <code>#UBIDI_MAP_NOWHERE</code> if the
   1561  * corresponding text characters are Bidi marks inserted in the visual output
   1562  * by the option <code>#UBIDI_OPTION_INSERT_MARKS</code>.
   1563  * <p>
   1564  * When the visual output is altered by using options of
   1565  * <code>ubidi_writeReordered()</code> such as <code>UBIDI_INSERT_LRM_FOR_NUMERIC</code>,
   1566  * <code>UBIDI_KEEP_BASE_COMBINING</code>, <code>UBIDI_OUTPUT_REVERSE</code>,
   1567  * <code>UBIDI_REMOVE_BIDI_CONTROLS</code>, the logical positions returned may not
   1568  * be correct. It is advised to use, when possible, reordering options
   1569  * such as <code>UBIDI_OPTION_INSERT_MARKS</code> and <code>UBIDI_OPTION_REMOVE_CONTROLS</code>.
   1570  *
   1571  * @param pBiDi is the paragraph or line <code>UBiDi</code> object.
   1572  *
   1573  * @param indexMap is a pointer to an array of <code>ubidi_getResultLength()</code>
   1574  *        indexes which will reflect the reordering of the characters.
   1575  *        If option <code>#UBIDI_OPTION_REMOVE_CONTROLS</code> is set, the number
   1576  *        of elements allocated in <code>indexMap</code> must be no less than
   1577  *        <code>ubidi_getProcessedLength()</code>.
   1578  *        The array does not need to be initialized.<br><br>
   1579  *        The index map will result in <code>indexMap[visualIndex]==logicalIndex</code>.
   1580  *
   1581  * @param pErrorCode must be a valid pointer to an error code value.
   1582  *
   1583  * @see ubidi_getLogicalMap
   1584  * @see ubidi_getLogicalIndex
   1585  * @see ubidi_getProcessedLength
   1586  * @see ubidi_getResultLength
   1587  * @stable ICU 2.0
   1588  */
   1589 U_STABLE void U_EXPORT2
   1590 ubidi_getVisualMap(UBiDi *pBiDi, int32_t *indexMap, UErrorCode *pErrorCode);
   1591 
   1592 /**
   1593  * This is a convenience function that does not use a UBiDi object.
   1594  * It is intended to be used for when an application has determined the levels
   1595  * of objects (character sequences) and just needs to have them reordered (L2).
   1596  * This is equivalent to using <code>ubidi_getLogicalMap()</code> on a
   1597  * <code>UBiDi</code> object.
   1598  *
   1599  * @param levels is an array with <code>length</code> levels that have been determined by
   1600  *        the application.
   1601  *
   1602  * @param length is the number of levels in the array, or, semantically,
   1603  *        the number of objects to be reordered.
   1604  *        It must be <code>length>0</code>.
   1605  *
   1606  * @param indexMap is a pointer to an array of <code>length</code>
   1607  *        indexes which will reflect the reordering of the characters.
   1608  *        The array does not need to be initialized.<p>
   1609  *        The index map will result in <code>indexMap[logicalIndex]==visualIndex</code>.
   1610  * @stable ICU 2.0
   1611  */
   1612 U_STABLE void U_EXPORT2
   1613 ubidi_reorderLogical(const UBiDiLevel *levels, int32_t length, int32_t *indexMap);
   1614 
   1615 /**
   1616  * This is a convenience function that does not use a UBiDi object.
   1617  * It is intended to be used for when an application has determined the levels
   1618  * of objects (character sequences) and just needs to have them reordered (L2).
   1619  * This is equivalent to using <code>ubidi_getVisualMap()</code> on a
   1620  * <code>UBiDi</code> object.
   1621  *
   1622  * @param levels is an array with <code>length</code> levels that have been determined by
   1623  *        the application.
   1624  *
   1625  * @param length is the number of levels in the array, or, semantically,
   1626  *        the number of objects to be reordered.
   1627  *        It must be <code>length>0</code>.
