michael@0: /*
michael@0: **********************************************************************
michael@0: *   Copyright (C) 1999-2013, International Business Machines
michael@0: *   Corporation and others.  All Rights Reserved.
michael@0: **********************************************************************
michael@0:  *  ucnv.h:
michael@0:  *  External APIs for the ICU's codeset conversion library
michael@0:  *  Bertrand A. Damiba
michael@0:  *
michael@0:  * Modification History:
michael@0:  *
michael@0:  *   Date        Name        Description
michael@0:  *   04/04/99    helena      Fixed internal header inclusion.
michael@0:  *   05/11/00    helena      Added setFallback and usesFallback APIs.
michael@0:  *   06/29/2000  helena      Major rewrite of the callback APIs.
michael@0:  *   12/07/2000  srl         Update of documentation
michael@0:  */
michael@0: 
michael@0: /**
michael@0:  * \file
michael@0:  * \brief C API: Character conversion 
michael@0:  *
michael@0:  * <h2>Character Conversion C API</h2>
michael@0:  *
michael@0:  * <p>This API is used to convert codepage or character encoded data to and
michael@0:  * from UTF-16. You can open a converter with {@link ucnv_open() }. With that
michael@0:  * converter, you can get its properties, set options, convert your data and
michael@0:  * close the converter.</p>
michael@0:  *
michael@0:  * <p>Since many software programs recogize different converter names for
michael@0:  * different types of converters, there are other functions in this API to
michael@0:  * iterate over the converter aliases. The functions {@link ucnv_getAvailableName() },
michael@0:  * {@link ucnv_getAlias() } and {@link ucnv_getStandardName() } are some of the
michael@0:  * more frequently used alias functions to get this information.</p>
michael@0:  *
michael@0:  * <p>When a converter encounters an illegal, irregular, invalid or unmappable character
michael@0:  * its default behavior is to use a substitution character to replace the
michael@0:  * bad byte sequence. This behavior can be changed by using {@link ucnv_setFromUCallBack() }
michael@0:  * or {@link ucnv_setToUCallBack() } on the converter. The header ucnv_err.h defines
michael@0:  * many other callback actions that can be used instead of a character substitution.</p>
michael@0:  *
michael@0:  * <p>More information about this API can be found in our 
michael@0:  * <a href="http://icu-project.org/userguide/conversion.html">User's
michael@0:  * Guide</a>.</p>
michael@0:  */
michael@0: 
michael@0: #ifndef UCNV_H
michael@0: #define UCNV_H
michael@0: 
michael@0: #include "unicode/ucnv_err.h"
michael@0: #include "unicode/uenum.h"
michael@0: #include "unicode/localpointer.h"
michael@0: 
michael@0: #ifndef __USET_H__
michael@0: 
michael@0: /**
michael@0:  * USet is the C API type for Unicode sets.
michael@0:  * It is forward-declared here to avoid including the header file if related
michael@0:  * conversion APIs are not used.
michael@0:  * See unicode/uset.h
michael@0:  *
michael@0:  * @see ucnv_getUnicodeSet
michael@0:  * @stable ICU 2.6
michael@0:  */
michael@0: struct USet;
michael@0: /** @stable ICU 2.6 */
michael@0: typedef struct USet USet;
michael@0: 
michael@0: #endif
michael@0: 
michael@0: #if !UCONFIG_NO_CONVERSION
michael@0: 
michael@0: U_CDECL_BEGIN
michael@0: 
michael@0: /** Maximum length of a converter name including the terminating NULL @stable ICU 2.0 */
michael@0: #define UCNV_MAX_CONVERTER_NAME_LENGTH 60
michael@0: /** Maximum length of a converter name including path and terminating NULL @stable ICU 2.0 */
michael@0: #define UCNV_MAX_FULL_FILE_NAME_LENGTH (600+UCNV_MAX_CONVERTER_NAME_LENGTH)
michael@0: 
michael@0: /** Shift in for EBDCDIC_STATEFUL and iso2022 states @stable ICU 2.0 */
michael@0: #define  UCNV_SI 0x0F
michael@0: /** Shift out for EBDCDIC_STATEFUL and iso2022 states @stable ICU 2.0 */
michael@0: #define  UCNV_SO 0x0E
michael@0: 
michael@0: /**
michael@0:  * Enum for specifying basic types of converters
michael@0:  * @see ucnv_getType
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: typedef enum {
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_UNSUPPORTED_CONVERTER = -1,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_SBCS = 0,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_DBCS = 1,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_MBCS = 2,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LATIN_1 = 3,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_UTF8 = 4,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_UTF16_BigEndian = 5,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_UTF16_LittleEndian = 6,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_UTF32_BigEndian = 7,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_UTF32_LittleEndian = 8,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_EBCDIC_STATEFUL = 9,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_ISO_2022 = 10,
michael@0: 
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LMBCS_1 = 11,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LMBCS_2, 
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LMBCS_3,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LMBCS_4,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LMBCS_5,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LMBCS_6,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LMBCS_8,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LMBCS_11,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LMBCS_16,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LMBCS_17,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LMBCS_18,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LMBCS_19,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_LMBCS_LAST = UCNV_LMBCS_19,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_HZ,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_SCSU,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_ISCII,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_US_ASCII,
michael@0:     /** @stable ICU 2.0 */
michael@0:     UCNV_UTF7,
michael@0:     /** @stable ICU 2.2 */
michael@0:     UCNV_BOCU1,
michael@0:     /** @stable ICU 2.2 */
michael@0:     UCNV_UTF16,
michael@0:     /** @stable ICU 2.2 */
michael@0:     UCNV_UTF32,
michael@0:     /** @stable ICU 2.2 */
michael@0:     UCNV_CESU8,
michael@0:     /** @stable ICU 2.4 */
michael@0:     UCNV_IMAP_MAILBOX,
michael@0:     /** @stable ICU 4.8 */
michael@0:     UCNV_COMPOUND_TEXT,
michael@0: 
michael@0:     /* Number of converter types for which we have conversion routines. */
michael@0:     UCNV_NUMBER_OF_SUPPORTED_CONVERTER_TYPES
michael@0: } UConverterType;
michael@0: 
michael@0: /**
michael@0:  * Enum for specifying which platform a converter ID refers to.
michael@0:  * The use of platform/CCSID is not recommended. See ucnv_openCCSID().
michael@0:  *
michael@0:  * @see ucnv_getPlatform
michael@0:  * @see ucnv_openCCSID
michael@0:  * @see ucnv_getCCSID
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: typedef enum {
michael@0:     UCNV_UNKNOWN = -1,
michael@0:     UCNV_IBM = 0
michael@0: } UConverterPlatform;
michael@0: 
michael@0: /**
michael@0:  * Function pointer for error callback in the codepage to unicode direction.
michael@0:  * Called when an error has occured in conversion to unicode, or on open/close of the callback (see reason).
michael@0:  * @param context Pointer to the callback's private data
michael@0:  * @param args Information about the conversion in progress
michael@0:  * @param codeUnits Points to 'length' bytes of the concerned codepage sequence
michael@0:  * @param length Size (in bytes) of the concerned codepage sequence
michael@0:  * @param reason Defines the reason the callback was invoked
michael@0:  * @param pErrorCode    ICU error code in/out parameter.
michael@0:  *                      For converter callback functions, set to a conversion error
michael@0:  *                      before the call, and the callback may reset it to U_ZERO_ERROR.
michael@0:  * @see ucnv_setToUCallBack
michael@0:  * @see UConverterToUnicodeArgs
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: typedef void (U_EXPORT2 *UConverterToUCallback) (
michael@0:                   const void* context,
michael@0:                   UConverterToUnicodeArgs *args,
michael@0:                   const char *codeUnits,
michael@0:                   int32_t length,
michael@0:                   UConverterCallbackReason reason,
michael@0:                   UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Function pointer for error callback in the unicode to codepage direction.
michael@0:  * Called when an error has occured in conversion from unicode, or on open/close of the callback (see reason).
michael@0:  * @param context Pointer to the callback's private data
michael@0:  * @param args Information about the conversion in progress
michael@0:  * @param codeUnits Points to 'length' UChars of the concerned Unicode sequence
michael@0:  * @param length Size (in bytes) of the concerned codepage sequence
michael@0:  * @param codePoint Single UChar32 (UTF-32) containing the concerend Unicode codepoint.
michael@0:  * @param reason Defines the reason the callback was invoked
michael@0:  * @param pErrorCode    ICU error code in/out parameter.
michael@0:  *                      For converter callback functions, set to a conversion error
michael@0:  *                      before the call, and the callback may reset it to U_ZERO_ERROR.
michael@0:  * @see ucnv_setFromUCallBack
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: typedef void (U_EXPORT2 *UConverterFromUCallback) (
michael@0:                     const void* context,
michael@0:                     UConverterFromUnicodeArgs *args,
michael@0:                     const UChar* codeUnits,
michael@0:                     int32_t length,
michael@0:                     UChar32 codePoint,
michael@0:                     UConverterCallbackReason reason,
michael@0:                     UErrorCode *pErrorCode);
michael@0: 
michael@0: U_CDECL_END
michael@0: 
michael@0: /**
michael@0:  * Character that separates converter names from options and options from each other.
michael@0:  * @see ucnv_open
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: #define UCNV_OPTION_SEP_CHAR ','
michael@0: 
michael@0: /**
michael@0:  * String version of UCNV_OPTION_SEP_CHAR. 
michael@0:  * @see ucnv_open
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: #define UCNV_OPTION_SEP_STRING ","
michael@0: 
michael@0: /**
michael@0:  * Character that separates a converter option from its value.
michael@0:  * @see ucnv_open
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: #define UCNV_VALUE_SEP_CHAR '='
michael@0: 
michael@0: /**
michael@0:  * String version of UCNV_VALUE_SEP_CHAR. 
michael@0:  * @see ucnv_open
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: #define UCNV_VALUE_SEP_STRING "="
michael@0: 
michael@0: /**
michael@0:  * Converter option for specifying a locale.
michael@0:  * For example, ucnv_open("SCSU,locale=ja", &errorCode);
michael@0:  * See convrtrs.txt.
michael@0:  *
michael@0:  * @see ucnv_open
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: #define UCNV_LOCALE_OPTION_STRING ",locale="
michael@0: 
michael@0: /**
michael@0:  * Converter option for specifying a version selector (0..9) for some converters.
michael@0:  * For example, 
michael@0:  * \code
michael@0:  *   ucnv_open("UTF-7,version=1", &errorCode);
michael@0:  * \endcode
michael@0:  * See convrtrs.txt.
michael@0:  *
michael@0:  * @see ucnv_open
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define UCNV_VERSION_OPTION_STRING ",version="
michael@0: 
michael@0: /**
michael@0:  * Converter option for EBCDIC SBCS or mixed-SBCS/DBCS (stateful) codepages.
michael@0:  * Swaps Unicode mappings for EBCDIC LF and NL codes, as used on
michael@0:  * S/390 (z/OS) Unix System Services (Open Edition).
michael@0:  * For example, ucnv_open("ibm-1047,swaplfnl", &errorCode);
michael@0:  * See convrtrs.txt.
michael@0:  *
michael@0:  * @see ucnv_open
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: #define UCNV_SWAP_LFNL_OPTION_STRING ",swaplfnl"
michael@0: 
michael@0: /**
michael@0:  * Do a fuzzy compare of two converter/alias names.
michael@0:  * The comparison is case-insensitive, ignores leading zeroes if they are not
michael@0:  * followed by further digits, and ignores all but letters and digits.
michael@0:  * Thus the strings "UTF-8", "utf_8", "u*T@f08" and "Utf 8" are exactly equivalent.
michael@0:  * See section 1.4, Charset Alias Matching in Unicode Technical Standard #22
michael@0:  * at http://www.unicode.org/reports/tr22/
michael@0:  *
michael@0:  * @param name1 a converter name or alias, zero-terminated
michael@0:  * @param name2 a converter name or alias, zero-terminated
michael@0:  * @return 0 if the names match, or a negative value if the name1
michael@0:  * lexically precedes name2, or a positive value if the name1
michael@0:  * lexically follows name2.
