path: root/include/rtl/ustring.hxx

/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
/*
 * This file is part of the LibreOffice project.
 *
 * This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/.
 *
 * This file incorporates work covered by the following license notice:
 *
 *   Licensed to the Apache Software Foundation (ASF) under one or more
 *   contributor license agreements. See the NOTICE file distributed
 *   with this work for additional information regarding copyright
 *   ownership. The ASF licenses this file to you under the Apache
 *   License, Version 2.0 (the "License"); you may not use this file
 *   except in compliance with the License. You may obtain a copy of
 *   the License at http://www.apache.org/licenses/LICENSE-2.0 .
 */

#ifndef INCLUDED_RTL_USTRING_HXX
#define INCLUDED_RTL_USTRING_HXX

#include <sal/config.h>

#include <cassert>
#include <new>
#include <ostream>
#include <string.h>

#include <rtl/ustring.h>
#include <rtl/string.hxx>
#include <rtl/stringutils.hxx>
#include <rtl/textenc.h>
#include <sal/log.hxx>

#ifdef LIBO_INTERNAL_ONLY // "RTL_FAST_STRING"
#include <rtl/stringconcat.hxx>
#endif

// The unittest uses slightly different code to help check that the proper
// calls are made. The class is put into a different namespace to make
// sure the compiler generates a different (if generating also non-inline)
// copy of the function and does not merge them together. The class
// is "brought" into the proper rtl namespace by a typedef below.
#ifdef RTL_STRING_UNITTEST
#define rtl rtlunittest
#endif

namespace rtl
{

#ifdef RTL_STRING_UNITTEST
#undef rtl
#endif

#if defined LIBO_INTERNAL_ONLY // "RTL_FAST_STRING"
/// @cond INTERNAL

/**
A simple wrapper around string literal. It is usually not necessary to use, can
be mostly used to force OUString operator+ working with operands that otherwise would
not trigger it.

This class is not part of public API and is meant to be used only in LibreOffice code.
@since LibreOffice 4.0
*/
struct SAL_WARN_UNUSED OUStringLiteral
{
    template< int N >
    explicit SAL_CONSTEXPR OUStringLiteral( const char (&str)[ N ] ) : size( N - 1 ), data( str )
    { /* only C++14 constexpr: assert( strlen( str ) == N - 1 ); */ }
    int size;
    const char* data;
};

/** A simple wrapper around an ASCII character literal, for use in certain
    OUString functions designed for efficient processing of string literals.

    @since LibreOffice 5.0
*/
template<char C> struct SAL_WARN_UNUSED OUStringLiteral1 {
    static_assert(
        static_cast<unsigned char>(C) < 0x80,
        "non-ASCII character in OUStringLiteral1");
};

/// @endcond
#endif

/* ======================================================================= */

/**
  This String class provides base functionality for C++ like Unicode
  character array handling. The advantage of this class is that it
  handles all the memory management for you - and it does it
  more efficiently. If you assign a string to another string, the
  data of both strings are shared (without any copy operation or
  memory allocation) as long as you do not change the string. This class
  also stores the length of the string, so that many operations are
  faster than the C-str-functions.

  This class provides only readonly string handling. So you could create
  a string and you could only query the content from this string.
  It provides also functionality to change the string, but this results
  in every case in a new string instance (in the most cases with a
  memory allocation). You don't have functionality to change the
  content of the string. If you want to change the string content, then
  you should use the OStringBuffer class, which provides these
  functionalities and avoids too much memory allocation.

  The design of this class is similar to the string classes in Java so
  less people should have understanding problems when they use this class.
*/

class SAL_WARN_UNUSED SAL_DLLPUBLIC_RTTI OUString
{
public:
    /// @cond INTERNAL
    rtl_uString * pData;
    /// @endcond

    /**
      New string containing no characters.
    */
    OUString()
    {
        pData = 0;
        rtl_uString_new( &pData );
    }

    /**
      New string from OUString.

      @param    str         a OUString.
    */
    OUString( const OUString & str )
    {
        pData = str.pData;
        rtl_uString_acquire( pData );
    }

    /**
      New string from OUString data.

      @param    str         a OUString data.
    */
    OUString( rtl_uString * str )
    {
        pData = str;
        rtl_uString_acquire( pData );
    }

    /** New OUString from OUString data without acquiring it.  Takeover of ownership.

        The SAL_NO_ACQUIRE dummy parameter is only there to distinguish this
        from other constructors.

        @param str
               OUString data
    */
    inline OUString( rtl_uString * str, __sal_NoAcquire )
        { pData = str; }

    /**
      New string from a single Unicode character.

      @param    value       a Unicode character.
    */
    explicit OUString( sal_Unicode value )
        : pData (0)
    {
        rtl_uString_newFromStr_WithLength( &pData, &value, 1 );
    }

#if defined LIBO_INTERNAL_ONLY && !defined RTL_STRING_UNITTEST_CONCAT
    /// @cond INTERNAL
    // Catch inadvertent conversions to the above ctor (but still allow
    // construction from char literals):
    OUString(int) = delete;
    explicit OUString(char c):
        OUString(sal_Unicode(static_cast<unsigned char>(c)))
    {}
    /// @endcond
#endif

    /**
      New string from a Unicode character buffer array.

      @param    value       a NULL-terminated Unicode character array.
    */
    OUString( const sal_Unicode * value )
    {
        pData = 0;
        rtl_uString_newFromStr( &pData, value );
    }

    /**
      New string from a Unicode character buffer array.

      @param    value       a Unicode character array.
      @param    length      the number of character which should be copied.
                            The character array length must be greater than
                            or equal to this value.
    */
    OUString( const sal_Unicode * value, sal_Int32 length )
    {
        pData = 0;
        rtl_uString_newFromStr_WithLength( &pData, value, length );
    }

    /**
      New string from an 8-Bit string literal that is expected to contain only
      characters in the ASCII set (i.e. first 128 characters). This constructor
      allows an efficient and convenient way to create OUString
      instances from ASCII literals. When creating strings from data that
      is not pure ASCII, it needs to be converted to OUString by explicitly
      providing the encoding to use for the conversion.

      If there are any embedded \0's in the string literal, the result is undefined.
      Use the overload that explicitly accepts length.

      @param    literal         the 8-bit ASCII string literal

      @since LibreOffice 3.6
    */
    template< typename T >
    OUString( T& literal, typename libreoffice_internal::ConstCharArrayDetector< T, libreoffice_internal::Dummy >::Type = libreoffice_internal::Dummy() )
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        pData = 0;
        if( libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 == 0 ) // empty string
            rtl_uString_new( &pData );
        else
            rtl_uString_newFromLiteral( &pData, literal, libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1, 0 );
#ifdef RTL_STRING_UNITTEST
        rtl_string_unittest_const_literal = true;
#endif
    }

#ifdef RTL_STRING_UNITTEST
    /**
     * Only used by unittests to detect incorrect conversions.
     * @internal
     */
    template< typename T >
    OUString( T&, typename libreoffice_internal::ExceptConstCharArrayDetector< T >::Type = libreoffice_internal::Dummy() )
    {
        pData = 0;
        rtl_uString_newFromLiteral( &pData, "!!br0ken!!", 10, 0 ); // set to garbage
        rtl_string_unittest_invalid_conversion = true;
    }
    /**
     * Only used by unittests to detect incorrect conversions.
     * @internal
     */
    template< typename T >
    OUString( const T&, typename libreoffice_internal::ExceptCharArrayDetector< T >::Type = libreoffice_internal::Dummy() )
    {
        pData = 0;
        rtl_uString_newFromLiteral( &pData, "!!br0ken!!", 10, 0 ); // set to garbage
        rtl_string_unittest_invalid_conversion = true;
    }
#endif

#ifdef LIBO_INTERNAL_ONLY // "RTL_FAST_STRING"
    /// @cond INTERNAL
    /**
      New string from an 8-Bit string literal that is expected to contain only
      characters in the ASCII set (i.e. first 128 characters).

      This constructor is similar to the "direct template" one, but can be
      useful in cases where the latter does not work, like in

        OUString(flag ? "a" : "bb")

      written as

        OUString(flag ? OUStringLiteral("a") : OUStringLiteral("bb"))

      @since LibreOffice 5.0
    */
    OUString(OUStringLiteral literal): pData(0) {
        rtl_uString_newFromLiteral(&pData, literal.data, literal.size, 0);
    }
    /// @endcond
#endif

    /**
      New string from an 8-Bit character buffer array.

      @param    value           An 8-Bit character array.
      @param    length          The number of character which should be converted.
                                The 8-Bit character array length must be
                                greater than or equal to this value.
      @param    encoding        The text encoding from which the 8-Bit character
                                sequence should be converted.
      @param    convertFlags    Flags which control the conversion.
                                see RTL_TEXTTOUNICODE_FLAGS_...

      @exception std::bad_alloc is thrown if an out-of-memory condition occurs
    */
    OUString( const sal_Char * value, sal_Int32 length,
              rtl_TextEncoding encoding,
              sal_uInt32 convertFlags = OSTRING_TO_OUSTRING_CVTFLAGS )
    {
        pData = 0;
        rtl_string2UString( &pData, value, length, encoding, convertFlags );
        if (pData == 0) {
            throw std::bad_alloc();
        }
    }

    /** Create a new string from an array of Unicode code points.

        @param codePoints
        an array of at least codePointCount code points, which each must be in
        the range from 0 to 0x10FFFF, inclusive.  May be null if codePointCount
        is zero.

        @param codePointCount
        the non-negative number of code points.

        @exception std::bad_alloc
        is thrown if either an out-of-memory condition occurs or the resulting
        number of UTF-16 code units would have been larger than SAL_MAX_INT32.

        @since UDK 3.2.7
    */
    inline explicit OUString(
        sal_uInt32 const * codePoints, sal_Int32 codePointCount):
        pData(NULL)
    {
        rtl_uString_newFromCodePoints(&pData, codePoints, codePointCount);
        if (pData == NULL) {
            throw std::bad_alloc();
        }
    }

#ifdef LIBO_INTERNAL_ONLY // "RTL_FAST_STRING"
    /**
     @overload
     @internal
    */
    template< typename T1, typename T2 >
    OUString( const OUStringConcat< T1, T2 >& c )
    {
        const sal_Int32 l = c.length();
        pData = rtl_uString_alloc( l );
        if (l != 0)
        {
            sal_Unicode* end = c.addData( pData->buffer );
            pData->length = l;
            *end = '\0';
            // TODO realloc in case pData->length is noticeably smaller than l?
        }
    }
#endif

    /**
      Release the string data.
    */
    ~OUString()
    {
        rtl_uString_release( pData );
    }

    /** Provides an OUString const & passing a storage pointer of an
        rtl_uString * handle.
        It is more convenient to use C++ OUString member functions when dealing
        with rtl_uString * handles.  Using this function avoids unnecessary
        acquire()/release() calls for a temporary OUString object.