   1628  *
   1629  * @param indexMap is a pointer to an array of <code>length</code>
   1630  *        indexes which will reflect the reordering of the characters.
   1631  *        The array does not need to be initialized.<p>
   1632  *        The index map will result in <code>indexMap[visualIndex]==logicalIndex</code>.
   1633  * @stable ICU 2.0
   1634  */
   1635 U_STABLE void U_EXPORT2
   1636 ubidi_reorderVisual(const UBiDiLevel *levels, int32_t length, int32_t *indexMap);
   1637 
   1638 /**
   1639  * Invert an index map.
   1640  * The index mapping of the first map is inverted and written to
   1641  * the second one.
   1642  *
   1643  * @param srcMap is an array with <code>length</code> elements
   1644  *        which defines the original mapping from a source array containing
   1645  *        <code>length</code> elements to a destination array.
   1646  *        Some elements of the source array may have no mapping in the
   1647  *        destination array. In that case, their value will be
   1648  *        the special value <code>UBIDI_MAP_NOWHERE</code>.
   1649  *        All elements must be >=0 or equal to <code>UBIDI_MAP_NOWHERE</code>.
   1650  *        Some elements may have a value >= <code>length</code>, if the
   1651  *        destination array has more elements than the source array.
   1652  *        There must be no duplicate indexes (two or more elements with the
   1653  *        same value except <code>UBIDI_MAP_NOWHERE</code>).
   1654  *
   1655  * @param destMap is an array with a number of elements equal to 1 + the highest
   1656  *        value in <code>srcMap</code>.
   1657  *        <code>destMap</code> will be filled with the inverse mapping.
   1658  *        If element with index i in <code>srcMap</code> has a value k different
   1659  *        from <code>UBIDI_MAP_NOWHERE</code>, this means that element i of
   1660  *        the source array maps to element k in the destination array.
   1661  *        The inverse map will have value i in its k-th element.
   1662  *        For all elements of the destination array which do not map to
   1663  *        an element in the source array, the corresponding element in the
   1664  *        inverse map will have a value equal to <code>UBIDI_MAP_NOWHERE</code>.
   1665  *
   1666  * @param length is the length of each array.
   1667  * @see UBIDI_MAP_NOWHERE
   1668  * @stable ICU 2.0
   1669  */
   1670 U_STABLE void U_EXPORT2
   1671 ubidi_invertMap(const int32_t *srcMap, int32_t *destMap, int32_t length);
   1672 
   1673 /** option flags for ubidi_writeReordered() */
   1674 
   1675 /**
   1676  * option bit for ubidi_writeReordered():
   1677  * keep combining characters after their base characters in RTL runs
   1678  *
   1679  * @see ubidi_writeReordered
   1680  * @stable ICU 2.0
   1681  */
   1682 #define UBIDI_KEEP_BASE_COMBINING       1
   1683 
   1684 /**
   1685  * option bit for ubidi_writeReordered():
   1686  * replace characters with the "mirrored" property in RTL runs
   1687  * by their mirror-image mappings
   1688  *
   1689  * @see ubidi_writeReordered
   1690  * @stable ICU 2.0
   1691  */
   1692 #define UBIDI_DO_MIRRORING              2
   1693 
   1694 /**
   1695  * option bit for ubidi_writeReordered():
   1696  * surround the run with LRMs if necessary;
   1697  * this is part of the approximate "inverse Bidi" algorithm
   1698  *
   1699  * <p>This option does not imply corresponding adjustment of the index
   1700  * mappings.</p>
   1701  *
   1702  * @see ubidi_setInverse
   1703  * @see ubidi_writeReordered
   1704  * @stable ICU 2.0
   1705  */
   1706 #define UBIDI_INSERT_LRM_FOR_NUMERIC    4
   1707 
   1708 /**
   1709  * option bit for ubidi_writeReordered():
   1710  * remove Bidi control characters
   1711  * (this does not affect #UBIDI_INSERT_LRM_FOR_NUMERIC)
   1712  *
   1713  * <p>This option does not imply corresponding adjustment of the index
   1714  * mappings.</p>
   1715  *
   1716  * @see ubidi_writeReordered
   1717  * @stable ICU 2.0
   1718  */
   1719 #define UBIDI_REMOVE_BIDI_CONTROLS      8
   1720 
   1721 /**
   1722  * option bit for ubidi_writeReordered():
   1723  * write the output in reverse order
   1724  *
   1725  * <p>This has the same effect as calling <code>ubidi_writeReordered()</code>
   1726  * first without this option, and then calling
   1727  * <code>ubidi_writeReverse()</code> without mirroring.
