X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/1545d942ecb905e4b3eb449c148908e94d0a1230..9c97f8ea883afa5db14c37bea63f6d4c327b4596:/include/wx/string.h diff --git a/include/wx/string.h b/include/wx/string.h index 48fbfb0484..f3afbd34ff 100644 --- a/include/wx/string.h +++ b/include/wx/string.h @@ -24,6 +24,7 @@ #include "wx/defs.h" // everybody should include this +#ifndef __WXPALMOS5__ #if defined(__WXMAC__) || defined(__VISAGECPP__) #include #endif @@ -46,12 +47,9 @@ #ifdef HAVE_STRCASECMP_IN_STRINGS_H #include // for strcasecmp() #endif // HAVE_STRCASECMP_IN_STRINGS_H +#endif // ! __WXPALMOS5__ -#ifdef __WXPALMOS__ - #include -#endif - -#include "wx/wxcrt.h" // for wxChar, wxStrlen() etc. +#include "wx/wxcrtbase.h" // for wxChar, wxStrlen() etc. #include "wx/strvararg.h" #include "wx/buffer.h" // for wxCharBuffer #include "wx/strconv.h" // for wxConvertXXX() macros and wxMBConv classes @@ -59,7 +57,7 @@ #include "wx/stringops.h" #include "wx/unichar.h" -class WXDLLIMPEXP_BASE wxString; +class WXDLLIMPEXP_FWD_BASE wxString; // unless this symbol is predefined to disable the compatibility functions, do // use them @@ -125,8 +123,6 @@ inline int Stricmp(const char *psz1, const char *psz2) return _stricmp(psz1, psz2); #elif defined(__SC__) return _stricmp(psz1, psz2); -#elif defined(__SALFORDC__) - return stricmp(psz1, psz2); #elif defined(__BORLANDC__) return stricmp(psz1, psz2); #elif defined(__WATCOMC__) @@ -239,22 +235,12 @@ public: wxCStrData operator+(size_t n) const { return wxCStrData(m_str, m_offset + n, m_owned); } - // and these for "str.c_str() + n - 2": - wxCStrData operator-(int n) const - { - wxASSERT_MSG( n <= (int)m_offset, - _T("attempt to construct address before the beginning of the string") ); - return wxCStrData(m_str, m_offset - n, m_owned); - } - wxCStrData operator-(long n) const - { - wxASSERT_MSG( n <= (int)m_offset, - _T("attempt to construct address before the beginning of the string") ); - return wxCStrData(m_str, m_offset - n, m_owned); - } - wxCStrData operator-(size_t n) const + // and these for "str.c_str() + (p2 - p1)" (it also works for any integer + // expression but it must be ptrdiff_t and not e.g. int to work in this + // example): + wxCStrData operator-(ptrdiff_t n) const { - wxASSERT_MSG( n <= m_offset, + wxASSERT_MSG( n <= (ptrdiff_t)m_offset, _T("attempt to construct address before the beginning of the string") ); return wxCStrData(m_str, m_offset - n, m_owned); } @@ -268,7 +254,7 @@ private: size_t m_offset; bool m_owned; - friend class WXDLLIMPEXP_BASE wxString; + friend class WXDLLIMPEXP_FWD_BASE wxString; }; // ---------------------------------------------------------------------------- @@ -334,14 +320,16 @@ public: // We have to implement the version without template arguments manually // because of the StringReturnType<> hack, although WX_DEFINE_VARARG_FUNC // normally does it itself. It has to be a template so that we can use - // the hack, even though there's no real template parameter: - struct FormatDummyArg {} ; - + // the hack, even though there's no real template parameter. We can't move + // it to wxStrig, because it would shadow these versions of Format() then. template inline static typename StringReturnType::type - Format(const wxFormatString& fmt, FormatDummyArg dummy = FormatDummyArg()) + Format(const T& fmt) { - return DoFormatWchar(fmt); + // NB: this doesn't compile if T is not (some form of) a string; + // this makes Format's prototype equivalent to + // Format(const wxFormatString& fmt) + return DoFormatWchar(wxFormatString(fmt)); } // int Printf(const wxString& format, ...); @@ -368,6 +356,43 @@ protected: #pragma warning (disable:4275) #endif +#if wxUSE_UNICODE_UTF8 +// see the comment near wxString::iterator for why we need this +class WXDLLIMPEXP_BASE wxStringIteratorNode +{ +public: + wxStringIteratorNode() + : m_str(NULL), m_citer(NULL), m_iter(NULL), m_prev(NULL), m_next(NULL) {} + wxStringIteratorNode(const wxString *str, + wxStringImpl::const_iterator *citer) + { DoSet(str, citer, NULL); } + wxStringIteratorNode(const wxString *str, wxStringImpl::iterator *iter) + { DoSet(str, NULL, iter); } + ~wxStringIteratorNode() + { clear(); } + + inline void set(const wxString *str, wxStringImpl::const_iterator *citer) + { clear(); DoSet(str, citer, NULL); } + inline void set(const wxString *str, wxStringImpl::iterator *iter) + { clear(); DoSet(str, NULL, iter); } + + const wxString *m_str; + wxStringImpl::const_iterator *m_citer; + wxStringImpl::iterator *m_iter; + wxStringIteratorNode *m_prev, *m_next; + +private: + inline void clear(); + inline void DoSet(const wxString *str, + wxStringImpl::const_iterator *citer, + wxStringImpl::iterator *iter); + + // the node belongs to a particular iterator instance, it's not copied + // when a copy of the iterator is made + DECLARE_NO_COPY_CLASS(wxStringIteratorNode) +}; +#endif // wxUSE_UNICODE_UTF8 + class WXDLLIMPEXP_BASE wxString #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN : public wxStringPrintfMixin @@ -485,9 +510,9 @@ private: { return ConvertStr(str, n, conv); } static wxCharBuffer ImplStr(const wchar_t* str) - { return