///////////////////////////////////////////////////////////////////////////// // Name: src/common/string.cpp // Purpose: wxString class // Author: Vadim Zeitlin, Ryan Norton // Modified by: // Created: 29/01/98 // RCS-ID: $Id$ // Copyright: (c) 1998 Vadim Zeitlin // (c) 2004 Ryan Norton // Licence: wxWindows licence ///////////////////////////////////////////////////////////////////////////// /* * About ref counting: * 1) all empty strings use g_strEmpty, nRefs = -1 (set in Init()) * 2) AllocBuffer() sets nRefs to 1, Lock() increments it by one * 3) Unlock() decrements nRefs and frees memory if it goes to 0 */ // =========================================================================== // headers, declarations, constants // =========================================================================== // For compilers that support precompilation, includes "wx.h". #include "wx/wxprec.h" #ifdef __BORLANDC__ #pragma hdrstop #endif #ifndef WX_PRECOMP #include "wx/string.h" #include "wx/intl.h" #include "wx/thread.h" #endif #include #ifndef __WXWINCE__ #include #endif #include #include #ifdef __SALFORDC__ #include #endif // allocating extra space for each string consumes more memory but speeds up // the concatenation operations (nLen is the current string's length) // NB: EXTRA_ALLOC must be >= 0! #define EXTRA_ALLOC (19 - nLen % 16) // string handling functions used by wxString: #if wxUSE_UNICODE_UTF8 #define wxStringMemcpy memcpy #define wxStringMemcmp memcmp #define wxStringMemchr memchr #define wxStringStrlen strlen #else #define wxStringMemcpy wxTmemcpy #define wxStringMemcmp wxTmemcmp #define wxStringMemchr wxTmemchr #define wxStringStrlen wxStrlen #endif // --------------------------------------------------------------------------- // static class variables definition // --------------------------------------------------------------------------- #if !wxUSE_STL_BASED_WXSTRING //According to STL _must_ be a -1 size_t const size_t wxStringImpl::npos = (size_t) -1; #endif const size_t wxString::npos = (size_t) -1; // ---------------------------------------------------------------------------- // static data // ---------------------------------------------------------------------------- #if wxUSE_STL_BASED_WXSTRING extern const wxChar WXDLLIMPEXP_BASE *wxEmptyString = _T(""); #else // for an empty string, GetStringData() will return this address: this // structure has the same layout as wxStringData and it's data() method will // return the empty string (dummy pointer) static const struct { wxStringData data; wxChar dummy; } g_strEmpty = { {-1, 0, 0}, wxT('\0') }; // empty C style string: points to 'string data' byte of g_strEmpty extern const wxChar WXDLLIMPEXP_BASE *wxEmptyString = &g_strEmpty.dummy; #endif // ---------------------------------------------------------------------------- // global functions // ---------------------------------------------------------------------------- #if wxUSE_STD_IOSTREAM #include wxSTD ostream& operator<<(wxSTD ostream& os, const wxString& str) { return os << str.c_str(); } wxSTD ostream& operator<<(wxSTD ostream& os, const wxCStrData& str) { #if wxUSE_UNICODE && !defined(__BORLANDC__) return os << str.AsWChar(); #else return os << str.AsChar(); #endif } #endif // wxUSE_STD_IOSTREAM // ---------------------------------------------------------------------------- // private classes // ---------------------------------------------------------------------------- // this small class is used to gather statistics for performance tuning //#define WXSTRING_STATISTICS #ifdef WXSTRING_STATISTICS class Averager { public: Averager(const wxChar *sz) { m_sz = sz; m_nTotal = m_nCount = 0; } ~Averager() { wxPrintf("wxString: average %s = %f\n", m_sz, ((float)m_nTotal)/m_nCount); } void Add(size_t n) { m_nTotal += n; m_nCount++; } private: size_t m_nCount, m_nTotal; const wxChar *m_sz; } g_averageLength("allocation size"), g_averageSummandLength("summand length"), g_averageConcatHit("hit probability in concat"), g_averageInitialLength("initial string length"); #define STATISTICS_ADD(av, val) g_average##av.Add(val) #else #define STATISTICS_ADD(av, val) #endif // WXSTRING_STATISTICS #if !wxUSE_STL_BASED_WXSTRING // =========================================================================== // wxStringData class deallocation // =========================================================================== #if defined(__VISUALC__) && defined(_MT) && !defined(_DLL) # pragma message (__FILE__ ": building with Multithreaded non DLL runtime has a performance impact on wxString!") void wxStringData::Free() { free(this); } #endif // =========================================================================== // wxStringImpl // =========================================================================== // takes nLength elements of psz starting at nPos void wxStringImpl::InitWith(const wxChar *psz, size_t nPos, size_t nLength) { Init(); // if the length is not given, assume the string to be NUL terminated if ( nLength == npos ) { wxASSERT_MSG( nPos <= wxStrlen(psz), _T("index out of bounds") ); nLength = wxStrlen(psz + nPos); } STATISTICS_ADD(InitialLength, nLength); if ( nLength > 0 ) { // trailing '\0' is written in AllocBuffer() if ( !AllocBuffer(nLength) ) { wxFAIL_MSG( _T("out of memory in wxStringImpl::InitWith") ); return; } wxStringMemcpy(m_pchData, psz + nPos, nLength); } } // poor man's iterators are "void *" pointers wxStringImpl::wxStringImpl(const void *pStart, const void *pEnd) { if ( pEnd >= pStart ) { InitWith((const wxChar *)pStart, 0, (const wxChar *)pEnd - (const wxChar *)pStart); } else { wxFAIL_MSG( _T("pStart is not before pEnd") ); Init(); } } wxStringImpl::wxStringImpl(size_type n, wxStringCharType ch) { Init(); append(n, ch); } // --------------------------------------------------------------------------- // memory allocation // --------------------------------------------------------------------------- // allocates memory needed to store a C string of length nLen bool wxStringImpl::AllocBuffer(size_t nLen) { // allocating 0 sized buffer doesn't make sense, all empty strings should // reuse g_strEmpty wxASSERT( nLen > 0 ); // make sure that we don't overflow wxASSERT( nLen < (INT_MAX / sizeof(wxChar)) - (sizeof(wxStringData) + EXTRA_ALLOC + 1) ); STATISTICS_ADD(Length, nLen); // allocate memory: // 1) one extra character for '\0' termination // 2) sizeof(wxStringData) for housekeeping info wxStringData* pData = (wxStringData*) malloc(sizeof(wxStringData) + (nLen + EXTRA_ALLOC + 1)*sizeof(wxChar)); if ( pData == NULL ) { // allocation failures are handled by the caller return false; } pData->nRefs = 1; pData->nDataLength = nLen; pData->nAllocLength = nLen + EXTRA_ALLOC; m_pchData = pData->data(); // data starts after wxStringData m_pchData[nLen] = wxT('\0'); return true; } // must be called before changing this string bool wxStringImpl::CopyBeforeWrite() { wxStringData* pData = GetStringData(); if ( pData->IsShared() ) { pData->Unlock(); // memory not freed because shared size_t nLen = pData->nDataLength; if ( !AllocBuffer(nLen) ) { // allocation failures are handled by the caller return false; } wxStringMemcpy(m_pchData, pData->data(), nLen); } wxASSERT( !GetStringData()->IsShared() ); // we must be the only owner return true; } // must be called before replacing contents of this string bool wxStringImpl::AllocBeforeWrite(size_t nLen) { wxASSERT( nLen != 0 ); // doesn't make any sense // must not share string and must have enough space wxStringData* pData = GetStringData(); if ( pData->IsShared() || pData->IsEmpty() ) { // can't