        @param ppHandle
               pointer to storage
        @return
               OUString const & based on given storage
    */
    static inline OUString const & unacquired( rtl_uString * const * ppHandle )
        { return * reinterpret_cast< OUString const * >( ppHandle ); }

    /**
      Assign a new string.

      @param    str         a OUString.
    */
    OUString & operator=( const OUString & str )
    {
        rtl_uString_assign( &pData, str.pData );
        return *this;
    }

    /**
      Assign a new string from an 8-Bit string literal that is expected to contain only
      characters in the ASCII set (i.e. first 128 characters). This operator
      allows an efficient and convenient way to assign OUString
      instances from ASCII literals. When assigning strings from data that
      is not pure ASCII, it needs to be converted to OUString by explicitly
      providing the encoding to use for the conversion.

      @param    literal         the 8-bit ASCII string literal

      @since LibreOffice 3.6
    */
    template< typename T >
    typename libreoffice_internal::ConstCharArrayDetector< T, OUString& >::Type operator=( T& literal )
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        if( libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 == 0 ) // empty string
            rtl_uString_new( &pData );
        else
            rtl_uString_newFromLiteral( &pData, literal, libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1, 0 );
        return *this;
    }

#if defined LIBO_INTERNAL_ONLY // "RTL_FAST_STRING"
    /// @cond INTERNAL
    /** Assign a new string from a single ASCII character literal.

        @since LibreOffice 5.0
    */
    template<char C> OUString & operator =(OUStringLiteral1<C>) {
        sal_Unicode const c = C;
        rtl_uString_newFromStr_WithLength(&pData, &c, 1);
        return *this;
    }
    /// @endcond
#endif

    /**
      Append a string to this string.

      @param    str         a OUString.
    */
    OUString & operator+=( const OUString & str )
    {
        rtl_uString_newConcat( &pData, pData, str.pData );
        return *this;
    }

#ifdef LIBO_INTERNAL_ONLY // "RTL_FAST_STRING"
    /**
     @overload
     @internal
    */
    template< typename T1, typename T2 >
    OUString& operator+=( const OUStringConcat< T1, T2 >& c )
    {
        sal_Int32 l = c.length();
        if( l == 0 )
            return *this;
        l += pData->length;
        rtl_uString_ensureCapacity( &pData, l );
        sal_Unicode* end = c.addData( pData->buffer + pData->length );
        *end = '\0';
        pData->length = l;
        return *this;
    }
#endif

    /**
      Clears the string, i.e, makes a zero-character string
      @since LibreOffice 4.4
    */
    void clear()
    {
        rtl_uString_new( &pData );
    }

    /**
      Returns the length of this string.

      The length is equal to the number of Unicode characters in this string.

      @return   the length of the sequence of characters represented by this
                object.
    */
    sal_Int32 getLength() const { return pData->length; }

    /**
      Checks if a string is empty.

      @return   true if the string is empty;
                false, otherwise.

      @since LibreOffice 3.4
    */
    bool isEmpty() const
    {
        return pData->length == 0;
    }

    /**
      Returns a pointer to the Unicode character buffer for this string.

      It isn't necessarily NULL terminated.

      @return   a pointer to the Unicode characters buffer for this object.
    */
    const sal_Unicode * getStr() const { return pData->buffer; }

    /**
      Access to individual characters.

      @param index must be non-negative and less than length.

      @return the character at the given index.

      @since LibreOffice 3.5
    */
    sal_Unicode operator [](sal_Int32 index) const {
        // silence spurious -Werror=strict-overflow warnings from GCC 4.8.2
        assert(index >= 0 && static_cast<sal_uInt32>(index) < static_cast<sal_uInt32>(getLength()));
        return getStr()[index];
    }

    /**
      Compares two strings.

      The comparison is based on the numeric value of each character in
      the strings and return a value indicating their relationship.
      This function can't be used for language specific sorting.

      @param    str         the object to be compared.
      @return   0 - if both strings are equal
                < 0 - if this string is less than the string argument
                > 0 - if this string is greater than the string argument
    */
    sal_Int32 compareTo( const OUString & str ) const
    {
        return rtl_ustr_compare_WithLength( pData->buffer, pData->length,
                                            str.pData->buffer, str.pData->length );
    }

    /**
      Compares two strings with a maximum count of characters.

      The comparison is based on the numeric value of each character in
      the strings and return a value indicating their relationship.
      This function can't be used for language specific sorting.

      @param    str         the object to be compared.
      @param    maxLength   the maximum count of characters to be compared.
      @return   0 - if both strings are equal
                < 0 - if this string is less than the string argument
                > 0 - if this string is greater than the string argument

      @since UDK 3.2.7
    */
    sal_Int32 compareTo( const OUString & str, sal_Int32 maxLength ) const
    {
        return rtl_ustr_shortenedCompare_WithLength( pData->buffer, pData->length,
                                                     str.pData->buffer, str.pData->length, maxLength );
    }

    /**
      Compares two strings in reverse order.

      The comparison is based on the numeric value of each character in
      the strings and return a value indicating their relationship.
      This function can't be used for language specific sorting.

      @param    str         the object to be compared.
      @return   0 - if both strings are equal
                < 0 - if this string is less than the string argument
                > 0 - if this string is greater than the string argument
    */
    sal_Int32 reverseCompareTo( const OUString & str ) const
    {
        return rtl_ustr_reverseCompare_WithLength( pData->buffer, pData->length,
                                                   str.pData->buffer, str.pData->length );
    }

    /**
     @overload
     This function accepts an ASCII string literal as its argument.
     @since LibreOffice 4.1
    */
    template< typename T >
    typename libreoffice_internal::ConstCharArrayDetector< T, sal_Int32 >::Type reverseCompareTo( T& literal ) const
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        return rtl_ustr_asciil_reverseCompare_WithLength( pData->buffer, pData->length,
            literal, libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 );
    }

    /**
      Perform a comparison of two strings.

      The result is true if and only if second string
      represents the same sequence of characters as the first string.
      This function can't be used for language specific comparison.

      @param    str         the object to be compared.
      @return   true if the strings are equal;
                false, otherwise.
    */
    bool equals( const OUString & str ) const
    {
        if ( pData->length != str.pData->length )
            return false;
        if ( pData == str.pData )
            return true;
        return rtl_ustr_reverseCompare_WithLength( pData->buffer, pData->length,
                                                   str.pData->buffer, str.pData->length ) == 0;
    }

    /**
      Perform a ASCII lowercase comparison of two strings.

      The result is true if and only if second string
      represents the same sequence of characters as the first string,
      ignoring the case.
      Character values between 65 and 90 (ASCII A-Z) are interpreted as
      values between 97 and 122 (ASCII a-z).
      This function can't be used for language specific comparison.

      @param    str         the object to be compared.
      @return   true if the strings are equal;
                false, otherwise.
    */
    bool equalsIgnoreAsciiCase( const OUString & str ) const
    {
        if ( pData->length != str.pData->length )
            return false;
        if ( pData == str.pData )
            return true;
        return rtl_ustr_compareIgnoreAsciiCase_WithLength( pData->buffer, pData->length,
                                                           str.pData->buffer, str.pData->length ) == 0;
    }

    /**
      Perform a ASCII lowercase comparison of two strings.

      Compare the two strings with uppercase ASCII
      character values between 65 and 90 (ASCII A-Z) interpreted as
      values between 97 and 122 (ASCII a-z).
      This function can't be used for language specific comparison.

      @param    str         the object to be compared.
      @return   0 - if both strings are equal
                < 0 - if this string is less than the string argument
                > 0 - if this string is greater than the string argument

      @since LibreOffice 4.0
    */
    sal_Int32 compareToIgnoreAsciiCase( const OUString & str ) const
    {
        return rtl_ustr_compareIgnoreAsciiCase_WithLength( pData->buffer, pData->length,
                                                           str.pData->buffer, str.pData->length );
    }


    /**
     @overload
     This function accepts an ASCII string literal as its argument.
     @since LibreOffice 3.6
    */
    template< typename T >
    typename libreoffice_internal::ConstCharArrayDetector< T, bool >::Type equalsIgnoreAsciiCase( T& literal ) const
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        if ( pData->length != libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 )
            return false;

        return rtl_ustr_ascii_compareIgnoreAsciiCase_WithLength( pData->buffer, pData->length, literal ) == 0;
    }

   /**
      Match against a substring appearing in this string.

      The result is true if and only if the second string appears as a substring
      of this string, at the given position.
      This function can't be used for language specific comparison.

      @param    str         the object (substring) to be compared.
      @param    fromIndex   the index to start the comparion from.
                            The index must be greater than or equal to 0
                            and less or equal as the string length.
      @return   true if str match with the characters in the string
                at the given position;
                false, otherwise.
    */
    bool match( const OUString & str, sal_Int32 fromIndex = 0 ) const
    {
        return rtl_ustr_shortenedCompare_WithLength( pData->buffer+fromIndex, pData->length-fromIndex,
                                                     str.pData->buffer, str.pData->length, str.pData->length ) == 0;
    }

    /**
     @overload
     This function accepts an ASCII string literal as its argument.
     @since LibreOffice 3.6
    */
    template< typename T >
    typename libreoffice_internal::ConstCharArrayDetector< T, bool >::Type match( T& literal, sal_Int32 fromIndex = 0 ) const
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        return rtl_ustr_ascii_shortenedCompare_WithLength( pData->buffer+fromIndex, pData->length-fromIndex,
            literal, libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 ) == 0;
    }

    /**
      Match against a substring appearing in this string, ignoring the case of
      ASCII letters.

      The result is true if and only if the second string appears as a substring
      of this string, at the given position.
      Character values between 65 and 90 (ASCII A-Z) are interpreted as
      values between 97 and 122 (ASCII a-z).
      This function can't be used for language specific comparison.

      @param    str         the object (substring) to be compared.
      @param    fromIndex   the index to start the comparion from.
                            The index must be greater than or equal to 0
                            and less than or equal to the string length.
      @return   true if str match with the characters in the string
                at the given position;
                false, otherwise.
    */
    bool matchIgnoreAsciiCase( const OUString & str, sal_Int32 fromIndex = 0 ) const
    {
        return rtl_ustr_shortenedCompareIgnoreAsciiCase_WithLength( pData->buffer+fromIndex, pData->length-fromIndex,
                                                                    str.pData->buffer, str.pData->length,
                                                                    str.pData->length ) == 0;
    }

    /**
     @overload
     This function accepts an ASCII string literal as its argument.
     @since LibreOffice 3.6
    */
    template< typename T >
    typename libreoffice_internal::ConstCharArrayDetector< T, bool >::Type matchIgnoreAsciiCase( T& literal, sal_Int32 fromIndex = 0 ) const
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        return rtl_ustr_ascii_shortenedCompareIgnoreAsciiCase_WithLength( pData->buffer+fromIndex, pData->length-fromIndex,
            literal, libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 ) == 0;
    }

    /**
      Compares two strings.

