+/*
+ Efficient string class [more or less] compatible with MFC CString,
+ wxWidgets version 1 wxString and std::string and some handy functions
+ missing from string.h.
+*/
+
+#ifndef _WX_WXSTRING_H__
+#define _WX_WXSTRING_H__
+
+// ----------------------------------------------------------------------------
+// headers
+// ----------------------------------------------------------------------------
+
+#include "wx/defs.h" // everybody should include this
+
+#ifndef __WXPALMOS5__
+#if defined(__WXMAC__) || defined(__VISAGECPP__)
+ #include <ctype.h>
+#endif
+
+#if defined(__VISAGECPP__) && __IBMCPP__ >= 400
+ // problem in VACPP V4 with including stdlib.h multiple times
+ // strconv includes it anyway
+# include <stdio.h>
+# include <string.h>
+# include <stdarg.h>
+# include <limits.h>
+#else
+# include <string.h>
+# include <stdio.h>
+# include <stdarg.h>
+# include <limits.h>
+# include <stdlib.h>
+#endif
+
+#ifdef HAVE_STRCASECMP_IN_STRINGS_H
+ #include <strings.h> // for strcasecmp()
+#endif // HAVE_STRCASECMP_IN_STRINGS_H
+#endif // ! __WXPALMOS5__
+
+#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
+// CS: under OSX release builds the string destructor/cache cleanup sometimes
+// crashes, disable until we find the true reason or a better workaround
+#if wxUSE_UNICODE_UTF8 && !defined(__WXMSW__) && !defined(__WXOSX__)
+ #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 <typename T> struct wxStringAsBufHelper;
+}
+
+// ---------------------------------------------------------------------------
+// macros
+// ---------------------------------------------------------------------------
+
+// casts [unfortunately!] needed to call some broken functions which require
+// "char *" instead of "const char *"
+#define WXSTRINGCAST (wxChar *)(const wxChar *)
+#define wxCSTRINGCAST (wxChar *)(const wxChar *)
+#define wxMBSTRINGCAST (char *)(const char *)
+#define wxWCSTRINGCAST (wchar_t *)(const wchar_t *)
+
+// ----------------------------------------------------------------------------
+// constants
+// ----------------------------------------------------------------------------
+
+#if WXWIN_COMPATIBILITY_2_6
+
+// deprecated in favour of wxString::npos, don't use in new code
+//
+// maximum possible length for a string means "take all string" everywhere
+#define wxSTRING_MAXLEN wxString::npos
+
+#endif // WXWIN_COMPATIBILITY_2_6
+
+// ---------------------------------------------------------------------------
+// global functions complementing standard C string library replacements for
+// strlen() and portable strcasecmp()
+//---------------------------------------------------------------------------
+
+#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(__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(__BORLANDC__)
+ return stricmp(psz1, psz2);
+#elif defined(__WATCOMC__)
+ return stricmp(psz1, psz2);
+#elif defined(__DJGPP__)
+ return stricmp(psz1, psz2);
+#elif defined(__EMX__)
+ return stricmp(psz1, psz2);
+#elif defined(__WXPM__)
+ return stricmp(psz1, psz2);
+#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;
+ do {
+ c1 = tolower(*psz1++);
+ c2 = tolower(*psz2++);
+ } while ( c1 && (c1 == c2) );
+
+ return c1 - c2;
+#else
+ // almost all compilers/libraries provide this function (unfortunately under
+ // different names), that's why we don't implement our own which will surely
+ // be more efficient than this code (uncomment to use):
+ /*
+ register char c1, c2;
+ do {
+ c1 = tolower(*psz1++);
+ c2 = tolower(*psz2++);
+ } while ( c1 && (c1 == c2) );
+
+ return c1 - c2;
+ */
+
+ #error "Please define string case-insensitive compare for your OS/compiler"
+#endif // OS/compiler
+}
+
+#endif // WXWIN_COMPATIBILITY_2_8
+
+// ----------------------------------------------------------------------------
+// wxCStrData
+// ----------------------------------------------------------------------------
+
+// Lightweight object returned by wxString::c_str() and implicitly convertible
+// to either const char* or const wchar_t*.