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE int U_EXPORT2
michael@0: ucnv_compareNames(const char *name1, const char *name2);
michael@0: 
michael@0: 
michael@0: /**
michael@0:  * Creates a UConverter object with the name of a coded character set specified as a C string.
michael@0:  * The actual name will be resolved with the alias file
michael@0:  * using a case-insensitive string comparison that ignores
michael@0:  * leading zeroes and all non-alphanumeric characters.
michael@0:  * E.g., the names "UTF8", "utf-8", "u*T@f08" and "Utf 8" are all equivalent.
michael@0:  * (See also ucnv_compareNames().)
michael@0:  * If <code>NULL</code> is passed for the converter name, it will create one with the
michael@0:  * getDefaultName return value.
michael@0:  *
michael@0:  * <p>A converter name for ICU 1.5 and above may contain options
michael@0:  * like a locale specification to control the specific behavior of
michael@0:  * the newly instantiated converter.
michael@0:  * The meaning of the options depends on the particular converter.
michael@0:  * If an option is not defined for or recognized by a given converter, then it is ignored.</p>
michael@0:  *
michael@0:  * <p>Options are appended to the converter name string, with a
michael@0:  * <code>UCNV_OPTION_SEP_CHAR</code> between the name and the first option and
michael@0:  * also between adjacent options.</p>
michael@0:  *
michael@0:  * <p>If the alias is ambiguous, then the preferred converter is used
michael@0:  * and the status is set to U_AMBIGUOUS_ALIAS_WARNING.</p>
michael@0:  *
michael@0:  * <p>The conversion behavior and names can vary between platforms. ICU may
michael@0:  * convert some characters differently from other platforms. Details on this topic
michael@0:  * are in the <a href="http://icu-project.org/userguide/conversion.html">User's
michael@0:  * Guide</a>. Aliases starting with a "cp" prefix have no specific meaning
michael@0:  * other than its an alias starting with the letters "cp". Please do not
michael@0:  * associate any meaning to these aliases.</p>
michael@0:  *
michael@0:  * \snippet samples/ucnv/convsamp.cpp ucnv_open
michael@0:  *
michael@0:  * @param converterName Name of the coded character set table.
michael@0:  *          This may have options appended to the string.
michael@0:  *          IANA alias character set names, IBM CCSIDs starting with "ibm-",
michael@0:  *          Windows codepage numbers starting with "windows-" are frequently
michael@0:  *          used for this parameter. See ucnv_getAvailableName and
michael@0:  *          ucnv_getAlias for a complete list that is available.
michael@0:  *          If this parameter is NULL, the default converter will be used.
michael@0:  * @param err outgoing error status <TT>U_MEMORY_ALLOCATION_ERROR, U_FILE_ACCESS_ERROR</TT>
michael@0:  * @return the created Unicode converter object, or <TT>NULL</TT> if an error occured
michael@0:  * @see ucnv_openU
michael@0:  * @see ucnv_openCCSID
michael@0:  * @see ucnv_getAvailableName
michael@0:  * @see ucnv_getAlias
michael@0:  * @see ucnv_getDefaultName
michael@0:  * @see ucnv_close
michael@0:  * @see ucnv_compareNames
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE UConverter* U_EXPORT2 
michael@0: ucnv_open(const char *converterName, UErrorCode *err);
michael@0: 
michael@0: 
michael@0: /**
michael@0:  * Creates a Unicode converter with the names specified as unicode string. 
michael@0:  * The name should be limited to the ASCII-7 alphanumerics range.
michael@0:  * The actual name will be resolved with the alias file
michael@0:  * using a case-insensitive string comparison that ignores
michael@0:  * leading zeroes and all non-alphanumeric characters.
michael@0:  * E.g., the names "UTF8", "utf-8", "u*T@f08" and "Utf 8" are all equivalent.
michael@0:  * (See also ucnv_compareNames().)
michael@0:  * If <TT>NULL</TT> is passed for the converter name, it will create 
michael@0:  * one with the ucnv_getDefaultName() return value.
michael@0:  * If the alias is ambiguous, then the preferred converter is used
michael@0:  * and the status is set to U_AMBIGUOUS_ALIAS_WARNING.
michael@0:  *
michael@0:  * <p>See ucnv_open for the complete details</p>
michael@0:  * @param name Name of the UConverter table in a zero terminated 
michael@0:  *        Unicode string
michael@0:  * @param err outgoing error status <TT>U_MEMORY_ALLOCATION_ERROR, 
michael@0:  *        U_FILE_ACCESS_ERROR</TT>
michael@0:  * @return the created Unicode converter object, or <TT>NULL</TT> if an 
michael@0:  *        error occured
michael@0:  * @see ucnv_open
michael@0:  * @see ucnv_openCCSID
michael@0:  * @see ucnv_close
michael@0:  * @see ucnv_compareNames
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE UConverter* U_EXPORT2 
michael@0: ucnv_openU(const UChar *name,
michael@0:            UErrorCode *err);
michael@0: 
michael@0: /**
michael@0:  * Creates a UConverter object from a CCSID number and platform pair.
michael@0:  * Note that the usefulness of this function is limited to platforms with numeric
michael@0:  * encoding IDs. Only IBM and Microsoft platforms use numeric (16-bit) identifiers for
michael@0:  * encodings.
michael@0:  *
michael@0:  * In addition, IBM CCSIDs and Unicode conversion tables are not 1:1 related.
michael@0:  * For many IBM CCSIDs there are multiple (up to six) Unicode conversion tables, and
michael@0:  * for some Unicode conversion tables there are multiple CCSIDs.
michael@0:  * Some "alternate" Unicode conversion tables are provided by the
michael@0:  * IBM CDRA conversion table registry.
michael@0:  * The most prominent example of a systematic modification of conversion tables that is
michael@0:  * not provided in the form of conversion table files in the repository is
michael@0:  * that S/390 Unix System Services swaps the codes for Line Feed and New Line in all
michael@0:  * EBCDIC codepages, which requires such a swap in the Unicode conversion tables as well.
michael@0:  *
michael@0:  * Only IBM default conversion tables are accessible with ucnv_openCCSID().
michael@0:  * ucnv_getCCSID() will return the same CCSID for all conversion tables that are associated
michael@0:  * with that CCSID.
michael@0:  *
michael@0:  * Currently, the only "platform" supported in the ICU converter API is UCNV_IBM.
michael@0:  *
michael@0:  * In summary, the use of CCSIDs and the associated API functions is not recommended.
michael@0:  *
michael@0:  * In order to open a converter with the default IBM CDRA Unicode conversion table,
michael@0:  * you can use this function or use the prefix "ibm-":
michael@0:  * \code
michael@0:  *     char name[20];
michael@0:  *     sprintf(name, "ibm-%hu", ccsid);
michael@0:  *     cnv=ucnv_open(name, &errorCode);
michael@0:  * \endcode
michael@0:  *
michael@0:  * In order to open a converter with the IBM S/390 Unix System Services variant
michael@0:  * of a Unicode/EBCDIC conversion table,
michael@0:  * you can use the prefix "ibm-" together with the option string UCNV_SWAP_LFNL_OPTION_STRING:
michael@0:  * \code
michael@0:  *     char name[20];
michael@0:  *     sprintf(name, "ibm-%hu" UCNV_SWAP_LFNL_OPTION_STRING, ccsid);
michael@0:  *     cnv=ucnv_open(name, &errorCode);
michael@0:  * \endcode
michael@0:  *
michael@0:  * In order to open a converter from a Microsoft codepage number, use the prefix "cp":
michael@0:  * \code
michael@0:  *     char name[20];
michael@0:  *     sprintf(name, "cp%hu", codepageID);
michael@0:  *     cnv=ucnv_open(name, &errorCode);
michael@0:  * \endcode
michael@0:  *
michael@0:  * If the alias is ambiguous, then the preferred converter is used
michael@0:  * and the status is set to U_AMBIGUOUS_ALIAS_WARNING.
michael@0:  *
michael@0:  * @param codepage codepage number to create
michael@0:  * @param platform the platform in which the codepage number exists
michael@0:  * @param err error status <TT>U_MEMORY_ALLOCATION_ERROR, U_FILE_ACCESS_ERROR</TT>
michael@0:  * @return the created Unicode converter object, or <TT>NULL</TT> if an error
michael@0:  *   occured.
michael@0:  * @see ucnv_open
michael@0:  * @see ucnv_openU
michael@0:  * @see ucnv_close
michael@0:  * @see ucnv_getCCSID
michael@0:  * @see ucnv_getPlatform
michael@0:  * @see UConverterPlatform
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE UConverter* U_EXPORT2
michael@0: ucnv_openCCSID(int32_t codepage,
michael@0:                UConverterPlatform platform,
michael@0:                UErrorCode * err);
michael@0: 
michael@0: /**
michael@0:  * <p>Creates a UConverter object specified from a packageName and a converterName.</p>
michael@0:  * 
michael@0:  * <p>The packageName and converterName must point to an ICU udata object, as defined by
michael@0:  *   <code> udata_open( packageName, "cnv", converterName, err) </code> or equivalent.
michael@0:  * Typically, packageName will refer to a (.dat) file, or to a package registered with
michael@0:  * udata_setAppData(). Using a full file or directory pathname for packageName is deprecated.</p>
michael@0:  * 
michael@0:  * <p>The name will NOT be looked up in the alias mechanism, nor will the converter be
michael@0:  * stored in the converter cache or the alias table. The only way to open further converters
michael@0:  * is call this function multiple times, or use the ucnv_safeClone() function to clone a 
michael@0:  * 'master' converter.</p>
michael@0:  *
michael@0:  * <p>A future version of ICU may add alias table lookups and/or caching
michael@0:  * to this function.</p>
michael@0:  * 
michael@0:  * <p>Example Use:
michael@0:  *      <code>cnv = ucnv_openPackage("myapp", "myconverter", &err);</code>
michael@0:  * </p>
michael@0:  *
michael@0:  * @param packageName name of the package (equivalent to 'path' in udata_open() call)
michael@0:  * @param converterName name of the data item to be used, without suffix.
michael@0:  * @param err outgoing error status <TT>U_MEMORY_ALLOCATION_ERROR, U_FILE_ACCESS_ERROR</TT>
michael@0:  * @return the created Unicode converter object, or <TT>NULL</TT> if an error occured
michael@0:  * @see udata_open
michael@0:  * @see ucnv_open
michael@0:  * @see ucnv_safeClone
michael@0:  * @see ucnv_close
michael@0:  * @stable ICU 2.2
michael@0:  */
michael@0: U_STABLE UConverter* U_EXPORT2 
michael@0: ucnv_openPackage(const char *packageName, const char *converterName, UErrorCode *err);
michael@0: 
michael@0: /**
michael@0:  * Thread safe converter cloning operation.
michael@0:  * For most efficient operation, pass in a stackBuffer (and a *pBufferSize)
michael@0:  * with at least U_CNV_SAFECLONE_BUFFERSIZE bytes of space.
michael@0:  * If the buffer size is sufficient, then the clone will use the stack buffer;
michael@0:  * otherwise, it will be allocated, and *pBufferSize will indicate
michael@0:  * the actual size. (This should not occur with U_CNV_SAFECLONE_BUFFERSIZE.)
michael@0:  *
michael@0:  * You must ucnv_close() the clone in any case.
michael@0:  *
michael@0:  * If *pBufferSize==0, (regardless of whether stackBuffer==NULL or not)
michael@0:  * then *pBufferSize will be changed to a sufficient size
michael@0:  * for cloning this converter,
michael@0:  * without actually cloning the converter ("pure pre-flighting").
michael@0:  *
michael@0:  * If *pBufferSize is greater than zero but not large enough for a stack-based
michael@0:  * clone, then the converter is cloned using newly allocated memory
michael@0:  * and *pBufferSize is changed to the necessary size.
michael@0:  *
michael@0:  * If the converter clone fits into the stack buffer but the stack buffer is not
michael@0:  * sufficiently aligned for the clone, then the clone will use an
michael@0:  * adjusted pointer and use an accordingly smaller buffer size.
michael@0:  *
michael@0:  * @param cnv converter to be cloned
michael@0:  * @param stackBuffer <em>Deprecated functionality as of ICU 52, use NULL.</em><br>
michael@0:  *  user allocated space for the new clone. If NULL new memory will be allocated. 
michael@0:  *  If buffer is not large enough, new memory will be allocated.
michael@0:  *  Clients can use the U_CNV_SAFECLONE_BUFFERSIZE. This will probably be enough to avoid memory allocations.
michael@0:  * @param pBufferSize <em>Deprecated functionality as of ICU 52, use NULL or 1.</em><br>
michael@0:  *  pointer to size of allocated space.
michael@0:  * @param status to indicate whether the operation went on smoothly or there were errors
michael@0:  *  An informational status value, U_SAFECLONE_ALLOCATED_WARNING,
michael@0:  *  is used if any allocations were necessary.
michael@0:  *  However, it is better to check if *pBufferSize grew for checking for
michael@0:  *  allocations because warning codes can be overridden by subsequent
michael@0:  *  function calls.
