X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/af49c4b8a2d3553e733e71c7dd3911881f4c1a2a..ac10ea91c9bff4a8daae3313063867647c1e1079:/include/wx/string.h diff --git a/include/wx/string.h b/include/wx/string.h index 1c6dbc1f0b..43dfbcb175 100644 --- a/include/wx/string.h +++ b/include/wx/string.h @@ -1,33 +1,22 @@ /////////////////////////////////////////////////////////////////////////////// -// Name: string.h -// Purpose: wxString and wxArrayString classes +// Name: wx/string.h +// Purpose: wxString class // Author: Vadim Zeitlin // Modified by: // Created: 29/01/98 // RCS-ID: $Id$ // Copyright: (c) 1998 Vadim Zeitlin -// Licence: wxWindows license +// Licence: wxWindows licence /////////////////////////////////////////////////////////////////////////////// /* Efficient string class [more or less] compatible with MFC CString, - wxWindows version 1 wxString and std::string and some handy functions + wxWidgets version 1 wxString and std::string and some handy functions missing from string.h. */ -#ifndef _WX_WXSTRINGH__ -#define _WX_WXSTRINGH__ - -#if defined(__GNUG__) && !defined(__APPLE__) - #pragma interface "string.h" -#endif - -// ---------------------------------------------------------------------------- -// conditinal compilation -// ---------------------------------------------------------------------------- - -// compile the std::string compatibility functions if defined -#define wxSTD_STRING_COMPATIBILITY +#ifndef _WX_WXSTRING_H__ +#define _WX_WXSTRING_H__ // ---------------------------------------------------------------------------- // headers @@ -35,14 +24,11 @@ #include "wx/defs.h" // everybody should include this +#ifndef __WXPALMOS5__ #if defined(__WXMAC__) || defined(__VISAGECPP__) #include #endif -#ifdef __EMX__ - #include -#endif - #if defined(__VISAGECPP__) && __IBMCPP__ >= 400 // problem in VACPP V4 with including stdlib.h multiple times // strconv includes it anyway @@ -58,13 +44,61 @@ # include #endif -#ifdef HAVE_STRINGS_H +#ifdef HAVE_STRCASECMP_IN_STRINGS_H #include // for strcasecmp() -#endif // HAVE_STRINGS_H +#endif // HAVE_STRCASECMP_IN_STRINGS_H +#endif // ! __WXPALMOS5__ -#include "wx/wxchar.h" // for wxChar +#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 +#include "wx/stringimpl.h" +#include "wx/stringops.h" +#include "wx/unichar.h" + +// by default we cache the mapping of the positions in UTF-8 string to the byte +// offset as this results in noticeable performance improvements for loops over +// strings using indices; comment out this line to disable this +// +// notice that this optimization is well worth using even in debug builds as it +// changes asymptotic complexity of algorithms using indices to iterate over +// wxString back to expected linear from quadratic +// +// also notice that wxTLS_TYPE() (__declspec(thread) in this case) is unsafe to +// use in DLL build under pre-Vista Windows so we disable this code for now, if +// anybody really needs to use UTF-8 build under Windows with this optimization +// it would have to be re-tested and probably corrected +#if wxUSE_UNICODE_UTF8 && !defined(__WXMSW__) + #define wxUSE_STRING_POS_CACHE 1 +#else + #define wxUSE_STRING_POS_CACHE 0 +#endif + +#if wxUSE_STRING_POS_CACHE + #include "wx/tls.h" + + // change this 0 to 1 to enable additional (very expensive) asserts + // verifying that string caching logic works as expected + #if 0 + #define wxSTRING_CACHE_ASSERT(cond) wxASSERT(cond) + #else + #define wxSTRING_CACHE_ASSERT(cond) + #endif +#endif // wxUSE_STRING_POS_CACHE + +class WXDLLIMPEXP_FWD_BASE wxString; + +// unless this symbol is predefined to disable the compatibility functions, do +// use them +#ifndef WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER + #define WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER 1 +#endif + +namespace wxPrivate +{ + template struct wxStringAsBufHelper; +} // --------------------------------------------------------------------------- // macros @@ -77,56 +111,53 @@ #define wxMBSTRINGCAST (char *)(const char *) #define wxWCSTRINGCAST (wchar_t *)(const wchar_t *) -// implementation only -#define wxASSERT_VALID_INDEX(i) \ - wxASSERT_MSG( (size_t)(i) <= Len(), _T("invalid index in wxString") ) - // ---------------------------------------------------------------------------- // constants // ---------------------------------------------------------------------------- -#if defined(__VISAGECPP__) && __IBMCPP__ >= 400 -// must define this static for VA or else you get multiply defined symbols everywhere -extern const unsigned int wxSTRING_MAXLEN; +#if WXWIN_COMPATIBILITY_2_6 -#else +// deprecated in favour of wxString::npos, don't use in new code +// // maximum possible length for a string means "take all string" everywhere -// (as sizeof(StringData) is unknown here, we substract 100) -const unsigned int wxSTRING_MAXLEN = UINT_MAX - 100; - -#endif - -// ---------------------------------------------------------------------------- -// global data -// ---------------------------------------------------------------------------- +#define wxSTRING_MAXLEN wxString::npos -// global pointer to empty string -WXDLLEXPORT_DATA(extern const wxChar*) wxEmptyString; +#endif // WXWIN_COMPATIBILITY_2_6 // --------------------------------------------------------------------------- // global functions complementing standard C string library replacements for // strlen() and portable strcasecmp() //--------------------------------------------------------------------------- -// Use wxXXX() functions from wxchar.h instead! These functions are for +#if WXWIN_COMPATIBILITY_2_8 +// Use wxXXX() functions from wxcrt.h instead! These functions are for // backwards compatibility only. // checks whether the passed in pointer is NULL and if the string is empty +wxDEPRECATED( inline bool IsEmpty(const char *p) ); inline bool IsEmpty(const char *p) { return (!p || !*p); } // safe version of strlen() (returns 0 if passed NULL pointer) +wxDEPRECATED( inline size_t Strlen(const char *psz) ); inline size_t Strlen(const char *psz) { return psz ? strlen(psz) : 0; } // portable strcasecmp/_stricmp +wxDEPRECATED( inline int Stricmp(const char *psz1, const char *psz2) ); inline int Stricmp(const char *psz1, const char *psz2) { -#if defined(__VISUALC__) || ( defined(__MWERKS__) && defined(__INTEL__) ) +#if defined(__VISUALC__) && defined(__WXWINCE__) + register char c1, c2; + do { + c1 = tolower(*psz1++); + c2 = tolower(*psz2++); + } while ( c1 && (c1 == c2) ); + + return c1 - c2; +#elif defined(__VISUALC__) || ( defined(__MWERKS__) && defined(__INTEL__) ) 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__) @@ -137,7 +168,10 @@ inline int Stricmp(const char *psz1, const char *psz2) return stricmp(psz1, psz2); #elif defined(__WXPM__) return stricmp(psz1, psz2); -#elif defined(__UNIX__) || defined(__GNUWIN32__) +#elif defined(__WXPALMOS__) || \ + defined(HAVE_STRCASECMP_IN_STRING_H) || \ + defined(HAVE_STRCASECMP_IN_STRINGS_H) || \ + defined(__GNUWIN32__) return strcasecmp(psz1, psz2); #elif defined(__MWERKS__) && !defined(__INTEL__) register char c1, c2; @@ -165,106 +199,248 @@ inline int Stricmp(const char *psz1, const char *psz2) #endif // OS/compiler } -// return an empty wxString -class WXDLLEXPORT wxString; // not yet defined -inline const wxString& wxGetEmptyString() { return *(wxString *)&wxEmptyString; } +#endif // WXWIN_COMPATIBILITY_2_8 -// --------------------------------------------------------------------------- -// string data prepended with some housekeeping info (used by wxString class), -// is never used directly (but had to be put here to allow inlining) -// --------------------------------------------------------------------------- +// ---------------------------------------------------------------------------- +// wxCStrData +// ---------------------------------------------------------------------------- -struct WXDLLEXPORT wxStringData +// Lightweight object returned by wxString::c_str() and implicitly convertible +// to either const char* or const wchar_t*. +class WXDLLIMPEXP_BASE wxCStrData { - int nRefs; // reference count - size_t nDataLength, // actual string length - nAllocLength; // allocated memory size +private: + // Ctors; for internal use by wxString and wxCStrData only + wxCStrData(const wxString *str, size_t offset = 0, bool owned = false) + : m_str(str), m_offset(offset), m_owned(owned) {} + +public: + // Ctor constructs the object from char literal; they are needed to make + // operator?: compile and they intentionally take char*, not const char* + inline wxCStrData(char *buf); + inline wxCStrData(wchar_t *buf); + inline wxCStrData(const wxCStrData& data); + + inline ~wxCStrData(); + + // methods defined inline below must be declared inline or mingw32 3.4.5 + // warns about " defined locally after being referenced with + // dllimport linkage" +#if wxUSE_UNICODE_WCHAR + inline +#endif + const wchar_t* AsWChar() const; + operator const wchar_t*() const { return AsWChar(); } - // mimics declaration 'wxChar data[nAllocLength]' - wxChar* data() const { return (wxChar*)(this + 1); } +#if !wxUSE_UNICODE || wxUSE_UTF8_LOCALE_ONLY + inline +#endif + const char* AsChar() const; + const unsigned char* AsUnsignedChar() const + { return (const unsigned char *) AsChar(); } + operator const char*() const { return AsChar(); } + operator const unsigned char*() const { return AsUnsignedChar(); } - // empty string has a special ref count so it's never deleted - bool IsEmpty() const { return (nRefs == -1); } - bool IsShared() const { return (nRefs > 1); } + operator const void*() const { return AsChar(); } - // lock/unlock - void Lock() { if ( !IsEmpty() ) nRefs++; } + inline const wxCharBuffer AsCharBuf() const; + inline const wxWCharBuffer AsWCharBuf() const; - // VC++ will refuse to inline this function but profiling shows that it - // is wrong -#if defined(__VISUALC__) && (__VISUALC__ >= 1200) - __forceinline -#endif - void Unlock() { if ( !IsEmpty() && --nRefs == 0) free(this); } + inline wxString AsString() const; - // if we had taken control over string memory (GetWriteBuf), it's - // intentionally put in invalid state - void Validate(bool b) { nRefs = (b ? 1 : 0); } - bool IsValid() const { return (nRefs != 0); } + // returns the value as C string in internal representation (equivalent + // to AsString().wx_str(), but more efficient) + const wxStringCharType *AsInternal() const; + + // allow expressions like "c_str()[0]": + inline wxUniChar operator[](size_t n) const; + wxUniChar operator[](int n) const { return operator[](size_t(n)); } + wxUniChar operator[](long n) const { return operator[](size_t(n)); } +#ifndef wxSIZE_T_IS_UINT + wxUniChar operator[](unsigned int n) const { return operator[](size_t(n)); } +#endif // size_t != unsigned int + + // these operators are needed to emulate the pointer semantics of c_str(): + // expressions like "wxChar *p = str.c_str() + 1;" should continue to work + // (we need both versions to resolve ambiguities): + wxCStrData operator+(int n) const + { return wxCStrData(m_str, m_offset + n, m_owned); } + wxCStrData operator+(long n) const + { return wxCStrData(m_str, m_offset + n, m_owned); } + wxCStrData operator+(size_t n) const + { return wxCStrData(m_str, m_offset + n, m_owned); } + + // 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 <= (ptrdiff_t)m_offset, + _T("attempt to construct address before the beginning of the string") ); + return wxCStrData(m_str, m_offset - n, m_owned); + } + + // this operator is needed to make expressions like "*c_str()" or + // "*(c_str() + 2)" work + inline wxUniChar operator*() const; + +private: + const wxString *m_str; + size_t m_offset; + bool m_owned; + + friend class WXDLLIMPEXP_FWD_BASE wxString; }; +// ---------------------------------------------------------------------------- +// wxStringPrintfMixin // --------------------------------------------------------------------------- -// This is (yet another one) String class for C++ programmers. It doesn't use -// any of "advanced" C++ features (i.e. templates, exceptions, namespaces...) -// thus you should be able to compile it with practicaly any C++ compiler. -// This class uses copy-on-write technique, i.e. identical strings share the -// same memory as long as neither of them is changed. -// -// This class aims to be as compatible as possible with the new standard -// std::string class, but adds some additional functions and should be at -// least as efficient than the standard implementation. -// -// Performance note: it's more efficient to write functions which take "const -// String&" arguments than "const char *" if you assign the argument to -// another string. -// -// It was compiled and tested under Win32, Linux (libc 5 & 6), Solaris 5.5. + +// NB: VC6 has a bug that causes linker errors if you have template methods +// in a class using __declspec(dllimport). The solution is to split such +// class into two classes, one that contains the template methods and does +// *not* use WXDLLIMPEXP_BASE and another class that contains the rest +// (with DLL linkage). // -// To do: -// - ressource support (string tables in ressources) -// - more wide character (UNICODE) support -// - regular expressions support +// We only do this for VC6 here, because the code is less efficient +// (Printf() has to use dynamic_cast<>) and because OpenWatcom compiler +// cannot compile this code. + +#if defined(__VISUALC__) && __VISUALC__ < 1300 + #define wxNEEDS_WXSTRING_PRINTF_MIXIN +#endif + +#ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN +// this class contains implementation of wxString's vararg methods, it's +// exported from wxBase DLL +class WXDLLIMPEXP_BASE wxStringPrintfMixinBase +{ +protected: + wxStringPrintfMixinBase() {} + +#if !wxUSE_UTF8_LOCALE_ONLY + int DoPrintfWchar(const wxChar *format, ...); + static wxString DoFormatWchar(const wxChar *format, ...); +#endif +#if wxUSE_UNICODE_UTF8 + int DoPrintfUtf8(const char *format, ...); + static wxString DoFormatUtf8(const char *format, ...); +#endif +}; + +// this class contains template wrappers for wxString's vararg methods, it's +// intentionally *not* exported from the DLL in order to fix the VC6 bug +// described above +class wxStringPrintfMixin : public wxStringPrintfMixinBase +{ +private: + // to further complicate things, we can't return wxString from + // wxStringPrintfMixin::Format() because wxString is not yet declared at + // this point; the solution is to use this fake type trait template - this + // way the compiler won't know the return type until Format() is used + // (this doesn't compile with Watcom, but VC6 compiles it just fine): + template struct StringReturnType + { + typedef wxString type; + }; + +public: + // these are duplicated wxString methods, they're also declared below + // if !wxNEEDS_WXSTRING_PRINTF_MIXIN: + + // static wxString Format(const wString& format, ...) ATTRIBUTE_PRINTF_1; + WX_DEFINE_VARARG_FUNC_SANS_N0(static typename StringReturnType::type, + Format, 1, (const wxFormatString&), + DoFormatWchar, DoFormatUtf8) + // 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. We can't move + // it to wxStrig, because it would shadow these versions of Format() then. + template + inline static typename StringReturnType::type + Format(const T& 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, ...); + WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const wxFormatString&), + DoPrintfWchar, DoPrintfUtf8) + // int sprintf(const wxString& format, ...) ATTRIBUTE_PRINTF_2; + WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const wxFormatString&), + DoPrintfWchar, DoPrintfUtf8) + +protected: + wxStringPrintfMixin() : wxStringPrintfMixinBase() {} +}; +#endif // wxNEEDS_WXSTRING_PRINTF_MIXIN + + +// ---------------------------------------------------------------------------- +// wxString: string class trying to be compatible with std::string, MFC +// CString and wxWindows 1.x wxString all at once // --------------------------------------------------------------------------- -class WXDLLEXPORT wxString +#ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN + // "non dll-interface class 'wxStringPrintfMixin' used as base interface + // for dll-interface class 'wxString'" -- this is OK in our case + #pragma warning (disable:4275) +#endif + +#if wxUSE_UNICODE_UTF8 +// see the comment near wxString::iterator for why we need this +class WXDLLIMPEXP_BASE wxStringIteratorNode { -friend class WXDLLEXPORT wxArrayString; +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 +#endif +{ // NB: special care was taken in arranging the member functions in such order // that all inline functions can be effectively inlined, verify that all - // performace critical functions are still inlined if you change order! -private: - // points to data preceded by wxStringData structure with ref count info - wxChar *m_pchData; - - // accessor to string data - wxStringData* GetStringData() const { return (wxStringData*)m_pchData - 1; } - - // string (re)initialization functions - // initializes the string to the empty value (must be called only from - // ctors, use Reinit() otherwise) - void Init() { m_pchData = (wxChar *)wxEmptyString; } - // initializaes the string with (a part of) C-string - void InitWith(const wxChar *psz, size_t nPos = 0, size_t nLen = wxSTRING_MAXLEN); - // as Init, but also frees old data - void Reinit() { GetStringData()->Unlock(); Init(); } - - // memory allocation - // allocates memory for string of length nLen - bool AllocBuffer(size_t nLen); - // copies data to another string - bool AllocCopy(wxString&, int, int) const; - // effectively copies data to string - bool AssignCopy(size_t, const wxChar *); - - // append a (sub)string - bool ConcatSelf(int nLen, const wxChar *src); - - // functions called before writing to the string: they copy it if there - // are other references to our data (should be the only owner when writing) - bool CopyBeforeWrite(); - bool AllocBeforeWrite(size_t); + // performance critical functions are still inlined if you change order! +public: + // an 'invalid' value for string index, moved to this place due to a CW bug + static const size_t npos; +private: // if we hadn't made these operators private, it would be possible to // compile "wxString s; s = 17;" without any warnings as 17 is implicitly // converted to char in C and we do have operator=(char) @@ -280,940 +456,3357 @@ private: // try `s << i' or `s.Printf("%d", i)' instead wxString(int); -public: - // constructors and destructor - // ctor for an empty string - wxString() : m_pchData(NULL) { Init(); } - // copy ctor - wxString(const wxString& stringSrc) : m_pchData(NULL) + + // buffer for holding temporary substring when using any of the methods + // that take (char*,size_t) or (wchar_t*,size_t) arguments: + template + struct SubstrBufFromType { - wxASSERT_MSG( stringSrc.GetStringData()->IsValid(), - _T("did you forget to call UngetWriteBuf()?") ); + T data; + size_t len; - if ( stringSrc.IsEmpty() ) { - // nothing to do for an empty string - Init(); - } - else { - m_pchData = stringSrc.m_pchData; // share same data - GetStringData()->Lock(); // => one more copy - } - } - // string containing nRepeat copies of ch - wxString(wxChar ch, size_t nRepeat = 1); - // ctor takes first nLength characters from C string - // (default value of wxSTRING_MAXLEN means take all the string) - wxString(const wxChar *psz, size_t nLength = wxSTRING_MAXLEN) - : m_pchData(NULL) - { InitWith(psz, 0, nLength); } - wxString(const wxChar *psz, wxMBConv& WXUNUSED(conv), size_t nLength = wxSTRING_MAXLEN) - : m_pchData(NULL) - { InitWith(psz, 0, nLength); } + SubstrBufFromType(const T& data_, size_t len_) + : data(data_), len(len_) + { + wxASSERT_MSG( len != npos, "must have real length" ); + } + }; + +#if wxUSE_UNICODE_UTF8 + // even char* -> char* needs conversion, from locale charset to UTF-8 + typedef SubstrBufFromType SubstrBufFromWC; + typedef SubstrBufFromType SubstrBufFromMB; +#elif wxUSE_UNICODE_WCHAR + typedef SubstrBufFromType SubstrBufFromWC; + typedef SubstrBufFromType SubstrBufFromMB; +#else + typedef SubstrBufFromType SubstrBufFromMB; + typedef SubstrBufFromType SubstrBufFromWC; +#endif -#if wxUSE_UNICODE - // from multibyte string - // (NB: nLength is right now number of Unicode characters, not - // characters in psz! So try not to use it yet!) - wxString(const char *psz, wxMBConv& conv = wxConvLibc, size_t nLength = wxSTRING_MAXLEN); - // from wxWCharBuffer (i.e. return from wxGetString) - wxString(const wxWCharBuffer& psz) - { InitWith(psz, 0, wxSTRING_MAXLEN); } -#else // ANSI - // from C string (for compilers using unsigned char) - wxString(const unsigned char* psz, size_t nLength = wxSTRING_MAXLEN) - : m_pchData(NULL) - { InitWith((const char*)psz, 0, nLength); } -#if wxUSE_WCHAR_T - // from wide (Unicode) string - wxString(const wchar_t *pwz, wxMBConv& conv = wxConvLibc, size_t nLength = wxSTRING_MAXLEN); -#endif // !wxUSE_WCHAR_T + // Functions implementing primitive operations on string data; wxString + // methods and iterators are implemented in terms of it. The differences + // between UTF-8 and wchar_t* representations of the string are mostly + // contained here. - // from wxCharBuffer - wxString(const wxCharBuffer& psz) - : m_pchData(NULL) - { InitWith(psz, 0, wxSTRING_MAXLEN); } -#endif // Unicode/ANSI +#if wxUSE_UNICODE_UTF8 + static SubstrBufFromMB ConvertStr(const char *psz, size_t nLength, + const wxMBConv& conv); + static SubstrBufFromWC ConvertStr(const wchar_t *pwz, size_t nLength, + const wxMBConv& conv); +#elif wxUSE_UNICODE_WCHAR + static SubstrBufFromMB ConvertStr(const char *psz, size_t nLength, + const wxMBConv& conv); +#else + static SubstrBufFromWC ConvertStr(const wchar_t *pwz, size_t nLength, + const wxMBConv& conv); +#endif - // dtor is not virtual, this class must not be inherited from! - ~wxString() { GetStringData()->Unlock(); } +#if !wxUSE_UNICODE_UTF8 // wxUSE_UNICODE_WCHAR or !wxUSE_UNICODE + // returns C string encoded as the implementation expects: + #if wxUSE_UNICODE + static const wchar_t* ImplStr(const wchar_t* str) + { return str ? str : wxT(""); } + static const SubstrBufFromWC ImplStr(const wchar_t* str, size_t n) + { return SubstrBufFromWC(str, (str && n == npos) ? wxWcslen(str) : n); } + static wxWCharBuffer ImplStr(const char* str, + const wxMBConv& conv = wxConvLibc) + { return ConvertStr(str, npos, conv).data; } + static SubstrBufFromMB ImplStr(const char* str, size_t n, + const wxMBConv& conv = wxConvLibc) + { return ConvertStr(str, n, conv); } + #else + static const char* ImplStr(const char* str, + const wxMBConv& WXUNUSED(conv) = wxConvLibc) + { return str ? str : ""; } + static const SubstrBufFromMB ImplStr(const char* str, size_t n, + const wxMBConv& WXUNUSED(conv) = wxConvLibc) + { return SubstrBufFromMB(str, (str && n == npos) ? wxStrlen(str) : n); } + static wxCharBuffer ImplStr(const wchar_t* str) + { return ConvertStr(str, npos, wxConvLibc).data; } + static SubstrBufFromWC ImplStr(const wchar_t* str, size_t n) + { return ConvertStr(str, n, wxConvLibc); } + #endif + + // translates position index in wxString to/from index in underlying + // wxStringImpl: + static size_t PosToImpl(size_t pos) { return pos; } + static void PosLenToImpl(size_t pos, size_t len, + size_t *implPos, size_t *implLen) + { *implPos = pos; *implLen = len; } + static size_t LenToImpl(size_t len) { return len; } + static size_t PosFromImpl(size_t pos) { return pos; } + + // we don't want to define these as empty inline functions as it could + // result in noticeable (and quite unnecessary in non-UTF-8 build) slowdown + // in debug build where the inline functions are not effectively inlined + #define wxSTRING_INVALIDATE_CACHE() + #define wxSTRING_INVALIDATE_CACHED_LENGTH() + #define wxSTRING_UPDATE_CACHED_LENGTH(n) + #define wxSTRING_SET_CACHED_LENGTH(n) + +#else // wxUSE_UNICODE_UTF8 + + static wxCharBuffer ImplStr(const char* str, + const wxMBConv& conv = wxConvLibc) + { return ConvertStr(str, npos, conv).data; } + static SubstrBufFromMB ImplStr(const char* str, size_t n, + const wxMBConv& conv = wxConvLibc) + { return ConvertStr(str, n, conv); } + + static wxCharBuffer ImplStr(const wchar_t* str) + { return ConvertStr(str, npos, wxMBConvUTF8()).data; } + static SubstrBufFromWC ImplStr(const wchar_t* str, size_t n) + { return ConvertStr(str, n, wxMBConvUTF8()); } + +#if wxUSE_STRING_POS_CACHE + // this is an extremely simple cache used by PosToImpl(): each cache element + // contains the string it applies to and the index corresponding to the last + // used position in this wxString in its m_impl string + // + // NB: notice that this struct (and nested Element one) must be a POD or we + // wouldn't be able to use a thread-local variable of this type, in + // particular it should have no ctor -- we rely on statics being + // initialized to 0 instead + struct Cache + { + enum { SIZE = 8 }; - // generic attributes & operations - // as standard strlen() - size_t Len() const { return GetStringData()->nDataLength; } - // string contains any characters? - bool IsEmpty() const { return Len() == 0; } - // empty string is "FALSE", so !str will return TRUE - bool operator!() const { return IsEmpty(); } - // truncate the string to given length - wxString& Truncate(size_t uiLen); - // empty string contents - void Empty() + struct Element + { + const wxString *str; // the string to which this element applies + size_t pos, // the cached index in this string + impl, // the corresponding position in its m_impl + len; // cached length or npos if unknown + + // reset cached index to 0 + void ResetPos() { pos = impl = 0; } + + // reset position and length + void Reset() { ResetPos(); len = npos; } + }; + + // cache the indices mapping for the last few string used + Element cached[SIZE]; + + // the last used index + unsigned lastUsed; + }; + +#ifndef wxHAS_COMPILER_TLS + // we must use an accessor function and not a static variable when the TLS + // variables support is implemented in the library (and not by the compiler) + // because the global s_cache variable could be not yet initialized when a + // ctor of another global object is executed and if that ctor uses any + // wxString methods, bad things happen + // + // however notice that this approach does not work when compiler TLS is used, + // at least not with g++ 4.1.2 under amd64 as it apparently compiles code + // using this accessor incorrectly when optimizations are enabled (-O2 is + // enough) -- luckily we don't need it then neither as static __thread + // variables are initialized by 0 anyhow then and so we can use the variable + // directly + WXEXPORT static Cache& GetCache() { - Truncate(0); + static wxTLS_TYPE(Cache) s_cache; - wxASSERT_MSG( IsEmpty(), _T("string not empty after call to Empty()?") ); + return wxTLS_VALUE(s_cache); } - // empty the string and free memory - void Clear() + + // this helper struct is used to ensure that GetCache() is called during + // static initialization time, i.e. before any threads creation, as otherwise + // the static s_cache construction inside GetCache() wouldn't be MT-safe + friend struct wxStrCacheInitializer; +#else // wxHAS_COMPILER_TLS + static wxTLS_TYPE(Cache) ms_cache; + static Cache& GetCache() { return wxTLS_VALUE(ms_cache); } +#endif // !wxHAS_COMPILER_TLS/wxHAS_COMPILER_TLS + + static Cache::Element *GetCacheBegin() { return GetCache().cached; } + static Cache::Element *GetCacheEnd() { return GetCacheBegin() + Cache::SIZE; } + static unsigned& LastUsedCacheElement() { return GetCache().lastUsed; } + + // this is used in debug builds only to provide a convenient function, + // callable from a debugger, to show the cache contents + friend struct wxStrCacheDumper; + + // uncomment this to have access to some profiling statistics on program + // termination + //#define wxPROFILE_STRING_CACHE + +#ifdef wxPROFILE_STRING_CACHE + static struct PosToImplCacheStats + { + unsigned postot, // total non-trivial calls to PosToImpl + poshits, // cache hits from PosToImpl() + mishits, // cached position beyond the needed one + sumpos, // sum of all positions, used to compute the + // average position after dividing by postot + sumofs, // sum of all offsets after using the cache, used to + // compute the average after dividing by hits + lentot, // number of total calls to length() + lenhits; // number of cache hits in length() + } ms_cacheStats; + + friend struct wxStrCacheStatsDumper; + + #define wxCACHE_PROFILE_FIELD_INC(field) ms_cacheStats.field++ + #define wxCACHE_PROFILE_FIELD_ADD(field, val) ms_cacheStats.field += (val) +#else // !wxPROFILE_STRING_CACHE + #define wxCACHE_PROFILE_FIELD_INC(field) + #define wxCACHE_PROFILE_FIELD_ADD(field, val) +#endif // wxPROFILE_STRING_CACHE/!wxPROFILE_STRING_CACHE + + // note: it could seem that the functions below shouldn't be inline because + // they are big, contain loops and so the compiler shouldn't be able to + // inline them anyhow, however moving them into string.cpp does decrease the + // code performance by ~5%, at least when using g++ 4.1 so do keep them here + // unless tests show that it's not advantageous any more + + // return the pointer to the cache element for this string or NULL if not + // cached + Cache::Element *FindCacheElement() const { - if ( !GetStringData()->IsEmpty() ) - Reinit(); + // profiling seems to show a small but consistent gain if we use this + // simple loop instead of starting from the last used element (there are + // a lot of misses in this function...) + Cache::Element * const cacheBegin = GetCacheBegin(); +#ifndef wxHAS_COMPILER_TLS + // during destruction tls calls may return NULL, in this case return NULL + // immediately without accessing anything else + if ( cacheBegin == NULL ) + return NULL; +#endif + Cache::Element * const cacheEnd = GetCacheEnd(); + for ( Cache::Element *c = cacheBegin; c != cacheEnd; c++ ) + { + if ( c->str == this ) + return c; + } - wxASSERT_MSG( !GetStringData()->nDataLength && - !GetStringData()->nAllocLength, - _T("string should be empty after Clear()") ); + return NULL; } - // contents test - // Is an ascii value - bool IsAscii() const; - // Is a number - bool IsNumber() const; - // Is a word - bool IsWord() const; - - // data access (all indexes are 0 based) - // read access - wxChar GetChar(size_t n) const - { wxASSERT_VALID_INDEX( n ); return m_pchData[n]; } - // read/write access - wxChar& GetWritableChar(size_t n) - { wxASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; } - // write access - void SetChar(size_t n, wxChar ch) - { wxASSERT_VALID_INDEX( n ); CopyBeforeWrite(); m_pchData[n] = ch; } - - // get last character - wxChar Last() const + // unlike FindCacheElement(), this one always returns a valid pointer to the + // cache element for this string, it may have valid last cached position and + // its corresponding index in the byte string or not + Cache::Element *GetCacheElement() const + { + Cache::Element * const cacheBegin = GetCacheBegin(); + Cache::Element * const cacheEnd = GetCacheEnd(); + Cache::Element * const cacheStart = cacheBegin + LastUsedCacheElement(); + + // check the last used first, this does no (measurable) harm for a miss + // but does help for simple loops addressing the same string all the time + if ( cacheStart->str == this ) + return cacheStart; + + // notice that we're going to check cacheStart again inside this call but + // profiling shows that it's still faster to use a simple loop like + // inside FindCacheElement() than manually looping with wrapping starting + // from the cache entry after the start one + Cache::Element *c = FindCacheElement(); + if ( !c ) { - wxASSERT_MSG( !IsEmpty(), _T("wxString: index out of bounds") ); + // claim the next cache entry for this string + c = cacheStart; + if ( ++c == cacheEnd ) + c = cacheBegin; - return m_pchData[Len() - 1]; - } + c->str = this; + c->Reset(); - // get writable last character - wxChar& Last() - { - wxASSERT_MSG( !IsEmpty(), _T("wxString: index out of bounds") ); - CopyBeforeWrite(); - return m_pchData[Len()-1]; + // and remember the last used element + LastUsedCacheElement() = c - cacheBegin; } - /* - So why do we have all these overloaded operator[]s? A bit of history: - initially there was only one of them, taking size_t. Then people - started complaining because they wanted to use ints as indices (I - wonder why) and compilers were giving warnings about it, so we had to - add the operator[](int). Then it became apparent that you couldn't - write str[0] any longer because there was ambiguity between two - overloads and so you now had to write str[0u] (or, of course, use the - explicit casts to either int or size_t but nobody did this). - - Finally, someone decided to compile wxWin on an Alpha machine and got - a surprize: str[0u] didn't compile there because it is of type - unsigned int and size_t is unsigned _long_ on Alpha and so there was - ambiguity between converting uint to int or ulong. To fix this one we - now add operator[](uint) for the machines where size_t is not already - the same as unsigned int - hopefully this fixes the problem (for some - time) - - The only real fix is, of course, to remove all versions but the one - taking size_t... - */ - - // operator version of GetChar - wxChar operator[](size_t n) const - { wxASSERT_VALID_INDEX( n ); return m_pchData[n]; } + return c; + } - // operator version of GetChar - wxChar operator[](int n) const - { wxASSERT_VALID_INDEX( n ); return m_pchData[n]; } + size_t DoPosToImpl(size_t pos) const + { + wxCACHE_PROFILE_FIELD_INC(postot); - // operator version of GetWriteableChar - wxChar& operator[](size_t n) - { wxASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; } + // NB: although the case of pos == 1 (and offset from cached position + // equal to 1) are common, nothing is gained by writing special code + // for handling them, the compiler (at least g++ 4.1 used) seems to + // optimize the code well enough on its own -#ifndef wxSIZE_T_IS_UINT - // operator version of GetChar - wxChar operator[](unsigned int n) const - { wxASSERT_VALID_INDEX( n ); return m_pchData[n]; } + wxCACHE_PROFILE_FIELD_ADD(sumpos, pos); - // operator version of GetWriteableChar - wxChar& operator[](unsigned int n) - { wxASSERT_VALID_INDEX( n ); CopyBeforeWrite(); return m_pchData[n]; } -#endif // size_t != unsigned int + Cache::Element * const cache = GetCacheElement(); - // implicit conversion to C string - operator const wxChar*() const { return m_pchData; } + // cached position can't be 0 so if it is, it means that this entry was + // used for length caching only so far, i.e. it doesn't count as a hit + // from our point of view + if ( cache->pos ) + { + wxCACHE_PROFILE_FIELD_INC(poshits); + } - // explicit conversion to C string (use this with printf()!) - const wxChar* c_str() const { return m_pchData; } - // identical to c_str(), for wxWin 1.6x compatibility - const wxChar* wx_str() const { return m_pchData; } - // identical to c_str(), for MFC compatibility - const wxChar* GetData() const { return m_pchData; } + if ( pos == cache->pos ) + return cache->impl; - // conversion to/from plain (i.e. 7 bit) ASCII: this is useful for - // converting numbers or strings which are certain not to contain special - // chars (typically system functions, X atoms, environment variables etc.) - // - // 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); - const wxCharBuffer ToAscii() const; -#else // ANSI - static wxString FromAscii(const char *ascii) { return wxString( ascii ); } - const char *ToAscii() const { return c_str(); } -#endif // Unicode/!Unicode + // this seems to happen only rarely so just reset the cache in this case + // instead of complicating code even further by seeking backwards in this + // case + if ( cache->pos > pos ) + { + wxCACHE_PROFILE_FIELD_INC(mishits); - // conversions with (possible) format convertions: have to return a - // buffer with temporary data - // - // the functions defined (in either Unicode or ANSI) mode are mb_str() to - // return an ANSI (multibyte) string, wc_str() to return a wide string and - // fn_str() to return a string which should be used with the OS APIs - // accepting the file names. The return value is always the same, but the - // type differs because a function may either return pointer to the buffer - // directly or have to use intermediate buffer for translation. -#if wxUSE_UNICODE - const wxCharBuffer mb_str(wxMBConv& conv = wxConvLibc) const - { return conv.cWC2MB(m_pchData); } + cache->ResetPos(); + } - const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); } + wxCACHE_PROFILE_FIELD_ADD(sumofs, pos - cache->pos); - const wxChar* wc_str() const { return m_pchData; } - // for compatibility with !wxUSE_UNICODE version - const wxChar* wc_str(wxMBConv& WXUNUSED(conv)) const { return m_pchData; } + wxStringImpl::const_iterator i(m_impl.begin() + cache->impl); + for ( size_t n = cache->pos; n < pos; n++ ) + wxStringOperations::IncIter(i); -#if wxMBFILES - const wxCharBuffer fn_str() const { return mb_str(wxConvFile); } -#else // !wxMBFILES - const wxChar* fn_str() const { return m_pchData; } -#endif // wxMBFILES/!wxMBFILES -#else // ANSI - const wxChar* mb_str() const { return m_pchData; } + cache->pos = pos; + cache->impl = i - m_impl.begin(); - // for compatibility with wxUSE_UNICODE version - const wxChar* mb_str(wxMBConv& WXUNUSED(conv)) const { return m_pchData; } + wxSTRING_CACHE_ASSERT( + (int)cache->impl == (begin() + pos).impl() - m_impl.begin() ); - const wxWX2MBbuf mbc_str() const { return mb_str(); } + return cache->impl; + } -#if wxUSE_WCHAR_T - const wxWCharBuffer wc_str(wxMBConv& conv) const - { return conv.cMB2WC(m_pchData); } -#endif // wxUSE_WCHAR_T + void InvalidateCache() + { + Cache::Element * const cache = FindCacheElement(); + if ( cache ) + cache->Reset(); + } - const wxChar* fn_str() const { return m_pchData; } -#endif // Unicode/ANSI + void InvalidateCachedLength() + { + Cache::Element * const cache = FindCacheElement(); + if ( cache ) + cache->len = npos; + } - // overloaded assignment - // from another wxString - wxString& operator=(const wxString& stringSrc); - // from a character - wxString& operator=(wxChar ch); - // from a C string - wxString& operator=(const wxChar *psz); -#if wxUSE_UNICODE - // from wxWCharBuffer - wxString& operator=(const wxWCharBuffer& psz) - { (void) operator=((const wchar_t *)psz); return *this; } -#else // ANSI - // from another kind of C string - wxString& operator=(const unsigned char* psz); -#if wxUSE_WCHAR_T - // from a wide string - wxString& operator=(const wchar_t *pwz); -#endif - // from wxCharBuffer - wxString& operator=(const wxCharBuffer& psz) - { (void) operator=((const char *)psz); return *this; } -#endif // Unicode/ANSI + void SetCachedLength(size_t len) + { + // we optimistically cache the length here even if the string wasn't + // present in the cache before, this seems to do no harm and the + // potential for avoiding length recomputation for long strings looks + // interesting + GetCacheElement()->len = len; + } - // string concatenation - // in place concatenation - /* - Concatenate and return the result. Note that the left to right - associativity of << allows to write things like "str << str1 << str2 - << ..." (unlike with +=) - */ - // string += string - wxString& operator<<(const wxString& s) + void UpdateCachedLength(ptrdiff_t delta) { - wxASSERT_MSG( s.GetStringData()->IsValid(), - _T("did you forget to call UngetWriteBuf()?") ); + Cache::Element * const cache = FindCacheElement(); + if ( cache && cache->len != npos ) + { + wxSTRING_CACHE_ASSERT( (ptrdiff_t)cache->len + delta >= 0 ); - ConcatSelf(s.Len(), s); - return *this; + cache->len += delta; + } } - // string += C string - wxString& operator<<(const wxChar *psz) - { ConcatSelf(wxStrlen(psz), psz); return *this; } - // string += char - wxString& operator<<(wxChar ch) { ConcatSelf(1, &ch); return *this; } - // string += string - void operator+=(const wxString& s) { (void)operator<<(s); } - // string += C string - void operator+=(const wxChar *psz) { (void)operator<<(psz); } - // string += char - void operator+=(wxChar ch) { (void)operator<<(ch); } + #define wxSTRING_INVALIDATE_CACHE() InvalidateCache() + #define wxSTRING_INVALIDATE_CACHED_LENGTH() InvalidateCachedLength() + #define wxSTRING_UPDATE_CACHED_LENGTH(n) UpdateCachedLength(n) + #define wxSTRING_SET_CACHED_LENGTH(n) SetCachedLength(n) +#else // !wxUSE_STRING_POS_CACHE + size_t DoPosToImpl(size_t pos) const + { + return (begin() + pos).impl() - m_impl.begin(); + } - // string += buffer (i.e. from wxGetString) -#if wxUSE_UNICODE - wxString& operator<<(const wxWCharBuffer& s) - { (void)operator<<((const wchar_t *)s); return *this; } - void operator+=(const wxWCharBuffer& s) - { (void)operator<<((const wchar_t *)s); } -#else // !wxUSE_UNICODE - wxString& operator<<(const wxCharBuffer& s) - { (void)operator<<((const char *)s); return *this; } - void operator+=(const wxCharBuffer& s) - { (void)operator<<((const char *)s); } -#endif // wxUSE_UNICODE/!wxUSE_UNICODE + #define wxSTRING_INVALIDATE_CACHE() + #define wxSTRING_INVALIDATE_CACHED_LENGTH() + #define wxSTRING_UPDATE_CACHED_LENGTH(n) + #define wxSTRING_SET_CACHED_LENGTH(n) +#endif // wxUSE_STRING_POS_CACHE/!wxUSE_STRING_POS_CACHE - // string += C string - wxString& Append(const wxString& s) - { - // test for IsEmpty() to share the string if possible - if ( IsEmpty() ) - *this = s; - else - ConcatSelf(s.Length(), s.c_str()); - return *this; - } - wxString& Append(const wxChar* psz) - { ConcatSelf(wxStrlen(psz), psz); return *this; } - // append count copies of given character - wxString& Append(wxChar ch, size_t count = 1u) - { wxString str(ch, count); return *this << str; } - wxString& Append(const wxChar* psz, size_t nLen) - { ConcatSelf(nLen, psz); return *this; } + size_t PosToImpl(size_t pos) const + { + return pos == 0 || pos == npos ? pos : DoPosToImpl(pos); + } - // prepend a string, return the string itself - wxString& Prepend(const wxString& str) - { *this = str + *this; return *this; } + void PosLenToImpl(size_t pos, size_t len, size_t *implPos, size_t *implLen) const; - // non-destructive concatenation - // - friend wxString WXDLLEXPORT operator+(const wxString& string1, const wxString& string2); - // - friend wxString WXDLLEXPORT operator+(const wxString& string, wxChar ch); - // - friend wxString WXDLLEXPORT operator+(wxChar ch, const wxString& string); - // - friend wxString WXDLLEXPORT operator+(const wxString& string, const wxChar *psz); - // - friend wxString WXDLLEXPORT operator+(const wxChar *psz, const wxString& string); + size_t LenToImpl(size_t len) const + { + size_t pos, len2; + PosLenToImpl(0, len, &pos, &len2); + return len2; + } - // stream-like functions - // insert an int into string - wxString& operator<<(int i) - { return (*this) << Format(_T("%d"), i); } - // insert an unsigned int into string - wxString& operator<<(unsigned int ui) - { return (*this) << Format(_T("%u"), ui); } - // insert a long into string - wxString& operator<<(long l) - { return (*this) << Format(_T("%ld"), l); } - // insert an unsigned long into string - wxString& operator<<(unsigned long ul) - { return (*this) << Format(_T("%lu"), ul); } - // insert a float into string - wxString& operator<<(float f) - { return (*this) << Format(_T("%f"), f); } - // insert a double into string - wxString& operator<<(double d) - { return (*this) << Format(_T("%g"), d); } + size_t PosFromImpl(size_t pos) const + { + if ( pos == 0 || pos == npos ) + return pos; + else + return const_iterator(this, m_impl.begin() + pos) - begin(); + } +#endif // !wxUSE_UNICODE_UTF8/wxUSE_UNICODE_UTF8 - // string comparison - // case-sensitive comparison (returns a value < 0, = 0 or > 0) - int Cmp(const wxChar *psz) const { return wxStrcmp(c_str(), psz); } - // same as Cmp() but not case-sensitive - int CmpNoCase(const wxChar *psz) const { return wxStricmp(c_str(), psz); } - // test for the string equality, either considering case or not - // (if compareWithCase then the case matters) - bool IsSameAs(const wxChar *psz, bool compareWithCase = TRUE) const - { return (compareWithCase ? Cmp(psz) : CmpNoCase(psz)) == 0; } - // comparison with a signle character: returns TRUE if equal - bool IsSameAs(wxChar c, bool compareWithCase = TRUE) const +public: + // standard types + typedef wxUniChar value_type; + typedef wxUniChar char_type; + typedef wxUniCharRef reference; + typedef wxChar* pointer; + typedef const wxChar* const_pointer; + + typedef size_t size_type; + typedef wxUniChar const_reference; + +#if wxUSE_STD_STRING + #if wxUSE_UNICODE_UTF8 + // random access is not O(1), as required by Random Access Iterator + #define WX_STR_ITERATOR_TAG std::bidirectional_iterator_tag + #else + #define WX_STR_ITERATOR_TAG std::random_access_iterator_tag + #endif + #define WX_DEFINE_ITERATOR_CATEGORY(cat) typedef cat iterator_category; +#else + // not defining iterator_category at all in this case is better than defining + // it as some dummy type -- at least it results in more intelligible error + // messages + #define WX_DEFINE_ITERATOR_CATEGORY(cat) +#endif + + #define WX_STR_ITERATOR_IMPL(iterator_name, pointer_type, reference_type) \ + private: \ + typedef wxStringImpl::iterator_name underlying_iterator; \ + public: \ + WX_DEFINE_ITERATOR_CATEGORY(WX_STR_ITERATOR_TAG) \ + typedef wxUniChar value_type; \ + typedef int difference_type; \ + typedef reference_type reference; \ + typedef pointer_type pointer; \ + \ + reference operator[](size_t n) const { return *(*this + n); } \ + \ + iterator_name& operator++() \ + { wxStringOperations::IncIter(m_cur); return *this; } \ + iterator_name& operator--() \ + { wxStringOperations::DecIter(m_cur); return *this; } \ + iterator_name operator++(int) \ + { \ + iterator_name tmp = *this; \ + wxStringOperations::IncIter(m_cur); \ + return tmp; \ + } \ + iterator_name operator--(int) \ + { \ + iterator_name tmp = *this; \ + wxStringOperations::DecIter(m_cur); \ + return tmp; \ + } \ + \ + iterator_name& operator+=(ptrdiff_t n) \ + { \ + m_cur = wxStringOperations::AddToIter(m_cur, n); \ + return *this; \ + } \ + iterator_name& operator-=(ptrdiff_t n) \ + { \ + m_cur = wxStringOperations::AddToIter(m_cur, -n); \ + return *this; \ + } \ + \ + difference_type operator-(const iterator_name& i) const \ + { return wxStringOperations::DiffIters(m_cur, i.m_cur); } \ + \ + bool operator==(const iterator_name& i) const \ + { return m_cur == i.m_cur; } \ + bool operator!=(const iterator_name& i) const \ + { return m_cur != i.m_cur; } \ + \ + bool operator<(const iterator_name& i) const \ + { return m_cur < i.m_cur; } \ + bool operator>(const iterator_name& i) const \ + { return m_cur > i.m_cur; } \ + bool operator<=(const iterator_name& i) const \ + { return m_cur <= i.m_cur; } \ + bool operator>=(const iterator_name& i) const \ + { return m_cur >= i.m_cur; } \ + \ + private: \ + /* for internal wxString use only: */ \ + underlying_iterator impl() const { return m_cur; } \ + \ + friend class wxString; \ + friend class wxCStrData; \ + \ + private: \ + underlying_iterator m_cur + + class WXDLLIMPEXP_FWD_BASE const_iterator; + +#if wxUSE_UNICODE_UTF8 + // 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) + + class WXDLLIMPEXP_BASE iterator + { + WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef); + + public: + 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(*str(), m_cur); } + + 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_node(str, &m_cur) {} + + wxString* str() const { return wx_const_cast(wxString*, m_node.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(); } + + iterator GetIterForNthChar(size_t n) + { return iterator(this, m_impl.begin() + PosToImpl(n)); } + const_iterator GetIterForNthChar(size_t n) const + { return const_iterator(this, m_impl.begin() + PosToImpl(n)); } +#else // !wxUSE_UNICODE_UTF8 + + class WXDLLIMPEXP_BASE iterator + { + WX_STR_ITERATOR_IMPL(iterator, wxChar*, wxUniCharRef); + + public: + iterator() {} + iterator(const iterator& i) : m_cur(i.m_cur) {} + + reference operator*() + { return wxUniCharRef::CreateForString(m_cur); } + + iterator operator+(ptrdiff_t n) const + { return iterator(wxStringOperations::AddToIter(m_cur, n)); } + iterator operator-(ptrdiff_t n) const + { return iterator(wxStringOperations::AddToIter(m_cur, -n)); } + + private: + // for internal wxString use only: + iterator(underlying_iterator ptr) : m_cur(ptr) {} + iterator(wxString *WXUNUSED(str), underlying_iterator ptr) : m_cur(ptr) {} + + 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) {} + const_iterator(const iterator& i) : m_cur(i.m_cur) {} + + 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-(ptrdiff_t n) const + { return const_iterator(wxStringOperations::AddToIter(m_cur, -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) {} + }; + + iterator GetIterForNthChar(size_t n) { return begin() + n; } + const_iterator GetIterForNthChar(size_t n) const { return begin() + n; } +#endif // wxUSE_UNICODE_UTF8/!wxUSE_UNICODE_UTF8 + + #undef WX_STR_ITERATOR_TAG + #undef WX_STR_ITERATOR_IMPL + + friend class iterator; + friend class const_iterator; + + template + class reverse_iterator_impl + { + public: + typedef T iterator_type; + + WX_DEFINE_ITERATOR_CATEGORY(typename T::iterator_category) + typedef typename T::value_type value_type; + typedef typename T::difference_type difference_type; + 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) {} + + iterator_type base() const { return m_cur; } + + reference operator*() const { return *(m_cur-1); } + reference operator[](size_t n) const { return *(*this + n); } + + reverse_iterator_impl& operator++() + { --m_cur; return *this; } + reverse_iterator_impl operator++(int) + { reverse_iterator_impl tmp = *this; --m_cur; return tmp; } + reverse_iterator_impl& operator--() + { ++m_cur; return *this; } + reverse_iterator_impl operator--(int) + { 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+(ptrdiff_t n) const + { return reverse_iterator_impl(m_cur - n); } + reverse_iterator_impl operator-(ptrdiff_t n) const + { return reverse_iterator_impl(m_cur + n); } + reverse_iterator_impl operator+=(ptrdiff_t n) + { m_cur -= n; return *this; } + reverse_iterator_impl operator-=(ptrdiff_t n) + { m_cur += n; return *this; } + + unsigned operator-(const reverse_iterator_impl& i) const + { return i.m_cur - m_cur; } + + bool operator==(const reverse_iterator_impl& ri) const + { return m_cur == ri.m_cur; } + bool operator!=(const reverse_iterator_impl& ri) const + { return !(*this == ri); } + + bool operator<(const reverse_iterator_impl& i) const + { return m_cur > i.m_cur; } + bool operator>(const reverse_iterator_impl& i) const + { return m_cur < i.m_cur; } + bool operator<=(const reverse_iterator_impl& i) const + { return m_cur >= i.m_cur; } + bool operator>=(const reverse_iterator_impl& i) const + { return m_cur <= i.m_cur; } + + private: + iterator_type m_cur; + }; + + typedef reverse_iterator_impl reverse_iterator; + typedef reverse_iterator_impl const_reverse_iterator; + +private: + // used to transform an expression built using c_str() (and hence of type + // wxCStrData) to an iterator into the string + static const_iterator CreateConstIterator(const wxCStrData& data) + { + 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 + // non-UTF8 locales, so we can't have and use wxString(wxStringImpl) ctor; + // instead we define dummy type that lets us have wxString ctor for creation + // from wxStringImpl that couldn't be used by user code (in all other builds, + // "standard" ctors can be used): +#if wxUSE_UNICODE_UTF8 && wxUSE_STL_BASED_WXSTRING + struct CtorFromStringImplTag {}; + + wxString(CtorFromStringImplTag* WXUNUSED(dummy), const wxStringImpl& src) + : m_impl(src) {} + + static wxString FromImpl(const wxStringImpl& src) + { return wxString((CtorFromStringImplTag*)NULL, src); } +#else + #if !wxUSE_STL_BASED_WXSTRING + wxString(const wxStringImpl& src) : m_impl(src) { } + // else: already defined as wxString(wxStdString) below + #endif + static wxString FromImpl(const wxStringImpl& src) { return wxString(src); } +#endif + +public: + // constructors and destructor + // ctor for an empty string + wxString() {} + + // copy ctor + wxString(const wxString& stringSrc) : m_impl(stringSrc.m_impl) { } + + // string containing nRepeat copies of ch + wxString(wxUniChar ch, size_t nRepeat = 1 ) + { assign(nRepeat, ch); } + wxString(size_t nRepeat, wxUniChar ch) + { assign(nRepeat, ch); } + wxString(wxUniCharRef ch, size_t nRepeat = 1) + { assign(nRepeat, ch); } + wxString(size_t nRepeat, wxUniCharRef ch) + { assign(nRepeat, ch); } + wxString(char ch, size_t nRepeat = 1) + { assign(nRepeat, ch); } + wxString(size_t nRepeat, char ch) + { assign(nRepeat, ch); } + wxString(wchar_t ch, size_t nRepeat = 1) + { assign(nRepeat, ch); } + wxString(size_t nRepeat, wchar_t ch) + { assign(nRepeat, ch); } + + // ctors from char* strings: + wxString(const char *psz) + : m_impl(ImplStr(psz)) {} + wxString(const char *psz, const wxMBConv& conv) + : m_impl(ImplStr(psz, conv)) {} + wxString(const char *psz, size_t nLength) + { assign(psz, nLength); } + wxString(const char *psz, const wxMBConv& conv, size_t nLength) + { + SubstrBufFromMB str(ImplStr(psz, nLength, conv)); + m_impl.assign(str.data, str.len); + } + + // and unsigned char*: + wxString(const unsigned char *psz) + : m_impl(ImplStr((const char*)psz)) {} + wxString(const unsigned char *psz, const wxMBConv& conv) + : m_impl(ImplStr((const char*)psz, conv)) {} + wxString(const unsigned char *psz, size_t nLength) + { assign((const char*)psz, nLength); } + wxString(const unsigned char *psz, const wxMBConv& conv, size_t nLength) + { + SubstrBufFromMB str(ImplStr((const char*)psz, nLength, conv)); + m_impl.assign(str.data, str.len); + } + + // ctors from wchar_t* strings: + wxString(const wchar_t *pwz) + : m_impl(ImplStr(pwz)) {} + wxString(const wchar_t *pwz, const wxMBConv& WXUNUSED(conv)) + : m_impl(ImplStr(pwz)) {} + wxString(const wchar_t *pwz, size_t nLength) + { assign(pwz, nLength); } + wxString(const wchar_t *pwz, const wxMBConv& WXUNUSED(conv), size_t nLength) + { assign(pwz, nLength); } + + wxString(const wxCharBuffer& buf) + { assign(buf.data()); } // FIXME-UTF8: fix for embedded NUL and buffer length + wxString(const wxWCharBuffer& buf) + { assign(buf.data()); } // FIXME-UTF8: fix for embedded NUL and buffer length + + // NB: this version uses m_impl.c_str() to force making a copy of the + // string, so that "wxString(str.c_str())" idiom for passing strings + // between threads works + wxString(const wxCStrData& cstr) + : m_impl(cstr.AsString().m_impl.c_str()) { } + + // as we provide both ctors with this signature for both char and unsigned + // char string, we need to provide one for wxCStrData to resolve ambiguity + wxString(const wxCStrData& cstr, size_t nLength) + : m_impl(cstr.AsString().Mid(0, nLength).m_impl) {} + + // and because wxString is convertible to wxCStrData and const wxChar * + // we also need to provide this one + wxString(const wxString& str, size_t nLength) + { assign(str, nLength); } + + +#if wxUSE_STRING_POS_CACHE + ~wxString() + { + // we need to invalidate our cache entry as another string could be + // recreated at the same address (unlikely, but still possible, with the + // heap-allocated strings but perfectly common with stack-allocated ones) + InvalidateCache(); + } +#endif // wxUSE_STRING_POS_CACHE + + // even if we're not built with wxUSE_STL == 1 it is very convenient to allow + // implicit conversions from std::string to wxString and vice verse as this + // allows to use the same strings in non-GUI and GUI code, however we don't + // want to unconditionally add this ctor as it would make wx lib dependent on + // libstdc++ on some Linux versions which is bad, so instead we ask the + // client code to define this wxUSE_STD_STRING symbol if they need it +#if wxUSE_STD_STRING + #if wxUSE_UNICODE_WCHAR + wxString(const wxStdWideString& str) : m_impl(str) {} + #else // UTF-8 or ANSI + wxString(const wxStdWideString& str) + { assign(str.c_str(), str.length()); } + #endif + + #if !wxUSE_UNICODE // ANSI build + // FIXME-UTF8: do this in UTF8 build #if wxUSE_UTF8_LOCALE_ONLY, too + wxString(const std::string& str) : m_impl(str) {} + #else // Unicode + 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; } + #else + // wxStringImpl is either not std::string or needs conversion + operator wxStdWideString() const + // FIXME-UTF8: broken for embedded NULs + { return wxStdWideString(wc_str()); } + #endif + + #if (!wxUSE_UNICODE || wxUSE_UTF8_LOCALE_ONLY) && wxUSE_STL_BASED_WXSTRING + // wxStringImpl is std::string in the encoding we want + operator const std::string&() const { return m_impl; } + #else + // wxStringImpl is either not std::string or needs conversion + operator std::string() const + // FIXME-UTF8: broken for embedded NULs + { return std::string(mb_str()); } + #endif +#endif // wxUSE_STL + + wxString Clone() const + { + // make a deep copy of the string, i.e. the returned string will have + // ref count = 1 with refcounted implementation + return wxString::FromImpl(wxStringImpl(m_impl.c_str(), m_impl.length())); + } + + // first valid index position + 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(this, m_impl.end()); } + iterator end() { return iterator(this, m_impl.end()); } + + // first element of the reversed string + const_reverse_iterator rbegin() const + { return const_reverse_iterator(end()); } + reverse_iterator rbegin() + { return reverse_iterator(end()); } + // one beyond the end of the reversed string + const_reverse_iterator rend() const + { return const_reverse_iterator(begin()); } + reverse_iterator rend() + { return reverse_iterator(begin()); } + + // std::string methods: +#if wxUSE_UNICODE_UTF8 + size_t length() const + { +#if wxUSE_STRING_POS_CACHE + wxCACHE_PROFILE_FIELD_INC(lentot); + + Cache::Element * const cache = GetCacheElement(); + + if ( cache->len == npos ) + { + // it's probably not worth trying to be clever and using cache->pos + // here as it's probably 0 anyhow -- you usually call length() before + // starting to index the string + cache->len = end() - begin(); + } + else + { + wxCACHE_PROFILE_FIELD_INC(lenhits); + + wxSTRING_CACHE_ASSERT( (int)cache->len == end() - begin() ); + } + + return cache->len; +#else // !wxUSE_STRING_POS_CACHE + return end() - begin(); +#endif // wxUSE_STRING_POS_CACHE/!wxUSE_STRING_POS_CACHE + } +#else + size_t length() const { return m_impl.length(); } +#endif + + size_type size() const { return length(); } + size_type max_size() const { return npos; } + + bool empty() const { return m_impl.empty(); } + + // NB: these methods don't have a well-defined meaning in UTF-8 case + size_type capacity() const { return m_impl.capacity(); } + void reserve(size_t sz) { m_impl.reserve(sz); } + + void resize(size_t nSize, wxUniChar ch = wxT('\0')) + { + const size_t len = length(); + if ( nSize == len) + return; + +#if wxUSE_UNICODE_UTF8 + if ( nSize < len ) { - return (Len() == 1) && (compareWithCase ? GetChar(0u) == c - : wxToupper(GetChar(0u)) == wxToupper(c)); + wxSTRING_INVALIDATE_CACHE(); + + // we can't use wxStringImpl::resize() for truncating the string as it + // counts in bytes, not characters + erase(nSize); + return; } - // simple sub-string extraction - // return substring starting at nFirst of length nCount (or till the end - // if nCount = default value) - wxString Mid(size_t nFirst, size_t nCount = wxSTRING_MAXLEN) const; + // 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 // can use (presumably faster) resize() version +#endif // wxUSE_UNICODE_UTF8 + { + wxSTRING_INVALIDATE_CACHED_LENGTH(); - // operator version of Mid() - wxString operator()(size_t start, size_t len) const - { return Mid(start, len); } + m_impl.resize(nSize, (wxStringCharType)ch); + } + } - // check that the string starts with prefix and return the rest of the - // string in the provided pointer if it is not NULL, otherwise return - // FALSE - bool StartsWith(const wxChar *prefix, wxString *rest = NULL) const; + wxString substr(size_t nStart = 0, size_t nLen = npos) const + { + size_t pos, len; + PosLenToImpl(nStart, nLen, &pos, &len); + return FromImpl(m_impl.substr(pos, len)); + } - // get first nCount characters - wxString Left(size_t nCount) const; - // get last nCount characters - wxString Right(size_t nCount) const; - // get all characters before the first occurance of ch - // (returns the whole string if ch not found) - wxString BeforeFirst(wxChar ch) const; - // get all characters before the last occurence of ch - // (returns empty string if ch not found) - wxString BeforeLast(wxChar ch) const; - // get all characters after the first occurence of ch - // (returns empty string if ch not found) - wxString AfterFirst(wxChar ch) const; - // get all characters after the last occurence of ch - // (returns the whole string if ch not found) - wxString AfterLast(wxChar ch) const; + // generic attributes & operations + // as standard strlen() + size_t Len() const { return length(); } + // string contains any characters? + bool IsEmpty() const { return empty(); } + // empty string is "false", so !str will return true + bool operator!