      The comparison is based on the numeric value of each character in
      the strings and return a value indicating their relationship.
      Since this method is optimized for performance, the ASCII character
      values are not converted in any way. The caller has to make sure that
      all ASCII characters are in the allowed range between 0 and
      127. The ASCII string must be NULL-terminated.
      This function can't be used for language specific sorting.

      @param  asciiStr      the 8-Bit ASCII character string to be compared.
      @return   0 - if both strings are equal
                < 0 - if this string is less than the string argument
                > 0 - if this string is greater than the string argument
    */
    sal_Int32 compareToAscii( const sal_Char* asciiStr ) const
    {
        return rtl_ustr_ascii_compare_WithLength( pData->buffer, pData->length, asciiStr );
    }

    /**
      Compares two strings with a maximum count of characters.

      The comparison is based on the numeric value of each character in
      the strings and return a value indicating their relationship.
      Since this method is optimized for performance, the ASCII character
      values are not converted in any way. The caller has to make sure that
      all ASCII characters are in the allowed range between 0 and
      127. The ASCII string must be NULL-terminated.
      This function can't be used for language specific sorting.

      @deprecated  This is a confusing overload with unexpectedly different
      semantics from the one-parameter form, so it is marked as deprecated.
      Practically all uses compare the return value against zero and can thus
      be replaced with uses of startsWith.

      @param  asciiStr          the 8-Bit ASCII character string to be compared.
      @param  maxLength         the maximum count of characters to be compared.
      @return   0 - if both strings are equal
                < 0 - if this string is less than the string argument
                > 0 - if this string is greater than the string argument
    */
    SAL_DEPRECATED(
        "replace s1.compareToAscii(s2, strlen(s2)) == 0 with s1.startsWith(s2)")
    sal_Int32 compareToAscii( const sal_Char * asciiStr, sal_Int32 maxLength ) const
    {
        return rtl_ustr_ascii_shortenedCompare_WithLength( pData->buffer, pData->length,
                                                           asciiStr, maxLength );
    }

    /**
      Compares two strings in reverse order.

      This could be useful, if normally both strings start with the same
      content. The comparison is based on the numeric value of each character
      in the strings and return a value indicating their relationship.
      Since this method is optimized for performance, the ASCII character
      values are not converted in any way. The caller has to make sure that
      all ASCII characters are in the allowed range between 0 and 127.
      The ASCII string must be NULL-terminated and must be greater than
      or equal to asciiStrLength.
      This function can't be used for language specific sorting.

      @param    asciiStr        the 8-Bit ASCII character string to be compared.
      @param    asciiStrLength  the length of the ascii string
      @return   0 - if both strings are equal
                < 0 - if this string is less than the string argument
                > 0 - if this string is greater than the string argument
    */
    sal_Int32 reverseCompareToAsciiL( const sal_Char * asciiStr, sal_Int32 asciiStrLength ) const
    {
        return rtl_ustr_asciil_reverseCompare_WithLength( pData->buffer, pData->length,
                                                          asciiStr, asciiStrLength );
    }

    /**
      Perform a comparison of two strings.

      The result is true if and only if second string
      represents the same sequence of characters as the first string.
      Since this method is optimized for performance, the ASCII character
      values are not converted in any way. The caller has to make sure that
      all ASCII characters are in the allowed range between 0 and
      127. The ASCII string must be NULL-terminated.
      This function can't be used for language specific comparison.

      @param    asciiStr        the 8-Bit ASCII character string to be compared.
      @return   true if the strings are equal;
                false, otherwise.
    */
    bool equalsAscii( const sal_Char* asciiStr ) const
    {
        return rtl_ustr_ascii_compare_WithLength( pData->buffer, pData->length,
                                                  asciiStr ) == 0;
    }

    /**
      Perform a comparison of two strings.

      The result is true if and only if second string
      represents the same sequence of characters as the first string.
      Since this method is optimized for performance, the ASCII character
      values are not converted in any way. The caller has to make sure that
      all ASCII characters are in the allowed range between 0 and
      127. The ASCII string must be NULL-terminated and must be greater than
      or equal to asciiStrLength.
      This function can't be used for language specific comparison.

      @param    asciiStr         the 8-Bit ASCII character string to be compared.
      @param    asciiStrLength   the length of the ascii string
      @return   true if the strings are equal;
                false, otherwise.
    */
    bool equalsAsciiL( const sal_Char* asciiStr, sal_Int32 asciiStrLength ) const
    {
        if ( pData->length != asciiStrLength )
            return false;

        return rtl_ustr_asciil_reverseEquals_WithLength(
                    pData->buffer, asciiStr, asciiStrLength );
    }

    /**
      Perform a ASCII lowercase comparison of two strings.

      The result is true if and only if second string
      represents the same sequence of characters as the first string,
      ignoring the case.
      Character values between 65 and 90 (ASCII A-Z) are interpreted as
      values between 97 and 122 (ASCII a-z).
      Since this method is optimized for performance, the ASCII character
      values are not converted in any way. The caller has to make sure that
      all ASCII characters are in the allowed range between 0 and
      127. The ASCII string must be NULL-terminated.
      This function can't be used for language specific comparison.

      @param    asciiStr        the 8-Bit ASCII character string to be compared.
      @return   true if the strings are equal;
                false, otherwise.
    */
    bool equalsIgnoreAsciiCaseAscii( const sal_Char * asciiStr ) const
    {
        return rtl_ustr_ascii_compareIgnoreAsciiCase_WithLength( pData->buffer, pData->length, asciiStr ) == 0;
    }

    /**
      Compares two ASCII strings ignoring case

      The comparison is based on the numeric value of each character in
      the strings and return a value indicating their relationship.
      Since this method is optimized for performance, the ASCII character
      values are not converted in any way. The caller has to make sure that
      all ASCII characters are in the allowed range between 0 and
      127. The ASCII string must be NULL-terminated.
      This function can't be used for language specific sorting.

      @param  asciiStr      the 8-Bit ASCII character string to be compared.
      @return   0 - if both strings are equal
                < 0 - if this string is less than the string argument
                > 0 - if this string is greater than the string argument

      @since LibreOffice 3.5
    */
    sal_Int32 compareToIgnoreAsciiCaseAscii( const sal_Char * asciiStr ) const
    {
        return rtl_ustr_ascii_compareIgnoreAsciiCase_WithLength( pData->buffer, pData->length, asciiStr );
    }

    /**
      Perform an ASCII lowercase comparison of two strings.

      The result is true if and only if second string
      represents the same sequence of characters as the first string,
      ignoring the case.
      Character values between 65 and 90 (ASCII A-Z) are interpreted as
      values between 97 and 122 (ASCII a-z).
      Since this method is optimized for performance, the ASCII character
      values are not converted in any way. The caller has to make sure that
      all ASCII characters are in the allowed range between 0 and 127.
      The ASCII string must be NULL-terminated and must be greater than
      or equal to asciiStrLength.
      This function can't be used for language specific comparison.

      @param    asciiStr        the 8-Bit ASCII character string to be compared.
      @param    asciiStrLength  the length of the ascii string
      @return   true if the strings are equal;
                false, otherwise.
    */
    bool equalsIgnoreAsciiCaseAsciiL( const sal_Char * asciiStr, sal_Int32 asciiStrLength ) const
    {
        if ( pData->length != asciiStrLength )
            return false;

        return rtl_ustr_ascii_compareIgnoreAsciiCase_WithLength( pData->buffer, pData->length, asciiStr ) == 0;
    }

    /**
      Match against a substring appearing in this string.

      The result is true if and only if the second string appears as a substring
      of this string, at the given position.
      Since this method is optimized for performance, the ASCII character
      values are not converted in any way. The caller has to make sure that
      all ASCII characters are in the allowed range between 0 and
      127. The ASCII string must be NULL-terminated and must be greater than or
      equal to asciiStrLength.
      This function can't be used for language specific comparison.

      @param    asciiStr    the object (substring) to be compared.
      @param    asciiStrLength the length of asciiStr.
      @param    fromIndex   the index to start the comparion from.
                            The index must be greater than or equal to 0
                            and less than or equal to the string length.
      @return   true if str match with the characters in the string
                at the given position;
                false, otherwise.
    */
    bool matchAsciiL( const sal_Char* asciiStr, sal_Int32 asciiStrLength, sal_Int32 fromIndex = 0 ) const
    {
        return rtl_ustr_ascii_shortenedCompare_WithLength( pData->buffer+fromIndex, pData->length-fromIndex,
                                                           asciiStr, asciiStrLength ) == 0;
    }

    // This overload is left undefined, to detect calls of matchAsciiL that
    // erroneously use RTL_CONSTASCII_USTRINGPARAM instead of
    // RTL_CONSTASCII_STRINGPARAM (but would lead to ambiguities on 32 bit
    // platforms):
#if SAL_TYPES_SIZEOFLONG == 8
    void matchAsciiL(char const *, sal_Int32, rtl_TextEncoding) const;
#endif

    /**
      Match against a substring appearing in this string, ignoring the case of
      ASCII letters.

      The result is true if and only if the second string appears as a substring
      of this string, at the given position.
      Character values between 65 and 90 (ASCII A-Z) are interpreted as
      values between 97 and 122 (ASCII a-z).
      Since this method is optimized for performance, the ASCII character
      values are not converted in any way. The caller has to make sure that
      all ASCII characters are in the allowed range between 0 and
      127. The ASCII string must be NULL-terminated and must be greater than or
      equal to asciiStrLength.
      This function can't be used for language specific comparison.

      @param    asciiStr        the 8-Bit ASCII character string to be compared.
      @param    asciiStrLength  the length of the ascii string
      @param    fromIndex       the index to start the comparion from.
                                The index must be greater than or equal to 0
                                and less than or equal to the string length.
      @return   true if str match with the characters in the string
                at the given position;
                false, otherwise.
    */
    bool matchIgnoreAsciiCaseAsciiL( const sal_Char* asciiStr, sal_Int32 asciiStrLength, sal_Int32 fromIndex = 0 ) const
    {
        return rtl_ustr_ascii_shortenedCompareIgnoreAsciiCase_WithLength( pData->buffer+fromIndex, pData->length-fromIndex,
                                                                          asciiStr, asciiStrLength ) == 0;
    }

    // This overload is left undefined, to detect calls of
    // matchIgnoreAsciiCaseAsciiL that erroneously use
    // RTL_CONSTASCII_USTRINGPARAM instead of RTL_CONSTASCII_STRINGPARAM (but
    // would lead to ambiguities on 32 bit platforms):
#if SAL_TYPES_SIZEOFLONG == 8
    void matchIgnoreAsciiCaseAsciiL(char const *, sal_Int32, rtl_TextEncoding)
        const;
#endif

    /**
      Check whether this string starts with a given substring.