+class wxCStrData
+{
+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 "<symbol> 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(); }
+
+#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(); }
+
+ operator const void*() const { return AsChar(); }
+
+ // returns buffers that are valid as long as the associated wxString exists
+ const wxScopedCharBuffer AsCharBuf() const
+ {
+ return wxScopedCharBuffer::CreateNonOwned(AsChar());
+ }
+
+ const wxScopedWCharBuffer AsWCharBuf() const
+ {
+ return wxScopedWCharBuffer::CreateNonOwned(AsWChar());
+ }
+
+ inline wxString AsString() const;
+
+ // 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). Note that this means
+ // the 'n' value is interpreted as addition to char*/wchar_t* pointer, it
+ // is *not* number of Unicode characters in wxString.
+ 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:
+ // the wxString this object was returned for
+ const wxString *m_str;
+ // Offset into c_str() return value. Note that this is *not* offset in
+ // m_str in Unicode characters. Instead, it is index into the
+ // char*/wchar_t* buffer returned by c_str(). It's interpretation depends
+ // on how is the wxCStrData instance used: if it is eventually cast to
+ // const char*, m_offset will be in bytes form string's start; if it is
+ // cast to const wchar_t*, it will be in wchar_t values.
+ size_t m_offset;
+ // should m_str be deleted, i.e. is it owned by us?
+ bool m_owned;
+
+ friend class WXDLLIMPEXP_FWD_BASE wxString;
+};
+
+// ----------------------------------------------------------------------------
+// wxStringPrintfMixin
+// ---------------------------------------------------------------------------
+
+// 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).
+//
+// 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<typename T> 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, ...) WX_ATTRIBUTE_PRINTF_1;
+ WX_DEFINE_VARARG_FUNC_SANS_N0(static typename StringReturnType<T1>::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<typename T>
+ inline static typename StringReturnType<T>::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, ...) WX_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
+// ---------------------------------------------------------------------------
+
+#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
+{
+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
+ wxDECLARE_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
+ // 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)
+ //
+ // NB: we don't need other versions (short/long and unsigned) as attempt
+ // to assign another numeric type to wxString will now result in
+ // ambiguity between operator=(char) and operator=(int)
+ wxString& operator=(int);
+
+ // these methods are not implemented - there is _no_ conversion from int to
+ // string, you're doing something wrong if the compiler wants to call it!
+ //
+ // try `s << i' or `s.Printf("%d", i)' instead
+ wxString(int);
+
+
+ // buffer for holding temporary substring when using any of the methods
+ // that take (char*,size_t) or (wchar_t*,size_t) arguments:
+ template<typename T>
+ struct SubstrBufFromType
+ {
+ T data;
+ size_t len;
+
+ 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<wxScopedCharBuffer> SubstrBufFromWC;
+ typedef SubstrBufFromType<wxScopedCharBuffer> SubstrBufFromMB;
+#elif wxUSE_UNICODE_WCHAR
+ typedef SubstrBufFromType<const wchar_t*> SubstrBufFromWC;
+ typedef SubstrBufFromType<wxScopedWCharBuffer> SubstrBufFromMB;
+#else
+ typedef SubstrBufFromType<const char*> SubstrBufFromMB;
+ typedef SubstrBufFromType<wxScopedCharBuffer> SubstrBufFromWC;
+#endif
+
+
+ // 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.