ConvertStr(str, npos, wxConvUTF8).data; } + { return ConvertStr(str, npos, wxMBConvUTF8()).data; } static SubstrBufFromWC ImplStr(const wchar_t* str, size_t n) - { return ConvertStr(str, n, wxConvUTF8); } + { return ConvertStr(str, n, wxMBConvUTF8()); } size_t PosToImpl(size_t pos) const { @@ -511,7 +536,7 @@ private: if ( pos == 0 || pos == npos ) return pos; else - return const_iterator(m_impl.begin() + pos) - begin(); + return const_iterator(this, m_impl.begin() + pos) - begin(); } #endif // !wxUSE_UNICODE_UTF8/wxUSE_UNICODE_UTF8 @@ -537,8 +562,7 @@ public: #define WX_STR_ITERATOR_TAG void /* dummy type */ #endif - #define WX_STR_ITERATOR_IMPL(iterator_name, pointer_type, \ - reference_type, reference_ctor) \ + #define WX_STR_ITERATOR_IMPL(iterator_name, pointer_type, reference_type) \ private: \ typedef wxStringImpl::iterator_name underlying_iterator; \ public: \ @@ -548,7 +572,6 @@ public: typedef reference_type reference; \ typedef pointer_type pointer; \ \ - reference operator*() const { return reference_ctor; } \ reference operator[](size_t n) const { return *(*this + n); } \ \ iterator_name& operator++() \ @@ -568,25 +591,15 @@ public: return tmp; \ } \ \ - iterator_name& operator+=(int n) \ + iterator_name& operator+=(ptrdiff_t n) \ { \ m_cur = wxStringOperations::AddToIter(m_cur, n); \ return *this; \ } \ - iterator_name& operator+=(size_t n) \ - { \ - m_cur = wxStringOperations::AddToIter(m_cur, (int)n); \ - return *this; \ - } \ - iterator_name& operator-=(int n) \ + iterator_name& operator-=(ptrdiff_t n) \ { \ m_cur = wxStringOperations::AddToIter(m_cur, -n); \ return *this; \ - } \ - iterator_name& operator-=(size_t n) \ - { \ - m_cur = wxStringOperations::AddToIter(m_cur, -(int)n); \ - return *this; \ } \ \ difference_type operator-(const iterator_name& i) const \ @@ -610,70 +623,129 @@ public: /* for internal wxString use only: */ \ underlying_iterator impl() const { return m_cur; } \ \ - friend class WXDLLIMPEXP_BASE wxString; \ - friend class WXDLLIMPEXP_BASE wxCStrData; \ + friend class wxString; \ + friend class wxCStrData; \ \ private: \ underlying_iterator m_cur - class const_iterator; + class WXDLLIMPEXP_FWD_BASE const_iterator; #if wxUSE_UNICODE_UTF8 - class iterator - { - // NB: In UTF-8 build, (non-const) iterator needs to keep reference - // to the underlying wxStringImpl, because UTF-8 is variable-length - // encoding and changing the value pointer to by an iterator using - // its operator* requires calling wxStringImpl::replace() if the old - // and new values differ in their encoding's length. + // NB: In UTF-8 build, (non-const) iterator needs to keep reference + // to the underlying wxStringImpl, because UTF-8 is variable-length + // encoding and changing the value pointer to by an iterator (using + // its operator*) requires calling wxStringImpl::replace() if the old + // and new values differ in their encoding's length. + // + // Furthermore, the replace() call may invalid all iterators for the + // string, so we have to keep track of outstanding iterators and update + // them if replace() happens. + // + // This is implemented by maintaining linked list of iterators for every + // string and traversing it in wxUniCharRef::operator=(). Head of the + // list is stored in wxString. (FIXME-UTF8) - WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef, - wxUniCharRef::CreateForString(m_str, m_cur)); + class WXDLLIMPEXP_BASE iterator + { + WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef); public: - iterator(const iterator& i) : m_cur(i.m_cur), m_str(i.m_str) {} + iterator() {} + iterator(const iterator& i) + : m_cur(i.m_cur), m_node(i.str(), &m_cur) {} + iterator& operator=(const iterator& i) + { + if (&i != this) + { + m_cur = i.m_cur; + m_node.set(i.str(), &m_cur); + } + return *this; + } + + reference operator*() + { return wxUniCharRef::CreateForString(m_node, m_cur); } - iterator operator+(int n) const - { return iterator(m_str, wxStringOperations::AddToIter(m_cur, n)); } - iterator operator+(size_t n) const - { return iterator(m_str, wxStringOperations::AddToIter(m_cur, (int)n)); } - iterator operator-(int n) const - { return iterator(m_str, wxStringOperations::AddToIter(m_cur, -n)); } - iterator operator-(size_t n) const - { return iterator(m_str, wxStringOperations::AddToIter(m_cur, -(int)n)); } + iterator operator+(ptrdiff_t n) const + { return iterator(str(), wxStringOperations::AddToIter(m_cur, n)); } + iterator operator-(ptrdiff_t n) const + { return iterator(str(), wxStringOperations::AddToIter(m_cur, -n)); } private: iterator(wxString *str, underlying_iterator ptr) - : m_cur(ptr), m_str(str->m_impl) {} - iterator(wxStringImpl& str, underlying_iterator ptr) - : m_cur(ptr), m_str(str) {} + : m_cur(ptr), m_node(str, &m_cur) {} + + wxString* str() const { return wx_const_cast(wxString*, m_node.m_str); } - wxStringImpl& m_str; + wxStringIteratorNode m_node; friend class const_iterator; }; + class WXDLLIMPEXP_BASE const_iterator + { + // NB: reference_type is intentionally value, not reference, the character + // may be encoded differently in wxString data: + WX_STR_ITERATOR_IMPL(const_iterator, const wxChar*, wxUniChar); + + public: + const_iterator() {} + const_iterator(const const_iterator& i) + : m_cur(i.m_cur), m_node(i.str(), &m_cur) {} + const_iterator(const iterator& i) + : m_cur(i.m_cur), m_node(i.str(), &m_cur) {} + + const_iterator& operator=(const const_iterator& i) + { + if (&i != this) + { + m_cur = i.m_cur; + m_node.set(i.str(), &m_cur); + } + return *this; + } + const_iterator& operator=(const iterator& i) + { m_cur = i.m_cur; m_node.set(i.str(), &m_cur); return *this; } + + reference operator*() const + { return wxStringOperations::DecodeChar(m_cur); } + + const_iterator operator+(ptrdiff_t n) const + { return const_iterator(str(), wxStringOperations::AddToIter(m_cur, n)); } + const_iterator operator-(ptrdiff_t n) const + { return const_iterator(str(), wxStringOperations::AddToIter(m_cur, -n)); } + + private: + // for internal wxString use only: + const_iterator(const wxString *str, underlying_iterator ptr) + : m_cur(ptr), m_node(str, &m_cur) {} + + const wxString* str() const { return m_node.m_str; } + + wxStringIteratorNode m_node; + }; + size_t IterToImplPos(wxString::iterator i) const { return wxStringImpl::const_iterator(i.impl()) - m_impl.begin(); } #else // !wxUSE_UNICODE_UTF8 - class iterator + class WXDLLIMPEXP_BASE iterator { - WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef, - wxUniCharRef::CreateForString(m_cur)); + WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef); public: + iterator() {} iterator(const iterator& i) : m_cur(i.m_cur) {} - iterator operator+(int n) const + reference operator*() + { return wxUniCharRef::CreateForString(m_cur); } + + iterator operator+(ptrdiff_t n) const { return iterator(wxStringOperations::AddToIter(m_cur, n)); } - iterator operator+(size_t n) const - { return iterator(wxStringOperations::AddToIter(m_cur, (int)n)); } - iterator operator-(int n) const + iterator operator-(ptrdiff_t n) const { return iterator(wxStringOperations::AddToIter(m_cur, -n)); } - iterator operator-(size_t n) const - { return iterator(wxStringOperations::AddToIter(m_cur, -(int)n)); } private: // for internal wxString use only: @@ -682,32 +754,33 @@ public: friend class const_iterator; }; -#endif // wxUSE_UNICODE_UTF8/!wxUSE_UNICODE_UTF8 - class const_iterator + class WXDLLIMPEXP_BASE const_iterator { // NB: reference_type is intentionally value, not reference, the character // may be encoded differently in wxString data: - WX_STR_ITERATOR_IMPL(const_iterator, const wxChar*, wxUniChar, - wxStringOperations::DecodeChar(m_cur)); + WX_STR_ITERATOR_IMPL(const_iterator, const wxChar*, wxUniChar); public: + const_iterator() {} const_iterator(const const_iterator& i) : m_cur(i.m_cur) {} const_iterator(const iterator& i) : m_cur(i.m_cur) {} - const_iterator operator+(int n) const + reference operator*() const + { return wxStringOperations::DecodeChar(m_cur); } + + const_iterator operator+(ptrdiff_t n) const { return const_iterator(wxStringOperations::AddToIter(m_cur, n)); } - const_iterator operator+(size_t n) const - { return const_iterator(wxStringOperations::AddToIter(m_cur, (int)n)); } - const_iterator operator-(int n) const + const_iterator operator-(ptrdiff_t n) const { return const_iterator(wxStringOperations::AddToIter(m_cur, -n)); } - const_iterator operator-(size_t n) const - { return const_iterator(wxStringOperations::AddToIter(m_cur, -(int)n)); } private: // for internal wxString use only: const_iterator(underlying_iterator ptr) : m_cur(ptr) {} + const_iterator(const wxString *WXUNUSED(str), underlying_iterator ptr) + : m_cur(ptr) {} }; +#endif // wxUSE_UNICODE_UTF8/!wxUSE_UNICODE_UTF8 #undef WX_STR_ITERATOR_TAG #undef WX_STR_ITERATOR_IMPL @@ -727,6 +800,7 @@ public: typedef typename T::reference reference; typedef typename T::pointer *pointer; + reverse_iterator_impl() {} reverse_iterator_impl(iterator_type i) : m_cur(i) {} reverse_iterator_impl(const reverse_iterator_impl& ri) : m_cur(ri.m_cur) {} @@ -746,21 +820,13 @@ public: { reverse_iterator_impl tmp = *this; ++m_cur; return tmp; } // NB: explicit in the functions below is to keep BCC 5.5 happy - reverse_iterator_impl operator+(int n) const - { return reverse_iterator_impl(m_cur - n); } - reverse_iterator_impl operator+(size_t n) const + reverse_iterator_impl operator+(ptrdiff_t n) const { return reverse_iterator_impl(m_cur - n); } - reverse_iterator_impl operator-(int n) const - { return reverse_iterator_impl(m_cur + n); } - reverse_iterator_impl operator-(size_t n) const + reverse_iterator_impl operator-(ptrdiff_t n) const { return reverse_iterator_impl(m_cur + n); } - reverse_iterator_impl operator+=(int n) + reverse_iterator_impl operator+=(ptrdiff_t n) { m_cur -= n; return *this; } - reverse_iterator_impl operator+=(size_t n) - { m_cur -= n; return *this; } - reverse_iterator_impl operator-=(int n) - { m_cur += n; return *this; } - reverse_iterator_impl operator-=(size_t n) + reverse_iterator_impl operator-=(ptrdiff_t n) { m_cur += n; return *this; } unsigned operator-(const reverse_iterator_impl& i) const @@ -792,7 +858,8 @@ private: // wxCStrData) to an iterator into the string static const_iterator CreateConstIterator(const wxCStrData& data) { - return