() const { return empty(); } + // truncate the string to given length + wxString& Truncate(size_t uiLen); + // empty string contents + void Empty() { clear(); } + // empty the string and free memory + void Clear() { clear(); } - // for compatibility only, use more explicitly named functions above - wxString Before(wxChar ch) const { return BeforeLast(ch); } - wxString After(wxChar ch) const { return AfterFirst(ch); } + // contents test + // Is an ascii value + bool IsAscii() const; + // Is a number + bool IsNumber() const; + // Is a word + bool IsWord() const; - // case conversion - // convert to upper case in place, return the string itself - wxString& MakeUpper(); - // convert to upper case, return the copy of the string - // Here's something to remember: BC++ doesn't like returns in inlines. - wxString Upper() const ; - // convert to lower case in place, return the string itself - wxString& MakeLower(); - // convert to lower case, return the copy of the string - wxString Lower() const ; + // data access (all indexes are 0 based) + // read access + wxUniChar at(size_t n) const + { return wxStringOperations::DecodeChar(m_impl.begin() + PosToImpl(n)); } + wxUniChar GetChar(size_t n) const + { return at(n); } + // read/write access + wxUniCharRef at(size_t n) + { return *GetIterForNthChar(n); } + wxUniCharRef GetWritableChar(size_t n) + { return at(n); } + // write access + void SetChar(size_t n, wxUniChar ch) + { at(n) = ch; } - // trimming/padding whitespace (either side) and truncating - // remove spaces from left or from right (default) side - wxString& Trim(bool bFromRight = TRUE); - // add nCount copies chPad in the beginning or at the end (default) - wxString& Pad(size_t nCount, wxChar chPad = wxT(' '), bool bFromRight = TRUE); + // get last character + wxUniChar Last() const + { + wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") ); + return *rbegin(); + } - // searching and replacing - // searching (return starting index, or -1 if not found) - int Find(wxChar ch, bool bFromEnd = FALSE) const; // like strchr/strrchr - // searching (return starting index, or -1 if not found) - int Find(const wxChar *pszSub) const; // like strstr - // replace first (or all of bReplaceAll) occurences of substring with - // another string, returns the number of replacements made - size_t Replace(const wxChar *szOld, - const wxChar *szNew, - bool bReplaceAll = TRUE); + // get writable last character + wxUniCharRef Last() + { + wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") ); + return *rbegin(); + } - // check if the string contents matches a mask containing '*' and '?' - bool Matches(const wxChar *szMask) const; + /* + Note that we we must define all of the overloads below to avoid + ambiguity when using str[0]. + */ + wxUniChar operator[](int n) const + { return at(n); } + wxUniChar operator[](long n) const + { return at(n); } + wxUniChar operator[](size_t n) const + { return at(n); } +#ifndef wxSIZE_T_IS_UINT + wxUniChar operator[](unsigned int n) const + { return at(n); } +#endif // size_t != unsigned int - // conversion to numbers: all functions return TRUE only if the whole - // string is a number and put the value of this number into the pointer - // provided, the base is the numeric base in which the conversion should be - // done and must be comprised between 2 and 36 or be 0 in which case the - // standard C rules apply (leading '0' => octal, "0x" => hex) - // convert to a signed integer - bool ToLong(long *val, int base = 10) const; - // convert to an unsigned integer - bool ToULong(unsigned long *val, int base = 10) const; - // convert to a double - bool ToDouble(double *val) const; + // operator versions of GetWriteableChar() + wxUniCharRef operator[](int n) + { return at(n); } + wxUniCharRef operator[](long n) + { return at(n); } + wxUniCharRef operator[](size_t n) + { return at(n); } +#ifndef wxSIZE_T_IS_UINT + wxUniCharRef operator[](unsigned int n) + { return at(n); } +#endif // size_t != unsigned int - // formated input/output - // as sprintf(), returns the number of characters written or < 0 on error - // (take 'this' into account in attribute parameter count) - int Printf(const wxChar *pszFormat, ...) ATTRIBUTE_PRINTF_2; - // as vprintf(), returns the number of characters written or < 0 on error - int PrintfV(const wxChar* pszFormat, va_list argptr); - // returns the string containing the result of Printf() to it - static wxString Format(const wxChar *pszFormat, ...) ATTRIBUTE_PRINTF_1; - // the same as above, but takes a va_list - static wxString FormatV(const wxChar *pszFormat, va_list argptr); + /* + Overview of wxString conversions, implicit and explicit: + + - wxString has a std::[w]string-like c_str() method, however it does + not return a C-style string directly but instead returns wxCStrData + helper object which is convertible to either "char *" narrow string + or "wchar_t *" wide string. Usually the correct conversion will be + applied by the compiler automatically but if this doesn't happen you + need to explicitly choose one using wxCStrData::AsChar() or AsWChar() + methods or another wxString conversion function. + + - One of the places where the conversion does *NOT* happen correctly is + when c_str() is passed to a vararg function such as printf() so you + must *NOT* use c_str() with them. Either use wxPrintf() (all wx + functions do handle c_str() correctly, even if they appear to be + vararg (but they're not, really)) or add an explicit AsChar() or, if + compatibility with previous wxWidgets versions is important, add a + cast to "const char *". + + - In non-STL mode only, wxString is also implicitly convertible to + wxCStrData. The same warning as above applies. + + - c_str() is polymorphic as it can be converted to either narrow or + wide string. If you explicitly need one or the other, choose to use + mb_str() (for narrow) or wc_str() (for wide) instead. Notice that + these functions can return either the pointer to string directly (if + this is what the string uses internally) or a temporary buffer + containing the string and convertible to it. Again, conversion will + usually be done automatically by the compiler but beware of the + vararg functions: you need an explicit cast when using them. + + - There are also non-const versions of mb_str() and wc_str() called + char_str() and wchar_str(). They are only meant to be used with + non-const-correct functions and they always return buffers. + + - Finally wx_str() returns whatever string representation is used by + wxString internally. It may be either a narrow or wide string + depending on wxWidgets build mode but it will always be a raw pointer + (and not a buffer). + */ - // raw access to string memory - // ensure that string has space for at least nLen characters - // only works if the data of this string is not shared - bool Alloc(size_t nLen); - // minimize the string's memory - // only works if the data of this string is not shared - bool Shrink(); - // get writable buffer of at least nLen bytes. Unget() *must* be called - // a.s.a.p. to put string back in a reasonable state! - wxChar *GetWriteBuf(size_t nLen); - // call this immediately after GetWriteBuf() has been used - void UngetWriteBuf(); - void UngetWriteBuf(size_t nLen); + // explicit conversion to wxCStrData + wxCStrData c_str() const { return wxCStrData(this); } + wxCStrData data() const { return c_str(); } - // wxWindows version 1 compatibility functions + // implicit conversion to wxCStrData + operator wxCStrData() const { return c_str(); } - // use Mid() - wxString SubString(size_t from, size_t to) const - { return Mid(from, (to - from + 1)); } - // values for second parameter of CompareTo function - enum caseCompare {exact, ignoreCase}; - // values for first parameter of Strip function - enum stripType {leading = 0x1, trailing = 0x2, both = 0x3}; + // 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(); } - // use Printf() - // (take 'this' into account in attribute parameter count) - int sprintf(const wxChar *pszFormat, ...) ATTRIBUTE_PRINTF_2; + // implicit conversion to untyped pointer for compatibility with previous + // 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 - // use Cmp() - inline int CompareTo(const wxChar* psz, caseCompare cmp = exact) const - { return cmp == exact ? Cmp(psz) : CmpNoCase(psz); } + // identical to c_str(), for MFC compatibility + const wxCStrData GetData() const { return c_str(); } - // use Len - size_t Length() const { return Len(); } - // Count the number of characters - int Freq(wxChar ch) const; - // use MakeLower - void LowerCase() { MakeLower(); } - // use MakeUpper - void UpperCase() { MakeUpper(); } - // use Trim except that it doesn't change this string - wxString Strip(stripType w = trailing) const; + // explicit conversion to C string in internal representation (char*, + // wchar_t*, UTF-8-encoded char*, depending on the build): + const wxStringCharType *wx_str() const { return m_impl.c_str(); } - // use Find (more general variants not yet supported) - size_t Index(const wxChar* psz) const { return Find(psz); } - size_t Index(wxChar ch) const { return Find(ch); } - // use Truncate - wxString& Remove(size_t pos) { return Truncate(pos); } - wxString& RemoveLast(size_t n = 1) { return Truncate(Len() - n); } + // conversion to *non-const* multibyte or widestring buffer; modifying + // returned buffer won't affect the string, these methods are only useful + // for passing values to const-incorrect functions + wxWritableCharBuffer char_str(const wxMBConv& conv = wxConvLibc) const + { return mb_str(conv); } + wxWritableWCharBuffer wchar_str() const { return wc_str(); } + + // conversion to the buffer of the given type T (= char or wchar_t) and + // also optionally return the buffer length + // + // this is mostly/only useful for the template functions + // + // FIXME-VC6: the second argument only exists for VC6 which doesn't support + // explicit template function selection, do not use it unless + // you must support VC6! + template + wxCharTypeBuffer tchar_str(size_t *len = NULL, + T * WXUNUSED(dummy) = NULL) const + { +#if wxUSE_UNICODE + // we need a helper dispatcher depending on type + return wxPrivate::wxStringAsBufHelper::Get(*this, len); +#else // ANSI + // T can only be char in ANSI build + if ( len ) + *len = length(); + + return wxCharTypeBuffer::CreateNonOwned(wx_str()); +#endif // Unicode build kind + } + + // conversion to/from plain (i.e. 7 bit) ASCII: this is useful for + // converting numbers or strings which are certain not to contain special + // chars (typically system functions, X atoms, environment variables etc.) + // + // 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, 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, 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 FromUTF8Unchecked(const char *utf8) + { + if ( !utf8 ) + return wxEmptyString; + + wxASSERT( wxStringOperations::IsValidUtf8String(utf8) ); + return FromImpl(wxStringImpl(utf8)); + } + static wxString FromUTF8Unchecked(const char *utf8, size_t len) + { + if ( !utf8 ) + return wxEmptyString; + if ( len == npos ) + return FromUTF8Unchecked(utf8); + + wxASSERT( wxStringOperations::IsValidUtf8String(utf8, len) ); + return FromImpl(wxStringImpl(utf8, len)); + } + + static wxString FromUTF8(const char *utf8) + { + if ( !utf8 || !wxStringOperations::IsValidUtf8String(utf8) ) + return ""; + + return FromImpl(wxStringImpl(utf8)); + } + static wxString FromUTF8(const char *utf8, size_t len) + { + if ( len == npos ) + return FromUTF8(utf8); + + if ( !utf8 || !wxStringOperations::IsValidUtf8String(utf8, len) ) + return ""; + + return FromImpl(wxStringImpl(utf8, len)); + } + + const char* utf8_str() const { return wx_str(); } + const char* ToUTF8() const { return wx_str(); } + + // this function exists in UTF-8 build only and returns the length of the + // internal UTF-8 representation + size_t utf8_length() const { return m_impl.length(); } +#elif wxUSE_UNICODE_WCHAR + static wxString FromUTF8(const char *utf8, size_t len = npos) + { return wxString(utf8, wxMBConvUTF8(), len); } + static wxString FromUTF8Unchecked(const char *utf8, size_t len = npos) + { + const wxString s(utf8, wxMBConvUTF8(), len); + wxASSERT_MSG( !utf8 || !*utf8 || !s.empty(), + "string must be valid UTF-8" ); + return s; + } + 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(wxMBConvUTF8().cMB2WC(utf8)); } + static wxString FromUTF8(const char *utf8, size_t len) + { + size_t wlen; + wxWCharBuffer buf(wxMBConvUTF8().cMB2WC(utf8, len == npos ? wxNO_LEN : len, &wlen)); + return wxString(buf.data(), wlen); + } + static wxString FromUTF8Unchecked(const char *utf8, size_t len = npos) + { + size_t wlen; + wxWCharBuffer buf(wxMBConvUTF8().cMB2WC(utf8, + len == npos ? wxNO_LEN : len, + &wlen)); + wxASSERT_MSG( !utf8 || !*utf8 || wlen, + "string must be valid UTF-8" ); + + return wxString(buf.data(), wlen); + } + const wxCharBuffer utf8_str() const + { return wxMBConvUTF8().cWC2MB(wc_str()); } + const wxCharBuffer ToUTF8() const { return utf8_str(); } +#endif + + // functions for storing binary data in wxString: +#if wxUSE_UNICODE + static wxString From8BitData(const char *data, size_t len) + { return wxString(data, wxConvISO8859_1, len); } + // version for NUL-terminated data: + static wxString From8BitData(const char *data) + { return wxString(data, wxConvISO8859_1); } + const wxCharBuffer To8BitData() const { return mb_str(wxConvISO8859_1); } +#else // ANSI + static wxString From8BitData(const char *data, size_t len) + { return wxString(data, len); } + // version for NUL-terminated data: + static wxString From8BitData(const char *data) + { return wxString(data); } + const char *To8BitData() const { return c_str(); } +#endif // Unicode/ANSI + + // conversions with (possible) format conversions: have to return a + // buffer with temporary data + // + // the functions defined (in either Unicode or ANSI) mode are mb_str() to + // return an ANSI (multibyte) string, wc_str() to return a wide string and + // fn_str() to return a string which should be used with the OS APIs + // accepting the file names. The return value is always the same, but the + // type differs because a function may either return pointer to the buffer + // directly or have to use intermediate buffer for translation. +#if wxUSE_UNICODE + +#if wxUSE_UTF8_LOCALE_ONLY + const char* mb_str() const { return wx_str(); } + const wxCharBuffer mb_str(const wxMBConv& conv) const; +#else + const wxCharBuffer mb_str(const wxMBConv& conv = wxConvLibc) const; +#endif + + const wxWX2MBbuf mbc_str() const { return mb_str(*wxConvCurrent); } + +#if wxUSE_UNICODE_WCHAR + const wchar_t* wc_str() const { return wx_str(); } +#elif wxUSE_UNICODE_UTF8 + const wxWCharBuffer wc_str() const; +#endif + // for compatibility with !wxUSE_UNICODE version + const wxWX2WCbuf wc_str(const wxMBConv& WXUNUSED(conv)) const + { return wc_str(); } + +#if wxMBFILES + const wxCharBuffer fn_str() const { return mb_str(wxConvFile); } +#else // !wxMBFILES + const wxWX2WCbuf fn_str() const { return wc_str(); } +#endif // wxMBFILES/!wxMBFILES + +#else // ANSI + const wxChar* mb_str() const { return wx_str(); } + + // for compatibility with wxUSE_UNICODE version + const char* mb_str(const wxMBConv& WXUNUSED(conv)) const { return wx_str(); } + + const wxWX2MBbuf mbc_str() const { return mb_str(); } + +#if wxUSE_WCHAR_T + const wxWCharBuffer wc_str(const wxMBConv& conv = wxConvLibc) const; +#endif // wxUSE_WCHAR_T + const wxCharBuffer fn_str() const { return wxConvFile.cWC2WX( wc_str( wxConvLibc ) ); } +#endif // Unicode/ANSI + +#if wxUSE_UNICODE_UTF8 + const wxWCharBuffer t_str() const { return wc_str(); } +#elif wxUSE_UNICODE_WCHAR + const wchar_t* t_str() const { return wx_str(); } +#else + const char* t_str() const { return wx_str(); } +#endif + + + // overloaded assignment + // from another wxString + wxString& operator=(const wxString& stringSrc) + { + if ( this != &stringSrc ) + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl = stringSrc.m_impl; + } + + return *this; + } + + wxString& operator=(const wxCStrData& cstr) + { return *this = cstr.AsString(); } + // from a character + wxString& operator=(wxUniChar ch) + { + wxSTRING_INVALIDATE_CACHE(); + +#if wxUSE_UNICODE_UTF8 + if ( !ch.IsAscii() ) + m_impl = wxStringOperations::EncodeChar(ch); + else +#endif // wxUSE_UNICODE_UTF8 + m_impl = (wxStringCharType)ch; + return *this; + } + + wxString& operator=(wxUniCharRef ch) + { return operator=((wxUniChar)ch); } + wxString& operator=(char ch) + { return operator=(wxUniChar(ch)); } + wxString& operator=(unsigned char ch) + { return operator=(wxUniChar(ch)); } + wxString& operator=(wchar_t ch) + { return operator=(wxUniChar(ch)); } + // from a C string - STL probably will crash on NULL, + // so we need to compensate in that case +#if wxUSE_STL_BASED_WXSTRING + wxString& operator=(const char *psz) + { + wxSTRING_INVALIDATE_CACHE(); + + if ( psz ) + m_impl = ImplStr(psz); + else + clear(); + + return *this; + } + + wxString& operator=(const wchar_t *pwz) + { + wxSTRING_INVALIDATE_CACHE(); + + if ( pwz ) + m_impl = ImplStr(pwz); + else + clear(); + + return *this; + } +#else // !wxUSE_STL_BASED_WXSTRING + wxString& operator=(const char *psz) + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl = ImplStr(psz); + + return *this; + } + + wxString& operator=(const wchar_t *pwz) + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl = ImplStr(pwz); + + return *this; + } +#endif // wxUSE_STL_BASED_WXSTRING/!wxUSE_STL_BASED_WXSTRING + + wxString& operator=(const unsigned char *psz) + { return operator=((const char*)psz); } + + // from wxWCharBuffer + wxString& operator=(const wxWCharBuffer& s) + { return operator=(s.data()); } // FIXME-UTF8: fix for embedded NULs + // from wxCharBuffer + wxString& operator=(const wxCharBuffer& s) + { return operator=(s.data()); } // FIXME-UTF8: fix for embedded NULs + + // string concatenation + // in place concatenation + /* + Concatenate and return the result. Note that the left to right + associativity of << allows to write things like "str << str1 << str2 + << ..." (unlike with +=) + */ + // string += string + wxString& operator<<(const wxString& s) + { +#if WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8 + wxASSERT_MSG( s.IsValid(), + _T("did you forget to call UngetWriteBuf()?") ); +#endif + + append(s); + return *this; + } + // string += C string + wxString& operator<<(const char *psz) + { append(psz); return *this; } + wxString& operator<<(const wchar_t *pwz) + { append(pwz); return *this; } + wxString& operator<<(const wxCStrData& psz) + { append(psz.AsString()); return *this; } + // string += char + wxString& operator<<(wxUniChar ch) { append(1, ch); return *this; } + wxString& operator<<(wxUniCharRef ch) { append(1, ch); return *this; } + wxString& operator<<(char ch) { append(1, ch); return *this; } + wxString& operator<<(unsigned char ch) { append(1, ch); return *this; } + wxString& operator<<(wchar_t ch) { append(1, ch); return *this; } + + // string += buffer (i.e. from wxGetString) + wxString& operator<<(const wxWCharBuffer& s) + { return operator<<((const wchar_t *)s); } + wxString& operator<<(const wxCharBuffer& s) + { return operator<<((const char *)s); } + + // string += C string + wxString& Append(const wxString& s) + { + // test for empty() to share the string if possible + if ( empty() ) + *this = s; + else + append(s); + return *this; + } + wxString& Append(const char* psz) + { append(psz); return *this; } + wxString& Append(const wchar_t* pwz) + { append(pwz); return *this; } + wxString& Append(const wxCStrData& psz) + { append(psz); return *this; } + wxString& Append(const wxCharBuffer& psz) + { append(psz); return *this; } + wxString& Append(const wxWCharBuffer& psz) + { append(psz); return *this; } + wxString& Append(const char* psz, size_t nLen) + { append(psz, nLen); return *this; } + wxString& Append(const wchar_t* pwz, size_t nLen) + { append(pwz, nLen); return *this; } + wxString& Append(const wxCStrData& psz, size_t nLen) + { append(psz, nLen); return *this; } + wxString& Append(const wxCharBuffer& psz, size_t nLen) + { append(psz, nLen); return *this; } + wxString& Append(const wxWCharBuffer& psz, size_t nLen) + { append(psz, nLen); return *this; } + // append count copies of given character + wxString& Append(wxUniChar ch, size_t count = 1u) + { append(count, ch); return *this; } + wxString& Append(wxUniCharRef ch, size_t count = 1u) + { append(count, ch); return *this; } + wxString& Append(char ch, size_t count = 1u) + { append(count, ch); return *this; } + wxString& Append(unsigned char ch, size_t count = 1u) + { append(count, ch); return *this; } + wxString& Append(wchar_t ch, size_t count = 1u) + { append(count, ch); return *this; } + + // prepend a string, return the string itself + wxString& Prepend(const wxString& str) + { *this = str + *this; return *this; } + + // non-destructive concatenation + // two strings + friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string1, + const wxString& string2); + // string with a single char + friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string, wxUniChar ch); + // char with a string + friend wxString WXDLLIMPEXP_BASE operator+(wxUniChar ch, const wxString& string); + // string with C string + friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string, + const char *psz); + friend wxString WXDLLIMPEXP_BASE operator+(const wxString& string, + const wchar_t *pwz); + // C string with string + friend wxString WXDLLIMPEXP_BASE operator+(const char *psz, + const wxString& string); + friend wxString WXDLLIMPEXP_BASE operator+(const wchar_t *pwz, + const wxString& string); + + // stream-like functions + // insert an int into string + wxString& operator<<(int i) + { return (*this) << Format(_T("%d"), i); } + // insert an unsigned int into string + wxString& operator<<(unsigned int ui) + { return (*this) << Format(_T("%u"), ui); } + // insert a long into string + wxString& operator<<(long l) + { return (*this) << Format(_T("%ld"), l); } + // insert an unsigned long into string + wxString& operator<<(unsigned long ul) + { return (*this) << Format(_T("%lu"), ul); } +#if defined wxLongLong_t && !defined wxLongLongIsLong + // insert a long long if they exist and aren't longs + wxString& operator<<(wxLongLong_t ll) + { + const wxChar *fmt = _T("%") wxLongLongFmtSpec _T("d"); + return (*this) << Format(fmt, ll); + } + // insert an unsigned long long + wxString& operator<<(wxULongLong_t ull) + { + const wxChar *fmt = _T("%") wxLongLongFmtSpec _T("u"); + return (*this) << Format(fmt , ull); + } +#endif // wxLongLong_t && !wxLongLongIsLong + // insert a float into string + wxString& operator<<(float f) + { return (*this) << Format(_T("%f"), f); } + // insert a double into string + wxString& operator<<(double d) + { return (*this) << Format(_T("%g"), d); } + + // string comparison + // case-sensitive comparison (returns a value < 0, = 0 or > 0) + int Cmp(const char *psz) const + { return compare(psz); } + int Cmp(const wchar_t *pwz) const + { return compare(pwz); } + int Cmp(const wxString& s) const + { return compare(s); } + int Cmp(const wxCStrData& s) const + { return compare(s); } + int Cmp(const wxCharBuffer& s) const + { return compare(s); } + int Cmp(const wxWCharBuffer& s) const + { 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 + { +#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 + { return IsSameAs(str.data(), compareWithCase); } + 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; + // FIXME-UTF8: remove these overloads + bool IsSameAs(wxUniCharRef c, bool compareWithCase = true) const + { return IsSameAs(wxUniChar(c), compareWithCase); } + bool IsSameAs(char c, bool compareWithCase = true) const + { return IsSameAs(wxUniChar(c), compareWithCase); } + bool IsSameAs(unsigned char c, bool compareWithCase = true) const + { return IsSameAs(wxUniChar(c), compareWithCase); } + bool IsSameAs(wchar_t c, bool compareWithCase = true) const + { return IsSameAs(wxUniChar(c), compareWithCase); } + bool IsSameAs(int c, bool compareWithCase = true) const + { return IsSameAs(wxUniChar(c), compareWithCase); } + + // simple sub-string extraction + // return substring starting at nFirst of length nCount (or till the end + // if nCount = default value) + wxString Mid(size_t nFirst, size_t nCount = npos) const; + + // operator version of Mid() + wxString operator()(size_t start, size_t len) const + { return Mid(start, len); } + + // check if the string starts with the given prefix and return the rest + // of the string in the provided pointer if it is not NULL; otherwise + // return false + bool StartsWith(const wxString& prefix, wxString *rest = NULL) const; + // check if the string ends with the given suffix and return the + // beginning of the string before the suffix in the provided pointer if + // it is not NULL; otherwise return false + bool EndsWith(const wxString& suffix, wxString *rest = NULL) const; + + // get first nCount characters + wxString Left(size_t nCount) const; + // get last nCount characters + wxString Right(size_t nCount) const; + // get all characters before the first occurrence of ch + // (returns the whole string if ch not found) + wxString BeforeFirst(wxUniChar ch) const; + // get all characters before the last occurrence of ch + // (returns empty string if ch not found) + wxString BeforeLast(wxUniChar ch) const; + // get all characters after the first occurrence of ch + // (returns empty string if ch not found) + wxString AfterFirst(wxUniChar ch) const; + // get all characters after the last occurrence of ch + // (returns the whole string if ch not found) + wxString AfterLast(wxUniChar ch) const; + + // for compatibility only, use more explicitly named functions above + wxString Before(wxUniChar ch) const { return BeforeLast(ch); } + wxString After(wxUniChar ch) const { return AfterFirst(ch); } + + // case conversion + // convert to upper case in place, return the string itself + wxString& MakeUpper(); + // convert to upper case, return the copy of the string + wxString Upper() const { return wxString(*this).MakeUpper(); } + // convert to lower case in place, return the string itself + wxString& MakeLower(); + // convert to lower case, return the copy of the string + wxString Lower() const { return wxString(*this).MakeLower(); } + // convert the first character to the upper case and the rest to the + // lower one, return the modified string itself + wxString& MakeCapitalized(); + // convert the first character to the upper case and the rest to the + // lower one, return the copy of the string + wxString Capitalize() const { return wxString(*this).MakeCapitalized(); } + + // trimming/padding whitespace (either side) and truncating + // remove spaces from left or from right (default) side + wxString& Trim(bool bFromRight = true); + // add nCount copies chPad in the beginning or at the end (default) + wxString& Pad(size_t nCount, wxUniChar chPad = wxT(' '), bool bFromRight = true); + + // searching and replacing + // searching (return starting index, or -1 if not found) + int Find(wxUniChar ch, bool bFromEnd = false) const; // like strchr/strrchr + int Find(wxUniCharRef ch, bool bFromEnd = false) const + { return Find(wxUniChar(ch), bFromEnd); } + int Find(char ch, bool bFromEnd = false) const + { return Find(wxUniChar(ch), bFromEnd); } + int Find(unsigned char ch, bool bFromEnd = false) const + { return Find(wxUniChar(ch), bFromEnd); } + int Find(wchar_t ch, bool bFromEnd = false) const + { return Find(wxUniChar(ch), bFromEnd); } + // searching (return starting index, or -1 if not found) + int Find(const wxString& sub) const // like strstr + { + size_type idx = find(sub); + return (idx == npos) ? wxNOT_FOUND : (int)idx; + } + int Find(const char *sub) const // like strstr + { + size_type idx = find(sub); + return (idx == npos) ? wxNOT_FOUND : (int)idx; + } + int Find(const wchar_t *sub) const // like strstr + { + size_type idx = find(sub); + return (idx == npos) ? wxNOT_FOUND : (int)idx; + } + + int Find(const wxCStrData& sub) const + { return Find(sub.AsString()); } + int Find(const wxCharBuffer& sub) const + { return Find(sub.data()); } + int Find(const wxWCharBuffer& sub) const + { return Find(sub.data()); } + + // replace first (or all of bReplaceAll) occurrences of substring with + // another string, returns the number of replacements made + size_t Replace(const wxString& strOld, + const wxString& strNew, + bool bReplaceAll = true); + + // check if the string contents matches a mask containing '*' and '?' + bool Matches(const wxString& mask) const; + + // conversion to numbers: all functions return true only if the whole + // string is a number and put the value of this number into the pointer + // provided, the base is the numeric base in which the conversion should be + // done and must be comprised between 2 and 36 or be 0 in which case the + // standard C rules apply (leading '0' => octal, "0x" => hex) + // convert to a signed integer + bool ToLong(long *val, int base = 10) const; + // convert to an unsigned integer + bool ToULong(unsigned long *val, int base = 10) const; + // convert to wxLongLong +#if defined(wxLongLong_t) + bool ToLongLong(wxLongLong_t *val, int base = 10) const; + // convert to wxULongLong + bool ToULongLong(wxULongLong_t *val, int base = 10) const; +#endif // wxLongLong_t + // convert to a double + bool ToDouble(double *val) const; + + +#ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN + // formatted input/output + // as sprintf(), returns the number of characters written or < 0 on error + // (take 'this' into account in attribute parameter count) + // int Printf(const wxString& format, ...); + WX_DEFINE_VARARG_FUNC(int, Printf, 1, (const wxFormatString&), + DoPrintfWchar, DoPrintfUtf8) +#ifdef __WATCOMC__ + // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351 + WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wxString&), + (wxFormatString(f1))); + WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wxCStrData&), + (wxFormatString(f1))); + WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const char*), + (wxFormatString(f1))); + WX_VARARG_WATCOM_WORKAROUND(int, Printf, 1, (const wchar_t*), + (wxFormatString(f1))); +#endif +#endif // !wxNEEDS_WXSTRING_PRINTF_MIXIN + // as vprintf(), returns the number of characters written or < 0 on error + int PrintfV(const wxString& format, va_list argptr); + +#ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN + // returns the string containing the result of Printf() to it + // static wxString Format(const wxString& format, ...) ATTRIBUTE_PRINTF_1; + WX_DEFINE_VARARG_FUNC(static wxString, Format, 1, (const wxFormatString&), + DoFormatWchar, DoFormatUtf8) +#ifdef __WATCOMC__ + // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351 + WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wxString&), + (wxFormatString(f1))); + WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wxCStrData&), + (wxFormatString(f1))); + WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const char*), + (wxFormatString(f1))); + WX_VARARG_WATCOM_WORKAROUND(static wxString, Format, 1, (const wchar_t*), + (wxFormatString(f1))); +#endif +#endif + // the same as above, but takes a va_list + static wxString FormatV(const wxString& format, va_list argptr); + + // raw access to string memory + // ensure that string has space for at least nLen characters + // only works if the data of this string is not shared + bool Alloc(size_t nLen) { reserve(nLen); return capacity() >= nLen; } + // minimize the string's memory + // only works if the data of this string is not shared + bool Shrink(); +#if WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8 + // These are deprecated, use wxStringBuffer or wxStringBufferLength instead + // + // get writable buffer of at least nLen bytes. Unget() *must* be called + // a.s.a.p. to put string back in a reasonable state! + wxDEPRECATED( wxStringCharType *GetWriteBuf(size_t nLen) ); + // call this immediately after GetWriteBuf() has been used + wxDEPRECATED( void UngetWriteBuf() ); + wxDEPRECATED( void UngetWriteBuf(size_t nLen) ); +#endif // WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && wxUSE_UNICODE_UTF8 + + // wxWidgets version 1 compatibility functions + + // use Mid() + wxString SubString(size_t from, size_t to) const + { return Mid(from, (to - from + 1)); } + // values for second parameter of CompareTo function + enum caseCompare {exact, ignoreCase}; + // values for first parameter of Strip function + enum stripType {leading = 0x1, trailing = 0x2, both = 0x3}; + +#ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN + // use Printf() + // (take 'this' into account in attribute parameter count) + // int sprintf(const wxString& format, ...) ATTRIBUTE_PRINTF_2; + WX_DEFINE_VARARG_FUNC(int, sprintf, 1, (const wxFormatString&), + DoPrintfWchar, DoPrintfUtf8) +#ifdef __WATCOMC__ + // workaround for http://bugzilla.openwatcom.org/show_bug.cgi?id=351 + WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wxString&), + (wxFormatString(f1))); + WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wxCStrData&), + (wxFormatString(f1))); + WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const char*), + (wxFormatString(f1))); + WX_VARARG_WATCOM_WORKAROUND(int, sprintf, 1, (const wchar_t*), + (wxFormatString(f1))); +#endif +#endif // wxNEEDS_WXSTRING_PRINTF_MIXIN + + // use Cmp() + int CompareTo(const wxChar* psz, caseCompare cmp = exact) const + { return cmp == exact ? Cmp(psz) : CmpNoCase(psz); } + + // use length() + size_t Length() const { return length(); } + // Count the number of characters + int Freq(wxUniChar ch) const; + // use MakeLower + void LowerCase() { MakeLower(); } + // use MakeUpper + void UpperCase() { MakeUpper(); } + // use Trim except that it doesn't change this string + wxString Strip(stripType w = trailing) const; + + // use Find (more general variants not yet supported) + size_t Index(const wxChar* psz) const { return Find(psz); } + size_t Index(wxUniChar ch) const { return Find(ch); } + // use Truncate + wxString& Remove(size_t pos) { return Truncate(pos); } + wxString& RemoveLast(size_t n = 1) { return Truncate(length() - n); } + + wxString& Remove(size_t nStart, size_t nLen) + { return (wxString&)erase( nStart, nLen ); } + + // use Find() + int First( wxUniChar ch ) const { return Find(ch); } + int First( wxUniCharRef ch ) const { return Find(ch); } + int First( char ch ) const { return Find(ch); } + int First( unsigned char ch ) const { return Find(ch); } + int First( wchar_t ch ) const { return Find(ch); } + int First( const wxString& str ) const { return Find(str); } + int Last( wxUniChar ch ) const { return Find(ch, true); } + bool Contains(const wxString& str) const { return Find(str) != wxNOT_FOUND; } + + // use empty() + bool IsNull() const { return empty(); } + + // std::string compatibility functions + + // take nLen chars starting at nPos + wxString(const wxString& str, size_t nPos, size_t nLen) + { assign(str, nPos, nLen); } + // take all characters from first to last + wxString(const_iterator first, const_iterator last) + : m_impl(first.impl(), last.impl()) { } +#if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER + // the 2 overloads below are for compatibility with the existing code using + // pointers instead of iterators + wxString(const char *first, const char *last) + { + SubstrBufFromMB str(ImplStr(first, last - first)); + m_impl.assign(str.data, str.len); + } + wxString(const wchar_t *first, const wchar_t *last) + { + SubstrBufFromWC str(ImplStr(first, last - first)); + m_impl.assign(str.data, str.len); + } + // and this one is needed to compile code adding offsets to c_str() result + wxString(const wxCStrData& first, const wxCStrData& last) + : m_impl(CreateConstIterator(first).impl(), + CreateConstIterator(last).impl()) + { + wxASSERT_MSG( first.m_str == last.m_str, + _T("pointers must be into the same string") ); + } +#endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER + + // lib.string.modifiers + // append elements str[pos], ..., str[pos+n] + wxString& append(const wxString& str, size_t pos, size_t n) + { + wxSTRING_UPDATE_CACHED_LENGTH(n); + + size_t from, len; + str.PosLenToImpl(pos, n, &from, &len); + m_impl.append(str.m_impl, from, len); + return *this; + } + // append a string + wxString& append(const wxString& str) + { + wxSTRING_UPDATE_CACHED_LENGTH(str.length()); + + m_impl.append(str.m_impl); + return *this; + } + + // append first n (or all if n == npos) characters of sz + wxString& append(const char *sz) + { + wxSTRING_INVALIDATE_CACHED_LENGTH(); + + m_impl.append(ImplStr(sz)); + return *this; + } + + wxString& append(const wchar_t *sz) + { + wxSTRING_INVALIDATE_CACHED_LENGTH(); + + m_impl.append(ImplStr(sz)); + return *this; + } + + wxString& append(const char *sz, size_t n) + { + wxSTRING_INVALIDATE_CACHED_LENGTH(); + + SubstrBufFromMB str(ImplStr(sz, n)); + m_impl.append(str.data, str.len); + return *this; + } + wxString& append(const wchar_t *sz, size_t n) + { + wxSTRING_UPDATE_CACHED_LENGTH(n); + + SubstrBufFromWC str(ImplStr(sz, n)); + m_impl.append(str.data, str.len); + return *this; + } + + wxString& append(const wxCStrData& str) + { return append(str.AsString()); } + wxString& append(const wxCharBuffer& str) + { return append(str.data()); } + wxString& append(const wxWCharBuffer& str) + { return append(str.data()); } + wxString& append(const wxCStrData& str, size_t n) + { return append(str.AsString(), 0, n); } + wxString& append(const wxCharBuffer& str, size_t n) + { return append(str.data(), n); } + wxString& append(const wxWCharBuffer& str, size_t n) + { return append(str.data(), n); } + + // append n copies of ch + wxString& append(size_t n, wxUniChar ch) + { +#if wxUSE_UNICODE_UTF8 + if ( !ch.IsAscii() ) + { + wxSTRING_INVALIDATE_CACHED_LENGTH(); + + m_impl.append(wxStringOperations::EncodeNChars(n, ch)); + } + else // ASCII +#endif + { + wxSTRING_UPDATE_CACHED_LENGTH(n); + + m_impl.append(n, (wxStringCharType)ch); + } + + return *this; + } + + wxString& append(size_t n, wxUniCharRef ch) + { return append(n, wxUniChar(ch)); } + wxString& append(size_t n, char ch) + { return append(n, wxUniChar(ch)); } + wxString& append(size_t n, unsigned char ch) + { return append(n, wxUniChar(ch)); } + wxString& append(size_t n, wchar_t ch) + { return append(n, wxUniChar(ch)); } + + // append from first to last + wxString& append(const_iterator first, const_iterator last) + { + wxSTRING_INVALIDATE_CACHED_LENGTH(); + + m_impl.append(first.impl(), last.impl()); + return *this; + } +#if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER + wxString& append(const char *first, const char *last) + { return append(first, last - first); } + wxString& append(const wchar_t *first, const wchar_t *last) + { return append(first, last - first); } + wxString& append(const wxCStrData& first, const wxCStrData& last) + { return append(CreateConstIterator(first), CreateConstIterator(last)); } +#endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER + + // same as `this_string = str' + wxString& assign(const wxString& str) + { + wxSTRING_SET_CACHED_LENGTH(str.length()); + + m_impl = str.m_impl; + + return *this; + } + + wxString& assign(const wxString& str, size_t len) + { + wxSTRING_SET_CACHED_LENGTH(len); + + m_impl.assign(str.m_impl, 0, str.LenToImpl(len)); + + return *this; + } + + // same as ` = str[pos..pos + n] + wxString& assign(const wxString& str, size_t pos, size_t n) + { + size_t from, len; + str.PosLenToImpl(pos, n, &from, &len); + m_impl.assign(str.m_impl, from, len); + + // it's important to call this after PosLenToImpl() above in case str is + // the same string as this one + wxSTRING_SET_CACHED_LENGTH(n); + + return *this; + } + + // same as `= first n (or all if n == npos) characters of sz' + wxString& assign(const char *sz) + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl.assign(ImplStr(sz)); + + return *this; + } + + wxString& assign(const wchar_t *sz) + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl.assign(ImplStr(sz)); + + return *this; + } + + wxString& assign(const char *sz, size_t n) + { + wxSTRING_SET_CACHED_LENGTH(n); + + SubstrBufFromMB str(ImplStr(sz, n)); + m_impl.assign(str.data, str.len); + + return *this; + } + + wxString& assign(const wchar_t *sz, size_t n) + { + wxSTRING_SET_CACHED_LENGTH(n); + + SubstrBufFromWC str(ImplStr(sz, n)); + m_impl.assign(str.data, str.len); + + return *this; + } + + wxString& assign(const wxCStrData& str) + { return assign(str.AsString()); } + wxString& assign(const wxCharBuffer& str) + { return assign(str.data()); } + wxString& assign(const wxWCharBuffer& str) + { return assign(str.data()); } + wxString& assign(const wxCStrData& str, size_t len) + { return assign(str.AsString(), len); } + wxString& assign(const