ICU 72.1 72.1
Namespaces | Typedefs | Enumerations | Functions
ubiditransform.h File Reference

C API: Bidi Transformations. More...

#include "unicode/utypes.h"
#include "unicode/ubidi.h"
#include "unicode/uchar.h"
#include "unicode/localpointer.h"

Go to the source code of this file.

Namespaces

namespace  icu
 File coll.h.
 

Typedefs

typedef struct UBiDiTransform UBiDiTransform
 Forward declaration of the UBiDiTransform structure that stores information used by the layout transformation engine. More...
 

Enumerations

enum  UBiDiOrder { UBIDI_LOGICAL = 0 , UBIDI_VISUAL }
 UBiDiOrder indicates the order of text. More...
 
enum  UBiDiMirroring { UBIDI_MIRRORING_OFF = 0 , UBIDI_MIRRORING_ON }
 UBiDiMirroring indicates whether or not characters with the "mirrored" property in RTL runs should be replaced with their mirror-image counterparts. More...
 

Functions

U_CAPI uint32_t ubiditransform_transform (UBiDiTransform *pBiDiTransform, const UChar *src, int32_t srcLength, UChar *dest, int32_t destSize, UBiDiLevel inParaLevel, UBiDiOrder inOrder, UBiDiLevel outParaLevel, UBiDiOrder outOrder, UBiDiMirroring doMirroring, uint32_t shapingOptions, UErrorCode *pErrorCode)
 Performs transformation of text from the bidi layout defined by the input ordering scheme to the bidi layout defined by the output ordering scheme, and applies character mirroring and Arabic shaping operations. More...
 
U_CAPI UBiDiTransformubiditransform_open (UErrorCode *pErrorCode)
 Allocates a UBiDiTransform object. More...
 
U_CAPI void ubiditransform_close (UBiDiTransform *pBidiTransform)
 Deallocates the given UBiDiTransform object. More...
 

Detailed Description

C API: Bidi Transformations.

Definition in file ubiditransform.h.

Typedef Documentation

◆ UBiDiTransform

Forward declaration of the UBiDiTransform structure that stores information used by the layout transformation engine.

Stable:
ICU 58

Definition at line 115 of file ubiditransform.h.

Enumeration Type Documentation

◆ UBiDiMirroring

UBiDiMirroring indicates whether or not characters with the "mirrored" property in RTL runs should be replaced with their mirror-image counterparts.

See also
UBIDI_DO_MIRRORING
ubidi_setReorderingOptions
ubidi_writeReordered
ubidi_writeReverse
Stable:
ICU 58
Enumerator
UBIDI_MIRRORING_OFF 

0: Constant indicating that character mirroring should not be performed.

This is the default.

Stable:
ICU 58
UBIDI_MIRRORING_ON 

1: Constant indicating that character mirroring should be performed.

This corresponds to calling ubidi_writeReordered or ubidi_writeReverse with the UBIDI_DO_MIRRORING option bit set.

Stable:
ICU 58

Definition at line 94 of file ubiditransform.h.

◆ UBiDiOrder

enum UBiDiOrder

UBiDiOrder indicates the order of text.

This bidi transformation engine supports all possible combinations (4 in total) of input and output text order:

  • <logical input, visual output>: unless the output direction is RTL, this corresponds to a normal operation of the Bidi algorithm as described in the Unicode Technical Report and implemented by UBiDi when the reordering mode is set to UBIDI_REORDER_DEFAULT. Visual RTL mode is not supported by UBiDi and is accomplished through reversing a visual LTR string,
  • <visual input, logical output>: unless the input direction is RTL, this corresponds to an "inverse bidi algorithm" in UBiDi with the reordering mode set to UBIDI_REORDER_INVERSE_LIKE_DIRECT. Visual RTL mode is not not supported by UBiDi and is accomplished through reversing a visual LTR string,
  • <logical input, logical output>: if the input and output base directions mismatch, this corresponds to the UBiDi implementation with the reordering mode set to UBIDI_REORDER_RUNS_ONLY; and if the input and output base directions are identical, the transformation engine will only handle character mirroring and Arabic shaping operations without reordering,
  • <visual input, visual output>: this reordering mode is not supported by the UBiDi engine; it implies character mirroring, Arabic shaping, and - if the input/output base directions mismatch - string reverse operations.
    See also
    ubidi_setInverse
    ubidi_setReorderingMode
    UBIDI_REORDER_DEFAULT
    UBIDI_REORDER_INVERSE_LIKE_DIRECT
    UBIDI_REORDER_RUNS_ONLY
    Stable:
    ICU 58
Enumerator
UBIDI_LOGICAL 

0: Constant indicating a logical order.