const_iterator(data.m_str->begin() + data.m_offset); + return const_iterator(data.m_str, + (data.m_str->begin() + data.m_offset).impl()); } // in UTF-8 STL build, creation from std::string requires conversion under @@ -825,7 +892,7 @@ public: wxString(const wxString& stringSrc) : m_impl(stringSrc.m_impl) { } // string containing nRepeat copies of ch - wxString(wxUniChar ch, size_t nRepeat = 1) + wxString(wxUniChar ch, size_t nRepeat = 1 ) { assign(nRepeat, ch); } wxString(size_t nRepeat, wxUniChar ch) { assign(nRepeat, ch); } @@ -917,7 +984,12 @@ public: wxString(const std::string& str) { assign(str.c_str(), str.length()); } #endif +#endif // wxUSE_STD_STRING + // Unlike ctor from std::string, we provide conversion to std::string only + // if wxUSE_STL and not merely wxUSE_STD_STRING (which is on by default), + // because it conflicts with operator const char/wchar_t*: +#if wxUSE_STL #if wxUSE_UNICODE_WCHAR && wxUSE_STL_BASED_WXSTRING // wxStringImpl is std::string in the encoding we want operator const wxStdWideString&() const { return m_impl; } @@ -940,10 +1012,10 @@ public: #endif // wxUSE_STL // first valid index position - const_iterator begin() const { return const_iterator(m_impl.begin()); } + const_iterator begin() const { return const_iterator(this, m_impl.begin()); } iterator begin() { return iterator(this, m_impl.begin()); } // position one after the last valid one - const_iterator end() const { return const_iterator(m_impl.end()); } + const_iterator end() const { return const_iterator(this, m_impl.end()); } iterator end() { return iterator(this, m_impl.end()); } // first element of the reversed string @@ -974,19 +1046,25 @@ public: void resize(size_t nSize, wxUniChar ch = wxT('\0')) { + const size_t len = length(); + if ( nSize == len) + return; + #if wxUSE_UNICODE_UTF8 - if ( !ch.IsAscii() ) + if ( nSize < len ) { - size_t len = length(); - if ( nSize == len) - return; - else if ( nSize < len ) - erase(nSize); - else - append(nSize - len, ch); + // we can't use wxStringImpl::resize() for truncating the string as it + // counts in bytes, not characters + erase(nSize); + return; } + + // we also can't use (presumably more efficient) resize() if we have to + // append characters taking more than one byte + if ( !ch.IsAscii() ) + append(nSize - len, ch); else -#endif +#endif // wxUSE_UNICODE_UTF8 m_impl.resize(nSize, (wxStringCharType)ch); } @@ -1045,18 +1123,17 @@ public: // get last character wxUniChar Last() const - { - wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") ); - - return at(length() - 1); - } + { + wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") ); + return *rbegin(); + } // get writable last character wxUniCharRef Last() - { - wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") ); - return at(length() - 1); - } + { + wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") ); + return *rbegin(); + } /* Note that we we must define all of the overloads below to avoid @@ -1091,6 +1168,13 @@ public: // implicit conversion to C string operator wxCStrData() const { return c_str(); } + + // the first two operators conflict with operators for conversion to + // std::string and they must be disabled in STL build; the next one only + // makes sense if conversions to char* are also defined and not defining it + // in STL build also helps us to get more clear error messages for the code + // which relies on implicit conversion to char* in STL build +#if !wxUSE_STL operator const char*() const { return c_str(); } operator const wchar_t*() const { return c_str(); } @@ -1098,6 +1182,7 @@ public: // wxWidgets versions: this is the same as conversion to const char * so it // may fail! operator const void*() const { return c_str(); } +#endif // wxUSE_STL // identical to c_str(), for MFC compatibility const wxCStrData GetData() const { return c_str(); } @@ -1120,24 +1205,42 @@ public: // the behaviour of these functions with the strings containing anything // else than 7 bit ASCII characters is undefined, use at your own risk. #if wxUSE_UNICODE - static wxString FromAscii(const char *ascii); // string - static wxString FromAscii(const char ascii); // char + static wxString FromAscii(const char *ascii, size_t len); + static wxString FromAscii(const char *ascii); + static wxString FromAscii(char ascii); const wxCharBuffer ToAscii() const; #else // ANSI static wxString FromAscii(const char *ascii) { return wxString( ascii ); } - static wxString FromAscii(const char ascii) { return wxString( ascii ); } + static wxString FromAscii(const char *ascii, size_t len) + { return wxString( ascii, len ); } + static wxString FromAscii(char ascii) { return wxString( ascii ); } const char *ToAscii() const { return c_str(); } #endif // Unicode/!Unicode + // also provide unsigned char overloads as signed/unsigned doesn't matter + // for 7 bit ASCII characters + static wxString FromAscii(const unsigned char *ascii) + { return FromAscii((const char *)ascii); } + static wxString FromAscii(const unsigned char *ascii, size_t len) + { return FromAscii((const char *)ascii, len); } + // conversion to/from UTF-8: #if wxUSE_UNICODE_UTF8 static wxString FromUTF8(const char *utf8) { + if ( !utf8 ) + return wxEmptyString; + wxASSERT( wxStringOperations::IsValidUtf8String(utf8) ); return FromImpl(wxStringImpl(utf8)); } static wxString FromUTF8(const char *utf8, size_t len) { + if ( !utf8 ) + return wxEmptyString; + if ( len == npos ) + return FromUTF8(utf8); + wxASSERT( wxStringOperations::IsValidUtf8String(utf8, len) ); return FromImpl(wxStringImpl(utf8, len)); } @@ -1145,21 +1248,22 @@ public: const char* ToUTF8() const { return wx_str(); } #elif wxUSE_UNICODE_WCHAR static wxString FromUTF8(const char *utf8) - { return wxString(utf8, wxConvUTF8); } + { return wxString(utf8, wxMBConvUTF8()); } static wxString FromUTF8(const char *utf8, size_t len) - { return wxString(utf8, wxConvUTF8, len); } - const wxCharBuffer utf8_str() const { return mb_str(wxConvUTF8); } + { return wxString(utf8, wxMBConvUTF8(), len); } + const wxCharBuffer utf8_str() const { return mb_str(wxMBConvUTF8()); } const wxCharBuffer ToUTF8() const { return utf8_str(); } #else // ANSI static wxString FromUTF8(const char *utf8) - { return wxString(wxConvUTF8.cMB2WC(utf8)); } + { return wxString(wxMBConvUTF8().cMB2WC(utf8)); } static wxString FromUTF8(const char *utf8, size_t len) { size_t wlen; - wxWCharBuffer buf(wxConvUTF8.cMB2WC(utf8, len == npos ? wxNO_LEN : len, &wlen)); + wxWCharBuffer buf(wxMBConvUTF8().cMB2WC(utf8, len == npos ? wxNO_LEN : len, &wlen)); return wxString(buf.data(), wlen); } - const wxCharBuffer utf8_str() const { return wxConvUTF8.cWC2MB(wc_str()); } + const wxCharBuffer utf8_str() const + { return wxMBConvUTF8().cWC2MB(wc_str()); } const wxCharBuffer ToUTF8() const { return utf8_str(); } #endif @@ -1226,22 +1330,26 @@ public: #if wxUSE_WCHAR_T const wxWCharBuffer wc_str(const wxMBConv& conv = wxConvLibc) const; #endif // wxUSE_WCHAR_T -#ifdef __WXOSX__ - const wxCharBuffer fn_str() const { return wxConvFile.cWC2WX( wc_str( wxConvLocal ) ); } -#else - const wxChar* fn_str() const { return c_str(); } -#endif + const wxCharBuffer fn_str() const { return wxConvFile.cWC2WX( wc_str( wxConvLibc ) ); } #endif // Unicode/ANSI // overloaded assignment // from another wxString wxString& operator=(const wxString& stringSrc) - { m_impl = stringSrc.m_impl; return *this; } + { if (&stringSrc != this) m_impl = stringSrc.m_impl; return *this; } wxString& operator=(const wxCStrData& cstr) { return *this = cstr.AsString(); } // from a character wxString& operator=(wxUniChar ch) - { m_impl = wxStringOperations::EncodeChar(ch); return *this; } + { +#if wxUSE_UNICODE_UTF8 + if ( !ch.IsAscii() ) + m_impl = wxStringOperations::EncodeChar(ch); + else +#endif + m_impl = (wxStringCharType)ch; + return *this; + } wxString& operator=(wxUniCharRef ch) { return operator=((wxUniChar)ch); } wxString& operator=(char ch) @@ -1420,14 +1528,23 @@ public: { return compare(s); } // same as Cmp() but not case-sensitive int CmpNoCase(const wxString& s) const; + // test for the string equality, either considering case or not // (if compareWithCase then the case matters) bool IsSameAs(const wxString& str, bool compareWithCase = true) const - { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; } + { +#if !wxUSE_UNICODE_UTF8 + // in UTF-8 build, length() is O(n) and doing this would be _slower_ + if ( length() != str.length() ) + return false; +#endif + return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; + } bool IsSameAs(const char *str, bool compareWithCase = true) const { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; } bool IsSameAs(const wchar_t *str, bool compareWithCase = true) const { return (compareWithCase ? Cmp(str) : CmpNoCase(str)) == 0; } + bool IsSameAs(const wxCStrData& str, bool compareWithCase = true) const { return IsSameAs(str.AsString(), compareWithCase); } bool IsSameAs(const wxCharBuffer& str, bool compareWithCase = true) const @@ -1435,11 +1552,7 @@ public: bool IsSameAs(const wxWCharBuffer& str, bool compareWithCase = true) const { return IsSameAs(str.data(), compareWithCase); } // comparison with a single character: returns true if equal - bool IsSameAs(wxUniChar c, bool compareWithCase = true) const - { - return (length() == 1) && (compareWithCase ? GetChar(0u) == c - : wxToupper(GetChar(0u)) == wxToupper(c)); - } + bool IsSameAs(wxUniChar c, bool compareWithCase = true) const; // FIXME-UTF8: remove these overloads bool IsSameAs(wxUniCharRef c, bool compareWithCase = true) const { return IsSameAs(wxUniChar(c), compareWithCase); } @@ -2135,8 +2248,15 @@ public: // find the first occurence of character ch after nStart size_t find(wxUniChar ch, size_t nStart = 0) const { - return PosFromImpl(m_impl.find(wxStringOperations::EncodeChar(ch), - PosToImpl(nStart))); +#if wxUSE_UNICODE_UTF8 + if ( !ch.IsAscii() ) + return PosFromImpl(m_impl.find(wxStringOperations::EncodeChar(ch), + PosToImpl(nStart))); + else +#endif + return PosFromImpl(m_impl.find((wxStringCharType)ch, + PosToImpl(nStart))); + } size_t find(wxUniCharRef ch, size_t nStart = 0) const { return find(wxUniChar(ch), nStart); } @@ -2173,8 +2293,14 @@ public: // as find, but from the end size_t rfind(wxUniChar ch, size_t nStart = npos) const { - return PosFromImpl(m_impl.rfind(wxStringOperations::EncodeChar(ch), - PosToImpl(nStart))); +#if