* *
limitof a sequence of characters is the position just after their * last character, i.e., one more than that position.
*
* Some of the API functions provide access to runs
.
* Such a run
is defined as a sequence of characters
* that are at the same embedding level
* after performing the BIDI algorithm.
*
* @author Markus W. Scherer. Ported to Mozilla by Simon Montagu
* @version 1.0
*/
/**
* Special value which can be returned by the mapping functions when a logical
* index has no corresponding visual index or vice-versa.
* @see GetVisualIndex
* @see GetVisualMap
* @see GetLogicalIndex
* @see GetLogicalMap
*/
#define NSBIDI_MAP_NOWHERE (-1)
/* miscellaneous definitions ------------------------------------------------ */
/* helper macros for each allocated array member */
#define GETDIRPROPSMEMORY(length) nsBidi::GetMemory((void **)&mDirPropsMemory, \
&mDirPropsSize, \
(length))
#define GETLEVELSMEMORY(length) nsBidi::GetMemory((void **)&mLevelsMemory, \
&mLevelsSize, \
(length))
#define GETRUNSMEMORY(length) nsBidi::GetMemory((void **)&mRunsMemory, \
&mRunsSize, \
(length)*sizeof(Run))
#define GETISOLATESMEMORY(length) nsBidi::GetMemory((void **)&mIsolatesMemory, \
&mIsolatesSize, \
(length)*sizeof(Isolate))
#define GETOPENINGSMEMORY(length) nsBidi::GetMemory((void **)&mOpeningsMemory, \
&mOpeningsSize, \
(length)*sizeof(Opening))
/*
* Sometimes, bit values are more appropriate
* to deal with directionality properties.
* Abbreviations in these macro names refer to names
* used in the Bidi algorithm.
*/
typedef uint8_t DirProp;
#define DIRPROP_FLAG(dir) (1UL<<(dir))
/* special flag for multiple runs from explicit embedding codes */
#define DIRPROP_FLAG_MULTI_RUNS (1UL<<31)
/* are there any characters that are LTR or RTL? */
#define MASK_LTR (DIRPROP_FLAG(L)|DIRPROP_FLAG(EN)|DIRPROP_FLAG(ENL)| \
DIRPROP_FLAG(ENR)|DIRPROP_FLAG(AN)|DIRPROP_FLAG(LRE)| \
DIRPROP_FLAG(LRO)|DIRPROP_FLAG(LRI))
#define MASK_RTL (DIRPROP_FLAG(R)|DIRPROP_FLAG(AL)|DIRPROP_FLAG(RLE)| \
DIRPROP_FLAG(RLO)|DIRPROP_FLAG(RLI))
#define MASK_R_AL (DIRPROP_FLAG(R)|DIRPROP_FLAG(AL))
/* explicit embedding codes */
#define MASK_EXPLICIT (DIRPROP_FLAG(LRE)|DIRPROP_FLAG(LRO)|DIRPROP_FLAG(RLE)|DIRPROP_FLAG(RLO)|DIRPROP_FLAG(PDF))
/* explicit isolate codes */
#define MASK_ISO (DIRPROP_FLAG(LRI)|DIRPROP_FLAG(RLI)|DIRPROP_FLAG(FSI)|DIRPROP_FLAG(PDI))
#define MASK_BN_EXPLICIT (DIRPROP_FLAG(BN)|MASK_EXPLICIT)
/* paragraph and segment separators */
#define MASK_B_S (DIRPROP_FLAG(B)|DIRPROP_FLAG(S))
/* all types that are counted as White Space or Neutral in some steps */
#define MASK_WS (MASK_B_S|DIRPROP_FLAG(WS)|MASK_BN_EXPLICIT|MASK_ISO)
/* types that are neutrals or could becomes neutrals in (Wn) */
#define MASK_POSSIBLE_N (DIRPROP_FLAG(O_N)|DIRPROP_FLAG(CS)|DIRPROP_FLAG(ES)|DIRPROP_FLAG(ET)|MASK_WS)
/*
* These types may be changed to "e",
* the embedding type (L or R) of the run,
* in the Bidi algorithm (N2)
*/
#define MASK_EMBEDDING (DIRPROP_FLAG(NSM)|MASK_POSSIBLE_N)
/* the dirProp's L and R are defined to 0 and 1 values in nsCharType */
#define GET_LR_FROM_LEVEL(level) ((DirProp)((level)&1))
#define IS_DEFAULT_LEVEL(level) (((level)&0xfe)==0xfe)
/*
* The following bit is used for the directional isolate status.
