chromium/third_party/icu/source/i18n/ucol_sit.cpp

// © 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
/*
*******************************************************************************
*   Copyright (C) 2004-2016, International Business Machines
*   Corporation and others.  All Rights Reserved.
*******************************************************************************
*   file name:  ucol_sit.cpp
*   encoding:   UTF-8
*   tab size:   8 (not used)
*   indentation:4
*
* Modification history
* Date        Name      Comments
* 03/12/2004  weiv      Creation
*/

#include "unicode/ustring.h"
#include "unicode/udata.h"
#include "unicode/utf16.h"
#include "utracimp.h"
#include "ucol_imp.h"
#include "ulocimp.h"
#include "bytesinkutil.h"
#include "cmemory.h"
#include "cstring.h"
#include "uresimp.h"
#include "unicode/coll.h"
#include "unicode/stringpiece.h"
#include "charstr.h"

U_NAMESPACE_USE

#ifdef UCOL_TRACE_SIT
# include <stdio.h>
#endif

#if !UCONFIG_NO_COLLATION

#include "unicode/tblcoll.h"

enum OptionsList { … };

/* option starters chars. */
static const char alternateHArg     = …;
static const char variableTopValArg = …;
static const char caseFirstArg      = …;
static const char numericCollArg    = …;
static const char caseLevelArg      = …;
static const char frenchCollArg     = …;
static const char hiraganaQArg      = …;
static const char keywordArg        = …;
static const char languageArg       = …;
static const char normArg           = …;
static const char providerArg       = …;
static const char regionArg         = …;
static const char strengthArg       = …;
static const char variableTopArg    = …;
static const char variantArg        = …;
static const char RFC3066Arg        = …;
static const char scriptArg         = …;

static const char collationKeyword[]  = …;
static const char providerKeyword[]  = …;


static const int32_t locElementCount = …;
static const int32_t locElementCapacity = …;
static const int32_t loc3066Capacity = …;

/* structure containing specification of a collator. Initialized
 * from a short string. Also used to construct a short string from a
 * collator instance
 */
struct CollatorSpec { … };

CollatorSpec::CollatorSpec() : … { … }


/* structure for converting between character attribute
 * representation and real collation attribute value.
 */
struct AttributeConversion { … };

static const AttributeConversion conversions[12] = …;


static UColAttributeValue
ucol_sit_letterToAttributeValue(char letter, UErrorCode *status) { … }

/* function prototype for functions used to parse a short string */
U_CDECL_BEGIN
ActionFunction;
U_CDECL_END

U_CDECL_BEGIN
static const char* U_CALLCONV
_processLocaleElement(CollatorSpec *spec, uint32_t value, const char* string,
                      UErrorCode *status)
{ … }
U_CDECL_END

U_CDECL_BEGIN
static const char* U_CALLCONV
_processRFC3066Locale(CollatorSpec *spec, uint32_t, const char* string,
                      UErrorCode *status)
{ … }

U_CDECL_END

U_CDECL_BEGIN
static const char* U_CALLCONV
_processCollatorOption(CollatorSpec *spec, uint32_t option, const char* string,
                       UErrorCode *status)
{ … }
U_CDECL_END


static char16_t
readHexCodeUnit(const char **string, UErrorCode *status)
{ … }

U_CDECL_BEGIN
static const char* U_CALLCONV
_processVariableTop(CollatorSpec *spec, uint32_t value1, const char* string, UErrorCode *status)
{ … }
U_CDECL_END


/* Table for parsing short strings */
struct ShortStringOptions { … };

static const ShortStringOptions options[UCOL_SIT_ITEMS_COUNT] = …;


static
const char* ucol_sit_readOption(const char *start, CollatorSpec *spec,
                            UErrorCode *status)
{ … }

static const char*
ucol_sit_readSpecs(CollatorSpec *s, const char *string,
                        UParseError *parseError, UErrorCode *status)
{ … }

static
int32_t ucol_sit_dumpSpecs(CollatorSpec *s, char *destination, int32_t capacity, UErrorCode *status)
{ … }

static void
ucol_sit_calculateWholeLocale(CollatorSpec *s, UErrorCode &status) { … }


U_CAPI void U_EXPORT2
ucol_prepareShortStringOpen( const char *definition,
                          UBool,
                          UParseError *parseError,
                          UErrorCode *status)
{ … }


U_CAPI UCollator* U_EXPORT2
ucol_openFromShortString( const char *definition,
                          UBool forceDefaults,
                          UParseError *parseError,
                          UErrorCode *status)
{ … }


U_CAPI int32_t U_EXPORT2
ucol_getShortDefinitionString(const UCollator *coll,
                              const char *locale,
                              char *dst,
                              int32_t capacity,
                              UErrorCode *status)
{ … }

U_CAPI int32_t U_EXPORT2
ucol_normalizeShortDefinitionString(const char *definition,
                                    char *destination,
                                    int32_t capacity,
                                    UParseError *parseError,
                                    UErrorCode *status)
{ … }

/**
 * Get a set containing the contractions defined by the collator. The set includes
 * both the UCA contractions and the contractions defined by the collator
 * @param coll collator
 * @param conts the set to hold the result
 * @param status to hold the error code
 * @return the size of the contraction set
 */
U_CAPI int32_t U_EXPORT2
ucol_getContractions( const UCollator *coll,
                  USet *contractions,
                  UErrorCode *status)
{ … }

/**
 * Get a set containing the expansions defined by the collator. The set includes
 * both the UCA expansions and the expansions defined by the tailoring
 * @param coll collator
 * @param conts the set to hold the result
 * @param addPrefixes add the prefix contextual elements to contractions
 * @param status to hold the error code
 *
 * @draft ICU 3.4
 */
U_CAPI void U_EXPORT2
ucol_getContractionsAndExpansions( const UCollator *coll,
                  USet *contractions,
                  USet *expansions,
                  UBool addPrefixes,
                  UErrorCode *status)
{ … }
#endif