wxUSE_UNICODE_UTF8 + if ( !ch.IsAscii() ) + return PosFromImpl(m_impl.rfind(wxStringOperations::EncodeChar(ch), + PosToImpl(nStart))); + else +#endif + return PosFromImpl(m_impl.rfind((wxStringCharType)ch, + PosToImpl(nStart))); } size_t rfind(wxUniCharRef ch, size_t nStart = npos) const { return rfind(wxUniChar(ch), nStart); } @@ -2412,7 +2538,15 @@ public: { return operator+=(s.data()); } // string += char wxString& operator+=(wxUniChar ch) - { m_impl += wxStringOperations::EncodeChar(ch); return *this; } + { +#if wxUSE_UNICODE_UTF8 + if ( !ch.IsAscii() ) + m_impl += wxStringOperations::EncodeChar(ch); + else +#endif + m_impl += (wxStringCharType)ch; + return *this; + } wxString& operator+=(wxUniCharRef ch) { return *this += wxUniChar(ch); } wxString& operator+=(int ch) { return *this += wxUniChar(ch); } wxString& operator+=(char ch) { return *this += wxUniChar(ch); } @@ -2477,9 +2611,30 @@ private: ConvertedBuffer m_convertedToWChar; #endif - friend class WXDLLIMPEXP_BASE wxCStrData; - friend class wxImplStringBuffer; - friend class wxImplStringBufferLength; +#if wxUSE_UNICODE_UTF8 + // FIXME-UTF8: (try to) move this elsewhere (TLS) or solve differently + // assigning to character pointer to by wxString::interator may + // change the underlying wxStringImpl iterator, so we have to + // keep track of all iterators and update them as necessary: + struct wxStringIteratorNodeHead + { + wxStringIteratorNodeHead() : ptr(NULL) {} + wxStringIteratorNode *ptr; + + // copying is disallowed as it would result in more than one pointer into + // the same linked list + DECLARE_NO_COPY_CLASS(wxStringIteratorNodeHead) + }; + + wxStringIteratorNodeHead m_iterators; + + friend class WXDLLIMPEXP_FWD_BASE wxStringIteratorNode; + friend class WXDLLIMPEXP_FWD_BASE wxUniCharRef; +#endif // wxUSE_UNICODE_UTF8 + + friend class WXDLLIMPEXP_FWD_BASE wxCStrData; + friend class wxStringInternalBuffer; + friend class wxStringInternalBufferLength; }; #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN @@ -2488,21 +2643,13 @@ private: // string iterator operators that satisfy STL Random Access Iterator // requirements: -inline wxString::iterator operator+(int n, wxString::iterator i) - { return i + n; } -inline wxString::iterator operator+(size_t n, wxString::iterator i) - { return i + n; } -inline wxString::const_iterator operator+(int n, wxString::const_iterator i) +inline wxString::iterator operator+(ptrdiff_t n, wxString::iterator i) { return i + n; } -inline wxString::const_iterator operator+(size_t n, wxString::const_iterator i) +inline wxString::const_iterator operator+(ptrdiff_t n, wxString::const_iterator i) { return i + n; } -inline wxString::reverse_iterator operator+(int n, wxString::reverse_iterator i) +inline wxString::reverse_iterator operator+(ptrdiff_t n, wxString::reverse_iterator i) { return i + n; } -inline wxString::reverse_iterator operator+(size_t n, wxString::reverse_iterator i) - { return i + n; } -inline wxString::const_reverse_iterator operator+(int n, wxString::const_reverse_iterator i) - { return i + n; } -inline wxString::const_reverse_iterator operator+(size_t n, wxString::const_reverse_iterator i) +inline wxString::const_reverse_iterator operator+(ptrdiff_t n, wxString::const_reverse_iterator i) { return i + n; } // notice that even though for many compilers the friend declarations above are @@ -2532,16 +2679,7 @@ inline wxString operator+(wchar_t ch, const wxString& string) { return wxUniChar(ch) + string; } -#if wxUSE_STL_BASED_WXSTRING - // return an empty wxString (not very useful with wxUSE_STL == 1) - inline const wxString wxGetEmptyString() { return wxString(); } -#else // !wxUSE_STL_BASED_WXSTRING - // return an empty wxString (more efficient than wxString() here) - inline const wxString& wxGetEmptyString() - { - return *(wxString *)&wxEmptyString; - } -#endif // wxUSE_STL_BASED_WXSTRING/!wxUSE_STL_BASED_WXSTRING +#define wxGetEmptyString() wxString() // ---------------------------------------------------------------------------- // wxStringBuffer: a tiny class allowing to get a writable pointer into string @@ -2550,16 +2688,16 @@ inline wxString operator+(wchar_t ch, const wxString& string) #if !wxUSE_STL_BASED_WXSTRING // string buffer for direct access to string data in their native // representation: -class wxImplStringBuffer +class wxStringInternalBuffer { public: typedef wxStringCharType CharType; - wxImplStringBuffer(wxString& str, size_t lenWanted = 1024) + wxStringInternalBuffer(wxString& str, size_t lenWanted = 1024) : m_str(str), m_buf(NULL) { m_buf = m_str.DoGetWriteBuf(lenWanted); } - ~wxImplStringBuffer() { m_str.DoUngetWriteBuf(); } + ~wxStringInternalBuffer() { m_str.DoUngetWriteBuf(); } operator wxStringCharType*() const { return m_buf; } @@ -2567,22 +2705,22 @@ private: wxString& m_str; wxStringCharType *m_buf; - DECLARE_NO_COPY_CLASS(wxImplStringBuffer) + DECLARE_NO_COPY_CLASS(wxStringInternalBuffer) }; -class wxImplStringBufferLength +class wxStringInternalBufferLength { public: typedef wxStringCharType CharType; - wxImplStringBufferLength(wxString& str, size_t lenWanted = 1024) + wxStringInternalBufferLength(wxString& str, size_t