michael@0:  * @return pointer to the new clone
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE UConverter * U_EXPORT2 
michael@0: ucnv_safeClone(const UConverter *cnv, 
michael@0:                void             *stackBuffer,
michael@0:                int32_t          *pBufferSize, 
michael@0:                UErrorCode       *status);
michael@0: 
michael@0: #ifndef U_HIDE_DEPRECATED_API
michael@0: 
michael@0: /**
michael@0:  * \def U_CNV_SAFECLONE_BUFFERSIZE
michael@0:  * Definition of a buffer size that is designed to be large enough for
michael@0:  * converters to be cloned with ucnv_safeClone().
michael@0:  * @deprecated ICU 52. Do not rely on ucnv_safeClone() cloning into any provided buffer.
michael@0:  */
michael@0: #define U_CNV_SAFECLONE_BUFFERSIZE  1024
michael@0: 
michael@0: #endif /* U_HIDE_DEPRECATED_API */
michael@0: 
michael@0: /**
michael@0:  * Deletes the unicode converter and releases resources associated
michael@0:  * with just this instance.
michael@0:  * Does not free up shared converter tables.
michael@0:  *
michael@0:  * @param converter the converter object to be deleted
michael@0:  * @see ucnv_open
michael@0:  * @see ucnv_openU
michael@0:  * @see ucnv_openCCSID
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void  U_EXPORT2
michael@0: ucnv_close(UConverter * converter);
michael@0: 
michael@0: #if U_SHOW_CPLUSPLUS_API
michael@0: 
michael@0: U_NAMESPACE_BEGIN
michael@0: 
michael@0: /**
michael@0:  * \class LocalUConverterPointer
michael@0:  * "Smart pointer" class, closes a UConverter via ucnv_close().
michael@0:  * For most methods see the LocalPointerBase base class.
michael@0:  *
michael@0:  * @see LocalPointerBase
michael@0:  * @see LocalPointer
michael@0:  * @stable ICU 4.4
michael@0:  */
michael@0: U_DEFINE_LOCAL_OPEN_POINTER(LocalUConverterPointer, UConverter, ucnv_close);
michael@0: 
michael@0: U_NAMESPACE_END
michael@0: 
michael@0: #endif
michael@0: 
michael@0: /**
michael@0:  * Fills in the output parameter, subChars, with the substitution characters
michael@0:  * as multiple bytes.
michael@0:  * If ucnv_setSubstString() set a Unicode string because the converter is
michael@0:  * stateful, then subChars will be an empty string.
michael@0:  *
michael@0:  * @param converter the Unicode converter
michael@0:  * @param subChars the subsitution characters
michael@0:  * @param len on input the capacity of subChars, on output the number 
michael@0:  * of bytes copied to it
michael@0:  * @param  err the outgoing error status code.
michael@0:  * If the substitution character array is too small, an
michael@0:  * <TT>U_INDEX_OUTOFBOUNDS_ERROR</TT> will be returned.
michael@0:  * @see ucnv_setSubstString
michael@0:  * @see ucnv_setSubstChars
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_getSubstChars(const UConverter *converter,
michael@0:                    char *subChars,
michael@0:                    int8_t *len,
michael@0:                    UErrorCode *err);
michael@0: 
michael@0: /**
michael@0:  * Sets the substitution chars when converting from unicode to a codepage. The
michael@0:  * substitution is specified as a string of 1-4 bytes, and may contain
michael@0:  * <TT>NULL</TT> bytes.
michael@0:  * The subChars must represent a single character. The caller needs to know the
michael@0:  * byte sequence of a valid character in the converter's charset.
michael@0:  * For some converters, for example some ISO 2022 variants, only single-byte
michael@0:  * substitution characters may be supported.
michael@0:  * The newer ucnv_setSubstString() function relaxes these limitations.
michael@0:  *
michael@0:  * @param converter the Unicode converter
michael@0:  * @param subChars the substitution character byte sequence we want set
michael@0:  * @param len the number of bytes in subChars
michael@0:  * @param err the error status code.  <TT>U_INDEX_OUTOFBOUNDS_ERROR </TT> if
michael@0:  * len is bigger than the maximum number of bytes allowed in subchars
michael@0:  * @see ucnv_setSubstString
michael@0:  * @see ucnv_getSubstChars
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_setSubstChars(UConverter *converter,
michael@0:                    const char *subChars,
michael@0:                    int8_t len,
michael@0:                    UErrorCode *err);
michael@0: 
michael@0: /**
michael@0:  * Set a substitution string for converting from Unicode to a charset.
michael@0:  * The caller need not know the charset byte sequence for each charset.
michael@0:  *
michael@0:  * Unlike ucnv_setSubstChars() which is designed to set a charset byte sequence
michael@0:  * for a single character, this function takes a Unicode string with
michael@0:  * zero, one or more characters, and immediately verifies that the string can be
michael@0:  * converted to the charset.
michael@0:  * If not, or if the result is too long (more than 32 bytes as of ICU 3.6),
michael@0:  * then the function returns with an error accordingly.
michael@0:  *
michael@0:  * Also unlike ucnv_setSubstChars(), this function works for stateful charsets
michael@0:  * by converting on the fly at the point of substitution rather than setting
michael@0:  * a fixed byte sequence.
michael@0:  *
michael@0:  * @param cnv The UConverter object.
michael@0:  * @param s The Unicode string.
michael@0:  * @param length The number of UChars in s, or -1 for a NUL-terminated string.
michael@0:  * @param err Pointer to a standard ICU error code. Its input value must
michael@0:  *            pass the U_SUCCESS() test, or else the function returns
michael@0:  *            immediately. Check for U_FAILURE() on output or use with
michael@0:  *            function chaining. (See User Guide for details.)
michael@0:  *
michael@0:  * @see ucnv_setSubstChars
michael@0:  * @see ucnv_getSubstChars
michael@0:  * @stable ICU 3.6
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_setSubstString(UConverter *cnv,
michael@0:                     const UChar *s,
michael@0:                     int32_t length,
michael@0:                     UErrorCode *err);
michael@0: 
michael@0: /**
michael@0:  * Fills in the output parameter, errBytes, with the error characters from the
michael@0:  * last failing conversion.
michael@0:  *
michael@0:  * @param converter the Unicode converter
michael@0:  * @param errBytes the codepage bytes which were in error
michael@0:  * @param len on input the capacity of errBytes, on output the number of
michael@0:  *  bytes which were copied to it
michael@0:  * @param err the error status code.
michael@0:  * If the substitution character array is too small, an
michael@0:  * <TT>U_INDEX_OUTOFBOUNDS_ERROR</TT> will be returned.
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_getInvalidChars(const UConverter *converter,
michael@0:                      char *errBytes,
michael@0:                      int8_t *len,
michael@0:                      UErrorCode *err);
michael@0: 
michael@0: /**
michael@0:  * Fills in the output parameter, errChars, with the error characters from the
michael@0:  * last failing conversion.
michael@0:  *
michael@0:  * @param converter the Unicode converter
michael@0:  * @param errUChars the UChars which were in error
michael@0:  * @param len on input the capacity of errUChars, on output the number of 
michael@0:  *  UChars which were copied to it
michael@0:  * @param err the error status code.
michael@0:  * If the substitution character array is too small, an
michael@0:  * <TT>U_INDEX_OUTOFBOUNDS_ERROR</TT> will be returned.
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_getInvalidUChars(const UConverter *converter,
michael@0:                       UChar *errUChars,
michael@0:                       int8_t *len,
michael@0:                       UErrorCode *err);
michael@0: 
michael@0: /**
michael@0:  * Resets the state of a converter to the default state. This is used
michael@0:  * in the case of an error, to restart a conversion from a known default state.
michael@0:  * It will also empty the internal output buffers.
michael@0:  * @param converter the Unicode converter
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_reset(UConverter *converter);
michael@0: 
michael@0: /**
michael@0:  * Resets the to-Unicode part of a converter state to the default state.
michael@0:  * This is used in the case of an error to restart a conversion to
michael@0:  * Unicode to a known default state. It will also empty the internal
michael@0:  * output buffers used for the conversion to Unicode codepoints.
michael@0:  * @param converter the Unicode converter
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2 
michael@0: ucnv_resetToUnicode(UConverter *converter);
michael@0: 
michael@0: /**
michael@0:  * Resets the from-Unicode part of a converter state to the default state.
michael@0:  * This is used in the case of an error to restart a conversion from
michael@0:  * Unicode to a known default state. It will also empty the internal output
michael@0:  * buffers used for the conversion from Unicode codepoints.
michael@0:  * @param converter the Unicode converter
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2 
michael@0: ucnv_resetFromUnicode(UConverter *converter);
michael@0: 
michael@0: /**
michael@0:  * Returns the maximum number of bytes that are output per UChar in conversion
michael@0:  * from Unicode using this converter.
michael@0:  * The returned number can be used with UCNV_GET_MAX_BYTES_FOR_STRING
michael@0:  * to calculate the size of a target buffer for conversion from Unicode.
michael@0:  *
michael@0:  * Note: Before ICU 2.8, this function did not return reliable numbers for
michael@0:  * some stateful converters (EBCDIC_STATEFUL, ISO-2022) and LMBCS.
michael@0:  *
michael@0:  * This number may not be the same as the maximum number of bytes per
michael@0:  * "conversion unit". In other words, it may not be the intuitively expected
michael@0:  * number of bytes per character that would be published for a charset,
michael@0:  * and may not fulfill any other purpose than the allocation of an output
michael@0:  * buffer of guaranteed sufficient size for a given input length and converter.
michael@0:  *
michael@0:  * Examples for special cases that are taken into account:
michael@0:  * - Supplementary code points may convert to more bytes than BMP code points.
michael@0:  *   This function returns bytes per UChar (UTF-16 code unit), not per
michael@0:  *   Unicode code point, for efficient buffer allocation.
michael@0:  * - State-shifting output (SI/SO, escapes, etc.) from stateful converters.
michael@0:  * - When m input UChars are converted to n output bytes, then the maximum m/n
michael@0:  *   is taken into account.
michael@0:  *
michael@0:  * The number returned here does not take into account
michael@0:  * (see UCNV_GET_MAX_BYTES_FOR_STRING):
michael@0:  * - callbacks which output more than one charset character sequence per call,
michael@0:  *   like escape callbacks
michael@0:  * - initial and final non-character bytes that are output by some converters
michael@0:  *   (automatic BOMs, initial escape sequence, final SI, etc.)
michael@0:  *
michael@0:  * Examples for returned values:
michael@0:  * - SBCS charsets: 1
michael@0:  * - Shift-JIS: 2
michael@0:  * - UTF-16: 2 (2 per BMP, 4 per surrogate _pair_, BOM not counted)
michael@0:  * - UTF-8: 3 (3 per BMP, 4 per surrogate _pair_)
michael@0:  * - EBCDIC_STATEFUL (EBCDIC mixed SBCS/DBCS): 3 (SO + DBCS)
michael@0:  * - ISO-2022: 3 (always outputs UTF-8)
michael@0:  * - ISO-2022-JP: 6 (4-byte escape sequences + DBCS)
michael@0:  * - ISO-2022-CN: 8 (4-byte designator sequences + 2-byte SS2/SS3 + DBCS)
michael@0:  *
michael@0:  * @param converter The Unicode converter.
michael@0:  * @return The maximum number of bytes per UChar that are output by ucnv_fromUnicode(),
michael@0:  *         to be used together with UCNV_GET_MAX_BYTES_FOR_STRING for buffer allocation.
michael@0:  *
michael@0:  * @see UCNV_GET_MAX_BYTES_FOR_STRING
michael@0:  * @see ucnv_getMinCharSize
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE int8_t U_EXPORT2
michael@0: ucnv_getMaxCharSize(const UConverter *converter);
michael@0: 
michael@0: /**
michael@0:  * Calculates the size of a buffer for conversion from Unicode to a charset.
michael@0:  * The calculated size is guaranteed to be sufficient for this conversion.
michael@0:  *
michael@0:  * It takes into account initial and final non-character bytes that are output
michael@0:  * by some converters.
michael@0:  * It does not take into account callbacks which output more than one charset
michael@0:  * character sequence per call, like escape callbacks.
michael@0:  * The default (substitution) callback only outputs one charset character sequence.
michael@0:  *
michael@0:  * @param length Number of UChars to be converted.
michael@0:  * @param maxCharSize Return value from ucnv_getMaxCharSize() for the converter
michael@0:  *                    that will be used.
michael@0:  * @return Size of a buffer that will be large enough to hold the output bytes of
michael@0:  *         converting length UChars with the converter that returned the maxCharSize.