   1728  * Doing this in the same step is faster and avoids a temporary buffer.
   1729  * An example for using this option is output to a character terminal that
   1730  * is designed for RTL scripts and stores text in reverse order.</p>
   1731  *
   1732  * @see ubidi_writeReordered
   1733  * @stable ICU 2.0
   1734  */
   1735 #define UBIDI_OUTPUT_REVERSE            16
   1736 
   1737 /**
   1738  * Get the length of the source text processed by the last call to
   1739  * <code>ubidi_setPara()</code>. This length may be different from the length
   1740  * of the source text if option <code>#UBIDI_OPTION_STREAMING</code>
   1741  * has been set.
   1742  * <br>
   1743  * Note that whenever the length of the text affects the execution or the
   1744  * result of a function, it is the processed length which must be considered,
   1745  * except for <code>ubidi_setPara</code> (which receives unprocessed source
   1746  * text) and <code>ubidi_getLength</code> (which returns the original length
   1747  * of the source text).<br>
   1748  * In particular, the processed length is the one to consider in the following
   1749  * cases:
   1750  * <ul>
   1751  * <li>maximum value of the <code>limit</code> argument of
   1752  * <code>ubidi_setLine</code></li>
   1753  * <li>maximum value of the <code>charIndex</code> argument of
   1754  * <code>ubidi_getParagraph</code></li>
   1755  * <li>maximum value of the <code>charIndex</code> argument of
   1756  * <code>ubidi_getLevelAt</code></li>
   1757  * <li>number of elements in the array returned by <code>ubidi_getLevels</code></li>
   1758  * <li>maximum value of the <code>logicalStart</code> argument of
   1759  * <code>ubidi_getLogicalRun</code></li>
   1760  * <li>maximum value of the <code>logicalIndex</code> argument of
   1761  * <code>ubidi_getVisualIndex</code></li>
   1762  * <li>number of elements filled in the <code>*indexMap</code> argument of
   1763  * <code>ubidi_getLogicalMap</code></li>
   1764  * <li>length of text processed by <code>ubidi_writeReordered</code></li>
   1765  * </ul>
   1766  *
   1767  * @param pBiDi is the paragraph <code>UBiDi</code> object.
   1768  *
   1769  * @return The length of the part of the source text processed by
   1770  *         the last call to <code>ubidi_setPara</code>.
   1771  * @see ubidi_setPara
   1772  * @see UBIDI_OPTION_STREAMING
   1773  * @stable ICU 3.6
   1774  */
   1775 U_STABLE int32_t U_EXPORT2
   1776 ubidi_getProcessedLength(const UBiDi *pBiDi);
   1777 
   1778 /**
   1779  * Get the length of the reordered text resulting from the last call to
   1780  * <code>ubidi_setPara()</code>. This length may be different from the length
   1781  * of the source text if option <code>#UBIDI_OPTION_INSERT_MARKS</code>
   1782  * or option <code>#UBIDI_OPTION_REMOVE_CONTROLS</code> has been set.
   1783  * <br>
   1784  * This resulting length is the one to consider in the following cases:
   1785  * <ul>
   1786  * <li>maximum value of the <code>visualIndex</code> argument of
   1787  * <code>ubidi_getLogicalIndex</code></li>
   1788  * <li>number of elements of the <code>*indexMap</code> argument of
   1789  * <code>ubidi_getVisualMap</code></li>
   1790  * </ul>
   1791  * Note that this length stays identical to the source text length if
   1792  * Bidi marks are inserted or removed using option bits of
   1793  * <code>ubidi_writeReordered</code>, or if option
   1794  * <code>#UBIDI_REORDER_INVERSE_NUMBERS_AS_L</code> has been set.