      @param str the substring to be compared

      @param rest if non-null, and this function returns true, then assign a
      copy of the remainder of this string to *rest. Available since
      LibreOffice 4.2

      @return true if and only if the given str appears as a substring at the
      start of this string

      @since LibreOffice 4.0
    */
    bool startsWith(OUString const & str, OUString * rest = 0) const {
        bool b = match(str, 0);
        if (b && rest != 0) {
            *rest = copy(str.getLength());
        }
        return b;
    }

    /**
     @overload
     This function accepts an ASCII string literal as its argument.
     @since LibreOffice 4.0
    */
    template< typename T >
    typename libreoffice_internal::ConstCharArrayDetector< T, bool >::Type startsWith(
        T & literal, OUString * rest = 0) const
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        bool b = libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 <= pData->length
            && rtl_ustr_asciil_reverseEquals_WithLength( pData->buffer, literal,
                libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1);
        if (b && rest != 0) {
            *rest = copy(libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1);
        }
        return b;
    }

    /**
      Check whether this string starts with a given string, ignoring the case of
      ASCII letters.

      Character values between 65 and 90 (ASCII A-Z) are interpreted as
      values between 97 and 122 (ASCII a-z).
      This function can't be used for language specific comparison.

      @param str the substring to be compared

      @param rest if non-null, and this function returns true, then assign a
      copy of the remainder of this string to *rest. Available since
      LibreOffice 4.2

      @return true if and only if the given str appears as a substring at the
      start of this string, ignoring the case of ASCII letters ("A"--"Z" and
      "a"--"z")

      @since LibreOffice 4.0
    */
    bool startsWithIgnoreAsciiCase(OUString const & str, OUString * rest = 0)
        const
    {
        bool b = matchIgnoreAsciiCase(str, 0);
        if (b && rest != 0) {
            *rest = copy(str.getLength());
        }
        return b;
    }

    /**
     @overload
     This function accepts an ASCII string literal as its argument.
     @since LibreOffice 4.0
    */
    template< typename T >
    typename libreoffice_internal::ConstCharArrayDetector< T, bool >::Type
    startsWithIgnoreAsciiCase(T & literal, OUString * rest = 0) const
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        bool b = (rtl_ustr_ascii_compareIgnoreAsciiCase_WithLengths(
                    pData->buffer,
                    libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1, literal,
                    libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1)
                == 0);
        if (b && rest != 0) {
            *rest = copy(libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1);
        }
        return b;
    }

    /**
      Check whether this string ends with a given substring.

      @param str the substring to be compared

      @param rest if non-null, and this function returns true, then assign a
      copy of the remainder of this string to *rest. Available since
      LibreOffice 4.2

      @return true if and only if the given str appears as a substring at the
      end of this string

      @since LibreOffice 3.6
    */
    bool endsWith(OUString const & str, OUString * rest = 0) const {
        bool b = str.getLength() <= getLength()
            && match(str, getLength() - str.getLength());
        if (b && rest != 0) {
            *rest = copy(0, getLength() - str.getLength());
        }
        return b;
    }

    /**
     @overload
     This function accepts an ASCII string literal as its argument.
     @since LibreOffice 3.6
    */
    template< typename T >
    typename libreoffice_internal::ConstCharArrayDetector< T, bool >::Type
    endsWith(T & literal, OUString * rest = 0) const
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        bool b = libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 <= pData->length
            && rtl_ustr_asciil_reverseEquals_WithLength(
                pData->buffer + pData->length - ( libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 ), literal,
                libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1);
        if (b && rest != 0) {
            *rest = copy(
                0,
                (getLength()
                 - (libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1)));
        }
        return b;
    }

    /**
      Check whether this string ends with a given ASCII string.

      @param asciiStr a sequence of at least asciiStrLength ASCII characters
          (bytes in the range 0x00--0x7F)
      @param asciiStrLength the length of asciiStr; must be non-negative
      @return true if this string ends with asciiStr; otherwise, false is
      returned

      @since UDK 3.2.7
     */
    inline bool endsWithAsciiL(char const * asciiStr, sal_Int32 asciiStrLength)
        const
    {
        return asciiStrLength <= pData->length
            && rtl_ustr_asciil_reverseEquals_WithLength(
                pData->buffer + pData->length - asciiStrLength, asciiStr,
                asciiStrLength);
    }

    /**
      Check whether this string ends with a given string, ignoring the case of
      ASCII letters.

      Character values between 65 and 90 (ASCII A-Z) are interpreted as
      values between 97 and 122 (ASCII a-z).
      This function can't be used for language specific comparison.

      @param str the substring to be compared

      @param rest if non-null, and this function returns true, then assign a
      copy of the remainder of this string to *rest. Available since
      LibreOffice 4.2

      @return true if and only if the given str appears as a substring at the
      end of this string, ignoring the case of ASCII letters ("A"--"Z" and
      "a"--"z")

      @since LibreOffice 3.6
    */
    bool endsWithIgnoreAsciiCase(OUString const & str, OUString * rest = 0) const
    {
        bool b =  str.getLength() <= getLength()
            && matchIgnoreAsciiCase(str, getLength() - str.getLength());
        if (b && rest != 0) {
            *rest = copy(0, getLength() - str.getLength());
        }
        return b;
    }

    /**
     @overload
     This function accepts an ASCII string literal as its argument.
     @since LibreOffice 3.6
    */
    template< typename T >
    typename libreoffice_internal::ConstCharArrayDetector< T, bool >::Type
    endsWithIgnoreAsciiCase(T & literal, OUString * rest = 0) const
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        bool b = libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 <= pData->length
            && (rtl_ustr_ascii_compareIgnoreAsciiCase_WithLengths(
                    pData->buffer + pData->length - ( libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 ),
                    libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1, literal,
                    libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1)
                == 0);
        if (b && rest != 0) {
            *rest = copy(
                0,
                (getLength()
                 - (libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1)));
        }
        return b;
    }

    /**
      Check whether this string ends with a given ASCII string, ignoring the
      case of ASCII letters.

      @param asciiStr a sequence of at least asciiStrLength ASCII characters
          (bytes in the range 0x00--0x7F)
      @param asciiStrLength the length of asciiStr; must be non-negative
      @return true if this string ends with asciiStr, ignoring the case of ASCII
      letters ("A"--"Z" and "a"--"z"); otherwise, false is returned
     */
    inline bool endsWithIgnoreAsciiCaseAsciiL(
        char const * asciiStr, sal_Int32 asciiStrLength) const
    {
        return asciiStrLength <= pData->length
            && (rtl_ustr_ascii_compareIgnoreAsciiCase_WithLengths(
                    pData->buffer + pData->length - asciiStrLength,
                    asciiStrLength, asciiStr, asciiStrLength)
                == 0);
    }

    friend bool     operator == ( const OUString& rStr1,    const OUString& rStr2 )
                        { return rStr1.equals(rStr2); }
    friend bool     operator == ( const OUString& rStr1,    const sal_Unicode * pStr2 )
                        { return rStr1.compareTo( pStr2 ) == 0; }
    friend bool     operator == ( const sal_Unicode * pStr1,    const OUString& rStr2 )
                        { return OUString( pStr1 ).compareTo( rStr2 ) == 0; }

    friend bool     operator != ( const OUString& rStr1,        const OUString& rStr2 )
                        { return !(operator == ( rStr1, rStr2 )); }
    friend bool     operator != ( const OUString& rStr1,    const sal_Unicode * pStr2 )
                        { return !(operator == ( rStr1, pStr2 )); }
    friend bool     operator != ( const sal_Unicode * pStr1,    const OUString& rStr2 )
                        { return !(operator == ( pStr1, rStr2 )); }

    friend bool     operator <  ( const OUString& rStr1,    const OUString& rStr2 )
                        { return rStr1.compareTo( rStr2 ) < 0; }
    friend bool     operator >  ( const OUString& rStr1,    const OUString& rStr2 )
                        { return rStr1.compareTo( rStr2 ) > 0; }
    friend bool     operator <= ( const OUString& rStr1,    const OUString& rStr2 )
                        { return rStr1.compareTo( rStr2 ) <= 0; }
    friend bool     operator >= ( const OUString& rStr1,    const OUString& rStr2 )
                        { return rStr1.compareTo( rStr2 ) >= 0; }

    /**
     * Compare string to an ASCII string literal.
     *
     * This operator is equal to calling equalsAsciiL().
     *
     * @since LibreOffice 3.6
     */
    template< typename T >
    friend inline typename libreoffice_internal::ConstCharArrayDetector< T, bool >::Type operator==( const OUString& string, T& literal )
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        return string.equalsAsciiL( literal, libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 );
    }
    /**
     * Compare string to an ASCII string literal.
     *
     * This operator is equal to calling equalsAsciiL().
     *
     * @since LibreOffice 3.6
     */
    template< typename T >
    friend inline typename libreoffice_internal::ConstCharArrayDetector< T, bool >::Type operator==( T& literal, const OUString& string )
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        return string.equalsAsciiL( literal, libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 );
    }
    /**
     * Compare string to an ASCII string literal.
     *
     * This operator is equal to calling !equalsAsciiL().
     *
     * @since LibreOffice 3.6
     */
    template< typename T >
    friend inline typename libreoffice_internal::ConstCharArrayDetector< T, bool >::Type operator!=( const OUString& string, T& literal )
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        return !string.equalsAsciiL( literal, libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 );
    }
    /**
     * Compare string to an ASCII string literal.
     *
     * This operator is equal to calling !equalsAsciiL().
     *
     * @since LibreOffice 3.6
     */
    template< typename T >
    friend inline typename libreoffice_internal::ConstCharArrayDetector< T, bool >::Type operator!=( T& literal, const OUString& string )
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        return !string.equalsAsciiL( literal, libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1 );
    }

#if defined LIBO_INTERNAL_ONLY
    /// @cond INTERNAL

    /* Comparison between OUString and OUStringLiteral.

       @since LibreOffice 5.0
    */

    friend bool operator ==(OUString const & lhs, OUStringLiteral const & rhs) {
        return lhs.equalsAsciiL(rhs.data, rhs.size);
    }

    friend bool operator !=(OUString const & lhs, OUStringLiteral const & rhs) {
        return !lhs.equalsAsciiL(rhs.data, rhs.size);
    }

    friend bool operator <(OUString const & lhs, OUStringLiteral const & rhs) {
        return
            (rtl_ustr_ascii_compare_WithLength(
                lhs.pData->buffer, lhs.pData->length, rhs.data))
            < 0;
    }

    friend bool operator <=(OUString const & lhs, OUStringLiteral const & rhs) {
        return
            (rtl_ustr_ascii_compare_WithLength(
                lhs.pData->buffer, lhs.pData->length, rhs.data))
            <= 0;
    }

    friend bool operator >(OUString const & lhs, OUStringLiteral const & rhs) {
        return
            (rtl_ustr_ascii_compare_WithLength(
                lhs.pData->buffer, lhs.pData->length, rhs.data))
            > 0;
    }

    friend bool operator >=(OUString const & lhs, OUStringLiteral const & rhs) {
        return
            (rtl_ustr_ascii_compare_WithLength(
                lhs.pData->buffer, lhs.pData->length, rhs.data))
            >= 0;
    }

    friend bool operator ==(OUStringLiteral const & lhs, OUString const & rhs) {
        return rhs.equalsAsciiL(lhs.data, lhs.size);
    }

    friend bool operator !=(OUStringLiteral const & lhs, OUString const & rhs) {
        return !rhs.equalsAsciiL(lhs.data, lhs.size);
    }

    friend bool operator <(OUStringLiteral const & lhs, OUString const & rhs) {
        return
            (rtl_ustr_ascii_compare_WithLength(
                rhs.pData->buffer, rhs.pData->length, lhs.data))
            >= 0;
    }

    friend bool operator <=(OUStringLiteral const & lhs, OUString const & rhs) {
        return
            (rtl_ustr_ascii_compare_WithLength(
                rhs.pData->buffer, rhs.pData->length, lhs.data))
            > 0;
    }

    friend bool operator >(OUStringLiteral const & lhs, OUString const & rhs) {
        return
            (rtl_ustr_ascii_compare_WithLength(
                rhs.pData->buffer, rhs.pData->length, lhs.data))
            <= 0;
    }

    friend bool operator >=(OUStringLiteral const & lhs, OUString const & rhs) {
        return
            (rtl_ustr_ascii_compare_WithLength(
                rhs.pData->buffer, rhs.pData->length, lhs.data))
            < 0;
    }

    /// @endcond
#endif

    /**
      Returns a hashcode for this string.