michael@0:  *
michael@0:  * @see ucnv_getMaxCharSize
michael@0:  * @stable ICU 2.8
michael@0:  */
michael@0: #define UCNV_GET_MAX_BYTES_FOR_STRING(length, maxCharSize) \
michael@0:      (((int32_t)(length)+10)*(int32_t)(maxCharSize))
michael@0: 
michael@0: /**
michael@0:  * Returns the minimum byte length for characters in this codepage. 
michael@0:  * This is usually either 1 or 2.
michael@0:  * @param converter the Unicode converter
michael@0:  * @return the minimum number of bytes allowed by this particular converter
michael@0:  * @see ucnv_getMaxCharSize
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE int8_t U_EXPORT2
michael@0: ucnv_getMinCharSize(const UConverter *converter);
michael@0: 
michael@0: /**
michael@0:  * Returns the display name of the converter passed in based on the Locale 
michael@0:  * passed in. If the locale contains no display name, the internal ASCII
michael@0:  * name will be filled in.
michael@0:  *
michael@0:  * @param converter the Unicode converter.
michael@0:  * @param displayLocale is the specific Locale we want to localised for
michael@0:  * @param displayName user provided buffer to be filled in
michael@0:  * @param displayNameCapacity size of displayName Buffer
michael@0:  * @param err error status code
michael@0:  * @return displayNameLength number of UChar needed in displayName
michael@0:  * @see ucnv_getName
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE int32_t U_EXPORT2
michael@0: ucnv_getDisplayName(const UConverter *converter,
michael@0:                     const char *displayLocale,
michael@0:                     UChar *displayName,
michael@0:                     int32_t displayNameCapacity,
michael@0:                     UErrorCode *err);
michael@0: 
michael@0: /**
michael@0:  * Gets the internal, canonical name of the converter (zero-terminated).
michael@0:  * The lifetime of the returned string will be that of the converter 
michael@0:  * passed to this function.
michael@0:  * @param converter the Unicode converter
michael@0:  * @param err UErrorCode status
michael@0:  * @return the internal name of the converter
michael@0:  * @see ucnv_getDisplayName
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE const char * U_EXPORT2 
michael@0: ucnv_getName(const UConverter *converter, UErrorCode *err);
michael@0: 
michael@0: /**
michael@0:  * Gets a codepage number associated with the converter. This is not guaranteed
michael@0:  * to be the one used to create the converter. Some converters do not represent
michael@0:  * platform registered codepages and return zero for the codepage number.
michael@0:  * The error code fill-in parameter indicates if the codepage number
michael@0:  * is available.
michael@0:  * Does not check if the converter is <TT>NULL</TT> or if converter's data
michael@0:  * table is <TT>NULL</TT>.
michael@0:  *
michael@0:  * Important: The use of CCSIDs is not recommended because it is limited
michael@0:  * to only two platforms in principle and only one (UCNV_IBM) in the current
michael@0:  * ICU converter API.
michael@0:  * Also, CCSIDs are insufficient to identify IBM Unicode conversion tables precisely.
michael@0:  * For more details see ucnv_openCCSID().
michael@0:  *
michael@0:  * @param converter the Unicode converter
michael@0:  * @param err the error status code.
michael@0:  * @return If any error occurrs, -1 will be returned otherwise, the codepage number
michael@0:  * will be returned
michael@0:  * @see ucnv_openCCSID
michael@0:  * @see ucnv_getPlatform
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE int32_t U_EXPORT2
michael@0: ucnv_getCCSID(const UConverter *converter,
michael@0:               UErrorCode *err);
michael@0: 
michael@0: /**
michael@0:  * Gets a codepage platform associated with the converter. Currently, 
michael@0:  * only <TT>UCNV_IBM</TT> will be returned.
michael@0:  * Does not test if the converter is <TT>NULL</TT> or if converter's data 
michael@0:  * table is <TT>NULL</TT>. 
michael@0:  * @param converter the Unicode converter
michael@0:  * @param err the error status code.
michael@0:  * @return The codepage platform
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE UConverterPlatform U_EXPORT2
michael@0: ucnv_getPlatform(const UConverter *converter,
michael@0:                  UErrorCode *err);
michael@0: 
michael@0: /**
michael@0:  * Gets the type of the converter
michael@0:  * e.g. SBCS, MBCS, DBCS, UTF8, UTF16_BE, UTF16_LE, ISO_2022, 
michael@0:  * EBCDIC_STATEFUL, LATIN_1
michael@0:  * @param converter a valid, opened converter
michael@0:  * @return the type of the converter
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE UConverterType U_EXPORT2
michael@0: ucnv_getType(const UConverter * converter);
michael@0: 
michael@0: /**
michael@0:  * Gets the "starter" (lead) bytes for converters of type MBCS.
michael@0:  * Will fill in an <TT>U_ILLEGAL_ARGUMENT_ERROR</TT> if converter passed in
michael@0:  * is not MBCS. Fills in an array of type UBool, with the value of the byte 
michael@0:  * as offset to the array. For example, if (starters[0x20] == TRUE) at return,
michael@0:  * it means that the byte 0x20 is a starter byte in this converter.
michael@0:  * Context pointers are always owned by the caller.
michael@0:  * 
michael@0:  * @param converter a valid, opened converter of type MBCS
michael@0:  * @param starters an array of size 256 to be filled in
michael@0:  * @param err error status, <TT>U_ILLEGAL_ARGUMENT_ERROR</TT> if the 
michael@0:  * converter is not a type which can return starters.
michael@0:  * @see ucnv_getType
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_getStarters(const UConverter* converter, 
michael@0:                  UBool starters[256],
michael@0:                  UErrorCode* err);
michael@0: 
michael@0: 
michael@0: /**
michael@0:  * Selectors for Unicode sets that can be returned by ucnv_getUnicodeSet().
michael@0:  * @see ucnv_getUnicodeSet
michael@0:  * @stable ICU 2.6
michael@0:  */
michael@0: typedef enum UConverterUnicodeSet {
michael@0:     /** Select the set of roundtrippable Unicode code points. @stable ICU 2.6 */
michael@0:     UCNV_ROUNDTRIP_SET,
michael@0:     /** Select the set of Unicode code points with roundtrip or fallback mappings. @stable ICU 4.0 */
michael@0:     UCNV_ROUNDTRIP_AND_FALLBACK_SET,
michael@0:     /** Number of UConverterUnicodeSet selectors. @stable ICU 2.6 */
michael@0:     UCNV_SET_COUNT
michael@0: } UConverterUnicodeSet;
michael@0: 
michael@0: 
michael@0: /**
michael@0:  * Returns the set of Unicode code points that can be converted by an ICU converter.
michael@0:  *
michael@0:  * Returns one of several kinds of set:
michael@0:  *
michael@0:  * 1. UCNV_ROUNDTRIP_SET
michael@0:  *
michael@0:  * The set of all Unicode code points that can be roundtrip-converted
michael@0:  * (converted without any data loss) with the converter (ucnv_fromUnicode()).
michael@0:  * This set will not include code points that have fallback mappings
michael@0:  * or are only the result of reverse fallback mappings.
michael@0:  * This set will also not include PUA code points with fallbacks, although
michael@0:  * ucnv_fromUnicode() will always uses those mappings despite ucnv_setFallback().
michael@0:  * See UTR #22 "Character Mapping Markup Language"
michael@0:  * at http://www.unicode.org/reports/tr22/
michael@0:  *
michael@0:  * This is useful for example for
michael@0:  * - checking that a string or document can be roundtrip-converted with a converter,
michael@0:  *   without/before actually performing the conversion
michael@0:  * - testing if a converter can be used for text for typical text for a certain locale,
michael@0:  *   by comparing its roundtrip set with the set of ExemplarCharacters from
michael@0:  *   ICU's locale data or other sources
michael@0:  *
michael@0:  * 2. UCNV_ROUNDTRIP_AND_FALLBACK_SET
michael@0:  *
michael@0:  * The set of all Unicode code points that can be converted with the converter (ucnv_fromUnicode())
michael@0:  * when fallbacks are turned on (see ucnv_setFallback()).
michael@0:  * This set includes all code points with roundtrips and fallbacks (but not reverse fallbacks).
michael@0:  *
michael@0:  * In the future, there may be more UConverterUnicodeSet choices to select
michael@0:  * sets with different properties.
michael@0:  *
michael@0:  * @param cnv The converter for which a set is requested.
michael@0:  * @param setFillIn A valid USet *. It will be cleared by this function before
michael@0:  *            the converter's specific set is filled into the USet.
michael@0:  * @param whichSet A UConverterUnicodeSet selector;
michael@0:  *              currently UCNV_ROUNDTRIP_SET is the only supported value.
michael@0:  * @param pErrorCode ICU error code in/out parameter.
michael@0:  *                   Must fulfill U_SUCCESS before the function call.
michael@0:  *
michael@0:  * @see UConverterUnicodeSet
michael@0:  * @see uset_open
michael@0:  * @see uset_close
michael@0:  * @stable ICU 2.6
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_getUnicodeSet(const UConverter *cnv,
michael@0:                    USet *setFillIn,
michael@0:                    UConverterUnicodeSet whichSet,
michael@0:                    UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Gets the current calback function used by the converter when an illegal
michael@0:  *  or invalid codepage sequence is found. 
michael@0:  * Context pointers are always owned by the caller.
michael@0:  *
michael@0:  * @param converter the unicode converter
michael@0:  * @param action fillin: returns the callback function pointer
michael@0:  * @param context fillin: returns the callback's private void* context
michael@0:  * @see ucnv_setToUCallBack
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_getToUCallBack (const UConverter * converter,
michael@0:                      UConverterToUCallback *action,
michael@0:                      const void **context);
michael@0: 
michael@0: /**
michael@0:  * Gets the current callback function used by the converter when illegal 
michael@0:  * or invalid Unicode sequence is found.
michael@0:  * Context pointers are always owned by the caller.
michael@0:  *
michael@0:  * @param converter the unicode converter
michael@0:  * @param action fillin: returns the callback function pointer
michael@0:  * @param context fillin: returns the callback's private void* context
michael@0:  * @see ucnv_setFromUCallBack
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_getFromUCallBack (const UConverter * converter,
michael@0:                        UConverterFromUCallback *action,
michael@0:                        const void **context);
michael@0: 
michael@0: /**
michael@0:  * Changes the callback function used by the converter when
michael@0:  * an illegal or invalid sequence is found.
michael@0:  * Context pointers are always owned by the caller.
michael@0:  * Predefined actions and contexts can be found in the ucnv_err.h header.
michael@0:  *
michael@0:  * @param converter the unicode converter
michael@0:  * @param newAction the new callback function
michael@0:  * @param newContext the new toUnicode callback context pointer. This can be NULL.
michael@0:  * @param oldAction fillin: returns the old callback function pointer. This can be NULL.
michael@0:  * @param oldContext fillin: returns the old callback's private void* context. This can be NULL.
michael@0:  * @param err The error code status
michael@0:  * @see ucnv_getToUCallBack
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_setToUCallBack (UConverter * converter,
michael@0:                      UConverterToUCallback newAction,
michael@0:                      const void* newContext,
michael@0:                      UConverterToUCallback *oldAction,
michael@0:                      const void** oldContext,
michael@0:                      UErrorCode * err);
michael@0: 
michael@0: /**
michael@0:  * Changes the current callback function used by the converter when
michael@0:  * an illegal or invalid sequence is found.
michael@0:  * Context pointers are always owned by the caller.
michael@0:  * Predefined actions and contexts can be found in the ucnv_err.h header.
michael@0:  *
michael@0:  * @param converter the unicode converter
michael@0:  * @param newAction the new callback function
michael@0:  * @param newContext the new fromUnicode callback context pointer. This can be NULL.
michael@0:  * @param oldAction fillin: returns the old callback function pointer. This can be NULL.
michael@0:  * @param oldContext fillin: returns the old callback's private void* context. This can be NULL.
michael@0:  * @param err The error code status
michael@0:  * @see ucnv_getFromUCallBack
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_setFromUCallBack (UConverter * converter,
michael@0:                        UConverterFromUCallback newAction,
michael@0:                        const void *newContext,
michael@0:                        UConverterFromUCallback *oldAction,
michael@0:                        const void **oldContext,
michael@0:                        UErrorCode * err);
michael@0: 
michael@0: /**
michael@0:  * Converts an array of unicode characters to an array of codepage
michael@0:  * characters. This function is optimized for converting a continuous
michael@0:  * stream of data in buffer-sized chunks, where the entire source and
michael@0:  * target does not fit in available buffers.
michael@0:  * 
michael@0:  * The source pointer is an in/out parameter. It starts out pointing where the 
michael@0:  * conversion is to begin, and ends up pointing after the last UChar consumed. 