wxCharBuffer& str, size_t len) + { return assign(str.data(), len); } + wxString& assign(const wxWCharBuffer& str, size_t len) + { return assign(str.data(), len); } + + // same as `= n copies of ch' + wxString& assign(size_t n, wxUniChar ch) + { + wxSTRING_SET_CACHED_LENGTH(n); + +#if wxUSE_UNICODE_UTF8 + if ( !ch.IsAscii() ) + m_impl.assign(wxStringOperations::EncodeNChars(n, ch)); + else +#endif + m_impl.assign(n, (wxStringCharType)ch); + + return *this; + } + + wxString& assign(size_t n, wxUniCharRef ch) + { return assign(n, wxUniChar(ch)); } + wxString& assign(size_t n, char ch) + { return assign(n, wxUniChar(ch)); } + wxString& assign(size_t n, unsigned char ch) + { return assign(n, wxUniChar(ch)); } + wxString& assign(size_t n, wchar_t ch) + { return assign(n, wxUniChar(ch)); } + + // assign from first to last + wxString& assign(const_iterator first, const_iterator last) + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl.assign(first.impl(), last.impl()); + + return *this; + } +#if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER + wxString& assign(const char *first, const char *last) + { return assign(first, last - first); } + wxString& assign(const wchar_t *first, const wchar_t *last) + { return assign(first, last - first); } + wxString& assign(const wxCStrData& first, const wxCStrData& last) + { return assign(CreateConstIterator(first), CreateConstIterator(last)); } +#endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER + + // string comparison + int compare(const wxString& str) const; + int compare(const char* sz) const; + int compare(const wchar_t* sz) const; + int compare(const wxCStrData& str) const + { return compare(str.AsString()); } + int compare(const wxCharBuffer& str) const + { return compare(str.data()); } + int compare(const wxWCharBuffer& str) const + { return compare(str.data()); } + // comparison with a substring + int compare(size_t nStart, size_t nLen, const wxString& str) const; + // comparison of 2 substrings + int compare(size_t nStart, size_t nLen, + const wxString& str, size_t nStart2, size_t nLen2) const; + // substring comparison with first nCount characters of sz + int compare(size_t nStart, size_t nLen, + const char* sz, size_t nCount = npos) const; + int compare(size_t nStart, size_t nLen, + const wchar_t* sz, size_t nCount = npos) const; + + // insert another string + wxString& insert(size_t nPos, const wxString& str) + { insert(GetIterForNthChar(nPos), str.begin(), str.end()); return *this; } + // insert n chars of str starting at nStart (in str) + wxString& insert(size_t nPos, const wxString& str, size_t nStart, size_t n) + { + wxSTRING_UPDATE_CACHED_LENGTH(n); + + size_t from, len; + str.PosLenToImpl(nStart, n, &from, &len); + m_impl.insert(PosToImpl(nPos), str.m_impl, from, len); + + return *this; + } + + // insert first n (or all if n == npos) characters of sz + wxString& insert(size_t nPos, const char *sz) + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl.insert(PosToImpl(nPos), ImplStr(sz)); + + return *this; + } + + wxString& insert(size_t nPos, const wchar_t *sz) + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl.insert(PosToImpl(nPos), ImplStr(sz)); return *this; + } + + wxString& insert(size_t nPos, const char *sz, size_t n) + { + wxSTRING_UPDATE_CACHED_LENGTH(n); + + SubstrBufFromMB str(ImplStr(sz, n)); + m_impl.insert(PosToImpl(nPos), str.data, str.len); + + return *this; + } + + wxString& insert(size_t nPos, const wchar_t *sz, size_t n) + { + wxSTRING_UPDATE_CACHED_LENGTH(n); + + SubstrBufFromWC str(ImplStr(sz, n)); + m_impl.insert(PosToImpl(nPos), str.data, str.len); + + return *this; + } + + // insert n copies of ch + wxString& insert(size_t nPos, size_t n, wxUniChar ch) + { + wxSTRING_UPDATE_CACHED_LENGTH(n); + +#if wxUSE_UNICODE_UTF8 + if ( !ch.IsAscii() ) + m_impl.insert(PosToImpl(nPos), wxStringOperations::EncodeNChars(n, ch)); + else +#endif + m_impl.insert(PosToImpl(nPos), n, (wxStringCharType)ch); + return *this; + } + + iterator insert(iterator it, wxUniChar ch) + { + wxSTRING_UPDATE_CACHED_LENGTH(1); + +#if wxUSE_UNICODE_UTF8 + if ( !ch.IsAscii() ) + { + size_t pos = IterToImplPos(it); + m_impl.insert(pos, wxStringOperations::EncodeChar(ch)); + return iterator(this, m_impl.begin() + pos); + } + else +#endif + return iterator(this, m_impl.insert(it.impl(), (wxStringCharType)ch)); + } + + void insert(iterator it, const_iterator first, const_iterator last) + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl.insert(it.impl(), first.impl(), last.impl()); + } + +#if WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER + void insert(iterator it, const char *first, const char *last) + { insert(it - begin(), first, last - first); } + void insert(iterator it, const wchar_t *first, const wchar_t *last) + { insert(it - begin(), first, last - first); } + void insert(iterator it, const wxCStrData& first, const wxCStrData& last) + { insert(it, CreateConstIterator(first), CreateConstIterator(last)); } +#endif // WXWIN_COMPATIBILITY_STRING_PTR_AS_ITER + + void insert(iterator it, size_type n, wxUniChar ch) + { + wxSTRING_UPDATE_CACHED_LENGTH(n); + +#if wxUSE_UNICODE_UTF8 + if ( !ch.IsAscii() ) + m_impl.insert(IterToImplPos(it), wxStringOperations::EncodeNChars(n, ch)); + else +#endif + m_impl.insert(it.impl(), n, (wxStringCharType)ch); + } + + // delete characters from nStart to nStart + nLen + wxString& erase(size_type pos = 0, size_type n = npos) + { + wxSTRING_INVALIDATE_CACHE(); + + size_t from, len; + PosLenToImpl(pos, n, &from, &len); + m_impl.erase(from, len); + + return *this; + } + + // delete characters from first up to last + iterator erase(iterator first, iterator last) + { + wxSTRING_INVALIDATE_CACHE(); + + return iterator(this, m_impl.erase(first.impl(), last.impl())); + } + + iterator erase(iterator first) + { + wxSTRING_UPDATE_CACHED_LENGTH(-1); + + return iterator(this, m_impl.erase(first.impl())); + } + +#ifdef wxSTRING_BASE_HASNT_CLEAR + void clear() { erase(); } +#else + void clear() + { + wxSTRING_SET_CACHED_LENGTH(0); + + m_impl.clear(); + } +#endif + + // replaces the substring of length nLen starting at nStart + wxString& replace(size_t nStart, size_t nLen, const char* sz) + { + wxSTRING_INVALIDATE_CACHE(); + + size_t from, len; + PosLenToImpl(nStart, nLen, &from, &len); + m_impl.replace(from, len, ImplStr(sz)); + + return *this; + } + + wxString& replace(size_t nStart, size_t nLen, const wchar_t* sz) + { + wxSTRING_INVALIDATE_CACHE(); + + size_t from, len; + PosLenToImpl(nStart, nLen, &from, &len); + m_impl.replace(from, len, ImplStr(sz)); + + return *this; + } + + // replaces the substring of length nLen starting at nStart + wxString& replace(size_t nStart, size_t nLen, const wxString& str) + { + wxSTRING_INVALIDATE_CACHE(); + + size_t from, len; + PosLenToImpl(nStart, nLen, &from, &len); + m_impl.replace(from, len, str.m_impl); + + return *this; + } + + // replaces the substring with nCount copies of ch + wxString& replace(size_t nStart, size_t nLen, size_t nCount, wxUniChar ch) + { + wxSTRING_INVALIDATE_CACHE(); + + size_t from, len; + PosLenToImpl(nStart, nLen, &from, &len); +#if wxUSE_UNICODE_UTF8 + if ( !ch.IsAscii() ) + m_impl.replace(from, len, wxStringOperations::EncodeNChars(nCount, ch)); + else +#endif + m_impl.replace(from, len, nCount, (wxStringCharType)ch); + + return *this; + } + + // replaces a substring with another substring + wxString& replace(size_t nStart, size_t nLen, + const wxString& str, size_t nStart2, size_t nLen2) + { + wxSTRING_INVALIDATE_CACHE(); + + size_t from, len; + PosLenToImpl(nStart, nLen, &from, &len); + + size_t from2, len2; + str.PosLenToImpl(nStart2, nLen2, &from2, &len2); + + m_impl.replace(from, len, str.m_impl, from2, len2); + + return *this; + } + + // replaces the substring with first nCount chars of sz + wxString& replace(size_t nStart, size_t nLen, + const char* sz, size_t nCount) + { + wxSTRING_INVALIDATE_CACHE(); + + size_t from, len; + PosLenToImpl(nStart, nLen, &from, &len); + + SubstrBufFromMB str(ImplStr(sz, nCount)); + + m_impl.replace(from, len, str.data, str.len); + + return *this; + } + + wxString& replace(size_t nStart, size_t nLen, + const wchar_t* sz, size_t nCount) + { + wxSTRING_INVALIDATE_CACHE(); + + size_t from, len; + PosLenToImpl(nStart, nLen, &from, &len); + + SubstrBufFromWC str(ImplStr(sz, nCount)); + + m_impl.replace(from, len, str.data, str.len); + + return *this; + } + + wxString& replace(size_t nStart, size_t nLen, + const wxString& s, size_t nCount) + { + wxSTRING_INVALIDATE_CACHE(); + + size_t from, len; + PosLenToImpl(nStart, nLen, &from, &len); + m_impl.replace(from, len, s.m_impl.c_str(), s.LenToImpl(nCount)); + + return *this; + } + + wxString& replace(iterator first, iterator last, const char* s) + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl.replace(first.impl(), last.impl(), ImplStr(s)); + + return *this; + } + + wxString& replace(iterator first, iterator last, const wchar_t* s) + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl.replace(first.impl(), last.impl(), ImplStr(s)); + + return *this; + } + + wxString& replace(iterator first, iterator last, const char* s, size_type n) + { + wxSTRING_INVALIDATE_CACHE(); + + SubstrBufFromMB str(ImplStr(s, n)); + m_impl.replace(first.impl(), last.impl(), str.data, str.len); + + return *this; + } + + wxString& replace(iterator first, iterator last, const wchar_t* s, size_type n) + { + wxSTRING_INVALIDATE_CACHE(); + + SubstrBufFromWC str(ImplStr(s, n)); + m_impl.replace(first.impl(), last.impl(), str.data, str.len); + + return *this; + } + + wxString& replace(iterator first, iterator last, const wxString& s) + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl.replace(first.impl(), last.impl(), s.m_impl); + + return *this; + } + + wxString& replace(iterator first, iterator last, size_type n, wxUniChar ch) + { + wxSTRING_INVALIDATE_CACHE(); + +#if wxUSE_UNICODE_UTF8 + if ( !ch.IsAscii() ) + m_impl.replace(first.impl(), last.impl(), + wxStringOperations::EncodeNChars(n, ch)); + else +#endif + m_impl.replace(first.impl(), last.impl(), n, (wxStringCharType)ch); + + return *this; + } + + wxString& replace(iterator first, iterator last, + const_iterator first1, const_iterator last1) + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl.replace(first.impl(), last.impl(), first1.impl(), last1.impl()); + + return *this; + } + + wxString& replace(iterator first, iterator last, + const char *first1, const char *last1) + { replace(first, last, first1, last1 - first1); return *this; } + wxString& replace(iterator first, iterator last, + const wchar_t *first1, const wchar_t *last1) + { replace(first, last, first1, last1 - first1); return *this; } + + // swap two strings + void swap(wxString& str) + { +#if wxUSE_STRING_POS_CACHE + // we modify not only this string but also the other one directly so we + // need to invalidate cache for both of them (we could also try to + // exchange their cache entries but it seems unlikely to be worth it) + InvalidateCache(); + str.InvalidateCache(); +#endif // wxUSE_STRING_POS_CACHE + + m_impl.swap(str.m_impl); + } + + // find a substring + size_t find(const wxString& str, size_t nStart = 0) const + { return PosFromImpl(m_impl.find(str.m_impl, PosToImpl(nStart))); } + + // find first n characters of sz + size_t find(const char* sz, size_t nStart = 0, size_t n = npos) const + { + SubstrBufFromMB str(ImplStr(sz, n)); + return PosFromImpl(m_impl.find(str.data, PosToImpl(nStart), str.len)); + } + size_t find(const wchar_t* sz, size_t nStart = 0, size_t n = npos) const + { + SubstrBufFromWC str(ImplStr(sz, n)); + return PosFromImpl(m_impl.find(str.data, PosToImpl(nStart), str.len)); + } + size_t find(const wxCharBuffer& s, size_t nStart = 0, size_t n = npos) const + { return find(s.data(), nStart, n); } + size_t find(const wxWCharBuffer& s, size_t nStart = 0, size_t n = npos) const + { return find(s.data(), nStart, n); } + size_t find(const wxCStrData& s, size_t nStart = 0, size_t n = npos) const + { return find(s.AsWChar(), nStart, n); } + + // find the first occurrence of character ch after nStart + size_t find(wxUniChar ch, size_t nStart = 0) const + { +#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); } + size_t find(char ch, size_t nStart = 0) const + { return find(wxUniChar(ch), nStart); } + size_t find(unsigned char ch, size_t nStart = 0) const + { return find(wxUniChar(ch), nStart); } + size_t find(wchar_t ch, size_t nStart = 0) const + { return find(wxUniChar(ch), nStart); } + + // rfind() family is exactly like find() but works right to left + + // as find, but from the end + size_t rfind(const wxString& str, size_t nStart = npos) const + { return PosFromImpl(m_impl.rfind(str.m_impl, PosToImpl(nStart))); } + + // as find, but from the end + size_t rfind(const char* sz, size_t nStart = npos, size_t n = npos) const + { + SubstrBufFromMB str(ImplStr(sz, n)); + return PosFromImpl(m_impl.rfind(str.data, PosToImpl(nStart), str.len)); + } + size_t rfind(const wchar_t* sz, size_t nStart = npos, size_t n = npos) const + { + SubstrBufFromWC str(ImplStr(sz, n)); + return PosFromImpl(m_impl.rfind(str.data, PosToImpl(nStart), str.len)); + } + size_t rfind(const wxCharBuffer& s, size_t nStart = npos, size_t n = npos) const + { return rfind(s.data(), nStart, n); } + size_t rfind(const wxWCharBuffer& s, size_t nStart = npos, size_t n = npos) const + { return rfind(s.data(), nStart, n); } + size_t rfind(const wxCStrData& s, size_t nStart = npos, size_t n = npos) const + { return rfind(s.AsWChar(), nStart, n); } + // as find, but from the end + size_t rfind(wxUniChar ch, size_t nStart = npos) const + { +#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); } + size_t rfind(char ch, size_t nStart = npos) const + { return rfind(wxUniChar(ch), nStart); } + size_t rfind(unsigned char ch, size_t nStart = npos) const + { return rfind(wxUniChar(ch), nStart); } + size_t rfind(wchar_t ch, size_t nStart = npos) const + { return rfind(wxUniChar(ch), nStart); } + + // find first/last occurrence of any character (not) in the set: +#if wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8 + // FIXME-UTF8: this is not entirely correct, because it doesn't work if + // sizeof(wchar_t)==2 and surrogates are present in the string; + // should we care? Probably not. + size_t find_first_of(const wxString& str, size_t nStart = 0) const + { return m_impl.find_first_of(str.m_impl, nStart); } + size_t find_first_of(const char* sz, size_t nStart = 0) const + { return m_impl.find_first_of(ImplStr(sz), nStart); } + size_t find_first_of(const wchar_t* sz, size_t nStart = 0) const + { return m_impl.find_first_of(ImplStr(sz), nStart); } + size_t find_first_of(const char* sz, size_t nStart, size_t n) const + { return m_impl.find_first_of(ImplStr(sz), nStart, n); } + size_t find_first_of(const wchar_t* sz, size_t nStart, size_t n) const + { return m_impl.find_first_of(ImplStr(sz), nStart, n); } + size_t find_first_of(wxUniChar c, size_t nStart = 0) const + { return m_impl.find_first_of((wxChar)c, nStart); } + + size_t find_last_of(const wxString& str, size_t nStart = npos) const + { return m_impl.find_last_of(str.m_impl, nStart); } + size_t find_last_of(const char* sz, size_t nStart = npos) const + { return m_impl.find_last_of(ImplStr(sz), nStart); } + size_t find_last_of(const wchar_t* sz, size_t nStart = npos) const + { return m_impl.find_last_of(ImplStr(sz), nStart); } + size_t find_last_of(const char* sz, size_t nStart, size_t n) const + { return m_impl.find_last_of(ImplStr(sz), nStart, n); } + size_t find_last_of(const wchar_t* sz, size_t nStart, size_t n) const + { return m_impl.find_last_of(ImplStr(sz), nStart, n); } + size_t find_last_of(wxUniChar c, size_t nStart = npos) const + { return m_impl.find_last_of((wxChar)c, nStart); } + + size_t find_first_not_of(const wxString& str, size_t nStart = 0) const + { return m_impl.find_first_not_of(str.m_impl, nStart); } + size_t find_first_not_of(const char* sz, size_t nStart = 0) const + { return m_impl.find_first_not_of(ImplStr(sz), nStart); } + size_t find_first_not_of(const wchar_t* sz, size_t nStart = 0) const + { return m_impl.find_first_not_of(ImplStr(sz), nStart); } + size_t find_first_not_of(const char* sz, size_t nStart, size_t n) const + { return m_impl.find_first_not_of(ImplStr(sz), nStart, n); } + size_t find_first_not_of(const wchar_t* sz, size_t nStart, size_t n) const + { return m_impl.find_first_not_of(ImplStr(sz), nStart, n); } + size_t find_first_not_of(wxUniChar c, size_t nStart = 0) const + { return m_impl.find_first_not_of((wxChar)c, nStart); } + + size_t find_last_not_of(const wxString& str, size_t nStart = npos) const + { return m_impl.find_last_not_of(str.m_impl, nStart); } + size_t find_last_not_of(const char* sz, size_t nStart = npos) const + { return m_impl.find_last_not_of(ImplStr(sz), nStart); } + size_t find_last_not_of(const wchar_t* sz, size_t nStart = npos) const + { return m_impl.find_last_not_of(ImplStr(sz), nStart); } + size_t find_last_not_of(const char* sz, size_t nStart, size_t n) const + { return m_impl.find_last_not_of(ImplStr(sz), nStart, n); } + size_t find_last_not_of(const wchar_t* sz, size_t nStart, size_t n) const + { return m_impl.find_last_not_of(ImplStr(sz), nStart, n); } + size_t find_last_not_of(wxUniChar c, size_t nStart = npos) const + { return m_impl.find_last_not_of((wxChar)c, nStart); } +#else + // we can't use std::string implementation in UTF-8 build, because the + // character sets would be interpreted wrongly: + + // as strpbrk() but starts at nStart, returns npos if not found + size_t find_first_of(const wxString& str, size_t nStart = 0) const +#if wxUSE_UNICODE // FIXME-UTF8: temporary + { return find_first_of(str.wc_str(), nStart); } +#else + { return find_first_of(str.mb_str(), nStart); } +#endif + // same as above + size_t find_first_of(const char* sz, size_t nStart = 0) const; + size_t find_first_of(const wchar_t* sz, size_t nStart = 0) const; + size_t find_first_of(const char* sz, size_t nStart, size_t n) const; + size_t find_first_of(const wchar_t* sz, size_t nStart, size_t n) const; + // same as find(char, size_t) + size_t find_first_of(wxUniChar c, size_t nStart = 0) const + { return find(c, nStart); } + // find the last (starting from nStart) char from str in this string + size_t find_last_of (const wxString& str, size_t nStart = npos) const +#if wxUSE_UNICODE // FIXME-UTF8: temporary + { return find_last_of(str.wc_str(), nStart); } +#else + { return find_last_of(str.mb_str(), nStart); } +#endif + // same as above + size_t find_last_of (const char* sz, size_t nStart = npos) const; + size_t find_last_of (const wchar_t* sz, size_t nStart = npos) const; + size_t find_last_of(const char* sz, size_t nStart, size_t n) const; + size_t find_last_of(const wchar_t* sz, size_t nStart, size_t n) const; + // same as above + size_t find_last_of(wxUniChar c, size_t nStart = npos) const + { return rfind(c, nStart); } + + // find first/last occurrence of any character not in the set + + // as strspn() (starting from nStart), returns npos on failure + size_t find_first_not_of(const wxString& str, size_t nStart = 0) const +#if wxUSE_UNICODE // FIXME-UTF8: temporary + { return find_first_not_of(str.wc_str(), nStart); } +#else + { return find_first_not_of(str.mb_str(), nStart); } +#endif + // same as above + size_t find_first_not_of(const char* sz, size_t nStart = 0) const; + size_t find_first_not_of(const wchar_t* sz, size_t nStart = 0) const; + size_t find_first_not_of(const char* sz, size_t nStart, size_t n) const; + size_t find_first_not_of(const wchar_t* sz, size_t nStart, size_t n) const; + // same as above + size_t find_first_not_of(wxUniChar ch, size_t nStart = 0) const; + // as strcspn() + size_t find_last_not_of(const wxString& str, size_t nStart = npos) const +#if wxUSE_UNICODE // FIXME-UTF8: temporary + { return find_last_not_of(str.wc_str(), nStart); } +#else + { return find_last_not_of(str.mb_str(), nStart); } +#endif + // same as above + size_t find_last_not_of(const char* sz, size_t nStart = npos) const; + size_t find_last_not_of(const wchar_t* sz, size_t nStart = npos) const; + size_t find_last_not_of(const char* sz, size_t nStart, size_t n) const; + size_t find_last_not_of(const wchar_t* sz, size_t nStart, size_t n) const; + // same as above + size_t find_last_not_of(wxUniChar ch, size_t nStart = npos) const; +#endif // wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8 