   1795  *
   1796  * @param pBiDi is the paragraph <code>UBiDi</code> object.
   1797  *
   1798  * @return The length of the reordered text resulting from
   1799  *         the last call to <code>ubidi_setPara</code>.
   1800  * @see ubidi_setPara
   1801  * @see UBIDI_OPTION_INSERT_MARKS
   1802  * @see UBIDI_OPTION_REMOVE_CONTROLS
   1803  * @stable ICU 3.6
   1804  */
   1805 U_STABLE int32_t U_EXPORT2
   1806 ubidi_getResultLength(const UBiDi *pBiDi);
   1807 
   1808 U_CDECL_BEGIN
   1809 /**
   1810  * value returned by <code>UBiDiClassCallback</code> callbacks when
   1811  * there is no need to override the standard Bidi class for a given code point.
   1812  * @see UBiDiClassCallback
   1813  * @stable ICU 3.6
   1814  */
   1815 #define U_BIDI_CLASS_DEFAULT  U_CHAR_DIRECTION_COUNT
   1816 
   1817 /**
   1818  * Callback type declaration for overriding default Bidi class values with
   1819  * custom ones.
   1820  * <p>Usually, the function pointer will be propagated to a <code>UBiDi</code>
   1821  * object by calling the <code>ubidi_setClassCallback()</code> function;
   1822  * then the callback will be invoked by the UBA implementation any time the
   1823  * class of a character is to be determined.</p>
   1824  *
   1825  * @param context is a pointer to the callback private data.
   1826  *
   1827  * @param c       is the code point to get a Bidi class for.
   1828  *
   1829  * @return The directional property / Bidi class for the given code point
   1830  *         <code>c</code> if the default class has been overridden, or
   1831  *         <code>#U_BIDI_CLASS_DEFAULT</code> if the standard Bidi class value
   1832  *         for <code>c</code> is to be used.
   1833  * @see ubidi_setClassCallback
   1834  * @see ubidi_getClassCallback
   1835  * @stable ICU 3.6
   1836  */
   1837 typedef UCharDirection U_CALLCONV
   1838 UBiDiClassCallback(const void *context, UChar32 c);
   1839 
   1840 U_CDECL_END
   1841 
   1842 /**
   1843  * Retrieve the Bidi class for a given code point.
   1844  * <p>If a <code>#UBiDiClassCallback</code> callback is defined and returns a
   1845  * value other than <code>#U_BIDI_CLASS_DEFAULT</code>, that value is used;
   1846  * otherwise the default class determination mechanism is invoked.</p>
   1847  *
   1848  * @param pBiDi is the paragraph <code>UBiDi</code> object.
   1849  *
   1850  * @param c     is the code point whose Bidi class must be retrieved.
   1851  *
   1852  * @return The Bidi class for character <code>c</code> based
   1853  *         on the given <code>pBiDi</code> instance.
   1854  * @see UBiDiClassCallback
   1855  * @stable ICU 3.6
   1856  */
   1857 U_STABLE UCharDirection U_EXPORT2
   1858 ubidi_getCustomizedClass(UBiDi *pBiDi, UChar32 c);
   1859 
   1860 /**
   1861  * Set the callback function and callback data used by the UBA
   1862  * implementation for Bidi class determination.
   1863  * <p>This may be useful for assigning Bidi classes to PUA characters, or
   1864  * for special application needs. For instance, an application may want to
   1865  * handle all spaces like L or R characters (according to the base direction)
   1866  * when creating the visual ordering of logical lines which are part of a report
   1867  * organized in columns: there should not be interaction between adjacent
   1868  * cells.<p>
   1869  *
   1870  * @param pBiDi is the paragraph <code>UBiDi</code> object.