lenWanted = 1024) : m_str(str), m_buf(NULL), m_len(0), m_lenSet(false) { m_buf = m_str.DoGetWriteBuf(lenWanted); wxASSERT(m_buf != NULL); } - ~wxImplStringBufferLength() + ~wxStringInternalBufferLength() { wxASSERT(m_lenSet); m_str.DoUngetWriteBuf(m_len); @@ -2597,13 +2735,13 @@ private: size_t m_len; bool m_lenSet; - DECLARE_NO_COPY_CLASS(wxImplStringBufferLength) + DECLARE_NO_COPY_CLASS(wxStringInternalBufferLength) }; #endif // !wxUSE_STL_BASED_WXSTRING template -class wxStringTypeBufferBase +class WXDLLIMPEXP_BASE wxStringTypeBufferBase { public: typedef T CharType; @@ -2621,7 +2759,7 @@ protected: }; template -class wxStringTypeBufferLengthBase +class WXDLLIMPEXP_BASE wxStringTypeBufferLengthBase { public: typedef T CharType; @@ -2630,12 +2768,6 @@ public: : m_str(str), m_buf(lenWanted), m_len(0), m_lenSet(false) { } - ~wxStringTypeBufferLengthBase() - { - wxASSERT(m_lenSet); - m_str.assign(m_buf.data(), m_len); - } - operator CharType*() { return m_buf.data(); } void SetLength(size_t length) { m_len = length; m_lenSet = true; } @@ -2677,30 +2809,37 @@ public: }; #if wxUSE_STL_BASED_WXSTRING -class wxImplStringBuffer : public wxStringTypeBufferBase + +WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferBase ) + +class wxStringInternalBuffer : public wxStringTypeBufferBase { public: - wxImplStringBuffer(wxString& str, size_t lenWanted = 1024) + wxStringInternalBuffer(wxString& str, size_t lenWanted = 1024) : wxStringTypeBufferBase(str, lenWanted) {} - ~wxImplStringBuffer() + ~wxStringInternalBuffer() { m_str.m_impl.assign(m_buf.data()); } - DECLARE_NO_COPY_CLASS(wxImplStringBuffer) + DECLARE_NO_COPY_CLASS(wxStringInternalBuffer) }; -class wxImplStringBufferLength : public wxStringTypeBufferLengthBase +WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( + wxStringTypeBufferLengthBase ) + +class wxStringInternalBufferLength + : public wxStringTypeBufferLengthBase { public: - wxImplStringBufferLength(wxString& str, size_t lenWanted = 1024) + wxStringInternalBufferLength(wxString& str, size_t lenWanted = 1024) : wxStringTypeBufferLengthBase(str, lenWanted) {} - ~wxImplStringBufferLength() + ~wxStringInternalBufferLength() { wxASSERT(m_lenSet); m_str.m_impl.assign(m_buf.data(), m_len); } - DECLARE_NO_COPY_CLASS(wxImplStringBufferLength) + DECLARE_NO_COPY_CLASS(wxStringInternalBufferLength) }; #endif // wxUSE_STL_BASED_WXSTRING @@ -2709,10 +2848,42 @@ public: typedef wxStringTypeBuffer wxStringBuffer; typedef wxStringTypeBufferLength wxStringBufferLength; #else // if !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8 -typedef wxImplStringBuffer wxStringBuffer; -typedef wxImplStringBufferLength wxStringBufferLength; +typedef wxStringInternalBuffer wxStringBuffer; +typedef wxStringInternalBufferLength wxStringBufferLength; #endif // !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8 +#if wxUSE_UNICODE_UTF8 +typedef wxStringInternalBuffer wxUTF8StringBuffer; +typedef wxStringInternalBufferLength wxUTF8StringBufferLength; +#elif wxUSE_UNICODE_WCHAR + +WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferBase ) + +class WXDLLIMPEXP_BASE wxUTF8StringBuffer : public wxStringTypeBufferBase +{ +public: + wxUTF8StringBuffer(wxString& str, size_t lenWanted = 1024) + : wxStringTypeBufferBase(str, lenWanted) {} + ~wxUTF8StringBuffer(); + + DECLARE_NO_COPY_CLASS(wxUTF8StringBuffer) +}; + +WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferLengthBase ) + +class WXDLLIMPEXP_BASE wxUTF8StringBufferLength + : public wxStringTypeBufferLengthBase +{ +public: + wxUTF8StringBufferLength(wxString& str, size_t lenWanted = 1024) + : wxStringTypeBufferLengthBase(str, lenWanted) {} + ~wxUTF8StringBufferLength(); + + DECLARE_NO_COPY_CLASS(wxUTF8StringBufferLength) +}; +#endif // wxUSE_UNICODE_UTF8/wxUSE_UNICODE_WCHAR + + // --------------------------------------------------------------------------- // wxString comparison functions: operator versions are always case sensitive // --------------------------------------------------------------------------- @@ -2723,12 +2894,10 @@ wxDEFINE_ALL_COMPARISONS(const wxChar *, const wxString&, wxCMP_WXCHAR_STRING) #undef wxCMP_WXCHAR_STRING -// note that there is an optimization in operator==() and !=(): we (quickly) -// checks the strings length first, before comparing their data inline bool operator==(const wxString& s1, const wxString& s2) - { return (s1.Len() == s2.Len()) && (s1.Cmp(s2) == 0); } + { return s1.IsSameAs(s2); } inline bool operator!=(const wxString& s1, const wxString& s2) - { return (s1.Len() != s2.Len()) || (s1.Cmp(s2) != 0); } + { return !s1.IsSameAs(s2); } inline bool operator< (const wxString& s1, const wxString& s2) { return s1.Cmp(s2) < 0; } inline bool operator> (const wxString& s1, const wxString& s2) @@ -2738,7 +2907,15 @@ inline bool operator<=(const wxString& s1, const wxString& s2) inline bool operator>=(const wxString& s1, const wxString& s2) { return s1.Cmp(s2) >= 0; } -#if wxUSE_UNICODE +inline bool operator==(const wxString& s1, const wxCStrData& s2) + { return s1 == s2.AsString(); } +inline bool operator==(const wxCStrData& s1, const wxString& s2) + { return s1.AsString() == s2; } +inline bool operator!