This is the default for input text.

Stable:
ICU 58
UBIDI_VISUAL 

1: Constant indicating a visual order.

This is a default for output text.

Stable:
ICU 58

Definition at line 71 of file ubiditransform.h.

Function Documentation

◆ ubiditransform_close()

U_CAPI void ubiditransform_close ( UBiDiTransform pBidiTransform)

Deallocates the given UBiDiTransform object.

Stable:
ICU 58

◆ ubiditransform_open()

U_CAPI UBiDiTransform * ubiditransform_open ( UErrorCode pErrorCode)

Allocates a UBiDiTransform object.

This object can be reused, e.g. with different ordering schemes, mirroring or shaping options.

Note:The object can only be reused in the same thread. All other threads should allocate a new UBiDiTransform object before using it.

Example of usage:

  
UErrorCode errorCode = U_ZERO_ERROR;
// Open a new UBiDiTransform.
UBiDiTransform* transform = ubiditransform_open(&errorCode);
// Run a transformation.
text1, -1, text2, -1,
&errorCode);
// Do something with the output text and invoke another transformation using
// that text as input.
text2, -1, text3, -1,
0, &errorCode);
*
@ UBIDI_LTR
Left-to-right text.
Definition: ubidi.h:441
@ UBIDI_RTL
Right-to-left text.
Definition: ubidi.h:453
U_CAPI uint32_t ubiditransform_transform(UBiDiTransform *pBiDiTransform, const UChar *src, int32_t srcLength, UChar *dest, int32_t destSize, UBiDiLevel inParaLevel, UBiDiOrder inOrder, UBiDiLevel outParaLevel, UBiDiOrder outOrder, UBiDiMirroring doMirroring, uint32_t shapingOptions, UErrorCode *pErrorCode)
Performs transformation of text from the bidi layout defined by the input ordering scheme to the bidi...
struct UBiDiTransform UBiDiTransform
Forward declaration of the UBiDiTransform structure that stores information used by the layout transf...
@ UBIDI_LOGICAL
0: Constant indicating a logical order.
@ UBIDI_VISUAL
1: Constant indicating a visual order.
@ UBIDI_MIRRORING_ON
1: Constant indicating that character mirroring should be performed.
U_CAPI UBiDiTransform * ubiditransform_open(UErrorCode *pErrorCode)
Allocates a UBiDiTransform object.
#define U_SHAPE_DIGITS_EN2AN
Digit shaping option: Replace European digits (U+0030...) by Arabic-Indic digits.
Definition: ushape.h:251
UErrorCode
Standard ICU4C error code type, a substitute for exceptions.
Definition: utypes.h:415
@ U_ZERO_ERROR
No error, no warning.
Definition: utypes.h:449

The UBiDiTransform object must be deallocated by calling ubiditransform_close().

Returns
An empty UBiDiTransform object.
Stable:
ICU 58

◆ ubiditransform_transform()

U_CAPI uint32_t ubiditransform_transform ( UBiDiTransform pBiDiTransform,
const UChar src,
int32_t  srcLength,
UChar dest,
int32_t  destSize,
UBiDiLevel  inParaLevel,
UBiDiOrder  inOrder,
UBiDiLevel  outParaLevel,
UBiDiOrder  outOrder,
UBiDiMirroring  doMirroring,
uint32_t  shapingOptions,
UErrorCode pErrorCode 
)

Performs transformation of text from the bidi layout defined by the input ordering scheme to the bidi layout defined by the output ordering scheme, and applies character mirroring and Arabic shaping operations.

In terms of UBiDi, such a transformation implies:

  • calling ubidi_setReorderingMode as needed (when the reordering mode is other than normal),
  • calling ubidi_setInverse as needed (when text should be transformed from a visual to a logical form),
  • resolving embedding levels of each character in the input text by calling ubidi_setPara,
  • reordering the characters based on the computed embedding levels, also performing character mirroring as needed, and streaming the result to the output, by calling ubidi_writeReordered,
  • performing Arabic digit and letter shaping on the output text by calling u_shapeArabic.

An "ordering scheme" encompasses the base direction and the order of text, and these characteristics must be defined by the caller for both input and output explicitly .