   1871  *
   1872  * @param newFn is the new callback function pointer.
   1873  *
   1874  * @param newContext is the new callback context pointer. This can be NULL.
   1875  *
   1876  * @param oldFn fillin: Returns the old callback function pointer. This can be
   1877  *                      NULL.
   1878  *
   1879  * @param oldContext fillin: Returns the old callback's context. This can be
   1880  *                           NULL.
   1881  *
   1882  * @param pErrorCode must be a valid pointer to an error code value.
   1883  *
   1884  * @see ubidi_getClassCallback
   1885  * @stable ICU 3.6
   1886  */
   1887 U_STABLE void U_EXPORT2
   1888 ubidi_setClassCallback(UBiDi *pBiDi, UBiDiClassCallback *newFn,
   1889                        const void *newContext, UBiDiClassCallback **oldFn,
   1890                        const void **oldContext, UErrorCode *pErrorCode);
   1891 
   1892 /**
   1893  * Get the current callback function used for Bidi class determination.
   1894  *
   1895  * @param pBiDi is the paragraph <code>UBiDi</code> object.
   1896  *
   1897  * @param fn fillin: Returns the callback function pointer.
   1898  *
   1899  * @param context fillin: Returns the callback's private context.
   1900  *
   1901  * @see ubidi_setClassCallback
   1902  * @stable ICU 3.6
   1903  */
   1904 U_STABLE void U_EXPORT2
   1905 ubidi_getClassCallback(UBiDi *pBiDi, UBiDiClassCallback **fn, const void **context);
   1906 
   1907 /**
   1908  * Take a <code>UBiDi</code> object containing the reordering
   1909  * information for a piece of text (one or more paragraphs) set by
   1910  * <code>ubidi_setPara()</code> or for a line of text set by
   1911  * <code>ubidi_setLine()</code> and write a reordered string to the
   1912  * destination buffer.
   1913  *
   1914  * This function preserves the integrity of characters with multiple
   1915  * code units and (optionally) combining characters.
   1916  * Characters in RTL runs can be replaced by mirror-image characters
   1917  * in the destination buffer. Note that "real" mirroring has
   1918  * to be done in a rendering engine by glyph selection
   1919  * and that for many "mirrored" characters there are no
   1920  * Unicode characters as mirror-image equivalents.
   1921  * There are also options to insert or remove Bidi control
   1922  * characters; see the description of the <code>destSize</code>
   1923  * and <code>options</code> parameters and of the option bit flags.
   1924  *
   1925  * @param pBiDi A pointer to a <code>UBiDi</code> object that
   1926  *              is set by <code>ubidi_setPara()</code> or
   1927  *              <code>ubidi_setLine()</code> and contains the reordering
   1928  *              information for the text that it was defined for,
   1929  *              as well as a pointer to that text.<br><br>
   1930  *              The text was aliased (only the pointer was stored
   1931  *              without copying the contents) and must not have been modified
   1932  *              since the <code>ubidi_setPara()</code> call.
   1933  *
   1934  * @param dest A pointer to where the reordered text is to be copied.
   1935  *             The source text and <code>dest[destSize]</code>
   1936  *             must not overlap.
   1937  *
   1938  * @param destSize The size of the <code>dest</code> buffer,
   1939  *                 in number of UChars.
   1940  *                 If the <code>UBIDI_INSERT_LRM_FOR_NUMERIC</code>
   1941  *                 option is set, then the destination length could be
   1942  *                 as large as
   1943  *                 <code>ubidi_getLength(pBiDi)+2*ubidi_countRuns(pBiDi)</code>.
   1944  *                 If the <code>UBIDI_REMOVE_BIDI_CONTROLS</code> option
   1945  *                 is set, then the destination length may be less than
   1946  *                 <code>ubidi_getLength(pBiDi)</code>.
   1947  *                 If none of these options is set, then the destination length
   1948  *                 will be exactly <code>ubidi_getProcessedLength(pBiDi)</code>.