+
+#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
+
+#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 wxScopedWCharBuffer 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 wxScopedCharBuffer 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 wxScopedCharBuffer 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 wxScopedCharBuffer 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 };
+
+ 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()
+ {
+ static wxTLS_TYPE(Cache) s_cache;
+
+ return wxTLS_VALUE(s_cache);
+ }
+
+ // 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
+ {
+ // 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;
+ }
+
+ return NULL;
+ }
+
+ // 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 )
+ {
+ // claim the next cache entry for this string
+ c = cacheStart;
+ if ( ++c == cacheEnd )
+ c = cacheBegin;
+
+ c->str = this;
+ c->Reset();
+
+ // and remember the last used element
+ LastUsedCacheElement() = c - cacheBegin;
+ }
+
+ return c;
+ }
+
+ size_t DoPosToImpl(size_t pos) const
+ {
+ wxCACHE_PROFILE_FIELD_INC(postot);
+
+ // 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
+
+ wxCACHE_PROFILE_FIELD_ADD(sumpos, pos);
+
+ Cache::Element * const cache = GetCacheElement();
+
+ // 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);
+ }
+
+ if ( pos == cache->pos )
+ return cache->impl;
+
+ // 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);
+
+ cache->ResetPos();
+ }
+
+ wxCACHE_PROFILE_FIELD_ADD(sumofs, pos - cache->pos);
+
+
+ wxStringImpl::const_iterator i(m_impl.begin() + cache->impl);
+ for ( size_t n = cache->pos; n < pos; n++ )
+ wxStringOperations::IncIter(i);
+
+ cache->pos = pos;
+ cache->impl = i - m_impl.begin();
+
+ wxSTRING_CACHE_ASSERT(
+ (int)cache->impl == (begin() + pos).impl() - m_impl.begin() );
+
+ return cache->impl;
+ }
+
+ void InvalidateCache()
+ {
+ Cache::Element * const cache = FindCacheElement();
+ if ( cache )
+ cache->Reset();
+ }
+
+ void InvalidateCachedLength()
+ {
+ Cache::Element * const cache = FindCacheElement();
+ if ( cache )
+ cache->len = npos;
+ }
+
+ 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;
+ }
+
+ void UpdateCachedLength(ptrdiff_t delta)
+ {
+ Cache::Element * const cache = FindCacheElement();
+ if ( cache && cache->len != npos )
+ {
+ wxSTRING_CACHE_ASSERT( (ptrdiff_t)cache->len + delta >= 0 );
+
+ cache->len += delta;
+ }
+ }
+
+ #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();
+ }
+
+ #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
+
+ size_t PosToImpl(size_t pos) const
+ {
+ return pos == 0 || pos == npos ? pos : DoPosToImpl(pos);
+ }
+
+ void PosLenToImpl(size_t pos, size_t len, size_t *implPos, size_t *implLen) const;
+
+ size_t LenToImpl(size_t len) const
+ {
+ size_t pos, len2;
+ PosLenToImpl(0, len, &pos, &len2);
+ return len2;
+ }
+
+ 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
+
+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 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 <typename T>
+ 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 <T> in the functions below is to keep BCC 5.5 happy
+ reverse_iterator_impl operator+(ptrdiff_t n) const
+ { return reverse_iterator_impl<T>(m_cur - n); }
+ reverse_iterator_impl operator-(ptrdiff_t n) const
+ { return reverse_iterator_impl<T>(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<iterator> reverse_iterator;
+ typedef reverse_iterator_impl<const_iterator> 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 wxScopedCharBuffer& buf)
+ { assign(buf.data(), buf.length()); }
+ wxString(const wxScopedWCharBuffer& buf)
+ { assign(buf.data(), buf.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
+ {
+ wxScopedWCharBuffer buf(wc_str());
+ return wxStdWideString(buf.data(), buf.length());
+ }
+ #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
+ {
+ wxScopedCharBuffer buf(mb_str());
+ return std::string(buf.data(), buf.length());
+ }
+ #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 )
+ {
+ wxSTRING_INVALIDATE_CACHE();
+
+ // we can't use wxStringImpl::resize() for truncating the string as it
+ // counts in bytes, not characters
+ erase(nSize);
+ return;
+ }
+
+ // we also can't use (presumably more efficient) resize() if we have to
+ // append characters taking more than one byte
+ if ( !ch.IsAscii() )
+ {
+ append(nSize - len, ch);
+ }
+ else // can use (presumably faster) resize() version
+#endif // wxUSE_UNICODE_UTF8
+ {
+ wxSTRING_INVALIDATE_CACHED_LENGTH();
+
+ m_impl.resize(nSize, (wxStringCharType)ch);
+ }
+ }
+
+ 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));
+ }
+
+ // 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(); }
+
+ // 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
+ 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; }
+
+ // get last character
+ wxUniChar Last() const
+ {
+ wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") );
+ return *rbegin();
+ }
+
+ // get writable last character
+ wxUniCharRef Last()
+ {
+ wxASSERT_MSG( !empty(), _T("wxString: index out of bounds") );
+ return *rbegin();
+ }
+
+ /*
+ 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
+
+ // 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
+
+
+ /*
+ 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).