* Stack entries corresponding to isolate sequences are greater than ISOLATE.
*/
#define ISOLATE 0x0100
/* number of isolate entries allocated initially without malloc */
#define SIMPLE_ISOLATES_SIZE 5
/* number of isolate run entries for paired brackets allocated initially without malloc */
#define SIMPLE_OPENINGS_COUNT 8
/* handle surrogate pairs --------------------------------------------------- */
#define IS_FIRST_SURROGATE(uchar) (((uchar)&0xfc00)==0xd800)
#define IS_SECOND_SURROGATE(uchar) (((uchar)&0xfc00)==0xdc00)
/* get the UTF-32 value directly from the surrogate pseudo-characters */
#define SURROGATE_OFFSET ((0xd800<<10UL)+0xdc00-0x10000)
#define GET_UTF_32(first, second) (((first)<<10UL)+(second)-SURROGATE_OFFSET)
#if !ENABLE_INTL_API // these are provided by ICU if present in the build
#define UTF_ERROR_VALUE 0xffff
/* definitions with forward iteration --------------------------------------- */
/*
* all the macros that go forward assume that
* the initial offset is 0<=i
* Reordering can be done on a line, or on a paragraph which is
* then interpreted as one single line.
*
* On construction, the class is initially empty. It is assigned
* the Bidi properties of a paragraph by
* A Bidi class can be reused for as long as it is not deallocated
* by calling its destructor.
*
* This object can be reused for as long as it is not destroyed.
*
*
* This function takes a single plain text paragraph with or without
* externally specified embedding levels from
*
* If the entire paragraph consists of text of only one direction, then
* the function may not perform all the steps described by the algorithm,
* i.e., some levels may not be the same as if all steps were performed.
* This is not relevant for unidirectional text.
*
* The text must be externally split into separate paragraphs (rule P1).
* Paragraph separators (B) should appear at most at the very end.
*
* @param aText is a pointer to the single-paragraph text that the
* Bidi algorithm will be performed on
* (step (P1) of the algorithm is performed externally).
* The text must be (at least)
* This is especially useful for line-breaking on a paragraph.
*
* @param aLogicalStart is the first character of the run.
*
* @param aLogicalLimit will receive the limit of the run.
* The l-value that you point to here may be the
* same expression (variable) as the one for
*
*
*
* Example:
* @code
* int32_t i, count, logicalStart, visualIndex=0, length;
* nsBidiDirection dir;
* pBidi->CountRuns(&count);
* for(i=0; i
* The index map will result in SetPara
* or the Bidi properties of a line of a paragraph by
* SetLine.SetPara will allocate additional memory for
* internal structures as necessary.
*/
class nsBidi
{
public:
/** @brief Default constructor.
*
* The nsBidi object is initially empty. It is assigned
* the Bidi properties of a paragraph by SetPara()
* or the Bidi properties of a line of a paragraph by
* GetLine().SetPara() will allocate additional memory for
* internal structures as necessary.
*
*/
nsBidi();
/** @brief Destructor. */
virtual ~nsBidi();
/**
* Perform the Unicode Bidi algorithm. It is defined in the
* Unicode Technical Report 9,
* version 5,
* also described in The Unicode Standard, Version 3.0 .styled
text
* and computes the left-right-directionality of each character.
* For example, in pure LTR text with numbers the numbers would get
* a resolved level of 2 higher than the surrounding text according to
* the algorithm. This implementation may set all resolved levels to
* the same value in such a case.aLength long.
*
* @param aLength is the length of the text; if aLength==-1 then
* the text must be zero-terminated.
*
* @param aParaLevel specifies the default level for the paragraph;
* it is typically 0 (LTR) or 1 (RTL).
* If the function shall determine the paragraph level from the text,
* then aParaLevel can be set to
* either NSBIDI_DEFAULT_LTR
* or NSBIDI_DEFAULT_RTL;
* if there is no strongly typed character, then
* the desired default is used (0 for LTR or 1 for RTL).