michael@0:  * 
michael@0:  * Target similarly starts out pointer at the first available byte in the output
michael@0:  * buffer, and ends up pointing after the last byte written to the output.
michael@0:  * 
michael@0:  * The converter always attempts to consume the entire source buffer, unless 
michael@0:  * (1.) the target buffer is full, or (2.) a failing error is returned from the
michael@0:  * current callback function.  When a successful error status has been
michael@0:  * returned, it means that all of the source buffer has been
michael@0:  *  consumed. At that point, the caller should reset the source and
michael@0:  *  sourceLimit pointers to point to the next chunk.
michael@0:  * 
michael@0:  * At the end of the stream (flush==TRUE), the input is completely consumed
michael@0:  * when *source==sourceLimit and no error code is set.
michael@0:  * The converter object is then automatically reset by this function.
michael@0:  * (This means that a converter need not be reset explicitly between data
michael@0:  * streams if it finishes the previous stream without errors.)
michael@0:  * 
michael@0:  * This is a <I>stateful</I> conversion. Additionally, even when all source data has
michael@0:  * been consumed, some data may be in the converters' internal state.
michael@0:  * Call this function repeatedly, updating the target pointers with
michael@0:  * the next empty chunk of target in case of a
michael@0:  * <TT>U_BUFFER_OVERFLOW_ERROR</TT>, and updating the source  pointers
michael@0:  *  with the next chunk of source when a successful error status is
michael@0:  * returned, until there are no more chunks of source data.
michael@0:  * @param converter the Unicode converter
michael@0:  * @param target I/O parameter. Input : Points to the beginning of the buffer to copy
michael@0:  *  codepage characters to. Output : points to after the last codepage character copied
michael@0:  *  to <TT>target</TT>.
michael@0:  * @param targetLimit the pointer just after last of the <TT>target</TT> buffer
michael@0:  * @param source I/O parameter, pointer to pointer to the source Unicode character buffer. 
michael@0:  * @param sourceLimit the pointer just after the last of the source buffer
michael@0:  * @param offsets if NULL is passed, nothing will happen to it, otherwise it needs to have the same number
michael@0:  * of allocated cells as <TT>target</TT>. Will fill in offsets from target to source pointer
michael@0:  * e.g: <TT>offsets[3]</TT> is equal to 6, it means that the <TT>target[3]</TT> was a result of transcoding <TT>source[6]</TT>
michael@0:  * For output data carried across calls, and other data without a specific source character
michael@0:  * (such as from escape sequences or callbacks)  -1 will be placed for offsets. 
michael@0:  * @param flush set to <TT>TRUE</TT> if the current source buffer is the last available
michael@0:  * chunk of the source, <TT>FALSE</TT> otherwise. Note that if a failing status is returned,
michael@0:  * this function may have to be called multiple times with flush set to <TT>TRUE</TT> until
michael@0:  * the source buffer is consumed.
michael@0:  * @param err the error status.  <TT>U_ILLEGAL_ARGUMENT_ERROR</TT> will be set if the
michael@0:  * converter is <TT>NULL</TT>.
michael@0:  * <code>U_BUFFER_OVERFLOW_ERROR</code> will be set if the target is full and there is 
michael@0:  * still data to be written to the target.
michael@0:  * @see ucnv_fromUChars
michael@0:  * @see ucnv_convert
michael@0:  * @see ucnv_getMinCharSize
michael@0:  * @see ucnv_setToUCallBack
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2 
michael@0: ucnv_fromUnicode (UConverter * converter,
michael@0:                   char **target,
michael@0:                   const char *targetLimit,
michael@0:                   const UChar ** source,
michael@0:                   const UChar * sourceLimit,
michael@0:                   int32_t* offsets,
michael@0:                   UBool flush,
michael@0:                   UErrorCode * err);
michael@0: 
michael@0: /**
michael@0:  * Converts a buffer of codepage bytes into an array of unicode UChars
michael@0:  * characters. This function is optimized for converting a continuous
michael@0:  * stream of data in buffer-sized chunks, where the entire source and
michael@0:  * target does not fit in available buffers.
michael@0:  * 
michael@0:  * The source pointer is an in/out parameter. It starts out pointing where the 
michael@0:  * conversion is to begin, and ends up pointing after the last byte of source consumed. 
michael@0:  * 
michael@0:  * Target similarly starts out pointer at the first available UChar in the output
michael@0:  * buffer, and ends up pointing after the last UChar written to the output. 
michael@0:  * It does NOT necessarily keep UChar sequences together.
michael@0:  * 
michael@0:  * The converter always attempts to consume the entire source buffer, unless 
michael@0:  * (1.) the target buffer is full, or (2.) a failing error is returned from the
michael@0:  * current callback function.  When a successful error status has been
michael@0:  * returned, it means that all of the source buffer has been
michael@0:  *  consumed. At that point, the caller should reset the source and
michael@0:  *  sourceLimit pointers to point to the next chunk.
michael@0:  *
michael@0:  * At the end of the stream (flush==TRUE), the input is completely consumed
michael@0:  * when *source==sourceLimit and no error code is set
michael@0:  * The converter object is then automatically reset by this function.
michael@0:  * (This means that a converter need not be reset explicitly between data
michael@0:  * streams if it finishes the previous stream without errors.)
michael@0:  * 
michael@0:  * This is a <I>stateful</I> conversion. Additionally, even when all source data has
michael@0:  * been consumed, some data may be in the converters' internal state.
michael@0:  * Call this function repeatedly, updating the target pointers with
michael@0:  * the next empty chunk of target in case of a
michael@0:  * <TT>U_BUFFER_OVERFLOW_ERROR</TT>, and updating the source  pointers
michael@0:  *  with the next chunk of source when a successful error status is
michael@0:  * returned, until there are no more chunks of source data.
michael@0:  * @param converter the Unicode converter
michael@0:  * @param target I/O parameter. Input : Points to the beginning of the buffer to copy
michael@0:  *  UChars into. Output : points to after the last UChar copied.
michael@0:  * @param targetLimit the pointer just after the end of the <TT>target</TT> buffer
michael@0:  * @param source I/O parameter, pointer to pointer to the source codepage buffer. 
michael@0:  * @param sourceLimit the pointer to the byte after the end of the source buffer
michael@0:  * @param offsets if NULL is passed, nothing will happen to it, otherwise it needs to have the same number
michael@0:  * of allocated cells as <TT>target</TT>. Will fill in offsets from target to source pointer
michael@0:  * e.g: <TT>offsets[3]</TT> is equal to 6, it means that the <TT>target[3]</TT> was a result of transcoding <TT>source[6]</TT>
michael@0:  * For output data carried across calls, and other data without a specific source character
michael@0:  * (such as from escape sequences or callbacks)  -1 will be placed for offsets. 
michael@0:  * @param flush set to <TT>TRUE</TT> if the current source buffer is the last available
michael@0:  * chunk of the source, <TT>FALSE</TT> otherwise. Note that if a failing status is returned,
michael@0:  * this function may have to be called multiple times with flush set to <TT>TRUE</TT> until
michael@0:  * the source buffer is consumed.
michael@0:  * @param err the error status.  <TT>U_ILLEGAL_ARGUMENT_ERROR</TT> will be set if the
michael@0:  * converter is <TT>NULL</TT>.
michael@0:  * <code>U_BUFFER_OVERFLOW_ERROR</code> will be set if the target is full and there is 
michael@0:  * still data to be written to the target. 
michael@0:  * @see ucnv_fromUChars
michael@0:  * @see ucnv_convert
michael@0:  * @see ucnv_getMinCharSize
michael@0:  * @see ucnv_setFromUCallBack
michael@0:  * @see ucnv_getNextUChar
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2 
michael@0: ucnv_toUnicode(UConverter *converter,
michael@0:                UChar **target,
michael@0:                const UChar *targetLimit,
michael@0:                const char **source,
michael@0:                const char *sourceLimit,
michael@0:                int32_t *offsets,
michael@0:                UBool flush,
michael@0:                UErrorCode *err);
michael@0: 
michael@0: /**
michael@0:  * Convert the Unicode string into a codepage string using an existing UConverter.
michael@0:  * The output string is NUL-terminated if possible.
michael@0:  *
michael@0:  * This function is a more convenient but less powerful version of ucnv_fromUnicode().
michael@0:  * It is only useful for whole strings, not for streaming conversion.
michael@0:  *
michael@0:  * The maximum output buffer capacity required (barring output from callbacks) will be
michael@0:  * UCNV_GET_MAX_BYTES_FOR_STRING(srcLength, ucnv_getMaxCharSize(cnv)).
michael@0:  *
michael@0:  * @param cnv the converter object to be used (ucnv_resetFromUnicode() will be called)
michael@0:  * @param src the input Unicode string
michael@0:  * @param srcLength the input string length, or -1 if NUL-terminated
michael@0:  * @param dest destination string buffer, can be NULL if destCapacity==0
michael@0:  * @param destCapacity the number of chars available at dest
michael@0:  * @param pErrorCode normal ICU error code;
michael@0:  *                  common error codes that may be set by this function include
michael@0:  *                  U_BUFFER_OVERFLOW_ERROR, U_STRING_NOT_TERMINATED_WARNING,
michael@0:  *                  U_ILLEGAL_ARGUMENT_ERROR, and conversion errors
michael@0:  * @return the length of the output string, not counting the terminating NUL;
michael@0:  *         if the length is greater than destCapacity, then the string will not fit
michael@0:  *         and a buffer of the indicated length would need to be passed in
michael@0:  * @see ucnv_fromUnicode
michael@0:  * @see ucnv_convert
michael@0:  * @see UCNV_GET_MAX_BYTES_FOR_STRING
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE int32_t U_EXPORT2
michael@0: ucnv_fromUChars(UConverter *cnv,
michael@0:                 char *dest, int32_t destCapacity,
michael@0:                 const UChar *src, int32_t srcLength,
michael@0:                 UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Convert the codepage string into a Unicode string using an existing UConverter.
michael@0:  * The output string is NUL-terminated if possible.
michael@0:  *
michael@0:  * This function is a more convenient but less powerful version of ucnv_toUnicode().
michael@0:  * It is only useful for whole strings, not for streaming conversion.
michael@0:  *
michael@0:  * The maximum output buffer capacity required (barring output from callbacks) will be
michael@0:  * 2*srcLength (each char may be converted into a surrogate pair).
michael@0:  *
michael@0:  * @param cnv the converter object to be used (ucnv_resetToUnicode() will be called)
michael@0:  * @param src the input codepage string
michael@0:  * @param srcLength the input string length, or -1 if NUL-terminated
michael@0:  * @param dest destination string buffer, can be NULL if destCapacity==0
michael@0:  * @param destCapacity the number of UChars available at dest
michael@0:  * @param pErrorCode normal ICU error code;
michael@0:  *                  common error codes that may be set by this function include
michael@0:  *                  U_BUFFER_OVERFLOW_ERROR, U_STRING_NOT_TERMINATED_WARNING,
michael@0:  *                  U_ILLEGAL_ARGUMENT_ERROR, and conversion errors
michael@0:  * @return the length of the output string, not counting the terminating NUL;
michael@0:  *         if the length is greater than destCapacity, then the string will not fit
michael@0:  *         and a buffer of the indicated length would need to be passed in
michael@0:  * @see ucnv_toUnicode
michael@0:  * @see ucnv_convert
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE int32_t U_EXPORT2
michael@0: ucnv_toUChars(UConverter *cnv,
michael@0:               UChar *dest, int32_t destCapacity,
michael@0:               const char *src, int32_t srcLength,
michael@0:               UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Convert a codepage buffer into Unicode one character at a time.
michael@0:  * The input is completely consumed when the U_INDEX_OUTOFBOUNDS_ERROR is set.
michael@0:  *
michael@0:  * Advantage compared to ucnv_toUnicode() or ucnv_toUChars():
michael@0:  * - Faster for small amounts of data, for most converters, e.g.,
michael@0:  *   US-ASCII, ISO-8859-1, UTF-8/16/32, and most "normal" charsets.
michael@0:  *   (For complex converters, e.g., SCSU, UTF-7 and ISO 2022 variants,
michael@0:  *    it uses ucnv_toUnicode() internally.)
michael@0:  * - Convenient.
michael@0:  *
michael@0:  * Limitations compared to ucnv_toUnicode():
michael@0:  * - Always assumes flush=TRUE.
michael@0:  *   This makes ucnv_getNextUChar() unsuitable for "streaming" conversion,
michael@0:  *   that is, for where the input is supplied in multiple buffers,
michael@0:  *   because ucnv_getNextUChar() will assume the end of the input at the end
michael@0:  *   of the first buffer.
michael@0:  * - Does not provide offset output.