+ */
+
+ // explicit conversion to wxCStrData
+ wxCStrData c_str() const { return wxCStrData(this); }
+ wxCStrData data() const { return c_str(); }
+
+ // implicit conversion to wxCStrData
+ operator wxCStrData() const { return c_str(); }
+
+ // the first two operators conflict with operators for conversion to
+ // std::string and they must be disabled in STL build; the next one only
+ // makes sense if conversions to char* are also defined and not defining it
+ // in STL build also helps us to get more clear error messages for the code
+ // which relies on implicit conversion to char* in STL build
+#if !wxUSE_STL
+ operator const char*() const { return c_str(); }
+ operator const wchar_t*() const { return c_str(); }
+
+ // 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
+
+ // identical to c_str(), for MFC compatibility
+ const wxCStrData GetData() const { return c_str(); }
+
+ // 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(); }
+
+ // 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 <typename T>
+ wxCharTypeBuffer<T> 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<T>::Get(*this, len);
+#else // ANSI
+ // T can only be char in ANSI build
+ if ( len )
+ *len = length();
+
+ return wxCharTypeBuffer<T>::CreateNonOwned(wx_str(), length());
+#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 wxScopedCharBuffer 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 wxScopedCharBuffer utf8_str() const
+ { return wxCharBuffer::CreateNonOwned(m_impl.c_str(), m_impl.length()); }
+
+ // 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 wxScopedCharBuffer utf8_str() const { return mb_str(wxMBConvUTF8()); }
+#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;
+ wxScopedWCharBuffer 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;
+ wxScopedWCharBuffer 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 wxScopedCharBuffer utf8_str() const
+ { return wxMBConvUTF8().cWC2MB(wc_str()); }
+#endif
+
+ const wxScopedCharBuffer ToUTF8() const { return utf8_str(); }
+
+ // 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 wxScopedCharBuffer 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
+
+ // this is an optimization: even though using mb_str(wxConvLibc) does the
+ // same thing (i.e. returns pointer to internal representation as locale is
+ // always an UTF-8 one) in wxUSE_UTF8_LOCALE_ONLY case, we can avoid the
+ // extra checks and the temporary buffer construction by providing a
+ // separate mb_str() overload
+#if wxUSE_UTF8_LOCALE_ONLY
+ const char* mb_str() const { return wx_str(); }
+ const wxScopedCharBuffer mb_str(const wxMBConv& conv) const
+ {
+ return AsCharBuf(conv);
+ }
+#else // !wxUSE_UTF8_LOCALE_ONLY
+ const wxScopedCharBuffer mb_str(const wxMBConv& conv = wxConvLibc) const
+ {
+ return AsCharBuf(conv);
+ }
+#endif // wxUSE_UTF8_LOCALE_ONLY/!wxUSE_UTF8_LOCALE_ONLY
+
+ 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 wxScopedWCharBuffer wc_str() const
+ { return AsWCharBuf(wxMBConvStrictUTF8()); }
+#endif
+ // for compatibility with !wxUSE_UNICODE version
+ const wxWX2WCbuf wc_str(const wxMBConv& WXUNUSED(conv)) const
+ { return wc_str(); }
+
+#if wxMBFILES
+ const wxScopedCharBuffer fn_str() const { return mb_str(wxConvFile); }
+#else // !wxMBFILES
+ const wxWX2WCbuf fn_str() const { return wc_str(); }
+#endif // wxMBFILES/!wxMBFILES