work with old buffer, get new one pData->Unlock(); if ( !AllocBuffer(nLen) ) { // allocation failures are handled by the caller return false; } } else { if ( nLen > pData->nAllocLength ) { // realloc the buffer instead of calling malloc() again, this is more // efficient STATISTICS_ADD(Length, nLen); nLen += EXTRA_ALLOC; pData = (wxStringData*) realloc(pData, sizeof(wxStringData) + (nLen + 1)*sizeof(wxChar)); if ( pData == NULL ) { // allocation failures are handled by the caller // keep previous data since reallocation failed return false; } pData->nAllocLength = nLen; m_pchData = pData->data(); } } wxASSERT( !GetStringData()->IsShared() ); // we must be the only owner // it doesn't really matter what the string length is as it's going to be // overwritten later but, for extra safety, set it to 0 for now as we may // have some junk in m_pchData GetStringData()->nDataLength = 0; return true; } wxStringImpl& wxStringImpl::append(size_t n, wxStringCharType ch) { size_type len = length(); if ( !Alloc(len + n) || !CopyBeforeWrite() ) { wxFAIL_MSG( _T("out of memory in wxStringImpl::append") ); } GetStringData()->nDataLength = len + n; m_pchData[len + n] = '\0'; for ( size_t i = 0; i < n; ++i ) m_pchData[len + i] = ch; return *this; } void wxStringImpl::resize(size_t nSize, wxStringCharType ch) { size_t len = length(); if ( nSize < len ) { erase(begin() + nSize, end()); } else if ( nSize > len ) { append(nSize - len, ch); } //else: we have exactly the specified length, nothing to do } // allocate enough memory for nLen characters bool wxStringImpl::Alloc(size_t nLen) { wxStringData *pData = GetStringData(); if ( pData->nAllocLength <= nLen ) { if ( pData->IsEmpty() ) { nLen += EXTRA_ALLOC; pData = (wxStringData *) malloc(sizeof(wxStringData) + (nLen + 1)*sizeof(wxChar)); if ( pData == NULL ) { // allocation failure handled by caller return false; } pData->nRefs = 1; pData->nDataLength = 0; pData->nAllocLength = nLen; m_pchData = pData->data(); // data starts after wxStringData m_pchData[0u] = wxT('\0'); } else if ( pData->IsShared() ) { pData->Unlock(); // memory not freed because shared size_t nOldLen = pData->nDataLength; if ( !AllocBuffer(nLen) ) { // allocation failure handled by caller return false; } // +1 to copy the terminator, too memcpy(m_pchData, pData->data(), (nOldLen+1)*sizeof(wxChar)); GetStringData()->nDataLength = nOldLen; } else { nLen += EXTRA_ALLOC; pData = (wxStringData *) realloc(pData, sizeof(wxStringData) + (nLen + 1)*sizeof(wxChar)); if ( pData == NULL ) { // allocation failure handled by caller // keep previous data since reallocation failed return false; } // it's not important if the pointer changed or not (the check for this // is not faster than assigning to m_pchData in all cases) pData->nAllocLength = nLen; m_pchData = pData->data(); } } //else: we've already got enough return true; } wxStringImpl::iterator wxStringImpl::begin() { if (length() > 0) CopyBeforeWrite(); return m_pchData; } wxStringImpl::iterator wxStringImpl::end() { if (length() > 0) CopyBeforeWrite(); return m_pchData + length(); } wxStringImpl::iterator wxStringImpl::erase(iterator it) { size_type idx = it - begin(); erase(idx, 1); return begin() + idx; } wxStringImpl& wxStringImpl::erase(size_t nStart, size_t nLen) { wxASSERT(nStart <= length()); size_t strLen = length() - nStart; // delete nLen or up to the end of the string characters nLen = strLen < nLen ? strLen : nLen; wxStringImpl strTmp(c_str(), nStart); strTmp.append(c_str() + nStart + nLen, length() - nStart - nLen); swap(strTmp); return *this; } wxStringImpl& wxStringImpl::insert(size_t nPos, const wxChar *sz, size_t n) { wxASSERT( nPos <= length() ); if ( n == npos ) n = wxStrlen(sz); if ( n == 0 ) return *this; if ( !Alloc(length() + n) || !CopyBeforeWrite() ) { wxFAIL_MSG( _T("out