=(const wxString& s1, const wxCStrData& s2) + { return s1 != s2.AsString(); } +inline bool operator!=(const wxCStrData& s1, const wxString& s2) + { return s1.AsString() != s2; } + inline bool operator==(const wxString& s1, const wxWCharBuffer& s2) { return (s1.Cmp((const wchar_t *)s2) == 0); } inline bool operator==(const wxWCharBuffer& s1, const wxString& s2) @@ -2747,7 +2924,7 @@ inline bool operator!=(const wxString& s1, const wxWCharBuffer& s2) { return (s1.Cmp((const wchar_t *)s2) != 0); } inline bool operator!=(const wxWCharBuffer& s1, const wxString& s2) { return (s2.Cmp((const wchar_t *)s1) != 0); } -#else // !wxUSE_UNICODE + inline bool operator==(const wxString& s1, const wxCharBuffer& s2) { return (s1.Cmp((const char *)s2) == 0); } inline bool operator==(const wxCharBuffer& s1, const wxString& s2) @@ -2756,19 +2933,16 @@ inline bool operator!=(const wxString& s1, const wxCharBuffer& s2) { return (s1.Cmp((const char *)s2) != 0); } inline bool operator!=(const wxCharBuffer& s1, const wxString& s2) { return (s2.Cmp((const char *)s1) != 0); } -#endif // wxUSE_UNICODE/!wxUSE_UNICODE -#if wxUSE_UNICODE inline wxString operator+(const wxString& string, const wxWCharBuffer& buf) { return string + (const wchar_t *)buf; } inline wxString operator+(const wxWCharBuffer& buf, const wxString& string) { return (const wchar_t *)buf + string; } -#else // !wxUSE_UNICODE + inline wxString operator+(const wxString& string, const wxCharBuffer& buf) { return string + (const char *)buf; } inline wxString operator+(const wxCharBuffer& buf, const wxString& string) { return (const char *)buf + string; } -#endif // wxUSE_UNICODE/!wxUSE_UNICODE // comparison with char inline bool operator==(const wxUniChar& c, const wxString& s) { return s.IsSameAs(c); } @@ -2830,7 +3004,15 @@ WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxCharBuffer&); WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxWCharBuffer&); #endif -#endif // wxSTD_STRING_COMPATIBILITY +#if wxUSE_UNICODE && defined(HAVE_WOSTREAM) + +WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxString&); +WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxCStrData&); +WXDLLIMPEXP_BASE wxSTD wostream& operator<<(wxSTD wostream&, const wxWCharBuffer&); + +#endif // wxUSE_UNICODE && defined(HAVE_WOSTREAM) + +#endif // wxUSE_STD_IOSTREAM // --------------------------------------------------------------------------- // wxCStrData implementation @@ -2851,7 +3033,7 @@ inline wxCStrData::wxCStrData(const wxCStrData& data) inline wxCStrData::~wxCStrData() { if ( m_owned ) - delete m_str; + delete wx_const_cast(wxString*, m_str); // cast to silence warnings } // simple cases for AsChar() and AsWChar(), the complicated ones are @@ -2863,13 +3045,20 @@ inline const wchar_t* wxCStrData::AsWChar() const } #endif // wxUSE_UNICODE_WCHAR -#if !wxUSE_UNICODE || wxUSE_UTF8_LOCALE_ONLY +#if !wxUSE_UNICODE inline const char* wxCStrData::AsChar() const { - return wxStringOperations::AddToIter(m_str->wx_str(), m_offset); + return m_str->wx_str() + m_offset; } #endif // !wxUSE_UNICODE +#if wxUSE_UTF8_LOCALE_ONLY +inline const char* wxCStrData::AsChar() const +{ + return wxStringOperations::AddToIter(m_str->wx_str(), m_offset); +} +#endif // wxUSE_UTF8_LOCALE_ONLY + inline const wxCharBuffer wxCStrData::AsCharBuf() const { #if !wxUSE_UNICODE @@ -2898,7 +3087,11 @@ inline wxString wxCStrData::AsString() const inline const wxStringCharType *wxCStrData::AsInternal() const { +#if wxUSE_UNICODE_UTF8 return wxStringOperations::AddToIter(m_str->wx_str(), m_offset); +#else + return m_str->wx_str() + m_offset; +#endif } inline wxUniChar wxCStrData::operator*() const @@ -2947,12 +3140,54 @@ inline wxWCharBuffer::wxWCharBuffer(const wxCStrData& cstr) { } +#if wxUSE_UNICODE_UTF8 +// ---------------------------------------------------------------------------- +// implementation of wxStringIteratorNode inline methods +// ---------------------------------------------------------------------------- + +void wxStringIteratorNode::DoSet(const wxString *str, + wxStringImpl::const_iterator *citer, + wxStringImpl::iterator *iter) +{ + m_prev = NULL; + m_iter = iter; + m_citer = citer; + m_str = str; + if ( str ) + { + m_next = str->m_iterators.ptr; + wx_const_cast(wxString*, m_str)->m_iterators.ptr = this; + if ( m_next ) + m_next->m_prev = this; + } + else + { + m_next = NULL; + } +} + +void wxStringIteratorNode::clear() +{ + if ( m_next ) + m_next->m_prev = m_prev; + if ( m_prev ) + m_prev->m_next = m_next; + else if ( m_str ) // first in the list + wx_const_cast(wxString*, m_str)->m_iterators.ptr = m_next; + + m_next = m_prev = NULL; + m_citer = NULL; + m_iter = NULL; + m_str = NULL; +} +#endif // wxUSE_UNICODE_UTF8 + #if WXWIN_COMPATIBILITY_2_8 - // lot of code out there doesn't explicitly include wx/wxchar.h, but uses + // lot of code out there doesn't explicitly include wx/crt.h, but uses // CRT wrappers that are now declared in wx/wxcrt.h and wx/wxcrtvararg.h, // so let's include this header now that wxString is defined and it's safe // to do it: - #include "wx/wxchar.h" + #include "wx/crt.h" #endif #endif // _WX_WXSTRING_H_