There are 36 possible combinations of <input, output> ordering schemes, which are partially supported by UBiDi already. Examples of the currently supported combinations:

  • <Logical LTR, Visual LTR>: this is equivalent to calling ubidi_setPara with paraLevel == UBIDI_LTR,
  • <Logical RTL, Visual LTR>: this is equivalent to calling ubidi_setPara with paraLevel == UBIDI_RTL,
  • <Logical Default ("Auto") LTR, Visual LTR>: this is equivalent to calling ubidi_setPara with paraLevel == UBIDI_DEFAULT_LTR,
  • <Logical Default ("Auto") RTL, Visual LTR>: this is equivalent to calling ubidi_setPara with paraLevel == UBIDI_DEFAULT_RTL,
  • <Visual LTR, Logical LTR>: this is equivalent to calling ubidi_setInverse(UBiDi*, true) and then ubidi_setPara with paraLevel == UBIDI_LTR,
  • <Visual LTR, Logical RTL>: this is equivalent to calling ubidi_setInverse(UBiDi*, true) and then ubidi_setPara with paraLevel == UBIDI_RTL.

All combinations that involve the Visual RTL scheme are unsupported by UBiDi, for instance:

  • <Logical LTR, Visual RTL>,
  • <Visual RTL, Logical RTL>.

Example of usage of the transformation engine:

  
UChar text1[] = {'a', 'b', 'c', 0x0625, '1', 0};
UChar text2[] = {'a', 'b', 'c', 0x0625, '1', 0};
UErrorCode errorCode = U_ZERO_ERROR;
// Run a transformation.
ubiditransform_transform(pBidiTransform,
text1, -1, text2, -1,
&errorCode);
// Do something with text2.
text2[4] = '2';
// Run a reverse transformation.
ubiditransform_transform(pBidiTransform,
text2, -1, text1, -1,
&errorCode);
*
@ UBIDI_MIRRORING_OFF
0: Constant indicating that character mirroring should not be performed.
char16_t UChar
The base type for UTF-16 code units and pointers.
Definition: umachine.h:412
#define U_SHAPE_DIGITS_AN2EN
Digit shaping option: Replace Arabic-Indic digits by European digits (U+0030...).
Definition: ushape.h:258
#define U_SHAPE_DIGIT_TYPE_AN_EXTENDED
Digit type option: Use Eastern (Extended) Arabic-Indic digits (U+06f0...U+06f9).
Definition: ushape.h:296
Parameters
pBiDiTransformA pointer to a UBiDiTransform object allocated with ubiditransform_open() or NULL.

This object serves for one-time setup to amortize initialization overheads. Use of this object is not thread-safe. All other threads should allocate a new UBiDiTransform object by calling ubiditransform_open() before using it. Alternatively, a caller can set this parameter to NULL, in which case the object will be allocated by the engine on the fly.

Parameters
srcA pointer to the text that the Bidi layout transformations will be performed on.

Note: the text must be (at least) srcLength long.

Parameters
srcLengthThe length of the text, in number of UChars. If length == -1 then the text must be zero-terminated.
destA pointer to where the processed text is to be copied.
destSizeThe size of the dest buffer, in number of UChars. If the U_SHAPE_LETTERS_UNSHAPE option is set, then the destination length could be as large as srcLength * 2. Otherwise, the destination length will not exceed srcLength. If the caller reserves the last position for zero-termination, it should be excluded from destSize.

destSize == -1 is allowed and makes sense when dest was holds some meaningful value, e.g. that of src. In this case dest must be zero-terminated.

Parameters
inParaLevelA base embedding level of the input as defined in ubidi_setPara documentation for the paraLevel parameter.
inOrderAn order of the input, which can be one of the UBiDiOrder values.
outParaLevelA base embedding level of the output as defined in ubidi_setPara documentation for the paraLevel parameter.
outOrderAn order of the output, which can be one of the UBiDiOrder values.
doMirroringIndicates whether or not to perform character mirroring, and can accept one of the UBiDiMirroring values.
shapingOptionsArabic digit and letter shaping options defined in the ushape.h documentation.

Note: Direction indicator options are computed by the transformation engine based on the effective ordering schemes, so user-defined direction indicators will be ignored.

Parameters
pErrorCodeA pointer to an error code value.
Returns
The destination length, i.e. the number of UChars written to dest. If the transformation fails, the return value will be 0 (and the error code will be written to pErrorCode).
See also
UBiDiLevel
UBiDiOrder
UBiDiMirroring
ubidi_setPara
u_shapeArabic
Stable:
ICU 58