      @return   a hash code value for this object.

      @see rtl::OUStringHash for convenient use of std::unordered_map
    */
    sal_Int32 hashCode() const
    {
        return rtl_ustr_hashCode_WithLength( pData->buffer, pData->length );
    }

    /**
      Returns the index within this string of the first occurrence of the
      specified character, starting the search at the specified index.

      @param    ch          character to be located.
      @param    fromIndex   the index to start the search from.
                            The index must be greater than or equal to 0
                            and less than or equal to the string length.
      @return   the index of the first occurrence of the character in the
                character sequence represented by this string that is
                greater than or equal to fromIndex, or
                -1 if the character does not occur.
    */
    sal_Int32 indexOf( sal_Unicode ch, sal_Int32 fromIndex = 0 ) const
    {
        sal_Int32 ret = rtl_ustr_indexOfChar_WithLength( pData->buffer+fromIndex, pData->length-fromIndex, ch );
        return (ret < 0 ? ret : ret+fromIndex);
    }

    /**
      Returns the index within this string of the last occurrence of the
      specified character, searching backward starting at the end.

      @param    ch          character to be located.
      @return   the index of the last occurrence of the character in the
                character sequence represented by this string, or
                -1 if the character does not occur.
    */
    sal_Int32 lastIndexOf( sal_Unicode ch ) const
    {
        return rtl_ustr_lastIndexOfChar_WithLength( pData->buffer, pData->length, ch );
    }

    /**
      Returns the index within this string of the last occurrence of the
      specified character, searching backward starting before the specified
      index.

      @param    ch          character to be located.
      @param    fromIndex   the index before which to start the search.
      @return   the index of the last occurrence of the character in the
                character sequence represented by this string that
                is less than fromIndex, or -1
                if the character does not occur before that point.
    */
    sal_Int32 lastIndexOf( sal_Unicode ch, sal_Int32 fromIndex ) const
    {
        return rtl_ustr_lastIndexOfChar_WithLength( pData->buffer, fromIndex, ch );
    }

    /**
      Returns the index within this string of the first occurrence of the
      specified substring, starting at the specified index.

      If str doesn't include any character, always -1 is
      returned. This is also the case, if both strings are empty.

      @param    str         the substring to search for.
      @param    fromIndex   the index to start the search from.
      @return   If the string argument occurs one or more times as a substring
                within this string at the starting index, then the index
                of the first character of the first such substring is
                returned. If it does not occur as a substring starting
                at fromIndex or beyond, -1 is returned.
    */
    sal_Int32 indexOf( const OUString & str, sal_Int32 fromIndex = 0 ) const
    {
        sal_Int32 ret = rtl_ustr_indexOfStr_WithLength( pData->buffer+fromIndex, pData->length-fromIndex,
                                                        str.pData->buffer, str.pData->length );
        return (ret < 0 ? ret : ret+fromIndex);
    }

    /**
     @overload
     This function accepts an ASCII string literal as its argument.
     @since LibreOffice 3.6
    */
    template< typename T >
    typename libreoffice_internal::ConstCharArrayDetector< T, sal_Int32 >::Type indexOf( T& literal, sal_Int32 fromIndex = 0 ) const
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        sal_Int32 ret = rtl_ustr_indexOfAscii_WithLength(
            pData->buffer + fromIndex, pData->length - fromIndex, literal,
            libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1);
        return ret < 0 ? ret : ret + fromIndex;
    }

    /**
       Returns the index within this string of the first occurrence of the
       specified ASCII substring, starting at the specified index.

       @param str
       the substring to be searched for.  Need not be null-terminated, but must
       be at least as long as the specified len.  Must only contain characters
       in the ASCII range 0x00--7F.

       @param len
       the length of the substring; must be non-negative.

       @param fromIndex
       the index to start the search from.  Must be in the range from zero to
       the length of this string, inclusive.

       @return
       the index (starting at 0) of the first character of the first occurrence
       of the substring within this string starting at the given fromIndex, or
       -1 if the substring does not occur.  If len is zero, -1 is returned.

       @since UDK 3.2.7
    */
    sal_Int32 indexOfAsciiL(
        char const * str, sal_Int32 len, sal_Int32 fromIndex = 0) const
    {
        sal_Int32 ret = rtl_ustr_indexOfAscii_WithLength(
            pData->buffer + fromIndex, pData->length - fromIndex, str, len);
        return ret < 0 ? ret : ret + fromIndex;
    }

    // This overload is left undefined, to detect calls of indexOfAsciiL that
    // erroneously use RTL_CONSTASCII_USTRINGPARAM instead of
    // RTL_CONSTASCII_STRINGPARAM (but would lead to ambiguities on 32 bit
    // platforms):
#if SAL_TYPES_SIZEOFLONG == 8
    void indexOfAsciiL(char const *, sal_Int32 len, rtl_TextEncoding) const;
#endif

    /**
      Returns the index within this string of the last occurrence of
      the specified substring, searching backward starting at the end.

      The returned index indicates the starting index of the substring
      in this string.
      If str doesn't include any character, always -1 is
      returned. This is also the case, if both strings are empty.

      @param    str         the substring to search for.
      @return   If the string argument occurs one or more times as a substring
                within this string, then the index of the first character of
                the last such substring is returned. If it does not occur as
                a substring, -1 is returned.
    */
    sal_Int32 lastIndexOf( const OUString & str ) const
    {
        return rtl_ustr_lastIndexOfStr_WithLength( pData->buffer, pData->length,
                                                   str.pData->buffer, str.pData->length );
    }

    /**
      Returns the index within this string of the last occurrence of
      the specified substring, searching backward starting before the specified
      index.

      The returned index indicates the starting index of the substring
      in this string.
      If str doesn't include any character, always -1 is
      returned. This is also the case, if both strings are empty.

      @param    str         the substring to search for.
      @param    fromIndex   the index before which to start the search.
      @return   If the string argument occurs one or more times as a substring
                within this string before the starting index, then the index
                of the first character of the last such substring is
                returned. Otherwise, -1 is returned.
    */
    sal_Int32 lastIndexOf( const OUString & str, sal_Int32 fromIndex ) const
    {
        return rtl_ustr_lastIndexOfStr_WithLength( pData->buffer, fromIndex,
                                                   str.pData->buffer, str.pData->length );
    }

    /**
     @overload
     This function accepts an ASCII string literal as its argument.
     @since LibreOffice 3.6
    */
    template< typename T >
    typename libreoffice_internal::ConstCharArrayDetector< T, sal_Int32 >::Type lastIndexOf( T& literal ) const
    {
        assert( strlen( literal ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        return rtl_ustr_lastIndexOfAscii_WithLength(
            pData->buffer, pData->length, literal, libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1);
    }

    /**
       Returns the index within this string of the last occurrence of the
       specified ASCII substring.

       @param str
       the substring to be searched for.  Need not be null-terminated, but must
       be at least as long as the specified len.  Must only contain characters
       in the ASCII range 0x00--7F.

       @param len
       the length of the substring; must be non-negative.

       @return
       the index (starting at 0) of the first character of the last occurrence
       of the substring within this string, or -1 if the substring does not
       occur.  If len is zero, -1 is returned.

       @since UDK 3.2.7
    */
    sal_Int32 lastIndexOfAsciiL(char const * str, sal_Int32 len) const
    {
        return rtl_ustr_lastIndexOfAscii_WithLength(
            pData->buffer, pData->length, str, len);
    }

    /**
      Returns a new string that is a substring of this string.

      The substring begins at the specified beginIndex. If
      beginIndex is negative or be greater than the length of
      this string, behaviour is undefined.

      @param     beginIndex   the beginning index, inclusive.
      @return    the specified substring.
    */
    SAL_WARN_UNUSED_RESULT OUString copy( sal_Int32 beginIndex ) const
    {
        rtl_uString *pNew = 0;
        rtl_uString_newFromSubString( &pNew, pData, beginIndex, getLength() - beginIndex );
        return OUString( pNew, SAL_NO_ACQUIRE );
    }

    /**
      Returns a new string that is a substring of this string.

      The substring begins at the specified beginIndex and contains count
      characters.  If either beginIndex or count are negative,
      or beginIndex + count are greater than the length of this string
      then behaviour is undefined.

      @param     beginIndex   the beginning index, inclusive.
      @param     count        the number of characters.
      @return    the specified substring.
    */
    SAL_WARN_UNUSED_RESULT OUString copy( sal_Int32 beginIndex, sal_Int32 count ) const
    {
        rtl_uString *pNew = 0;
        rtl_uString_newFromSubString( &pNew, pData, beginIndex, count );
        return OUString( pNew, SAL_NO_ACQUIRE );
    }

    /**
      Concatenates the specified string to the end of this string.

      @param    str   the string that is concatenated to the end
                      of this string.
      @return   a string that represents the concatenation of this string
                followed by the string argument.
    */
    SAL_WARN_UNUSED_RESULT OUString concat( const OUString & str ) const
    {
        rtl_uString* pNew = 0;
        rtl_uString_newConcat( &pNew, pData, str.pData );
        return OUString( pNew, SAL_NO_ACQUIRE );
    }

#ifndef LIBO_INTERNAL_ONLY // "RTL_FAST_STRING"
    friend OUString operator+( const OUString& rStr1, const OUString& rStr2  )
    {
        return rStr1.concat( rStr2 );
    }
#endif

    /**
      Returns a new string resulting from replacing n = count characters
      from position index in this string with newStr.

      @param  index   the replacing index in str.
                      The index must be greater than or equal to 0 and
                      less than or equal to the length of the string.
      @param  count   the count of characters that will be replaced
                      The count must be greater than or equal to 0 and
                      less than or equal to the length of the string minus index.
      @param  newStr  the new substring.
      @return the new string.
    */
    SAL_WARN_UNUSED_RESULT OUString replaceAt( sal_Int32 index, sal_Int32 count, const OUString& newStr ) const
    {
        rtl_uString* pNew = 0;
        rtl_uString_newReplaceStrAt( &pNew, pData, index, count, newStr.pData );
        return OUString( pNew, SAL_NO_ACQUIRE );
    }

    /**
      Returns a new string resulting from replacing all occurrences of
      oldChar in this string with newChar.