* Any other value between 0 and NSBIDI_MAX_EXPLICIT_LEVEL is also valid,
* with odd levels indicating RTL.
*/
nsresult SetPara(const char16_t *aText, int32_t aLength, nsBidiLevel aParaLevel);
/**
* Get the directionality of the text.
*
* @param aDirection receives a NSBIDI_XXX value that indicates if the entire text
* represented by this object is unidirectional,
* and which direction, or if it is mixed-directional.
*
* @see nsBidiDirection
*/
nsresult GetDirection(nsBidiDirection* aDirection);
/**
* Get the paragraph level of the text.
*
* @param aParaLevel receives a NSBIDI_XXX value indicating the paragraph level
*
* @see nsBidiLevel
*/
nsresult GetParaLevel(nsBidiLevel* aParaLevel);
/**
* Get a logical run.
* This function returns information about a run and is used
* to retrieve runs in logical order.aLogicalStart.
* This pointer can be nullptr if this
* value is not necessary.
*
* @param aLevel will receive the level of the run.
* This pointer can be nullptr if this
* value is not necessary.
*/
nsresult GetLogicalRun(int32_t aLogicalStart, int32_t* aLogicalLimit, nsBidiLevel* aLevel);
/**
* Get the number of runs.
* This function may invoke the actual reordering on the
* nsBidi object, after SetPara
* may have resolved only the levels of the text. Therefore,
* CountRuns may have to allocate memory,
* and may fail doing so.
*
* @param aRunCount will receive the number of runs.
*/
nsresult CountRuns(int32_t* aRunCount);
/**
* Get one run's logical start, length, and directionality,
* which can be 0 for LTR or 1 for RTL.
* In an RTL run, the character at the logical start is
* visually on the right of the displayed run.
* The length is the number of characters in the run.CountRuns should be called
* before the runs are retrieved.
*
* @param aRunIndex is the number of the run in visual order, in the
* range [0..CountRuns-1].
*
* @param aLogicalStart is the first logical character index in the text.
* The pointer may be nullptr if this index is not needed.
*
* @param aLength is the number of characters (at least one) in the run.
* The pointer may be nullptr if this is not needed.
*
* @param aDirection will receive the directionality of the run,
* NSBIDI_LTR==0 or NSBIDI_RTL==1,
* never NSBIDI_MIXED.
*
* @see CountRunsGetVisualMap on a
* nsBidi object.
*
* @param aLevels is an array with aLength levels that have been determined by
* the application.
*
* @param aLength is the number of levels in the array, or, semantically,
* the number of objects to be reordered.
* It must be aLength>0.
*
* @param aIndexMap is a pointer to an array of aLength
* indexes which will reflect the reordering of the characters.
* The array does not need to be initialized.aIndexMap[aVisualIndex]==aLogicalIndex.
*/
static nsresult ReorderVisual(const nsBidiLevel *aLevels, int32_t aLength, int32_t *aIndexMap);
/**
* Reverse a Right-To-Left run of Unicode text.
*
* This function preserves the integrity of characters with multiple
* code units and (optionally) modifier letters.
* Characters can be replaced by mirror-image characters
* in the destination buffer. Note that "real" mirroring has
* to be done in a rendering engine by glyph selection
* and that for many "mirrored" characters there are no
* Unicode characters as mirror-image equivalents.
* There are also options to insert or remove Bidi control
* characters; see the description of the aDestSize
* and aOptions parameters and of the option bit flags.
*
* Since no Bidi controls are inserted here, this function will never
* write more than aSrcLength characters to aDest.
*
* @param aSrc A pointer to the RTL run text.
*
* @param aSrcLength The length of the RTL run.
* If the NSBIDI_REMOVE_BIDI_CONTROLS option
* is set, then the destination length may be less than
* aSrcLength.
* If this option is not set, then the destination length
* will be exactly aSrcLength.
*
* @param aDest A pointer to where the reordered text is to be copied.
* aSrc[aSrcLength] and aDest[aSrcLength]
* must not overlap.
*
* @param aOptions A bit set of options for the reordering that control
* how the reordered text is written.
*
* @param aDestSize will receive the number of characters that were written to aDest.