+
+#else // ANSI
+ const char* 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(); }
+
+ const wxScopedWCharBuffer wc_str(const wxMBConv& conv = wxConvLibc) const
+ { return AsWCharBuf(conv); }
+
+ const wxScopedCharBuffer fn_str() const
+ { return wxConvFile.cWC2WX( wc_str( wxConvLibc ) ); }
+#endif // Unicode/ANSI
+
+#if wxUSE_UNICODE_UTF8
+ const wxScopedWCharBuffer 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 wxScopedWCharBuffer
+ wxString& operator=(const wxScopedWCharBuffer& s)
+ { return assign(s); }
+ // from wxScopedCharBuffer
+ wxString& operator=(const wxScopedCharBuffer& s)
+ { return assign(s); }
+
+ // 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 wxScopedWCharBuffer& s)
+ { return append(s); }
+ wxString& operator<<(const wxScopedCharBuffer& s)
+ { return append(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 wxScopedCharBuffer& psz)
+ { append(psz); return *this; }
+ wxString& Append(const wxScopedWCharBuffer& 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 wxScopedCharBuffer& psz, size_t nLen)
+ { append(psz, nLen); return *this; }
+ wxString& Append(const wxScopedWCharBuffer& 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 wxScopedCharBuffer& s) const
+ { return compare(s); }
+ int Cmp(const wxScopedWCharBuffer& 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 wxScopedCharBuffer& str, bool compareWithCase = true) const
+ { return IsSameAs(str.data(), compareWithCase); }
+ bool IsSameAs(const wxScopedWCharBuffer& 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 wxScopedCharBuffer& sub) const
+ { return Find(sub.data()); }
+ int Find(const wxScopedWCharBuffer& 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;
+
+#if wxUSE_XLOCALE
+ // conversions to numbers using C locale
+ // convert to a signed integer
+ bool ToCLong(long *val, int base = 10) const;
+ // convert to an unsigned integer
+ bool ToCULong(unsigned long *val, int base = 10) const;
+ // convert to a double
+ bool ToCDouble(double *val) const;
+#endif
+
+#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, ...) WX_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, ...) WX_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 wxScopedCharBuffer& str)
+ { return append(str.data(), str.length()); }
+ wxString& append(const wxScopedWCharBuffer& str)
+ { return append(str.data(), str.length()); }
+ wxString& append(const wxCStrData& str, size_t n)
+ { return append(str.AsString(), 0, n); }
+ wxString& append(const wxScopedCharBuffer& str, size_t n)
+ { return append(str.data(), n); }
+ wxString& append(const wxScopedWCharBuffer& 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 wxScopedCharBuffer& str)
+ { return assign(str.data(), str.length()); }
+ wxString& assign(const wxScopedWCharBuffer& str)
+ { return assign(str.data(), str.length()); }
+ wxString& assign(const wxCStrData& str, size_t len)
+ { return assign(str.AsString(), len); }
+ wxString& assign(const wxScopedCharBuffer& str, size_t len)
+ { return assign(str.data(), len); }
+ wxString& assign(const wxScopedWCharBuffer& 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 wxScopedCharBuffer& str) const
+ { return compare(str.data()); }
+ int compare(const wxScopedWCharBuffer& 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));
projects / wxWidgets.git / blobdiff