ucnvsel.cpp | Explore in Territory

// © 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
/*
*******************************************************************************
*
*   Copyright (C) 2008-2011, International Business Machines
*   Corporation, Google and others.  All Rights Reserved.
*
*******************************************************************************
*/
// Author : [email protected] (Mohamed Eldawy)
// ucnvsel.cpp
//
// Purpose: To generate a list of encodings capable of handling
// a given Unicode text
//
// Started 09-April-2008

/**
 * \file
 *
 * This is an implementation of an encoding selector.
 * The goal is, given a unicode string, find the encodings
 * this string can be mapped to. To make processing faster
 * a trie is built when you call ucnvsel_open() that
 * stores all encodings a codepoint can map to
 */

#include "unicode/ucnvsel.h"

#if !UCONFIG_NO_CONVERSION

#include <string.h>

#include "unicode/uchar.h"
#include "unicode/uniset.h"
#include "unicode/ucnv.h"
#include "unicode/ustring.h"
#include "unicode/uchriter.h"
#include "utrie2.h"
#include "propsvec.h"
#include "uassert.h"
#include "ucmndata.h"
#include "udataswp.h"
#include "uenumimp.h"
#include "cmemory.h"
#include "cstring.h"

U_NAMESPACE_USE

struct UConverterSelector { … };

static void generateSelectorData(UConverterSelector* result,
                                 UPropsVectors *upvec,
                                 const USet* excludedCodePoints,
                                 const UConverterUnicodeSet whichSet,
                                 UErrorCode* status) { … }

/* open a selector. If converterListSize is 0, build for all converters.
   If excludedCodePoints is nullptr, don't exclude any codepoints */
U_CAPI UConverterSelector* U_EXPORT2
ucnvsel_open(const char* const*  converterList, int32_t converterListSize,
             const USet* excludedCodePoints,
             const UConverterUnicodeSet whichSet, UErrorCode* status) { … }

/* close opened selector */
U_CAPI void U_EXPORT2
ucnvsel_close(UConverterSelector *sel) { … }

static const UDataInfo dataInfo = …;

enum { … };

/*
 * Serialized form of a UConverterSelector, formatVersion 1:
 *
 * The serialized form begins with a standard ICU DataHeader with a UDataInfo
 * as the template above.
 * This is followed by:
 *   int32_t indexes[UCNVSEL_INDEX_COUNT];          // see index entry constants above
 *   serialized UTrie2;                             // indexes[UCNVSEL_INDEX_TRIE_SIZE] bytes
 *   uint32_t pv[indexes[UCNVSEL_INDEX_PV_COUNT]];  // bit vectors
 *   char* encodingNames[indexes[UCNVSEL_INDEX_NAMES_LENGTH]];  // NUL-terminated strings + padding
 */

/* serialize a selector */
U_CAPI int32_t U_EXPORT2
ucnvsel_serialize(const UConverterSelector* sel,
                  void* buffer, int32_t bufferCapacity, UErrorCode* status) { … }

/**
 * swap a selector into the desired Endianness and Asciiness of
 * the system. Just as FYI, selectors are always saved in the format
 * of the system that created them. They are only converted if used
 * on another system. In other words, selectors created on different
 * system can be different even if the params are identical (endianness
 * and Asciiness differences only)
 *
 * @param ds pointer to data swapper containing swapping info
 * @param inData pointer to incoming data
 * @param length length of inData in bytes
 * @param outData pointer to output data. Capacity should
 *                be at least equal to capacity of inData
 * @param status an in/out ICU UErrorCode
 * @return 0 on failure, number of bytes swapped on success
 *         number of bytes swapped can be smaller than length
 */
static int32_t
ucnvsel_swap(const UDataSwapper *ds,
             const void *inData, int32_t length,
             void *outData, UErrorCode *status) { … }

/* unserialize a selector */
U_CAPI UConverterSelector* U_EXPORT2
ucnvsel_openFromSerialized(const void* buffer, int32_t length, UErrorCode* status) { … }

// a bunch of functions for the enumeration thingie! Nothing fancy here. Just
// iterate over the selected encodings
struct Enumerator { … };

U_CDECL_BEGIN

static void U_CALLCONV
ucnvsel_close_selector_iterator(UEnumeration *enumerator) { … }


static int32_t U_CALLCONV
ucnvsel_count_encodings(UEnumeration *enumerator, UErrorCode *status) { … }


static const char* U_CALLCONV ucnvsel_next_encoding(UEnumeration* enumerator,
                                                 int32_t* resultLength,
                                                 UErrorCode* status) { … }

static void U_CALLCONV ucnvsel_reset_iterator(UEnumeration* enumerator,
                                           UErrorCode* status) { … }

U_CDECL_END


static const UEnumeration defaultEncodings = …;


// internal fn to intersect two sets of masks
// returns whether the mask has reduced to all zeros
static UBool intersectMasks(uint32_t* dest, const uint32_t* source1, int32_t len) { … }

// internal fn to count how many 1's are there in a mask
// algorithm taken from  http://graphics.stanford.edu/~seander/bithacks.html
static int16_t countOnes(uint32_t* mask, int32_t len) { … }


/* internal function! */
static UEnumeration *selectForMask(const UConverterSelector* sel,
                                   uint32_t *theMask, UErrorCode *status) { … }

/* check a string against the selector - UTF16 version */
U_CAPI UEnumeration * U_EXPORT2
ucnvsel_selectForString(const UConverterSelector* sel,
                        const char16_t *s, int32_t length, UErrorCode *status) { … }

/* check a string against the selector - UTF8 version */
U_CAPI UEnumeration * U_EXPORT2
ucnvsel_selectForUTF8(const UConverterSelector* sel,
                      const char *s, int32_t length, UErrorCode *status) { … }

#endif  // !UCONFIG_NO_CONVERSION
chromium/third_party/icu/source/common/ucnvsel.cpp