      If the character oldChar does not occur in the character sequence
      represented by this object, then the string is assigned with
      str.

      @param    oldChar     the old character.
      @param    newChar     the new character.
      @return   a string derived from this string by replacing every
                occurrence of oldChar with newChar.
    */
    SAL_WARN_UNUSED_RESULT OUString replace( sal_Unicode oldChar, sal_Unicode newChar ) const
    {
        rtl_uString* pNew = 0;
        rtl_uString_newReplace( &pNew, pData, oldChar, newChar );
        return OUString( pNew, SAL_NO_ACQUIRE );
    }

    /**
      Returns a new string resulting from replacing the first occurrence of a
      given substring with another substring.

      @param from  the substring to be replaced

      @param to  the replacing substring

      @param[in,out] index  pointer to a start index; if the pointer is
      non-null: upon entry to the function, its value is the index into this
      string at which to start searching for the \p from substring, the value
      must be non-negative and not greater than this string's length; upon exiting
      the function its value is the index into this string at which the
      replacement took place or -1 if no replacement took place; if the pointer
      is null, searching always starts at index 0

      @since LibreOffice 3.6
    */
    SAL_WARN_UNUSED_RESULT OUString replaceFirst(
        OUString const & from, OUString const & to, sal_Int32 * index = 0) const
    {
        rtl_uString * s = 0;
        sal_Int32 i = 0;
        rtl_uString_newReplaceFirst(
            &s, pData, from.pData, to.pData, index == 0 ? &i : index);
        return OUString(s, SAL_NO_ACQUIRE);
    }

    /**
      Returns a new string resulting from replacing the first occurrence of a
      given substring with another substring.

      @param from  ASCII string literal, the substring to be replaced

      @param to  the replacing substring

      @param[in,out] index  pointer to a start index; if the pointer is
      non-null: upon entry to the function, its value is the index into the this
      string at which to start searching for the \p from substring, the value
      must be non-negative and not greater than this string's length; upon exiting
      the function its value is the index into this string at which the
      replacement took place or -1 if no replacement took place; if the pointer
      is null, searching always starts at index 0

      @since LibreOffice 3.6
    */
    template< typename T >
    SAL_WARN_UNUSED_RESULT typename libreoffice_internal::ConstCharArrayDetector< T, OUString >::Type replaceFirst( T& from, OUString const & to,
                                                                                                        sal_Int32 * index = 0) const
    {
        rtl_uString * s = 0;
        sal_Int32 i = 0;
        assert( strlen( from ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        rtl_uString_newReplaceFirstAsciiL(
            &s, pData, from, libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1, to.pData, index == 0 ? &i : index);
        return OUString(s, SAL_NO_ACQUIRE);
    }

    /**
      Returns a new string resulting from replacing the first occurrence of a
      given substring with another substring.

      @param from  ASCII string literal, the substring to be replaced

      @param to  ASCII string literal, the substring to be replaced

      @param[in,out] index  pointer to a start index; if the pointer is
      non-null: upon entry to the function, its value is the index into the this
      string at which to start searching for the \p from substring, the value
      must be non-negative and not greater than this string's length; upon exiting
      the function its value is the index into this string at which the
      replacement took place or -1 if no replacement took place; if the pointer
      is null, searching always starts at index 0

      @since LibreOffice 3.6
    */
    template< typename T1, typename T2 >
        SAL_WARN_UNUSED_RESULT typename libreoffice_internal::ConstCharArrayDetector< T1, typename libreoffice_internal::ConstCharArrayDetector< T2, OUString >::Type >::Type
        replaceFirst( T1& from, T2& to, sal_Int32 * index = 0) const
    {
        rtl_uString * s = 0;
        sal_Int32 i = 0;
        assert( strlen( from ) == libreoffice_internal::ConstCharArrayDetector< T1 >::size - 1 );
        assert( strlen( to ) == libreoffice_internal::ConstCharArrayDetector< T2 >::size - 1 );
        rtl_uString_newReplaceFirstAsciiLAsciiL(
            &s, pData, from, libreoffice_internal::ConstCharArrayDetector< T1, void >::size - 1, to,
            libreoffice_internal::ConstCharArrayDetector< T2, void >::size - 1, index == 0 ? &i : index);
        return OUString(s, SAL_NO_ACQUIRE);
    }

    /**
      Returns a new string resulting from replacing all occurrences of a given
      substring with another substring.

      Replacing subsequent occurrences picks up only after a given replacement.
      That is, replacing from "xa" to "xx" in "xaa" results in "xxa", not "xxx".

      @param from  the substring to be replaced

      @param to  the replacing substring

      @param fromIndex  the position in the string where we will begin searching

      @since LibreOffice 4.0
    */
    SAL_WARN_UNUSED_RESULT OUString replaceAll(
        OUString const & from, OUString const & to, sal_Int32 fromIndex = 0) const
    {
        rtl_uString * s = 0;
        rtl_uString_newReplaceAllFromIndex(&s, pData, from.pData, to.pData, fromIndex);
        return OUString(s, SAL_NO_ACQUIRE);
    }

    /**
      Returns a new string resulting from replacing all occurrences of a given
      substring with another substring.

      Replacing subsequent occurrences picks up only after a given replacement.
      That is, replacing from "xa" to "xx" in "xaa" results in "xxa", not "xxx".

      @param from ASCII string literal, the substring to be replaced

      @param to  the replacing substring

      @since LibreOffice 3.6
    */
    template< typename T >
    SAL_WARN_UNUSED_RESULT typename libreoffice_internal::ConstCharArrayDetector< T, OUString >::Type replaceAll( T& from, OUString const & to) const
    {
        rtl_uString * s = 0;
        assert( strlen( from ) == libreoffice_internal::ConstCharArrayDetector< T >::size - 1 );
        rtl_uString_newReplaceAllAsciiL(&s, pData, from, libreoffice_internal::ConstCharArrayDetector< T, void >::size - 1, to.pData);
        return OUString(s, SAL_NO_ACQUIRE);
    }

    /**
      Returns a new string resulting from replacing all occurrences of a given
      substring with another substring.

      Replacing subsequent occurrences picks up only after a given replacement.
      That is, replacing from "xa" to "xx" in "xaa" results in "xxa", not "xxx".

      @param from  ASCII string literal, the substring to be replaced

      @param to  ASCII string literal, the substring to be replaced

      @since LibreOffice 3.6
    */
    template< typename T1, typename T2 >
    SAL_WARN_UNUSED_RESULT typename libreoffice_internal::ConstCharArrayDetector< T1, typename libreoffice_internal::ConstCharArrayDetector< T2, OUString >::Type >::Type
        replaceAll( T1& from, T2& to ) const
    {
        rtl_uString * s = 0;
        assert( strlen( from ) == libreoffice_internal::ConstCharArrayDetector< T1 >::size - 1 );
        assert( strlen( to ) == libreoffice_internal::ConstCharArrayDetector< T2 >::size - 1 );
        rtl_uString_newReplaceAllAsciiLAsciiL(
            &s, pData, from, libreoffice_internal::ConstCharArrayDetector< T1, void >::size - 1,
            to, libreoffice_internal::ConstCharArrayDetector< T2, void >::size - 1);
        return OUString(s, SAL_NO_ACQUIRE);
    }

    /**
      Converts from this string all ASCII uppercase characters (65-90)
      to ASCII lowercase characters (97-122).

      This function can't be used for language specific conversion.
      If the string doesn't contain characters which must be converted,
      then the new string is assigned with str.

      @return   the string, converted to ASCII lowercase.
    */
    SAL_WARN_UNUSED_RESULT OUString toAsciiLowerCase() const
    {
        rtl_uString* pNew = 0;
        rtl_uString_newToAsciiLowerCase( &pNew, pData );
        return OUString( pNew, SAL_NO_ACQUIRE );
    }

    /**
      Converts from this string all ASCII lowercase characters (97-122)
      to ASCII uppercase characters (65-90).

      This function can't be used for language specific conversion.
      If the string doesn't contain characters which must be converted,
      then the new string is assigned with str.

      @return   the string, converted to ASCII uppercase.
    */
    SAL_WARN_UNUSED_RESULT OUString toAsciiUpperCase() const
    {
        rtl_uString* pNew = 0;
        rtl_uString_newToAsciiUpperCase( &pNew, pData );
        return OUString( pNew, SAL_NO_ACQUIRE );
    }

    /**
      Returns a new string resulting from removing white space from both ends
      of the string.

      All characters that have codes less than or equal to
      32 (the space character) are considered to be white space.
      If the string doesn't contain white spaces at both ends,
      then the new string is assigned with str.

      @return   the string, with white space removed from the front and end.
    */
    SAL_WARN_UNUSED_RESULT OUString trim() const
    {
        rtl_uString* pNew = 0;
        rtl_uString_newTrim( &pNew, pData );
        return OUString( pNew, SAL_NO_ACQUIRE );
    }

    /**
      Returns a token in the string.

      Example:
        sal_Int32 nIndex = 0;
        do
        {
            ...
            OUString aToken = aStr.getToken( 0, ';', nIndex );
            ...
        }
        while ( nIndex >= 0 );

      @param    token       the number of the token to return
      @param    cTok        the character which separate the tokens.
      @param    index       the position at which the token is searched in the
                            string.
                            The index must not be greater than the length of the
                            string.
                            This param is set to the position of the
                            next token or to -1, if it is the last token.
      @return   the token; if either token or index is negative, an empty token
                is returned (and index is set to -1)
    */
    OUString getToken( sal_Int32 token, sal_Unicode cTok, sal_Int32& index ) const
    {
        rtl_uString * pNew = 0;
        index = rtl_uString_getToken( &pNew, pData, token, cTok, index );
        return OUString( pNew, SAL_NO_ACQUIRE );
    }

    /**
      Returns a token from the string.

      The same as getToken(sal_Int32, sal_Unicode, sal_Int32 &), but always
      passing in 0 as the start index in the third argument.

      @param count  the number of the token to return, starting with 0
      @param separator  the character which separates the tokens

      @return  the given token, or an empty string

      @since LibreOffice 3.6
     */
    OUString getToken(sal_Int32 count, sal_Unicode separator) const {
        sal_Int32 n = 0;
        return getToken(count, separator, n);
    }

    /**
      Returns the Boolean value from this string.

      This function can't be used for language specific conversion.

      @return   true, if the string is 1 or "True" in any ASCII case.
                false in any other case.
    */
    bool toBoolean() const
    {
        return rtl_ustr_toBoolean( pData->buffer );
    }

    /**
      Returns the first character from this string.

      @return   the first character from this string or 0, if this string
                is emptry.
    */
    sal_Unicode toChar() const
    {
        return pData->buffer[0];
    }

    /**
      Returns the int32 value from this string.