   1949  *
   1950  * @param options A bit set of options for the reordering that control
   1951  *                how the reordered text is written.
   1952  *                The options include mirroring the characters on a code
   1953  *                point basis and inserting LRM characters, which is used
   1954  *                especially for transforming visually stored text
   1955  *                to logically stored text (although this is still an
   1956  *                imperfect implementation of an "inverse Bidi" algorithm
   1957  *                because it uses the "forward Bidi" algorithm at its core).
   1958  *                The available options are:
   1959  *                <code>#UBIDI_DO_MIRRORING</code>,
   1960  *                <code>#UBIDI_INSERT_LRM_FOR_NUMERIC</code>,
   1961  *                <code>#UBIDI_KEEP_BASE_COMBINING</code>,
   1962  *                <code>#UBIDI_OUTPUT_REVERSE</code>,
   1963  *                <code>#UBIDI_REMOVE_BIDI_CONTROLS</code>
   1964  *
   1965  * @param pErrorCode must be a valid pointer to an error code value.
   1966  *
   1967  * @return The length of the output string.
   1968  *
   1969  * @see ubidi_getProcessedLength
   1970  * @stable ICU 2.0
   1971  */
   1972 U_STABLE int32_t U_EXPORT2
   1973 ubidi_writeReordered(UBiDi *pBiDi,
   1974                      UChar *dest, int32_t destSize,
   1975                      uint16_t options,
   1976                      UErrorCode *pErrorCode);
   1977 
   1978 /**
   1979  * Reverse a Right-To-Left run of Unicode text.
   1980  *
   1981  * This function preserves the integrity of characters with multiple
   1982  * code units and (optionally) combining characters.
   1983  * Characters can be replaced by mirror-image characters
   1984  * in the destination buffer. Note that "real" mirroring has
   1985  * to be done in a rendering engine by glyph selection
   1986  * and that for many "mirrored" characters there are no
   1987  * Unicode characters as mirror-image equivalents.
   1988  * There are also options to insert or remove Bidi control
   1989  * characters.
   1990  *
   1991  * This function is the implementation for reversing RTL runs as part
   1992  * of <code>ubidi_writeReordered()</code>. For detailed descriptions
   1993  * of the parameters, see there.
   1994  * Since no Bidi controls are inserted here, the output string length
   1995  * will never exceed <code>srcLength</code>.
   1996  *
   1997  * @see ubidi_writeReordered
   1998  *
   1999  * @param src A pointer to the RTL run text.
   2000  *
   2001  * @param srcLength The length of the RTL run.
   2002  *
   2003  * @param dest A pointer to where the reordered text is to be copied.
   2004  *             <code>src[srcLength]</code> and <code>dest[destSize]</code>
   2005  *             must not overlap.
   2006  *
   2007  * @param destSize The size of the <code>dest</code> buffer,
   2008  *                 in number of UChars.
   2009  *                 If the <code>UBIDI_REMOVE_BIDI_CONTROLS</code> option
   2010  *                 is set, then the destination length may be less than
   2011  *                 <code>srcLength</code>.
   2012  *                 If this option is not set, then the destination length
   2013  *                 will be exactly <code>srcLength</code>.
   2014  *
   2015  * @param options A bit set of options for the reordering that control
   2016  *                how the reordered text is written.
   2017  *                See the <code>options</code> parameter in <code>ubidi_writeReordered()</code>.
   2018  *
   2019  * @param pErrorCode must be a valid pointer to an error code value.
   2020  *
   2021  * @return The length of the output string.
   2022  * @stable ICU 2.0
   2023  */
   2024 U_STABLE int32_t U_EXPORT2
   2025 ubidi_writeReverse(const UChar *src, int32_t srcLength,
   2026                    UChar *dest, int32_t destSize,
   2027                    uint16_t options,
   2028                    UErrorCode *pErrorCode);
   2029 
   2030 /*#define BIDI_SAMPLE_CODE*/
   2031 /*@}*/
   2032 
   2033 #endif
   2034