or not + + // provide char/wchar_t/wxUniCharRef overloads for char-finding functions + // above to resolve ambiguities: + size_t find_first_of(wxUniCharRef ch, size_t nStart = 0) const + { return find_first_of(wxUniChar(ch), nStart); } + size_t find_first_of(char ch, size_t nStart = 0) const + { return find_first_of(wxUniChar(ch), nStart); } + size_t find_first_of(unsigned char ch, size_t nStart = 0) const + { return find_first_of(wxUniChar(ch), nStart); } + size_t find_first_of(wchar_t ch, size_t nStart = 0) const + { return find_first_of(wxUniChar(ch), nStart); } + size_t find_last_of(wxUniCharRef ch, size_t nStart = npos) const + { return find_last_of(wxUniChar(ch), nStart); } + size_t find_last_of(char ch, size_t nStart = npos) const + { return find_last_of(wxUniChar(ch), nStart); } + size_t find_last_of(unsigned char ch, size_t nStart = npos) const + { return find_last_of(wxUniChar(ch), nStart); } + size_t find_last_of(wchar_t ch, size_t nStart = npos) const + { return find_last_of(wxUniChar(ch), nStart); } + size_t find_first_not_of(wxUniCharRef ch, size_t nStart = 0) const + { return find_first_not_of(wxUniChar(ch), nStart); } + size_t find_first_not_of(char ch, size_t nStart = 0) const + { return find_first_not_of(wxUniChar(ch), nStart); } + size_t find_first_not_of(unsigned char ch, size_t nStart = 0) const + { return find_first_not_of(wxUniChar(ch), nStart); } + size_t find_first_not_of(wchar_t ch, size_t nStart = 0) const + { return find_first_not_of(wxUniChar(ch), nStart); } + size_t find_last_not_of(wxUniCharRef ch, size_t nStart = npos) const + { return find_last_not_of(wxUniChar(ch), nStart); } + size_t find_last_not_of(char ch, size_t nStart = npos) const + { return find_last_not_of(wxUniChar(ch), nStart); } + size_t find_last_not_of(unsigned char ch, size_t nStart = npos) const + { return find_last_not_of(wxUniChar(ch), nStart); } + size_t find_last_not_of(wchar_t ch, size_t nStart = npos) const + { return find_last_not_of(wxUniChar(ch), nStart); } + + // and additional overloads for the versions taking strings: + size_t find_first_of(const wxCStrData& sz, size_t nStart = 0) const + { return find_first_of(sz.AsString(), nStart); } + size_t find_first_of(const wxCharBuffer& sz, size_t nStart = 0) const + { return find_first_of(sz.data(), nStart); } + size_t find_first_of(const wxWCharBuffer& sz, size_t nStart = 0) const + { return find_first_of(sz.data(), nStart); } + size_t find_first_of(const wxCStrData& sz, size_t nStart, size_t n) const + { return find_first_of(sz.AsWChar(), nStart, n); } + size_t find_first_of(const wxCharBuffer& sz, size_t nStart, size_t n) const + { return find_first_of(sz.data(), nStart, n); } + size_t find_first_of(const wxWCharBuffer& sz, size_t nStart, size_t n) const + { return find_first_of(sz.data(), nStart, n); } + + size_t find_last_of(const wxCStrData& sz, size_t nStart = 0) const + { return find_last_of(sz.AsString(), nStart); } + size_t find_last_of(const wxCharBuffer& sz, size_t nStart = 0) const + { return find_last_of(sz.data(), nStart); } + size_t find_last_of(const wxWCharBuffer& sz, size_t nStart = 0) const + { return find_last_of(sz.data(), nStart); } + size_t find_last_of(const wxCStrData& sz, size_t nStart, size_t n) const + { return find_last_of(sz.AsWChar(), nStart, n); } + size_t find_last_of(const wxCharBuffer& sz, size_t nStart, size_t n) const + { return find_last_of(sz.data(), nStart, n); } + size_t find_last_of(const wxWCharBuffer& sz, size_t nStart, size_t n) const + { return find_last_of(sz.data(), nStart, n); } + + size_t find_first_not_of(const wxCStrData& sz, size_t nStart = 0) const + { return find_first_not_of(sz.AsString(), nStart); } + size_t find_first_not_of(const wxCharBuffer& sz, size_t nStart = 0) const + { return find_first_not_of(sz.data(), nStart); } + size_t find_first_not_of(const wxWCharBuffer& sz, size_t nStart = 0) const + { return find_first_not_of(sz.data(), nStart); } + size_t find_first_not_of(const wxCStrData& sz, size_t nStart, size_t n) const + { return find_first_not_of(sz.AsWChar(), nStart, n); } + size_t find_first_not_of(const wxCharBuffer& sz, size_t nStart, size_t n) const + { return find_first_not_of(sz.data(), nStart, n); } + size_t find_first_not_of(const wxWCharBuffer& sz, size_t nStart, size_t n) const + { return find_first_not_of(sz.data(), nStart, n); } + + size_t find_last_not_of(const wxCStrData& sz, size_t nStart = 0) const + { return find_last_not_of(sz.AsString(), nStart); } + size_t find_last_not_of(const wxCharBuffer& sz, size_t nStart = 0) const + { return find_last_not_of(sz.data(), nStart); } + size_t find_last_not_of(const wxWCharBuffer& sz, size_t nStart = 0) const + { return find_last_not_of(sz.data(), nStart); } + size_t find_last_not_of(const wxCStrData& sz, size_t nStart, size_t n) const + { return find_last_not_of(sz.AsWChar(), nStart, n); } + size_t find_last_not_of(const wxCharBuffer& sz, size_t nStart, size_t n) const + { return find_last_not_of(sz.data(), nStart, n); } + size_t find_last_not_of(const wxWCharBuffer& sz, size_t nStart, size_t n) const + { return find_last_not_of(sz.data(), nStart, n); } + + // string += string + wxString& operator+=(const wxString& s) + { + wxSTRING_INVALIDATE_CACHED_LENGTH(); + + m_impl += s.m_impl; + return *this; + } + // string += C string + wxString& operator+=(const char *psz) + { + wxSTRING_INVALIDATE_CACHED_LENGTH(); + + m_impl += ImplStr(psz); + return *this; + } + wxString& operator+=(const wchar_t *pwz) + { + wxSTRING_INVALIDATE_CACHED_LENGTH(); + + m_impl += ImplStr(pwz); + return *this; + } + wxString& operator+=(const wxCStrData& s) + { + wxSTRING_INVALIDATE_CACHED_LENGTH(); + + m_impl += s.AsString().m_impl; + return *this; + } + wxString& operator+=(const wxCharBuffer& s) + { return operator+=(s.data()); } + wxString& operator+=(const wxWCharBuffer& s) + { return operator+=(s.data()); } + // string += char + wxString& operator+=(wxUniChar ch) + { + wxSTRING_UPDATE_CACHED_LENGTH(1); + +#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); } + wxString& operator+=(unsigned char ch) { return *this += wxUniChar(ch); } + wxString& operator+=(wchar_t ch) { return *this += wxUniChar(ch); } + +private: +#if !wxUSE_STL_BASED_WXSTRING + // helpers for wxStringBuffer and wxStringBufferLength + wxStringCharType *DoGetWriteBuf(size_t nLen) + { + return m_impl.DoGetWriteBuf(nLen); + } + + void DoUngetWriteBuf() + { + wxSTRING_INVALIDATE_CACHE(); + + m_impl.DoUngetWriteBuf(); + } + + void DoUngetWriteBuf(size_t nLen) + { + wxSTRING_SET_CACHED_LENGTH(nLen); + + m_impl.DoUngetWriteBuf(nLen); + } +#endif // !wxUSE_STL_BASED_WXSTRING + +#ifndef wxNEEDS_WXSTRING_PRINTF_MIXIN + #if !wxUSE_UTF8_LOCALE_ONLY + int DoPrintfWchar(const wxChar *format, ...); + static wxString DoFormatWchar(const wxChar *format, ...); + #endif + #if wxUSE_UNICODE_UTF8 + int DoPrintfUtf8(const char *format, ...); + static wxString DoFormatUtf8(const char *format, ...); + #endif +#endif + +#if !wxUSE_STL_BASED_WXSTRING + // check string's data validity + bool IsValid() const { return m_impl.GetStringData()->IsValid(); } +#endif + +private: + wxStringImpl m_impl; + + // buffers for compatibility conversion from (char*)c_str() and + // (wchar_t*)c_str(): + // FIXME-UTF8: bechmark various approaches to keeping compatibility buffers + template + struct ConvertedBuffer + { + ConvertedBuffer() : m_buf(NULL) {} + ~ConvertedBuffer() + { free(m_buf); } + + operator T*() const { return m_buf; } + + ConvertedBuffer& operator=(T *str) + { + free(m_buf); + m_buf = str; + return *this; + } + + T *m_buf; + }; +#if wxUSE_UNICODE && !wxUSE_UTF8_LOCALE_ONLY + ConvertedBuffer m_convertedToChar; +#endif +#if !wxUSE_UNICODE_WCHAR + ConvertedBuffer m_convertedToWChar; +#endif + +#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; - wxString& Remove(size_t nStart, size_t nLen) { return erase( nStart, nLen ); } + // copying is disallowed as it would result in more than one pointer into + // the same linked list + DECLARE_NO_COPY_CLASS(wxStringIteratorNodeHead) + }; - // use Find() - int First( const wxChar ch ) const { return Find(ch); } - int First( const wxChar* psz ) const { return Find(psz); } - int First( const wxString &str ) const { return Find(str); } - int Last( const wxChar ch ) const { return Find(ch, TRUE); } - bool Contains(const wxString& str) const { return Find(str) != -1; } + wxStringIteratorNodeHead m_iterators; - // use IsEmpty() - bool IsNull() const { return IsEmpty(); } + friend class WXDLLIMPEXP_FWD_BASE wxStringIteratorNode; + friend class WXDLLIMPEXP_FWD_BASE wxUniCharRef; +#endif // wxUSE_UNICODE_UTF8 -#ifdef wxSTD_STRING_COMPATIBILITY - // std::string compatibility functions + friend class WXDLLIMPEXP_FWD_BASE wxCStrData; + friend class wxStringInternalBuffer; + friend class wxStringInternalBufferLength; +}; - // standard types - typedef wxChar value_type; - typedef const value_type *const_iterator; +#ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN + #pragma warning (default:4275) +#endif - // an 'invalid' value for string index - static const size_t npos; +// string iterator operators that satisfy STL Random Access Iterator +// requirements: +inline wxString::iterator operator+(ptrdiff_t n, wxString::iterator i) + { return i + n; } +inline wxString::const_iterator operator+(ptrdiff_t n, wxString::const_iterator i) + { return i + n; } +inline wxString::reverse_iterator operator+(ptrdiff_t n, wxString::reverse_iterator i) + { return i + n; } +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 +// enough, from the point of view of C++ standard we must have the declarations +// here as friend ones are not injected in the enclosing namespace and without +// them the code fails to compile with conforming compilers such as xlC or g++4 +wxString WXDLLIMPEXP_BASE operator+(const wxString& string1, const wxString& string2); +wxString WXDLLIMPEXP_BASE operator+(const wxString& string, const char *psz); +wxString WXDLLIMPEXP_BASE operator+(const wxString& string, const wchar_t *pwz); +wxString WXDLLIMPEXP_BASE operator+(const char *psz, const wxString& string); +wxString WXDLLIMPEXP_BASE operator+(const wchar_t *pwz, const wxString& string); + +wxString WXDLLIMPEXP_BASE operator+(const wxString& string, wxUniChar ch); +wxString WXDLLIMPEXP_BASE operator+(wxUniChar ch, const wxString& string); + +inline wxString operator+(const wxString& string, wxUniCharRef ch) + { return string + (wxUniChar)ch; } +inline wxString operator+(const wxString& string, char ch) + { return string + wxUniChar(ch); } +inline wxString operator+(const wxString& string, wchar_t ch) + { return string + wxUniChar(ch); } +inline wxString operator+(wxUniCharRef ch, const wxString& string) + { return (wxUniChar)ch + string; } +inline wxString operator+(char ch, const wxString& string) + { return wxUniChar(ch) + string; } +inline wxString operator+(wchar_t ch, const wxString& string) + { return wxUniChar(ch) + string; } + + +#define wxGetEmptyString() wxString() - // constructors - // take nLen chars starting at nPos - wxString(const wxString& str, size_t nPos, size_t nLen) - : m_pchData(NULL) - { - wxASSERT_MSG( str.GetStringData()->IsValid(), - _T("did you forget to call UngetWriteBuf()?") ); +// ---------------------------------------------------------------------------- +// helper functions which couldn't be defined inline +// ---------------------------------------------------------------------------- - InitWith(str.c_str(), nPos, nLen == npos ? 0 : nLen); - } - // take all characters from pStart to pEnd - wxString(const void *pStart, const void *pEnd); - - // lib.string.capacity - // return the length of the string - size_t size() const { return Len(); } - // return the length of the string - size_t length() const { return Len(); } - // return the maximum size of the string - size_t max_size() const { return wxSTRING_MAXLEN; } - // resize the string, filling the space with c if c != 0 - void resize(size_t nSize, wxChar ch = wxT('\0')); - // delete the contents of the string - void clear() { Empty(); } - // returns true if the string is empty - bool empty() const { return IsEmpty(); } - // inform string about planned change in size - void reserve(size_t size) { Alloc(size); } - - // lib.string.access - // return the character at position n - wxChar at(size_t n) const { return GetChar(n); } - // returns the writable character at position n - wxChar& at(size_t n) { return GetWritableChar(n); } - - // first valid index position - const_iterator begin() const { return wx_str(); } - // position one after the last valid one - const_iterator end() const { return wx_str() + length(); } +namespace wxPrivate +{ - // lib.string.modifiers - // append a string - wxString& append(const wxString& str) - { *this += str; return *this; } - // append elements str[pos], ..., str[pos+n] - wxString& append(const wxString& str, size_t pos, size_t n) - { ConcatSelf(n, str.c_str() + pos); return *this; } - // append first n (or all if n == npos) characters of sz - wxString& append(const wxChar *sz, size_t n = npos) - { ConcatSelf(n == npos ? wxStrlen(sz) : n, sz); return *this; } +#if wxUSE_UNICODE_WCHAR - // append n copies of ch - wxString& append(size_t n, wxChar ch) { return Pad(n, ch); } +template <> +struct wxStringAsBufHelper +{ + static wxCharBuffer Get(const wxString& s, size_t *len) + { + wxCharBuffer buf(s.mb_str()); + if ( len ) + *len = buf ? strlen(buf) : 0; + return buf; + } +}; - // same as `this_string = str' - wxString& assign(const wxString& str) - { return *this = str; } - // same as ` = str[pos..pos + n] - wxString& assign(const wxString& str, size_t pos, size_t n) - { Empty(); return Append(str.c_str() + pos, n); } - // same as `= first n (or all if n == npos) characters of sz' - wxString& assign(const wxChar *sz, size_t n = npos) - { Empty(); return Append(sz, n == npos ? wxStrlen(sz) : n); } - // same as `= n copies of ch' - wxString& assign(size_t n, wxChar ch) - { Empty(); return Append(ch, n); } +template <> +struct wxStringAsBufHelper +{ + static wxWCharBuffer Get(const wxString& s, size_t *len) + { + if ( len ) + *len = s.length(); + return wxWCharBuffer::CreateNonOwned(s.wx_str()); + } +}; - // insert another string - wxString& insert(size_t nPos, const wxString& str); - // insert n chars of str starting at nStart (in str) - wxString& insert(size_t nPos, const wxString& str, size_t nStart, size_t n) - { return insert(nPos, wxString((const wxChar *)str + nStart, n)); } +#elif wxUSE_UNICODE_UTF8 - // insert first n (or all if n == npos) characters of sz - wxString& insert(size_t nPos, const wxChar *sz, size_t n = npos) - { return insert(nPos, wxString(sz, n)); } - // insert n copies of ch - wxString& insert(size_t nPos, size_t n, wxChar ch) - { return insert(nPos, wxString(ch, n)); } +template <> +struct wxStringAsBufHelper +{ + static wxCharBuffer Get(const wxString& s, size_t *len) + { + if ( len ) + *len = s.utf8_length(); + return wxCharBuffer::CreateNonOwned(s.wx_str()); + } +}; - // delete characters from nStart to nStart + nLen - wxString& erase(size_t nStart = 0, size_t nLen = npos); +template <> +struct wxStringAsBufHelper +{ + static wxWCharBuffer Get(const wxString& s, size_t *len) + { + wxWCharBuffer wbuf(s.wc_str()); + if ( len ) + *len = wxWcslen(wbuf); + return wbuf; + } +}; - // replaces the substring of length nLen starting at nStart - wxString& replace(size_t nStart, size_t nLen, const wxChar* sz); - // replaces the substring with nCount copies of ch - wxString& replace(size_t nStart, size_t nLen, size_t nCount, wxChar ch); - // replaces a substring with another substring - wxString& replace(size_t nStart, size_t nLen, - const wxString& str, size_t nStart2, size_t nLen2); - // replaces the substring with first nCount chars of sz - wxString& replace(size_t nStart, size_t nLen, - const wxChar* sz, size_t nCount); +#endif // Unicode build kind - // swap two strings - void swap(wxString& str); +} // namespace wxPrivate - // All find() functions take the nStart argument which specifies the - // position to start the search on, the default value is 0. All functions - // return npos if there were no match. +// ---------------------------------------------------------------------------- +// wxStringBuffer: a tiny class allowing to get a writable pointer into string +// ---------------------------------------------------------------------------- - // find a substring - size_t find(const wxString& str, size_t nStart = 0) const; +#if !wxUSE_STL_BASED_WXSTRING +// string buffer for direct access to string data in their native +// representation: +class wxStringInternalBuffer +{ +public: + typedef wxStringCharType CharType; - // VC++ 1.5 can't cope with this syntax. -#if !defined(__VISUALC__) || defined(__WIN32__) - // find first n characters of sz - size_t find(const wxChar* sz, size_t nStart = 0, size_t n = npos) const; -#endif // VC++ 1.5 + wxStringInternalBuffer(wxString& str, size_t lenWanted = 1024) + : m_str(str), m_buf(NULL) + { m_buf = m_str.DoGetWriteBuf(lenWanted); } - // Gives a duplicate symbol (presumably a case-insensitivity problem) -#if !defined(__BORLANDC__) - // find the first occurence of character ch after nStart - size_t find(wxChar ch, size_t nStart = 0) const; -#endif - // rfind() family is exactly like find() but works right to left + ~wxStringInternalBuffer() { m_str.DoUngetWriteBuf(); } - // as find, but from the end - size_t rfind(const wxString& str, size_t nStart = npos) const; + operator wxStringCharType*() const { return m_buf; } - // VC++ 1.5 can't cope with this syntax. -#if !defined(__VISUALC__) || defined(__WIN32__) - // as find, but from the end - size_t rfind(const wxChar* sz, size_t nStart = npos, - size_t n = npos) const; - // as find, but from the end - size_t rfind(wxChar ch, size_t nStart = npos) const; -#endif // VC++ 1.5 +private: + wxString& m_str; + wxStringCharType *m_buf; - // find first/last occurence of any character in the set + DECLARE_NO_COPY_CLASS(wxStringInternalBuffer) +}; - // as strpbrk() but starts at nStart, returns npos if not found - size_t find_first_of(const wxString& str, size_t nStart = 0) const - { return find_first_of(str.c_str(), nStart); } - // same as above - size_t find_first_of(const wxChar* sz, size_t nStart = 0) const; - // same as find(char, size_t) - size_t find_first_of(wxChar c, size_t nStart = 0) const - { return find(c, nStart); } - // find the last (starting from nStart) char from str in this string - size_t find_last_of (const wxString& str, size_t nStart = npos) const - { return find_last_of(str.c_str(), nStart); } - // same as above - size_t find_last_of (const wxChar* sz, size_t nStart = npos) const; - // same as above - size_t find_last_of(wxChar c, size_t nStart = npos) const - { return rfind(c, nStart); } +class wxStringInternalBufferLength +{ +public: + typedef wxStringCharType CharType; - // find first/last occurence of any character not in the set + 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); + } - // as strspn() (starting from nStart), returns npos on failure - size_t find_first_not_of(const wxString& str, size_t nStart = 0) const - { return find_first_not_of(str.c_str(), nStart); } - // same as above - size_t find_first_not_of(const wxChar* sz, size_t nStart = 0) const; - // same as above - size_t find_first_not_of(wxChar ch, size_t nStart = 0) const; - // as strcspn() - size_t find_last_not_of(const wxString& str, size_t nStart = npos) const - { return find_first_not_of(str.c_str(), nStart); } - // same as above - size_t find_last_not_of(const wxChar* sz, size_t nStart = npos) const; - // same as above - size_t find_last_not_of(wxChar ch, size_t nStart = npos) const; + ~wxStringInternalBufferLength() + { + wxASSERT(m_lenSet); + m_str.DoUngetWriteBuf(m_len); + } - // All compare functions return -1, 0 or 1 if the [sub]string is less, - // equal or greater than the compare() argument. + operator wxStringCharType*() const { return m_buf; } + void SetLength(size_t length) { m_len = length; m_lenSet = true; } - // just like strcmp() - int compare(const wxString& str) const { return Cmp(str); } - // comparison with a substring - int compare(size_t nStart, size_t nLen, const wxString& str) const - { return Mid(nStart, nLen).Cmp(str); } - // comparison of 2 substrings - int compare(size_t nStart, size_t nLen, - const wxString& str, size_t nStart2, size_t nLen2) const - { return