      This function can't be used for language specific conversion.

      @param    radix       the radix (between 2 and 36)
      @return   the int32 represented from this string.
                0 if this string represents no number or one of too large
                magnitude.
    */
    sal_Int32 toInt32( sal_Int16 radix = 10 ) const
    {
        return rtl_ustr_toInt32( pData->buffer, radix );
    }

    /**
      Returns the uint32 value from this string.

      This function can't be used for language specific conversion.

      @param    radix       the radix (between 2 and 36)
      @return   the uint32 represented from this string.
                0 if this string represents no number or one of too large
                magnitude.

      @since LibreOffice 4.2
    */
    sal_uInt32 toUInt32( sal_Int16 radix = 10 ) const
    {
        return rtl_ustr_toUInt32( pData->buffer, radix );
    }

    /**
      Returns the int64 value from this string.

      This function can't be used for language specific conversion.

      @param    radix       the radix (between 2 and 36)
      @return   the int64 represented from this string.
                0 if this string represents no number or one of too large
                magnitude.
    */
    sal_Int64 toInt64( sal_Int16 radix = 10 ) const
    {
        return rtl_ustr_toInt64( pData->buffer, radix );
    }

    /**
      Returns the uint64 value from this string.

      This function can't be used for language specific conversion.

      @param    radix       the radix (between 2 and 36)
      @return   the uint64 represented from this string.
                0 if this string represents no number or one of too large
                magnitude.

      @since LibreOffice 4.1
    */
    sal_uInt64 toUInt64( sal_Int16 radix = 10 ) const
    {
        return rtl_ustr_toUInt64( pData->buffer, radix );
    }

    /**
      Returns the float value from this string.

      This function can't be used for language specific conversion.

      @return   the float represented from this string.
                0.0 if this string represents no number.
    */
    float toFloat() const
    {
        return rtl_ustr_toFloat( pData->buffer );
    }

    /**
      Returns the double value from this string.

      This function can't be used for language specific conversion.

      @return   the double represented from this string.
                0.0 if this string represents no number.
    */
    double toDouble() const
    {
        return rtl_ustr_toDouble( pData->buffer );
    }


    /**
       Return a canonical representation for a string.

       A pool of strings, initially empty is maintained privately
       by the string class. On invocation, if present in the pool
       the original string will be returned. Otherwise this string,
       or a copy thereof will be added to the pool and returned.

       @return
       a version of the string from the pool.

       @exception std::bad_alloc is thrown if an out-of-memory condition occurs

       @since UDK 3.2.7
    */
    OUString intern() const
    {
        rtl_uString * pNew = 0;
        rtl_uString_intern( &pNew, pData );
        if (pNew == 0) {
            throw std::bad_alloc();
        }
        return OUString( pNew, SAL_NO_ACQUIRE );
    }

    /**
       Return a canonical representation for a converted string.

       A pool of strings, initially empty is maintained privately
       by the string class. On invocation, if present in the pool
       the original string will be returned. Otherwise this string,
       or a copy thereof will be added to the pool and returned.

       @param    value           a 8-Bit character array.
       @param    length          the number of character which should be converted.
                                 The 8-Bit character array length must be
                                 greater than or equal to this value.
       @param    encoding        the text encoding from which the 8-Bit character
                                 sequence should be converted.
       @param    convertFlags    flags which controls the conversion.
                                 see RTL_TEXTTOUNICODE_FLAGS_...
       @param    pInfo           pointer to return conversion status or NULL.

       @return
       a version of the converted string from the pool.

       @exception std::bad_alloc is thrown if an out-of-memory condition occurs

       @since UDK 3.2.7
    */
    static OUString intern( const sal_Char * value, sal_Int32 length,
                            rtl_TextEncoding encoding,
                            sal_uInt32 convertFlags = OSTRING_TO_OUSTRING_CVTFLAGS,
                            sal_uInt32 *pInfo = NULL )
    {
        rtl_uString * pNew = 0;
        rtl_uString_internConvert( &pNew, value, length, encoding,
                                   convertFlags, pInfo );
        if (pNew == 0) {
            throw std::bad_alloc();
        }
        return OUString( pNew, SAL_NO_ACQUIRE );
    }

    /**
      Converts to an OString, signalling failure.

      @param pTarget
      An out parameter receiving the converted OString.  Must not be null; the
      contents are not modified if conversion fails (convertToOString returns
      false).

      @param nEncoding
      The text encoding to convert into.  Must be an octet encoding (i.e.,
      rtl_isOctetTextEncoding(nEncoding) must return true).

      @param nFlags
      A combination of RTL_UNICODETOTEXT_FLAGS that detail how to do the
      conversion (see rtl_convertUnicodeToText).  RTL_UNICODETOTEXT_FLAGS_FLUSH
      need not be included, it is implicitly assumed.  Typical uses are either
      RTL_UNICODETOTEXT_FLAGS_UNDEFINED_ERROR |
      RTL_UNICODETOTEXT_FLAGS_INVALID_ERROR (fail if a Unicode character cannot
      be converted to the target nEncoding) or OUSTRING_TO_OSTRING_CVTFLAGS
      (make a best efforts conversion).

      @return
      True if the conversion succeeded, false otherwise.
     */
    inline bool convertToString(OString * pTarget, rtl_TextEncoding nEncoding,
                                sal_uInt32 nFlags) const
    {
        return rtl_convertUStringToString(&pTarget->pData, pData->buffer,
                                            pData->length, nEncoding, nFlags);
    }

    /** Iterate through this string based on code points instead of UTF-16 code
        units.

        See Chapter 3 of The Unicode Standard 5.0 (Addison--Wesley, 2006) for
        definitions of the various terms used in this description.

        This string is interpreted as a sequence of zero or more UTF-16 code
        units.  For each index into this sequence (from zero to one less than
        the length of the sequence, inclusive), a code point represented
        starting at the given index is computed as follows:

        - If the UTF-16 code unit addressed by the index constitutes a
        well-formed UTF-16 code unit sequence, the computed code point is the
        scalar value encoded by that UTF-16 code unit sequence.

        - Otherwise, if the index is at least two UTF-16 code units away from
        the end of the sequence, and the sequence of two UTF-16 code units
        addressed by the index constitutes a well-formed UTF-16 code unit
        sequence, the computed code point is the scalar value encoded by that
        UTF-16 code unit sequence.

        - Otherwise, the computed code point is the UTF-16 code unit addressed
        by the index.  (This last case catches unmatched surrogates as well as
        indices pointing into the middle of surrogate pairs.)

        @param indexUtf16
        pointer to a UTF-16 based index into this string; must not be null.  On
        entry, the index must be in the range from zero to the length of this
        string (in UTF-16 code units), inclusive.  Upon successful return, the
        index will be updated to address the UTF-16 code unit that is the given
        incrementCodePoints away from the initial index.

        @param incrementCodePoints
        the number of code points to move the given *indexUtf16.  If
        non-negative, moving is done after determining the code point at the
        index.  If negative, moving is done before determining the code point
        at the (then updated) index.  The value must be such that the resulting
        UTF-16 based index is in the range from zero to the length of this
        string (in UTF-16 code units), inclusive.

        @return
        the code point (an integer in the range from 0 to 0x10FFFF, inclusive)
        that is represented within this string starting at the index computed as
        follows:  If incrementCodePoints is non-negative, the index is the
        initial value of *indexUtf16; if incrementCodePoints is negative, the
        index is the updated value of *indexUtf16.  In either case, the computed
        index must be in the range from zero to one less than the length of this
        string (in UTF-16 code units), inclusive.

        @since UDK 3.2.7
    */
    inline sal_uInt32 iterateCodePoints(
        sal_Int32 * indexUtf16, sal_Int32 incrementCodePoints = 1) const
    {
        return rtl_uString_iterateCodePoints(
            pData, indexUtf16, incrementCodePoints);
    }

    /**
     * Convert an OString to an OUString, assuming that the OString is
     * UTF-8-encoded.
     *
     * @param rSource
     * an OString to convert
     *
     * @since LibreOffice 4.4
     */
    static inline OUString fromUtf8(const OString& rSource)
    {
        OUString aTarget;
        bool bSuccess = rtl_convertStringToUString(&aTarget.pData,
                                                   rSource.getStr(),
                                                   rSource.getLength(),
                                                   RTL_TEXTENCODING_UTF8,
                                                   RTL_TEXTTOUNICODE_FLAGS_UNDEFINED_ERROR|RTL_TEXTTOUNICODE_FLAGS_MBUNDEFINED_ERROR|RTL_TEXTTOUNICODE_FLAGS_INVALID_ERROR);
        (void) bSuccess;
        assert(bSuccess);
        return aTarget;
    }

    /**
     * Convert this string to an OString, assuming that the string can be
     * UTF-8-encoded successfully.
     *
     * In other words, you must not use this method on a random sequence of
     * UTF-16 code units, but only at places where it is assumed that the
     * content is a proper string.
     *
     * @since LibreOffice 4.4
     */
    inline OString toUtf8() const
    {
        OString aTarget;
        bool bSuccess = rtl_convertUStringToString(&aTarget.pData,
                                                   getStr(),
                                                   getLength(),
                                                   RTL_TEXTENCODING_UTF8,
                                                   RTL_UNICODETOTEXT_FLAGS_UNDEFINED_ERROR|RTL_UNICODETOTEXT_FLAGS_INVALID_ERROR);
        (void) bSuccess;
        assert(bSuccess);
        return aTarget;
    }

    /**
      Returns the string representation of the integer argument.

      This function can't be used for language specific conversion.

      @param    i           an integer value
      @param    radix       the radix (between 2 and 36)
      @return   a string with the string representation of the argument.
      @since LibreOffice 4.1
    */
    static OUString number( int i, sal_Int16 radix = 10 )
    {
        sal_Unicode aBuf[RTL_USTR_MAX_VALUEOFINT32];
        rtl_uString* pNewData = 0;
        rtl_uString_newFromStr_WithLength( &pNewData, aBuf, rtl_ustr_valueOfInt32( aBuf, i, radix ) );
        return OUString( pNewData, SAL_NO_ACQUIRE );
    }
    /// @overload
    /// @since LibreOffice 4.1
    static OUString number( unsigned int i, sal_Int16 radix = 10 )
    {
        return number( static_cast< unsigned long long >( i ), radix );
    }
    /// @overload
    /// @since LibreOffice 4.1
    static OUString number( long i, sal_Int16 radix = 10)
    {
        return number( static_cast< long long >( i ), radix );
    }
    /// @overload
    /// @since LibreOffice 4.1
    static OUString number( unsigned long i, sal_Int16 radix = 10 )
    {
        return number( static_cast< unsigned long long >( i ), radix );
    }
    /// @overload
    /// @since LibreOffice 4.1
    static OUString number( long long ll, sal_Int16 radix = 10 )
    {
        sal_Unicode aBuf[RTL_STR_MAX_VALUEOFINT64];
        rtl_uString* pNewData = 0;
        rtl_uString_newFromStr_WithLength( &pNewData, aBuf, rtl_ustr_valueOfInt64( aBuf, ll, radix ) );
        return OUString( pNewData, SAL_NO_ACQUIRE );
    }
    /// @overload
    /// @since LibreOffice 4.1
    static OUString number( unsigned long long ll, sal_Int16 radix = 10 )
    {
        sal_Unicode aBuf[RTL_STR_MAX_VALUEOFUINT64];
        rtl_uString* pNewData = 0;
        rtl_uString_newFromStr_WithLength( &pNewData, aBuf, rtl_ustr_valueOfUInt64( aBuf, ll, radix ) );
        return OUString( pNewData, SAL_NO_ACQUIRE );
    }

    /**
      Returns the string representation of the float argument.