michael@0:  *
michael@0:  * It is possible to "mix" ucnv_getNextUChar() and ucnv_toUnicode() because
michael@0:  * ucnv_getNextUChar() uses the current state of the converter
michael@0:  * (unlike ucnv_toUChars() which always resets first).
michael@0:  * However, if ucnv_getNextUChar() is called after ucnv_toUnicode()
michael@0:  * stopped in the middle of a character sequence (with flush=FALSE),
michael@0:  * then ucnv_getNextUChar() will always use the slower ucnv_toUnicode()
michael@0:  * internally until the next character boundary.
michael@0:  * (This is new in ICU 2.6. In earlier releases, ucnv_getNextUChar() had to
michael@0:  * start at a character boundary.)
michael@0:  *
michael@0:  * Instead of using ucnv_getNextUChar(), it is recommended
michael@0:  * to convert using ucnv_toUnicode() or ucnv_toUChars()
michael@0:  * and then iterate over the text using U16_NEXT() or a UCharIterator (uiter.h)
michael@0:  * or a C++ CharacterIterator or similar.
michael@0:  * This allows streaming conversion and offset output, for example.
michael@0:  *
michael@0:  * <p>Handling of surrogate pairs and supplementary-plane code points:<br>
michael@0:  * There are two different kinds of codepages that provide mappings for surrogate characters:
michael@0:  * <ul>
michael@0:  *   <li>Codepages like UTF-8, UTF-32, and GB 18030 provide direct representations for Unicode
michael@0:  *       code points U+10000-U+10ffff as well as for single surrogates U+d800-U+dfff.
michael@0:  *       Each valid sequence will result in exactly one returned code point.
michael@0:  *       If a sequence results in a single surrogate, then that will be returned
michael@0:  *       by itself, even if a neighboring sequence encodes the matching surrogate.</li>
michael@0:  *   <li>Codepages like SCSU and LMBCS (and UTF-16) provide direct representations only for BMP code points
michael@0:  *       including surrogates. Code points in supplementary planes are represented with
michael@0:  *       two sequences, each encoding a surrogate.
michael@0:  *       For these codepages, matching pairs of surrogates will be combined into single
michael@0:  *       code points for returning from this function.
michael@0:  *       (Note that SCSU is actually a mix of these codepage types.)</li>
michael@0:  * </ul></p>
michael@0:  *
michael@0:  * @param converter an open UConverter
michael@0:  * @param source the address of a pointer to the codepage buffer, will be
michael@0:  *  updated to point after the bytes consumed in the conversion call.
michael@0:  * @param sourceLimit points to the end of the input buffer
michael@0:  * @param err fills in error status (see ucnv_toUnicode)
michael@0:  * <code>U_INDEX_OUTOFBOUNDS_ERROR</code> will be set if the input 
michael@0:  * is empty or does not convert to any output (e.g.: pure state-change 
michael@0:  * codes SI/SO, escape sequences for ISO 2022,
michael@0:  * or if the callback did not output anything, ...).
michael@0:  * This function will not set a <code>U_BUFFER_OVERFLOW_ERROR</code> because
michael@0:  *  the "buffer" is the return code. However, there might be subsequent output
michael@0:  *  stored in the converter object
michael@0:  * that will be returned in following calls to this function.
michael@0:  * @return a UChar32 resulting from the partial conversion of source
michael@0:  * @see ucnv_toUnicode
michael@0:  * @see ucnv_toUChars
michael@0:  * @see ucnv_convert
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE UChar32 U_EXPORT2
michael@0: ucnv_getNextUChar(UConverter * converter,
michael@0:                   const char **source,
michael@0:                   const char * sourceLimit,
michael@0:                   UErrorCode * err);
michael@0: 
michael@0: /**
michael@0:  * Convert from one external charset to another using two existing UConverters.
michael@0:  * Internally, two conversions - ucnv_toUnicode() and ucnv_fromUnicode() -
michael@0:  * are used, "pivoting" through 16-bit Unicode.
michael@0:  *
michael@0:  * Important: For streaming conversion (multiple function calls for successive
michael@0:  * parts of a text stream), the caller must provide a pivot buffer explicitly,
michael@0:  * and must preserve the pivot buffer and associated pointers from one
michael@0:  * call to another. (The buffer may be moved if its contents and the relative
michael@0:  * pointer positions are preserved.)
michael@0:  *
michael@0:  * There is a similar function, ucnv_convert(),
michael@0:  * which has the following limitations:
michael@0:  * - it takes charset names, not converter objects, so that
michael@0:  *   - two converters are opened for each call
michael@0:  *   - only single-string conversion is possible, not streaming operation
michael@0:  * - it does not provide enough information to find out,
michael@0:  *   in case of failure, whether the toUnicode or
michael@0:  *   the fromUnicode conversion failed
michael@0:  *
michael@0:  * By contrast, ucnv_convertEx()
michael@0:  * - takes UConverter parameters instead of charset names
michael@0:  * - fully exposes the pivot buffer for streaming conversion and complete error handling
michael@0:  *
michael@0:  * ucnv_convertEx() also provides further convenience:
michael@0:  * - an option to reset the converters at the beginning
michael@0:  *   (if reset==TRUE, see parameters;
michael@0:  *    also sets *pivotTarget=*pivotSource=pivotStart)
michael@0:  * - allow NUL-terminated input
michael@0:  *   (only a single NUL byte, will not work for charsets with multi-byte NULs)
michael@0:  *   (if sourceLimit==NULL, see parameters)
michael@0:  * - terminate with a NUL on output
michael@0:  *   (only a single NUL byte, not useful for charsets with multi-byte NULs),
michael@0:  *   or set U_STRING_NOT_TERMINATED_WARNING if the output exactly fills
michael@0:  *   the target buffer
michael@0:  * - the pivot buffer can be provided internally;
michael@0:  *   possible only for whole-string conversion, not streaming conversion;
michael@0:  *   in this case, the caller will not be able to get details about where an
michael@0:  *   error occurred
michael@0:  *   (if pivotStart==NULL, see below)
michael@0:  *
michael@0:  * The function returns when one of the following is true:
michael@0:  * - the entire source text has been converted successfully to the target buffer
michael@0:  * - a target buffer overflow occurred (U_BUFFER_OVERFLOW_ERROR)
michael@0:  * - a conversion error occurred
michael@0:  *   (other U_FAILURE(), see description of pErrorCode)
michael@0:  *
michael@0:  * Limitation compared to the direct use of
michael@0:  * ucnv_fromUnicode() and ucnv_toUnicode():
michael@0:  * ucnv_convertEx() does not provide offset information.
michael@0:  *
michael@0:  * Limitation compared to ucnv_fromUChars() and ucnv_toUChars():
michael@0:  * ucnv_convertEx() does not support preflighting directly.
michael@0:  *
michael@0:  * Sample code for converting a single string from
michael@0:  * one external charset to UTF-8, ignoring the location of errors:
michael@0:  *
michael@0:  * \code
michael@0:  * int32_t
michael@0:  * myToUTF8(UConverter *cnv,
michael@0:  *          const char *s, int32_t length,
michael@0:  *          char *u8, int32_t capacity,
michael@0:  *          UErrorCode *pErrorCode) {
michael@0:  *     UConverter *utf8Cnv;
michael@0:  *     char *target;
michael@0:  *
michael@0:  *     if(U_FAILURE(*pErrorCode)) {
michael@0:  *         return 0;
michael@0:  *     }
michael@0:  *
michael@0:  *     utf8Cnv=myGetCachedUTF8Converter(pErrorCode);
michael@0:  *     if(U_FAILURE(*pErrorCode)) {
michael@0:  *         return 0;
michael@0:  *     }
michael@0:  *
michael@0:  *     if(length<0) {
michael@0:  *         length=strlen(s);
michael@0:  *     }
michael@0:  *     target=u8;
michael@0:  *     ucnv_convertEx(utf8Cnv, cnv,
michael@0:  *                    &target, u8+capacity,
michael@0:  *                    &s, s+length,
michael@0:  *                    NULL, NULL, NULL, NULL,
michael@0:  *                    TRUE, TRUE,
michael@0:  *                    pErrorCode);
michael@0:  * 
michael@0:  *     myReleaseCachedUTF8Converter(utf8Cnv);
michael@0:  *
michael@0:  *     // return the output string length, but without preflighting
michael@0:  *     return (int32_t)(target-u8);
michael@0:  * }
michael@0:  * \endcode
michael@0:  *
michael@0:  * @param targetCnv     Output converter, used to convert from the UTF-16 pivot
michael@0:  *                      to the target using ucnv_fromUnicode().
michael@0:  * @param sourceCnv     Input converter, used to convert from the source to
michael@0:  *                      the UTF-16 pivot using ucnv_toUnicode().
michael@0:  * @param target        I/O parameter, same as for ucnv_fromUChars().
michael@0:  *                      Input: *target points to the beginning of the target buffer.
michael@0:  *                      Output: *target points to the first unit after the last char written.
michael@0:  * @param targetLimit   Pointer to the first unit after the target buffer.
michael@0:  * @param source        I/O parameter, same as for ucnv_toUChars().
michael@0:  *                      Input: *source points to the beginning of the source buffer.
michael@0:  *                      Output: *source points to the first unit after the last char read.
michael@0:  * @param sourceLimit   Pointer to the first unit after the source buffer.
michael@0:  * @param pivotStart    Pointer to the UTF-16 pivot buffer. If pivotStart==NULL,
michael@0:  *                      then an internal buffer is used and the other pivot
michael@0:  *                      arguments are ignored and can be NULL as well.
michael@0:  * @param pivotSource   I/O parameter, same as source in ucnv_fromUChars() for
michael@0:  *                      conversion from the pivot buffer to the target buffer.
michael@0:  * @param pivotTarget   I/O parameter, same as target in ucnv_toUChars() for
michael@0:  *                      conversion from the source buffer to the pivot buffer.
michael@0:  *                      It must be pivotStart<=*pivotSource<=*pivotTarget<=pivotLimit
michael@0:  *                      and pivotStart<pivotLimit (unless pivotStart==NULL).
michael@0:  * @param pivotLimit    Pointer to the first unit after the pivot buffer.
michael@0:  * @param reset         If TRUE, then ucnv_resetToUnicode(sourceCnv) and
michael@0:  *                      ucnv_resetFromUnicode(targetCnv) are called, and the
michael@0:  *                      pivot pointers are reset (*pivotTarget=*pivotSource=pivotStart).
michael@0:  * @param flush         If true, indicates the end of the input.
michael@0:  *                      Passed directly to ucnv_toUnicode(), and carried over to
michael@0:  *                      ucnv_fromUnicode() when the source is empty as well.
michael@0:  * @param pErrorCode    ICU error code in/out parameter.
michael@0:  *                      Must fulfill U_SUCCESS before the function call.
michael@0:  *                      U_BUFFER_OVERFLOW_ERROR always refers to the target buffer
michael@0:  *                      because overflows into the pivot buffer are handled internally.
michael@0:  *                      Other conversion errors are from the source-to-pivot
michael@0:  *                      conversion if *pivotSource==pivotStart, otherwise from
michael@0:  *                      the pivot-to-target conversion.
michael@0:  *
michael@0:  * @see ucnv_convert
michael@0:  * @see ucnv_fromAlgorithmic
michael@0:  * @see ucnv_toAlgorithmic
michael@0:  * @see ucnv_fromUnicode
michael@0:  * @see ucnv_toUnicode
michael@0:  * @see ucnv_fromUChars
michael@0:  * @see ucnv_toUChars
michael@0:  * @stable ICU 2.6
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_convertEx(UConverter *targetCnv, UConverter *sourceCnv,
michael@0:                char **target, const char *targetLimit,
michael@0:                const char **source, const char *sourceLimit,
michael@0:                UChar *pivotStart, UChar **pivotSource,
michael@0:                UChar **pivotTarget, const UChar *pivotLimit,
michael@0:                UBool reset, UBool flush,
michael@0:                UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Convert from one external charset to another.
michael@0:  * Internally, two converters are opened according to the name arguments,
michael@0:  * then the text is converted to and from the 16-bit Unicode "pivot"
michael@0:  * using ucnv_convertEx(), then the converters are closed again.