Mid(nStart, nLen).Cmp(str.Mid(nStart2, nLen2)); } - // just like strcmp() - int compare(const wxChar* sz) const { return Cmp(sz); } - // substring comparison with first nCount characters of sz - int compare(size_t nStart, size_t nLen, - const wxChar* sz, size_t nCount = npos) const - { return Mid(nStart, nLen).Cmp(wxString(sz, nCount)); } +private: + wxString& m_str; + wxStringCharType *m_buf; + size_t m_len; + bool m_lenSet; - // substring extraction - wxString substr(size_t nStart = 0, size_t nLen = npos) const - { return Mid(nStart, nLen); } -#endif // wxSTD_STRING_COMPATIBILITY + DECLARE_NO_COPY_CLASS(wxStringInternalBufferLength) }; -// ---------------------------------------------------------------------------- -// The string array uses it's knowledge of internal structure of the wxString -// class to optimize string storage. Normally, we would store pointers to -// string, but as wxString is, in fact, itself a pointer (sizeof(wxString) is -// sizeof(char *)) we store these pointers instead. The cast to "wxString *" is -// really all we need to turn such pointer into a string! -// -// Of course, it can be called a dirty hack, but we use twice less memory and -// this approach is also more speed efficient, so it's probably worth it. -// -// Usage notes: when a string is added/inserted, a new copy of it is created, -// so the original string may be safely deleted. When a string is retrieved -// from the array (operator[] or Item() method), a reference is returned. -// ---------------------------------------------------------------------------- +#endif // !wxUSE_STL_BASED_WXSTRING -class WXDLLEXPORT wxArrayString +template +class WXDLLIMPEXP_BASE wxStringTypeBufferBase { public: - // type of function used by wxArrayString::Sort() - typedef int (*CompareFunction)(const wxString& first, - const wxString& second); + typedef T CharType; - // constructors and destructor - // default ctor - wxArrayString() - : m_nSize(0), m_nCount(0), m_pItems(NULL), m_autoSort(FALSE) - { Init(FALSE); } - // if autoSort is TRUE, the array is always sorted (in alphabetical order) - // - // NB: the reason for using int and not bool is that like this we can avoid - // using this ctor for implicit conversions from "const char *" (which - // we'd like to be implicitly converted to wxString instead!) - // - // of course, using explicit would be even better - if all compilers - // supported it... - wxArrayString(int autoSort) - : m_nSize(0), m_nCount(0), m_pItems(NULL), m_autoSort(FALSE) - { Init(autoSort != 0); } - // copy ctor - wxArrayString(const wxArrayString& array); - // assignment operator - wxArrayString& operator=(const wxArrayString& src); - // not virtual, this class should not be derived from - ~wxArrayString(); - - // memory management - // empties the list, but doesn't release memory - void Empty(); - // empties the list and releases memory - void Clear(); - // preallocates memory for given number of items - void Alloc(size_t nCount); - // minimzes the memory usage (by freeing all extra memory) - void Shrink(); - - // simple accessors - // number of elements in the array - size_t GetCount() const { return m_nCount; } - // is it empty? - bool IsEmpty() const { return m_nCount == 0; } - // number of elements in the array (GetCount is preferred API) - size_t Count() const { return m_nCount; } - - // items access (range checking is done in debug version) - // get item at position uiIndex - wxString& Item(size_t nIndex) const + wxStringTypeBufferBase(wxString& str, size_t lenWanted = 1024) + : m_str(str), m_buf(lenWanted) { - wxASSERT_MSG( nIndex < m_nCount, - _T("wxArrayString: index out of bounds") ); + // for compatibility with old wxStringBuffer which provided direct + // access to wxString internal buffer, initialize ourselves with the + // string initial contents + + // FIXME-VC6: remove the ugly (CharType *)NULL and use normal + // tchar_str + size_t len; + const wxCharTypeBuffer buf(str.tchar_str(&len, (CharType *)NULL)); + if ( buf ) + { + if ( len > lenWanted ) + { + // in this case there is not enough space for terminating NUL, + // ensure that we still put it there + m_buf.data()[lenWanted] = 0; + len = lenWanted - 1; + } + + memcpy(m_buf.data(), buf, (len + 1)*sizeof(CharType)); + } + //else: conversion failed, this can happen when trying to get Unicode + // string contents into a char string + } + + operator CharType*() { return m_buf.data(); } + +protected: + wxString& m_str; + wxCharTypeBuffer m_buf; +}; + +template +class WXDLLIMPEXP_BASE wxStringTypeBufferLengthBase + : public wxStringTypeBufferBase +{ +public: + wxStringTypeBufferLengthBase(wxString& str, size_t lenWanted = 1024) + : wxStringTypeBufferBase(str, lenWanted), + m_len(0), + m_lenSet(false) + { } - return *(wxString *)&(m_pItems[nIndex]); + ~wxStringTypeBufferLengthBase() + { + wxASSERT_MSG( this->m_lenSet, "forgot to call SetLength()" ); } - // same as Item() - wxString& operator[](size_t nIndex) const { return Item(nIndex); } - // get last item - wxString& Last() const - { - wxASSERT_MSG( !IsEmpty(), - _T("wxArrayString: index out of bounds") ); - return Item(Count() - 1); - } - - // return a wxString[], useful for the controls which - // take one in their ctor. You must delete[] it yourself - // once you are done with it. Will return NULL if the - // ArrayString was empty. - wxString* GetStringArray() const; - - // item management - // Search the element in the array, starting from the beginning if - // bFromEnd is FALSE or from end otherwise. If bCase, comparison is case - // sensitive (default). Returns index of the first item matched or - // wxNOT_FOUND - int Index (const wxChar *sz, bool bCase = TRUE, bool bFromEnd = FALSE) const; - // add new element at the end (if the array is not sorted), return its - // index - size_t Add(const wxString& str, size_t nInsert = 1); - // add new element at given position - void Insert(const wxString& str, size_t uiIndex, size_t nInsert = 1); - // expand the array to have count elements - void SetCount(size_t count); - // remove first item matching this value - void Remove(const wxChar *sz); - // remove item by index - void Remove(size_t nIndex, size_t nRemove = 1); - void RemoveAt(size_t nIndex, size_t nRemove = 1) { Remove(nIndex, nRemove); } - - // sorting - // sort array elements in alphabetical order (or reversed alphabetical - // order if reverseOrder parameter is TRUE) - void Sort(bool reverseOrder = FALSE); - // sort array elements using specified comparaison function - void Sort(CompareFunction compareFunction); - - // comparison - // compare two arrays case sensitively - bool operator==(const wxArrayString& a) const; - // compare two arrays case sensitively - bool operator!=(const wxArrayString& a) const { return !(*this == a); } + void SetLength(size_t length) { m_len = length; m_lenSet = true; } protected: - void Init(bool autoSort); // common part of all ctors - void Copy(const wxArrayString& src); // copies the contents of another array + size_t m_len; + bool m_lenSet; +}; -private: - void Grow(size_t nIncrement = 0); // makes array bigger if needed - void Free(); // free all the strings stored +template +class wxStringTypeBuffer : public wxStringTypeBufferBase +{ +public: + wxStringTypeBuffer(wxString& str, size_t lenWanted = 1024) + : wxStringTypeBufferBase(str, lenWanted) + { } + + ~wxStringTypeBuffer() + { + this->m_str.assign(this->m_buf.data()); + } + + DECLARE_NO_COPY_CLASS(wxStringTypeBuffer) +}; + +template +class wxStringTypeBufferLength : public wxStringTypeBufferLengthBase +{ +public: + wxStringTypeBufferLength(wxString& str, size_t lenWanted = 1024) + : wxStringTypeBufferLengthBase(str, lenWanted) + { } + + ~wxStringTypeBufferLength() + { + this->m_str.assign(this->m_buf.data(), this->m_len); + } + + DECLARE_NO_COPY_CLASS(wxStringTypeBufferLength) +}; - void DoSort(); // common part of all Sort() variants +#if wxUSE_STL_BASED_WXSTRING - size_t m_nSize, // current size of the array - m_nCount; // current number of elements +WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferBase ) - wxChar **m_pItems; // pointer to data +class wxStringInternalBuffer : public wxStringTypeBufferBase +{ +public: + wxStringInternalBuffer(wxString& str, size_t lenWanted = 1024) + : wxStringTypeBufferBase(str, lenWanted) {} + ~wxStringInternalBuffer() + { m_str.m_impl.assign(m_buf.data()); } - bool m_autoSort; // if TRUE, keep the array always sorted + DECLARE_NO_COPY_CLASS(wxStringInternalBuffer) }; -class WXDLLEXPORT wxSortedArrayString : public wxArrayString +WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( + wxStringTypeBufferLengthBase ) + +class wxStringInternalBufferLength + : public wxStringTypeBufferLengthBase { public: - wxSortedArrayString() : wxArrayString(TRUE) - { } - wxSortedArrayString(const wxArrayString& array) : wxArrayString(TRUE) - { Copy(array); } + wxStringInternalBufferLength(wxString& str, size_t lenWanted = 1024) + : wxStringTypeBufferLengthBase(str, lenWanted) {} + + ~wxStringInternalBufferLength() + { + m_str.m_impl.assign(m_buf.data(), m_len); + } + + DECLARE_NO_COPY_CLASS(wxStringInternalBufferLength) }; -// ---------------------------------------------------------------------------- -// wxStringBuffer: a tiny class allowing to get a writable pointer into string -// ---------------------------------------------------------------------------- +#endif // wxUSE_STL_BASED_WXSTRING + + +#if wxUSE_STL_BASED_WXSTRING || wxUSE_UNICODE_UTF8 +typedef wxStringTypeBuffer wxStringBuffer; +typedef wxStringTypeBufferLength wxStringBufferLength; +#else // if !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8 +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 -class WXDLLEXPORT wxStringBuffer +WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferBase ) + +class WXDLLIMPEXP_BASE wxUTF8StringBuffer : public wxStringTypeBufferBase { public: - wxStringBuffer(wxString& str, size_t lenWanted = 1024) - : m_str(str), m_buf(NULL) - { m_buf = m_str.GetWriteBuf(lenWanted); } + wxUTF8StringBuffer(wxString& str, size_t lenWanted = 1024) + : wxStringTypeBufferBase(str, lenWanted) {} + ~wxUTF8StringBuffer(); - ~wxStringBuffer() { m_str.UngetWriteBuf(); } + DECLARE_NO_COPY_CLASS(wxUTF8StringBuffer) +}; - operator wxChar*() const { return m_buf; } +WXDLLIMPEXP_TEMPLATE_INSTANCE_BASE( wxStringTypeBufferLengthBase ) -private: - wxString& m_str; - wxChar *m_buf; +class WXDLLIMPEXP_BASE wxUTF8StringBufferLength + : public wxStringTypeBufferLengthBase +{ +public: + wxUTF8StringBufferLength(wxString& str, size_t lenWanted = 1024) + : wxStringTypeBufferLengthBase(str, lenWanted) {} + ~wxUTF8StringBufferLength(); - DECLARE_NO_COPY_CLASS(wxStringBuffer) + DECLARE_NO_COPY_CLASS(wxUTF8StringBufferLength) }; +#endif // wxUSE_UNICODE_UTF8/wxUSE_UNICODE_WCHAR + // --------------------------------------------------------------------------- // wxString comparison functions: operator versions are always case sensitive // --------------------------------------------------------------------------- +#define wxCMP_WXCHAR_STRING(p, s, op) 0 op s.Cmp(p) + +wxDEFINE_ALL_COMPARISONS(const wxChar *, const wxString&, wxCMP_WXCHAR_STRING) + +#undef wxCMP_WXCHAR_STRING + inline bool operator==(const wxString& s1, const wxString& s2) - { return (s1.Len() == s2.Len()) && (s1.Cmp(s2) == 0); } -inline bool operator==(const wxString& s1, const wxChar * s2) - { return s1.Cmp(s2) == 0; } -inline bool operator==(const wxChar * s1, const wxString& s2) - { return s2.Cmp(s1) == 0; } + { return s1.IsSameAs(s2); } inline bool operator!=(const wxString& s1, const wxString& s2) - { return (s1.Len() != s2.Len()) || (s1.Cmp(s2) != 0); } -inline bool operator!=(const wxString& s1, const wxChar * s2) - { return s1.Cmp(s2) != 0; } -inline bool operator!=(const wxChar * s1, const wxString& s2) - { return s2.Cmp(s1) != 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 wxChar * s2) - { return s1.Cmp(s2) < 0; } -inline bool operator< (const wxChar * s1, const wxString& s2) - { return s2.Cmp(s1) > 0; } inline bool operator> (const wxString& s1, const wxString& s2) { return s1.Cmp(s2) > 0; } -inline bool operator> (const wxString& s1, const wxChar * s2) - { return s1.Cmp(s2) > 0; } -inline bool operator> (const wxChar * s1, const wxString& s2) - { return s2.Cmp(s1) < 0; } inline bool operator<=(const wxString& s1, const wxString& s2) { return s1.Cmp(s2) <= 0; } -inline bool operator<=(const wxString& s1, const wxChar * s2) - { return s1.Cmp(s2) <= 0; } -inline bool operator<=(const wxChar * s1, const wxString& s2) - { return s2.Cmp(s1) >= 0; } inline bool operator>=(const wxString& s1, const wxString& s2) { return s1.Cmp(s2) >= 0; } -inline bool operator>=(const wxString& s1, const wxChar * s2) - { return s1.Cmp(s2) >= 0; } -inline bool operator>=(const wxChar * s1, const wxString& s2) - { return s2.Cmp(s1) <= 0; } -// comparison with char -inline bool operator==(wxChar c, const wxString& s) { return s.IsSameAs(c); } -inline bool operator==(const wxString& s, wxChar c) { return s.IsSameAs(c); } -inline bool operator!=(wxChar c, const wxString& s) { return !s.IsSameAs(c); } -inline bool operator!=(const wxString& s, wxChar c) { return !s.IsSameAs(c); } +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; } -#if wxUSE_UNICODE 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) @@ -1222,7 +3815,7 @@ 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) @@ -1231,39 +3824,261 @@ 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 -wxString WXDLLEXPORT operator+(const wxString& string1, const wxString& string2); -wxString WXDLLEXPORT operator+(const wxString& string, wxChar ch); -wxString WXDLLEXPORT operator+(wxChar ch, const wxString& string); -wxString WXDLLEXPORT operator+(const wxString& string, const wxChar *psz); -wxString WXDLLEXPORT operator+(const wxChar *psz, const wxString& string); -#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); } +inline bool operator==(const wxUniCharRef& c, const wxString& s) { return s.IsSameAs(c); } +inline bool operator==(char c, const wxString& s) { return s.IsSameAs(c); } +inline bool operator==(wchar_t c, const wxString& s) { return s.IsSameAs(c); } +inline bool operator==(int c, const wxString& s) { return s.IsSameAs(c); } +inline bool operator==(const wxString& s, const wxUniChar& c) { return s.IsSameAs(c); } +inline bool operator==(const wxString& s, const wxUniCharRef& c) { return s.IsSameAs(c); } +inline bool operator==(const wxString& s, char c) { return s.IsSameAs(c); } +inline bool operator==(const wxString& s, wchar_t c) { return s.IsSameAs(c); } +inline bool operator!=(const wxUniChar& c, const wxString& s) { return !s.IsSameAs(c); } +inline bool operator!=(const wxUniCharRef& c, const wxString& s) { return !s.IsSameAs(c); } +inline bool operator!=(char c, const wxString& s) { return !s.IsSameAs(c); } +inline bool operator!=(wchar_t c, const wxString& s) { return !s.IsSameAs(c); } +inline bool operator!=(int c, const wxString& s) { return !s.IsSameAs(c); } +inline bool operator!=(const wxString& s, const wxUniChar& c) { return !s.IsSameAs(c); } +inline bool operator!=(const wxString& s, const wxUniCharRef& c) { return !s.IsSameAs(c); } +inline bool operator!=(const wxString& s, char c) { return !s.IsSameAs(c); } +inline bool operator!=(const wxString& s, wchar_t c) { return !s.IsSameAs(c); } + +// comparison with C string in Unicode build +#if wxUSE_UNICODE + +#define wxCMP_CHAR_STRING(p, s, op) wxString(p) op s + +wxDEFINE_ALL_COMPARISONS(const char *, const wxString&, wxCMP_CHAR_STRING) + +#undef wxCMP_CHAR_STRING + +#endif // wxUSE_UNICODE + +// we also need to provide the operators for comparison with wxCStrData to +// resolve ambiguity between operator(const wxChar *,const wxString &) and +// operator(const wxChar *, const wxChar *) for "p == s.c_str()" +// +// notice that these are (shallow) pointer comparisons, not (deep) string ones +#define wxCMP_CHAR_CSTRDATA(p, s, op) p op s.AsChar() +#define wxCMP_WCHAR_CSTRDATA(p, s, op) p op s.AsWChar() + +wxDEFINE_ALL_COMPARISONS(const wchar_t *, const wxCStrData&, wxCMP_WCHAR_CSTRDATA) +wxDEFINE_ALL_COMPARISONS(const char *, const wxCStrData&, wxCMP_CHAR_CSTRDATA) + +#undef wxCMP_CHAR_CSTRDATA +#undef wxCMP_WCHAR_CSTRDATA // --------------------------------------------------------------------------- // Implementation only from here until the end of file // --------------------------------------------------------------------------- -// don't pollute the library user's name space -#undef wxASSERT_VALID_INDEX +#if wxUSE_STD_IOSTREAM + +#include "wx/iosfwrap.h" + +WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxString&); +WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxCStrData&); +WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxCharBuffer&); +#ifndef __BORLANDC__ +WXDLLIMPEXP_BASE wxSTD ostream& operator<<(wxSTD ostream&, const wxWCharBuffer&); +#endif + +#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 +// --------------------------------------------------------------------------- + +inline wxCStrData::wxCStrData(char *buf) + : m_str(new wxString(buf)), m_offset(0), m_owned(true) {} +inline wxCStrData::wxCStrData(wchar_t *buf) + : m_str(new wxString(buf)), m_offset(0), m_owned(true) {} + +inline wxCStrData::wxCStrData(const wxCStrData& data) + : m_str(data.m_owned ? new wxString(*data.m_str) : data.m_str), + m_offset(data.m_offset), + m_owned(data.m_owned) +{ +} + +inline wxCStrData::~wxCStrData() +{ + if ( m_owned ) + delete wx_const_cast(wxString*, m_str); // cast to silence warnings +} + +// simple cases for AsChar() and AsWChar(), the complicated ones are +// in string.cpp +#if wxUSE_UNICODE_WCHAR +inline const wchar_t* wxCStrData::AsWChar() const +{ + return m_str->wx_str() + m_offset; +} +#endif // wxUSE_UNICODE_WCHAR + +#if !wxUSE_UNICODE +inline const char* wxCStrData::AsChar() const +{ + 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 + return wxCharBuffer::CreateNonOwned(AsChar()); +#else + return AsString().mb_str(); +#endif +} + +inline const wxWCharBuffer wxCStrData::AsWCharBuf() const +{ +#if wxUSE_UNICODE_WCHAR + return wxWCharBuffer::CreateNonOwned(AsWChar()); +#else + return AsString().wc_str(); +#endif +} + +inline wxString wxCStrData::AsString() const +{ + if ( m_offset == 0 ) + return *m_str; + else + return m_str->Mid(m_offset); +} + +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 +{ + if ( m_str->empty() ) + return wxUniChar(_T('\0')); + else + return (*m_str)[m_offset]; +} + +inline wxUniChar wxCStrData::operator[](size_t n) const +{ + // NB: we intentionally use operator[] and not at() here because the former + // works for the terminating NUL while the latter does not + return (*m_str)[m_offset + n]; +} + +// ---------------------------------------------------------------------------- +// more wxCStrData operators +// ---------------------------------------------------------------------------- + +// we need to define those to allow "size_t pos = p - s.c_str()" where p is +// some pointer into the string +inline size_t operator-(const char *p, const wxCStrData& cs) +{ + return p - cs.AsChar(); +} + +inline size_t operator-(const wchar_t *p, const wxCStrData& cs) +{ + return p - cs.AsWChar(); +} + +// ---------------------------------------------------------------------------- +// implementation of wx[W]CharBuffer inline methods using wxCStrData +// ---------------------------------------------------------------------------- + +// FIXME-UTF8: move this to buffer.h +inline wxCharBuffer::wxCharBuffer(const wxCStrData& cstr) + : wxCharTypeBufferBase(cstr.AsCharBuf()) +{ +} -#if defined(wxSTD_STRING_COMPATIBILITY) && wxUSE_STD_IOSTREAM +inline wxWCharBuffer::wxWCharBuffer(const wxCStrData& cstr) + : wxCharTypeBufferBase(cstr.AsWCharBuf()) +{ +} -#include "wx/ioswrap.h" +#if wxUSE_UNICODE_UTF8 +// ---------------------------------------------------------------------------- +// implementation of wxStringIteratorNode inline methods +// ---------------------------------------------------------------------------- -WXDLLEXPORT wxSTD istream& operator>>(wxSTD istream&, wxString&); -WXDLLEXPORT wxSTD ostream& operator<<(wxSTD ostream&, const wxString&); +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; + } +} -#endif // wxSTD_STRING_COMPATIBILITY +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/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/crt.h" +#endif -#endif // _WX_WXSTRINGH__ +#endif // _WX_WXSTRING_H_