      This function can't be used for language specific conversion.

      @param    f           a float.
      @return   a string with the string representation of the argument.
      @since LibreOffice 4.1
    */
    static OUString number( float f )
    {
        sal_Unicode aBuf[RTL_USTR_MAX_VALUEOFFLOAT];
        rtl_uString* pNewData = 0;
        rtl_uString_newFromStr_WithLength( &pNewData, aBuf, rtl_ustr_valueOfFloat( aBuf, f ) );
        return OUString( pNewData, SAL_NO_ACQUIRE );
    }

    /**
      Returns the string representation of the double argument.

      This function can't be used for language specific conversion.

      @param    d           a double.
      @return   a string with the string representation of the argument.
      @since LibreOffice 4.1
    */
    static OUString number( double d )
    {
        sal_Unicode aBuf[RTL_USTR_MAX_VALUEOFDOUBLE];
        rtl_uString* pNewData = 0;
        rtl_uString_newFromStr_WithLength( &pNewData, aBuf, rtl_ustr_valueOfDouble( aBuf, d ) );
        return OUString( pNewData, SAL_NO_ACQUIRE );
    }

    /**
      Returns the string representation of the sal_Bool argument.

      If the sal_Bool is true, the string "true" is returned.
      If the sal_Bool is false, the string "false" is returned.
      This function can't be used for language specific conversion.

      @param    b   a sal_Bool.
      @return   a string with the string representation of the argument.
      @deprecated use boolean()
    */
    SAL_DEPRECATED("use boolean()") static OUString valueOf( sal_Bool b )
    {
        return boolean(b);
    }

    /**
      Returns the string representation of the boolean argument.

      If the argument is true, the string "true" is returned.
      If the argument is false, the string "false" is returned.
      This function can't be used for language specific conversion.

      @param    b   a bool.
      @return   a string with the string representation of the argument.
      @since LibreOffice 4.1
    */
    static OUString boolean( bool b )
    {
        sal_Unicode aBuf[RTL_USTR_MAX_VALUEOFBOOLEAN];
        rtl_uString* pNewData = 0;
        rtl_uString_newFromStr_WithLength( &pNewData, aBuf, rtl_ustr_valueOfBoolean( aBuf, b ) );
        return OUString( pNewData, SAL_NO_ACQUIRE );
    }

    /**
      Returns the string representation of the char argument.

      @param    c   a character.
      @return   a string with the string representation of the argument.
      @deprecated use operator, function or constructor taking char or sal_Unicode argument
    */
    SAL_DEPRECATED("convert to OUString or use directly") static OUString valueOf( sal_Unicode c )
    {
        return OUString( &c, 1 );
    }

    /**
      Returns the string representation of the int argument.

      This function can't be used for language specific conversion.

      @param    i           a int32.
      @param    radix       the radix (between 2 and 36)
      @return   a string with the string representation of the argument.
      @deprecated use number()
    */
    SAL_DEPRECATED("use number()") static OUString valueOf( sal_Int32 i, sal_Int16 radix = 10 )
    {
        return number( i, radix );
    }

    /**
      Returns the string representation of the long argument.

      This function can't be used for language specific conversion.

      @param    ll          a int64.
      @param    radix       the radix (between 2 and 36)
      @return   a string with the string representation of the argument.
      @deprecated use number()
    */
    SAL_DEPRECATED("use number()") static OUString valueOf( sal_Int64 ll, sal_Int16 radix = 10 )
    {
        return number( ll, radix );
    }

    /**
      Returns the string representation of the float argument.

      This function can't be used for language specific conversion.

      @param    f           a float.
      @return   a string with the string representation of the argument.
      @deprecated use number()
    */
    SAL_DEPRECATED("use number()") static OUString valueOf( float f )
    {
        return number(f);
    }

    /**
      Returns the string representation of the double argument.

      This function can't be used for language specific conversion.

      @param    d           a double.
      @return   a string with the string representation of the argument.
      @deprecated use number()
    */
    SAL_DEPRECATED("use number()") static OUString valueOf( double d )
    {
        return number(d);
    }

    /**
      Returns a OUString copied without conversion from an ASCII
      character string.

      Since this method is optimized for performance, the ASCII character
      values are not converted in any way. The caller has to make sure that
      all ASCII characters are in the allowed range between 0 and
      127. The ASCII string must be NULL-terminated.

      Note that for string literals it is simpler and more efficient
      to directly use the OUString constructor.

      @param    value       the 8-Bit ASCII character string
      @return   a string with the string representation of the argument.
     */
    static OUString createFromAscii( const sal_Char * value )
    {
        rtl_uString* pNew = 0;
        rtl_uString_newFromAscii( &pNew, value );
        return OUString( pNew, SAL_NO_ACQUIRE );
    }
};

#if defined LIBO_INTERNAL_ONLY // "RTL_FAST_STRING"
/// @cond INTERNAL

/** Compare a string and an ASCII character literal for equality.

    @since LibreOffice 5.0
*/
template<char C> bool operator ==(OUString const & string, OUStringLiteral1<C>)
{
    char const c = C;
    return string.equalsAsciiL(&c, 1);
}

/** Compare a string and an ASCII character literal for inequality.

    @since LibreOffice 5.0
*/
template<char C> bool operator !=(
    OUString const & string, OUStringLiteral1<C> literal)
{
    return !(string == literal);
}

/**
 @internal
*/
template<>
struct ToStringHelper< OUString >
    {
    static int length( const OUString& s ) { return s.getLength(); }
    static sal_Unicode* addData( sal_Unicode* buffer, const OUString& s ) { return addDataHelper( buffer, s.getStr(), s.getLength()); }
    static const bool allowOStringConcat = false;
    static const bool allowOUStringConcat = true;
    };

/**
 @internal
*/
template<>
struct ToStringHelper< OUStringLiteral >
    {
    static int length( const OUStringLiteral& str ) { return str.size; }
    static sal_Unicode* addData( sal_Unicode* buffer, const OUStringLiteral& str ) { return addDataLiteral( buffer, str.data, str.size ); }
    static const bool allowOStringConcat = false;
    static const bool allowOUStringConcat = true;
    };

/**
 @internal
*/
template<char C> struct ToStringHelper<OUStringLiteral1<C>>
{
    static int length(OUStringLiteral1<C>) { return 1; }
    static sal_Unicode * addData(sal_Unicode * buffer, OUStringLiteral1<C>)
    { *buffer++ = C; return buffer; }
    static const bool allowOStringConcat = false;
    static const bool allowOUStringConcat = true;
};

/**
 @internal
*/
template< typename charT, typename traits, typename T1, typename T2 >
inline std::basic_ostream<charT, traits> & operator <<(
    std::basic_ostream<charT, traits> & stream, const OUStringConcat< T1, T2 >& concat)
{
    return stream << OUString( concat );
}

/// @endcond
#endif

/** A helper to use OUStrings with hash maps.

    Instances of this class are unary function objects that can be used as
    hash function arguments to std::unordered_map and similar constructs.
 */
struct OUStringHash
{
    /** Compute a hash code for a string.

        @param rString
        a string.

        @return
        a hash code for the string.  This hash code should not be stored
        persistently, as its computation may change in later revisions.
     */
    size_t operator()(const OUString& rString) const
        { return (size_t)rString.hashCode(); }
};

/* ======================================================================= */

/** Convert an OString to an OUString, using a specific text encoding.

    The lengths of the two strings may differ (e.g., for double-byte
    encodings, UTF-7, UTF-8).

    @param rStr
    an OString to convert.

    @param encoding
    the text encoding to use for conversion.

    @param convertFlags
    flags which control the conversion.  Either use
    OSTRING_TO_OUSTRING_CVTFLAGS, or see
    <http://udk.openoffice.org/cpp/man/spec/textconversion.html> for more
    details.
 */
inline OUString OStringToOUString( const OString & rStr,
                                   rtl_TextEncoding encoding,
                                   sal_uInt32 convertFlags = OSTRING_TO_OUSTRING_CVTFLAGS )
{
    return OUString( rStr.getStr(), rStr.getLength(), encoding, convertFlags );
}

/** Convert an OUString to an OString, using a specific text encoding.

    The lengths of the two strings may differ (e.g., for double-byte
    encodings, UTF-7, UTF-8).

    @param rUnicode
    an OUString to convert.

    @param encoding
    the text encoding to use for conversion.

    @param convertFlags
    flags which control the conversion.  Either use
    OUSTRING_TO_OSTRING_CVTFLAGS, or see
    <http://udk.openoffice.org/cpp/man/spec/textconversion.html> for more
    details.
 */
inline OString OUStringToOString( const OUString & rUnicode,
                                  rtl_TextEncoding encoding,
                                  sal_uInt32 convertFlags = OUSTRING_TO_OSTRING_CVTFLAGS )
{
    return OString( rUnicode.getStr(), rUnicode.getLength(), encoding, convertFlags );
}

/* ======================================================================= */

/**
    Support for rtl::OUString in std::ostream (and thus in
    CPPUNIT_ASSERT or SAL_INFO macros, for example).

    The rtl::OUString is converted to UTF-8.

    @since LibreOffice 3.5.
*/
template< typename charT, typename traits >
inline std::basic_ostream<charT, traits> & operator <<(
    std::basic_ostream<charT, traits> & stream, OUString const & string)
{
    return stream <<
        OUStringToOString(string, RTL_TEXTENCODING_UTF8).getStr();
        // best effort; potentially loses data due to conversion failures
        // (stray surrogate halves) and embedded null characters
}

} // namespace

#ifdef RTL_STRING_UNITTEST
namespace rtl
{
typedef rtlunittest::OUString OUString;
}
#endif

// In internal code, allow to use classes like OUString without having to
// explicitly refer to the rtl namespace, which is kind of superfluous given
// that OUString itself is namespaced by its OU prefix:
#if defined LIBO_INTERNAL_ONLY && !defined RTL_STRING_UNITTEST
using ::rtl::OUString;
using ::rtl::OUStringHash;
using ::rtl::OStringToOUString;
using ::rtl::OUStringToOString;
using ::rtl::OUStringLiteral;
using ::rtl::OUStringLiteral1;
#endif

#endif /* _RTL_USTRING_HXX */

/* vim:set shiftwidth=4 softtabstop=4 expandtab: */