*/
nsresult WriteReverse(const char16_t *aSrc, int32_t aSrcLength, char16_t *aDest, uint16_t aOptions, int32_t *aDestSize);
protected:
friend class nsBidiPresUtils;
class BracketData {
public:
explicit BracketData(const nsBidi* aBidi);
~BracketData();
void ProcessBoundary(int32_t aLastDirControlCharPos,
nsBidiLevel aContextLevel,
nsBidiLevel aEmbeddingLevel,
const DirProp* aDirProps);
void ProcessLRI_RLI(nsBidiLevel aLevel);
void ProcessPDI();
bool AddOpening(char16_t aMatch, int32_t aPosition);
void FixN0c(int32_t aOpeningIndex, int32_t aNewPropPosition,
DirProp aNewProp, DirProp* aDirProps);
DirProp ProcessClosing(int32_t aOpenIdx, int32_t aPosition,
DirProp* aDirProps);
bool ProcessChar(int32_t aPosition, char16_t aCh, DirProp* aDirProps,
nsBidiLevel* aLevels);
private:
// array of opening entries which should be enough in most cases;
// no malloc() needed
Opening mSimpleOpenings[SIMPLE_OPENINGS_COUNT];
Opening* mOpenings; // pointer to current array of entries,
// either mSimpleOpenings or malloced array
Opening* mOpeningsMemory;
size_t mOpeningsSize;
// array of nested isolated sequence entries; can never exceed
// UBIDI_MAX_EXPLICIT_LEVEL
// + 1 for index 0
// + 1 for before the first isolated sequence
IsoRun mIsoRuns[NSBIDI_MAX_EXPLICIT_LEVEL+2];
int32_t mIsoRunLast; // index of last used entry in mIsoRuns
int32_t mOpeningsCount; // number of allocated entries in mOpenings
};
/** length of the current text */
int32_t mLength;
/** memory sizes in bytes */
size_t mDirPropsSize, mLevelsSize, mRunsSize;
size_t mIsolatesSize;
/** allocated memory */
DirProp* mDirPropsMemory;
nsBidiLevel* mLevelsMemory;
Run* mRunsMemory;
Isolate* mIsolatesMemory;
DirProp* mDirProps;
nsBidiLevel* mLevels;
/** the paragraph level */
nsBidiLevel mParaLevel;
/** flags is a bit set for which directional properties are in the text */
Flags mFlags;
/** the overall paragraph or line directionality - see nsBidiDirection */
nsBidiDirection mDirection;
/** characters after trailingWSStart are WS and are */
/* implicitly at the paraLevel (rule (L1)) - levels may not reflect that */
int32_t mTrailingWSStart;
/** fields for line reordering */
int32_t mRunCount; /* ==-1: runs not set up yet */
Run* mRuns;
/** for non-mixed text, we only need a tiny array of runs (no malloc()) */
Run mSimpleRuns[1];
/* maxium of current nesting depth of isolate sequences */
/* Within ResolveExplicitLevels() and checkExpicitLevels(), this is the maximal
nesting encountered.
Within ResolveImplicitLevels(), this is the index of the current isolates
stack entry. */
int32_t mIsolateCount;
Isolate* mIsolates;
/** for simple text, have a small stack (no malloc()) */
Isolate mSimpleIsolates[SIMPLE_ISOLATES_SIZE];
private:
void Init();
static bool GetMemory(void **aMemory, size_t* aSize, size_t aSizeNeeded);
void Free();
void GetDirProps(const char16_t *aText);
void ResolveExplicitLevels(nsBidiDirection *aDirection, const char16_t *aText);
nsBidiDirection DirectionFromFlags(Flags aFlags);
void ProcessPropertySeq(LevState *pLevState, uint8_t _prop, int32_t start, int32_t limit);
void ResolveImplicitLevels(int32_t aStart, int32_t aLimit, DirProp aSOR, DirProp aEOR);
void AdjustWSLevels();
void SetTrailingWSStart();
bool GetRuns();
void GetSingleRun(nsBidiLevel aLevel);
void ReorderLine(nsBidiLevel aMinLevel, nsBidiLevel aMaxLevel);
static bool PrepareReorder(const nsBidiLevel *aLevels, int32_t aLength, int32_t *aIndexMap, nsBidiLevel *aMinLevel, nsBidiLevel *aMaxLevel);
};
#endif // _nsBidi_noICU_h_