michael@0:  *
michael@0:  * This is a convenience function, not an efficient way to convert a lot of text:
michael@0:  * ucnv_convert()
michael@0:  * - takes charset names, not converter objects, so that
michael@0:  *   - two converters are opened for each call
michael@0:  *   - only single-string conversion is possible, not streaming operation
michael@0:  * - does not provide enough information to find out,
michael@0:  *   in case of failure, whether the toUnicode or
michael@0:  *   the fromUnicode conversion failed
michael@0:  * - allows NUL-terminated input
michael@0:  *   (only a single NUL byte, will not work for charsets with multi-byte NULs)
michael@0:  *   (if sourceLength==-1, see parameters)
michael@0:  * - terminate with a NUL on output
michael@0:  *   (only a single NUL byte, not useful for charsets with multi-byte NULs),
michael@0:  *   or set U_STRING_NOT_TERMINATED_WARNING if the output exactly fills
michael@0:  *   the target buffer
michael@0:  * - a pivot buffer is provided internally
michael@0:  *
michael@0:  * The function returns when one of the following is true:
michael@0:  * - the entire source text has been converted successfully to the target buffer
michael@0:  *   and either the target buffer is terminated with a single NUL byte
michael@0:  *   or the error code is set to U_STRING_NOT_TERMINATED_WARNING
michael@0:  * - a target buffer overflow occurred (U_BUFFER_OVERFLOW_ERROR)
michael@0:  *   and the full output string length is returned ("preflighting")
michael@0:  * - a conversion error occurred
michael@0:  *   (other U_FAILURE(), see description of pErrorCode)
michael@0:  *
michael@0:  * @param toConverterName   The name of the converter that is used to convert
michael@0:  *                          from the UTF-16 pivot buffer to the target.
michael@0:  * @param fromConverterName The name of the converter that is used to convert
michael@0:  *                          from the source to the UTF-16 pivot buffer.
michael@0:  * @param target            Pointer to the output buffer.
michael@0:  * @param targetCapacity    Capacity of the target, in bytes.
michael@0:  * @param source            Pointer to the input buffer.
michael@0:  * @param sourceLength      Length of the input text, in bytes, or -1 for NUL-terminated input.
michael@0:  * @param pErrorCode        ICU error code in/out parameter.
michael@0:  *                          Must fulfill U_SUCCESS before the function call.
michael@0:  * @return Length of the complete output text in bytes, even if it exceeds the targetCapacity
michael@0:  *         and a U_BUFFER_OVERFLOW_ERROR is set.
michael@0:  *
michael@0:  * @see ucnv_convertEx
michael@0:  * @see ucnv_fromAlgorithmic
michael@0:  * @see ucnv_toAlgorithmic
michael@0:  * @see ucnv_fromUnicode
michael@0:  * @see ucnv_toUnicode
michael@0:  * @see ucnv_fromUChars
michael@0:  * @see ucnv_toUChars
michael@0:  * @see ucnv_getNextUChar
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE int32_t U_EXPORT2
michael@0: ucnv_convert(const char *toConverterName,
michael@0:              const char *fromConverterName,
michael@0:              char *target,
michael@0:              int32_t targetCapacity,
michael@0:              const char *source,
michael@0:              int32_t sourceLength,
michael@0:              UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Convert from one external charset to another.
michael@0:  * Internally, the text is converted to and from the 16-bit Unicode "pivot"
michael@0:  * using ucnv_convertEx(). ucnv_toAlgorithmic() works exactly like ucnv_convert()
michael@0:  * except that the two converters need not be looked up and opened completely.
michael@0:  *
michael@0:  * The source-to-pivot conversion uses the cnv converter parameter.
michael@0:  * The pivot-to-target conversion uses a purely algorithmic converter
michael@0:  * according to the specified type, e.g., UCNV_UTF8 for a UTF-8 converter.
michael@0:  *
michael@0:  * Internally, the algorithmic converter is opened and closed for each
michael@0:  * function call, which is more efficient than using the public ucnv_open()
michael@0:  * but somewhat less efficient than only resetting an existing converter
michael@0:  * and using ucnv_convertEx().
michael@0:  *
michael@0:  * This function is more convenient than ucnv_convertEx() for single-string
michael@0:  * conversions, especially when "preflighting" is desired (returning the length
michael@0:  * of the complete output even if it does not fit into the target buffer;
michael@0:  * see the User Guide Strings chapter). See ucnv_convert() for details.
michael@0:  *
michael@0:  * @param algorithmicType   UConverterType constant identifying the desired target
michael@0:  *                          charset as a purely algorithmic converter.
michael@0:  *                          Those are converters for Unicode charsets like
michael@0:  *                          UTF-8, BOCU-1, SCSU, UTF-7, IMAP-mailbox-name, etc.,
michael@0:  *                          as well as US-ASCII and ISO-8859-1.
michael@0:  * @param cnv               The converter that is used to convert
michael@0:  *                          from the source to the UTF-16 pivot buffer.
michael@0:  * @param target            Pointer to the output buffer.
michael@0:  * @param targetCapacity    Capacity of the target, in bytes.
michael@0:  * @param source            Pointer to the input buffer.
michael@0:  * @param sourceLength      Length of the input text, in bytes
michael@0:  * @param pErrorCode        ICU error code in/out parameter.
michael@0:  *                          Must fulfill U_SUCCESS before the function call.
michael@0:  * @return Length of the complete output text in bytes, even if it exceeds the targetCapacity
michael@0:  *         and a U_BUFFER_OVERFLOW_ERROR is set.
michael@0:  *
michael@0:  * @see ucnv_fromAlgorithmic
michael@0:  * @see ucnv_convert
michael@0:  * @see ucnv_convertEx
michael@0:  * @see ucnv_fromUnicode
michael@0:  * @see ucnv_toUnicode
michael@0:  * @see ucnv_fromUChars
michael@0:  * @see ucnv_toUChars
michael@0:  * @stable ICU 2.6
michael@0:  */
michael@0: U_STABLE int32_t U_EXPORT2
michael@0: ucnv_toAlgorithmic(UConverterType algorithmicType,
michael@0:                    UConverter *cnv,
michael@0:                    char *target, int32_t targetCapacity,
michael@0:                    const char *source, int32_t sourceLength,
michael@0:                    UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Convert from one external charset to another.
michael@0:  * Internally, the text is converted to and from the 16-bit Unicode "pivot"
michael@0:  * using ucnv_convertEx(). ucnv_fromAlgorithmic() works exactly like ucnv_convert()
michael@0:  * except that the two converters need not be looked up and opened completely.
michael@0:  *
michael@0:  * The source-to-pivot conversion uses a purely algorithmic converter
michael@0:  * according to the specified type, e.g., UCNV_UTF8 for a UTF-8 converter.
michael@0:  * The pivot-to-target conversion uses the cnv converter parameter.
michael@0:  *
michael@0:  * Internally, the algorithmic converter is opened and closed for each
michael@0:  * function call, which is more efficient than using the public ucnv_open()
michael@0:  * but somewhat less efficient than only resetting an existing converter
michael@0:  * and using ucnv_convertEx().
michael@0:  *
michael@0:  * This function is more convenient than ucnv_convertEx() for single-string
michael@0:  * conversions, especially when "preflighting" is desired (returning the length
michael@0:  * of the complete output even if it does not fit into the target buffer;
michael@0:  * see the User Guide Strings chapter). See ucnv_convert() for details.
michael@0:  *
michael@0:  * @param cnv               The converter that is used to convert
michael@0:  *                          from the UTF-16 pivot buffer to the target.
michael@0:  * @param algorithmicType   UConverterType constant identifying the desired source
michael@0:  *                          charset as a purely algorithmic converter.
michael@0:  *                          Those are converters for Unicode charsets like
michael@0:  *                          UTF-8, BOCU-1, SCSU, UTF-7, IMAP-mailbox-name, etc.,
michael@0:  *                          as well as US-ASCII and ISO-8859-1.
michael@0:  * @param target            Pointer to the output buffer.
michael@0:  * @param targetCapacity    Capacity of the target, in bytes.
michael@0:  * @param source            Pointer to the input buffer.
michael@0:  * @param sourceLength      Length of the input text, in bytes
michael@0:  * @param pErrorCode        ICU error code in/out parameter.
michael@0:  *                          Must fulfill U_SUCCESS before the function call.
michael@0:  * @return Length of the complete output text in bytes, even if it exceeds the targetCapacity
michael@0:  *         and a U_BUFFER_OVERFLOW_ERROR is set.
michael@0:  *
michael@0:  * @see ucnv_fromAlgorithmic
michael@0:  * @see ucnv_convert
michael@0:  * @see ucnv_convertEx
michael@0:  * @see ucnv_fromUnicode
michael@0:  * @see ucnv_toUnicode
michael@0:  * @see ucnv_fromUChars
michael@0:  * @see ucnv_toUChars
michael@0:  * @stable ICU 2.6
michael@0:  */
michael@0: U_STABLE int32_t U_EXPORT2
michael@0: ucnv_fromAlgorithmic(UConverter *cnv,
michael@0:                      UConverterType algorithmicType,
michael@0:                      char *target, int32_t targetCapacity,
michael@0:                      const char *source, int32_t sourceLength,
michael@0:                      UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Frees up memory occupied by unused, cached converter shared data.
michael@0:  *
michael@0:  * @return the number of cached converters successfully deleted
michael@0:  * @see ucnv_close
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE int32_t U_EXPORT2
michael@0: ucnv_flushCache(void);
michael@0: 
michael@0: /**
michael@0:  * Returns the number of available converters, as per the alias file.
michael@0:  *
michael@0:  * @return the number of available converters
michael@0:  * @see ucnv_getAvailableName
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE int32_t U_EXPORT2
michael@0: ucnv_countAvailable(void);
michael@0: 
michael@0: /**
michael@0:  * Gets the canonical converter name of the specified converter from a list of
michael@0:  * all available converters contaied in the alias file. All converters
michael@0:  * in this list can be opened.
michael@0:  *
michael@0:  * @param n the index to a converter available on the system (in the range <TT>[0..ucnv_countAvaiable()]</TT>)
michael@0:  * @return a pointer a string (library owned), or <TT>NULL</TT> if the index is out of bounds.
michael@0:  * @see ucnv_countAvailable
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE const char* U_EXPORT2
michael@0: ucnv_getAvailableName(int32_t n);
michael@0: 
michael@0: /**
michael@0:  * Returns a UEnumeration to enumerate all of the canonical converter
michael@0:  * names, as per the alias file, regardless of the ability to open each
michael@0:  * converter.
michael@0:  *
michael@0:  * @return A UEnumeration object for getting all the recognized canonical
michael@0:  *   converter names.
michael@0:  * @see ucnv_getAvailableName
michael@0:  * @see uenum_close
michael@0:  * @see uenum_next
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: U_STABLE UEnumeration * U_EXPORT2
michael@0: ucnv_openAllNames(UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Gives the number of aliases for a given converter or alias name.
michael@0:  * If the alias is ambiguous, then the preferred converter is used
michael@0:  * and the status is set to U_AMBIGUOUS_ALIAS_WARNING.
michael@0:  * This method only enumerates the listed entries in the alias file.
michael@0:  * @param alias alias name
michael@0:  * @param pErrorCode error status
michael@0:  * @return number of names on alias list for given alias
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE uint16_t U_EXPORT2 
michael@0: ucnv_countAliases(const char *alias, UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Gives the name of the alias at given index of alias list.
michael@0:  * This method only enumerates the listed entries in the alias file.
michael@0:  * If the alias is ambiguous, then the preferred converter is used
michael@0:  * and the status is set to U_AMBIGUOUS_ALIAS_WARNING.
michael@0:  * @param alias alias name
michael@0:  * @param n index in alias list
michael@0:  * @param pErrorCode result of operation
michael@0:  * @return returns the name of the alias at given index
michael@0:  * @see ucnv_countAliases
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE const char * U_EXPORT2 
michael@0: ucnv_getAlias(const char *alias, uint16_t n, UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Fill-up the list of alias names for the given alias.
michael@0:  * This method only enumerates the listed entries in the alias file.
michael@0:  * If the alias is ambiguous, then the preferred converter is used
michael@0:  * and the status is set to U_AMBIGUOUS_ALIAS_WARNING.
michael@0:  * @param alias alias name
michael@0:  * @param aliases fill-in list, aliases is a pointer to an array of
michael@0:  *        <code>ucnv_countAliases()</code> string-pointers
michael@0:  *        (<code>const char *</code>) that will be filled in.
michael@0:  *        The strings themselves are owned by the library.
michael@0:  * @param pErrorCode result of operation
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2 
michael@0: ucnv_getAliases(const char *alias, const char **aliases, UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Return a new UEnumeration object for enumerating all the
michael@0:  * alias names for a given converter that are recognized by a standard.
michael@0:  * This method only enumerates the listed entries in the alias file.
michael@0:  * The convrtrs.txt file can be modified to change the results of
michael@0:  * this function.
michael@0:  * The first result in this list is the same result given by
michael@0:  * <code>ucnv_getStandardName</code>, which is the default alias for
michael@0:  * the specified standard name. The returned object must be closed with
michael@0:  * <code>uenum_close</code> when you are done with the object.
michael@0:  *
michael@0:  * @param convName original converter name
michael@0:  * @param standard name of the standard governing the names; MIME and IANA
michael@0:  *      are such standards
michael@0:  * @param pErrorCode The error code
michael@0:  * @return A UEnumeration object for getting all aliases that are recognized
michael@0:  *      by a standard. If any of the parameters are invalid, NULL
michael@0:  *      is returned.
michael@0:  * @see ucnv_getStandardName
michael@0:  * @see uenum_close
michael@0:  * @see uenum_next
michael@0:  * @stable ICU 2.2
michael@0:  */
michael@0: U_STABLE UEnumeration * U_EXPORT2
michael@0: ucnv_openStandardNames(const char *convName,
michael@0:                        const char *standard,
michael@0:                        UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Gives the number of standards associated to converter names.
michael@0:  * @return number of standards
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE uint16_t U_EXPORT2
michael@0: ucnv_countStandards(void);
michael@0: 
michael@0: /**
michael@0:  * Gives the name of the standard at given index of standard list.
michael@0:  * @param n index in standard list
michael@0:  * @param pErrorCode result of operation
michael@0:  * @return returns the name of the standard at given index. Owned by the library.
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE const char * U_EXPORT2
michael@0: ucnv_getStandard(uint16_t n, UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Returns a standard name for a given converter name.
michael@0:  * <p>
michael@0:  * Example alias table:<br>
michael@0:  * conv alias1 { STANDARD1 } alias2 { STANDARD1* }
michael@0:  * <p>
michael@0:  * Result of ucnv_getStandardName("conv", "STANDARD1") from example
michael@0:  * alias table:<br>
michael@0:  * <b>"alias2"</b>
michael@0:  *
michael@0:  * @param name original converter name
michael@0:  * @param standard name of the standard governing the names; MIME and IANA
michael@0:  *        are such standards
michael@0:  * @param pErrorCode result of operation
michael@0:  * @return returns the standard converter name;
michael@0:  *         if a standard converter name cannot be determined,
michael@0:  *         then <code>NULL</code> is returned. Owned by the library.
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE const char * U_EXPORT2
michael@0: ucnv_getStandardName(const char *name, const char *standard, UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * This function will return the internal canonical converter name of the
michael@0:  * tagged alias. This is the opposite of ucnv_openStandardNames, which
michael@0:  * returns the tagged alias given the canonical name.
michael@0:  * <p>
michael@0:  * Example alias table:<br>
michael@0:  * conv alias1 { STANDARD1 } alias2 { STANDARD1* }
michael@0:  * <p>
michael@0:  * Result of ucnv_getStandardName("alias1", "STANDARD1") from example
michael@0:  * alias table:<br>
michael@0:  * <b>"conv"</b>
michael@0:  *
michael@0:  * @return returns the canonical converter name;
michael@0:  *         if a standard or alias name cannot be determined,
michael@0:  *         then <code>NULL</code> is returned. The returned string is
michael@0:  *         owned by the library.
michael@0:  * @see ucnv_getStandardName
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: U_STABLE const char * U_EXPORT2
michael@0: ucnv_getCanonicalName(const char *alias, const char *standard, UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Returns the current default converter name. If you want to open
michael@0:  * a default converter, you do not need to use this function.
michael@0:  * It is faster if you pass a NULL argument to ucnv_open the
michael@0:  * default converter.
michael@0:  *
michael@0:  * If U_CHARSET_IS_UTF8 is defined to 1 in utypes.h then this function
michael@0:  * always returns "UTF-8".
michael@0:  *
michael@0:  * @return returns the current default converter name.
michael@0:  *         Storage owned by the library
michael@0:  * @see ucnv_setDefaultName
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE const char * U_EXPORT2
michael@0: ucnv_getDefaultName(void);
michael@0: 
michael@0: #ifndef U_HIDE_SYSTEM_API
michael@0: /**
michael@0:  * This function is not thread safe. DO NOT call this function when ANY ICU
michael@0:  * function is being used from more than one thread! This function sets the
michael@0:  * current default converter name. If this function needs to be called, it
michael@0:  * should be called during application initialization. Most of the time, the
michael@0:  * results from ucnv_getDefaultName() or ucnv_open with a NULL string argument
michael@0:  * is sufficient for your application.
michael@0:  *
michael@0:  * If U_CHARSET_IS_UTF8 is defined to 1 in utypes.h then this function
michael@0:  * does nothing.
michael@0:  *
michael@0:  * @param name the converter name to be the default (must be known by ICU).
michael@0:  * @see ucnv_getDefaultName
michael@0:  * @system
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_setDefaultName(const char *name);
michael@0: #endif  /* U_HIDE_SYSTEM_API */
michael@0: 
michael@0: /**
michael@0:  * Fixes the backslash character mismapping.  For example, in SJIS, the backslash 
michael@0:  * character in the ASCII portion is also used to represent the yen currency sign.  
michael@0:  * When mapping from Unicode character 0x005C, it's unclear whether to map the 
michael@0:  * character back to yen or backslash in SJIS.  This function will take the input
michael@0:  * buffer and replace all the yen sign characters with backslash.  This is necessary
michael@0:  * when the user tries to open a file with the input buffer on Windows.
michael@0:  * This function will test the converter to see whether such mapping is
michael@0:  * required.  You can sometimes avoid using this function by using the correct version
michael@0:  * of Shift-JIS.
michael@0:  *
michael@0:  * @param cnv The converter representing the target codepage.
michael@0:  * @param source the input buffer to be fixed
michael@0:  * @param sourceLen the length of the input buffer
michael@0:  * @see ucnv_isAmbiguous
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE void U_EXPORT2
michael@0: ucnv_fixFileSeparator(const UConverter *cnv, UChar *source, int32_t sourceLen);
michael@0: 
michael@0: /**
michael@0:  * Determines if the converter contains ambiguous mappings of the same
michael@0:  * character or not.
michael@0:  * @param cnv the converter to be tested
michael@0:  * @return TRUE if the converter contains ambiguous mapping of the same 
michael@0:  * character, FALSE otherwise.
michael@0:  * @stable ICU 2.0
michael@0:  */
michael@0: U_STABLE UBool U_EXPORT2
michael@0: ucnv_isAmbiguous(const UConverter *cnv);
michael@0: 
michael@0: /**
michael@0:  * Sets the converter to use fallback mappings or not.
michael@0:  * Regardless of this flag, the converter will always use
michael@0:  * fallbacks from Unicode Private Use code points, as well as
michael@0:  * reverse fallbacks (to Unicode).
michael@0:  * For details see ".ucm File Format"
michael@0:  * in the Conversion Data chapter of the ICU User Guide:
michael@0:  * http://www.icu-project.org/userguide/conversion-data.html#ucmformat
michael@0:  *
michael@0:  * @param cnv The converter to set the fallback mapping usage on.
michael@0:  * @param usesFallback TRUE if the user wants the converter to take advantage of the fallback 
michael@0:  * mapping, FALSE otherwise.
michael@0:  * @stable ICU 2.0
michael@0:  * @see ucnv_usesFallback
michael@0:  */
michael@0: U_STABLE void U_EXPORT2 
michael@0: ucnv_setFallback(UConverter *cnv, UBool usesFallback);
michael@0: 
michael@0: /**
michael@0:  * Determines if the converter uses fallback mappings or not.
michael@0:  * This flag has restrictions, see ucnv_setFallback().
michael@0:  *
michael@0:  * @param cnv The converter to be tested
michael@0:  * @return TRUE if the converter uses fallback, FALSE otherwise.
michael@0:  * @stable ICU 2.0
michael@0:  * @see ucnv_setFallback
michael@0:  */
michael@0: U_STABLE UBool U_EXPORT2 
michael@0: ucnv_usesFallback(const UConverter *cnv);
michael@0: 
michael@0: /**
michael@0:  * Detects Unicode signature byte sequences at the start of the byte stream
michael@0:  * and returns the charset name of the indicated Unicode charset.
michael@0:  * NULL is returned when no Unicode signature is recognized.
michael@0:  * The number of bytes in the signature is output as well.
michael@0:  *
michael@0:  * The caller can ucnv_open() a converter using the charset name.
michael@0:  * The first code unit (UChar) from the start of the stream will be U+FEFF
michael@0:  * (the Unicode BOM/signature character) and can usually be ignored.
michael@0:  *
michael@0:  * For most Unicode charsets it is also possible to ignore the indicated
michael@0:  * number of initial stream bytes and start converting after them.
michael@0:  * However, there are stateful Unicode charsets (UTF-7 and BOCU-1) for which
michael@0:  * this will not work. Therefore, it is best to ignore the first output UChar
michael@0:  * instead of the input signature bytes.
michael@0:  * <p>
michael@0:  * Usage:
michael@0:  * \snippet samples/ucnv/convsamp.cpp ucnv_detectUnicodeSignature
michael@0:  *
michael@0:  * @param source            The source string in which the signature should be detected.
michael@0:  * @param sourceLength      Length of the input string, or -1 if terminated with a NUL byte.
michael@0:  * @param signatureLength   A pointer to int32_t to receive the number of bytes that make up the signature 
michael@0:  *                          of the detected UTF. 0 if not detected.
michael@0:  *                          Can be a NULL pointer.
michael@0:  * @param pErrorCode        ICU error code in/out parameter.
michael@0:  *                          Must fulfill U_SUCCESS before the function call.
michael@0:  * @return The name of the encoding detected. NULL if encoding is not detected. 
michael@0:  * @stable ICU 2.4
michael@0:  */
michael@0: U_STABLE const char* U_EXPORT2
michael@0: ucnv_detectUnicodeSignature(const char* source,
michael@0:                             int32_t sourceLength,
michael@0:                             int32_t *signatureLength,
michael@0:                             UErrorCode *pErrorCode);
michael@0: 
michael@0: /**
michael@0:  * Returns the number of UChars held in the converter's internal state 
michael@0:  * because more input is needed for completing the conversion. This function is 
michael@0:  * useful for mapping semantics of ICU's converter interface to those of iconv,
michael@0:  * and this information is not needed for normal conversion.
michael@0:  * @param cnv       The converter in which the input is held
michael@0:  * @param status    ICU error code in/out parameter.
michael@0:  *                  Must fulfill U_SUCCESS before the function call.
michael@0:  * @return The number of UChars in the state. -1 if an error is encountered.
michael@0:  * @stable ICU 3.4
michael@0:  */
michael@0: U_STABLE int32_t U_EXPORT2
michael@0: ucnv_fromUCountPending(const UConverter* cnv, UErrorCode* status);
michael@0: 
michael@0: /**
michael@0:  * Returns the number of chars held in the converter's internal state
michael@0:  * because more input is needed for completing the conversion. This function is 
michael@0:  * useful for mapping semantics of ICU's converter interface to those of iconv,
michael@0:  * and this information is not needed for normal conversion.
michael@0:  * @param cnv       The converter in which the input is held as internal state
michael@0:  * @param status    ICU error code in/out parameter.
michael@0:  *                  Must fulfill U_SUCCESS before the function call.
michael@0:  * @return The number of chars in the state. -1 if an error is encountered.
michael@0:  * @stable ICU 3.4
michael@0:  */
michael@0: U_STABLE int32_t U_EXPORT2
michael@0: ucnv_toUCountPending(const UConverter* cnv, UErrorCode* status);
michael@0: 
michael@0: /**
michael@0:  * Returns whether or not the charset of the converter has a fixed number of bytes
michael@0:  * per charset character.
michael@0:  * An example of this are converters that are of the type UCNV_SBCS or UCNV_DBCS.
michael@0:  * Another example is UTF-32 which is always 4 bytes per character.
michael@0:  * A Unicode code point may be represented by more than one UTF-8 or UTF-16 code unit
michael@0:  * but a UTF-32 converter encodes each code point with 4 bytes.
michael@0:  * Note: This method is not intended to be used to determine whether the charset has a
michael@0:  * fixed ratio of bytes to Unicode codes <i>units</i> for any particular Unicode encoding form.
michael@0:  * FALSE is returned with the UErrorCode if error occurs or cnv is NULL.
michael@0:  * @param cnv       The converter to be tested
michael@0:  * @param status    ICU error code in/out paramter
michael@0:  * @return TRUE if the converter is fixed-width
michael@0:  * @stable ICU 4.8
michael@0:  */
michael@0: U_STABLE UBool U_EXPORT2
michael@0: ucnv_isFixedWidth(UConverter *cnv, UErrorCode *status);
michael@0: 
michael@0: #endif
michael@0: 
michael@0: #endif
michael@0: /*_UCNV*/