of memory in wxStringImpl::insert") ); } memmove(m_pchData + nPos + n, m_pchData + nPos, (length() - nPos) * sizeof(wxChar)); memcpy(m_pchData + nPos, sz, n * sizeof(wxChar)); GetStringData()->nDataLength = length() + n; m_pchData[length()] = '\0'; return *this; } void wxStringImpl::swap(wxStringImpl& str) { wxStringCharType* tmp = str.m_pchData; str.m_pchData = m_pchData; m_pchData = tmp; } size_t wxStringImpl::find(const wxStringImpl& str, size_t nStart) const { // deal with the special case of empty string first const size_t nLen = length(); const size_t nLenOther = str.length(); if ( !nLenOther ) { // empty string is a substring of anything return 0; } if ( !nLen ) { // the other string is non empty so can't be our substring return npos; } wxASSERT( str.GetStringData()->IsValid() ); wxASSERT( nStart <= nLen ); const wxStringCharType * const other = str.c_str(); // anchor const wxStringCharType* p = (const wxStringCharType*)wxStringMemchr(c_str() + nStart, *other, nLen - nStart); if ( !p ) return npos; while ( p - c_str() + nLenOther <= nLen && wxStringMemcmp(p, other, nLenOther) ) { p++; // anchor again p = (const wxStringCharType*) wxStringMemchr(p, *other, nLen - (p - c_str())); if ( !p ) return npos; } return p - c_str() + nLenOther <= nLen ? p - c_str() : npos; } size_t wxStringImpl::find(const wxChar* sz, size_t nStart, size_t n) const { return find(wxStringImpl(sz, n), nStart); } size_t wxStringImpl::find(wxStringCharType ch, size_t nStart) const { wxASSERT( nStart <= length() ); const wxStringCharType *p = (const wxStringCharType*) wxStringMemchr(c_str() + nStart, ch, length() - nStart); return p == NULL ? npos : p - c_str(); } size_t wxStringImpl::rfind(const wxStringImpl& str, size_t nStart) const { wxASSERT( str.GetStringData()->IsValid() ); wxASSERT( nStart == npos || nStart <= length() ); if ( length() >= str.length() ) { // avoids a corner case later if ( length() == 0 && str.length() == 0 ) return 0; // "top" is the point where search starts from size_t top = length() - str.length(); if ( nStart == npos ) nStart = length() - 1; if ( nStart < top ) top = nStart; const wxStringCharType *cursor = c_str() + top; do { if ( wxStringMemcmp(cursor, str.c_str(), str.length()) == 0 ) { return cursor - c_str(); } } while ( cursor-- > c_str() ); } return npos; } size_t wxStringImpl::rfind(const wxChar* sz, size_t nStart, size_t n) const { return rfind(wxStringImpl(sz, n), nStart); } size_t wxStringImpl::rfind(wxStringCharType ch, size_t nStart) const { if ( nStart == npos ) { nStart = length(); } else { wxASSERT( nStart <= length() ); } const wxStringCharType *actual; for ( actual = c_str() + ( nStart == npos ? length() : nStart + 1 ); actual > c_str(); --actual ) { if ( *(actual - 1) == ch ) return (actual - 1) - c_str(); } return npos; } wxStringImpl& wxStringImpl::replace(size_t nStart, size_t nLen, const wxChar *sz) { wxASSERT_MSG( nStart <= length(), _T("index out of bounds in wxStringImpl::replace") ); size_t strLen = length() - nStart; nLen = strLen < nLen ? strLen : nLen; wxStringImpl strTmp; strTmp.reserve(length()); // micro optimisation to avoid multiple mem allocs //This is kind of inefficient, but its pretty good considering... //we don't want to use character access operators here because on STL //it will freeze the reference count of strTmp, which means a deep copy //at the end when swap is called // //Also, we can't use append with the full character pointer and must //do it manually because this string can contain null characters for(size_t i1 = 0; i1 < nStart; ++i1) strTmp.append(1, this->c_str()[i1]); //its safe to do the full version here because //sz must be a normal c string strTmp.append(sz); for(size_t i2 = nStart + nLen; i2 < length(); ++i2) strTmp.append(1, this->c_str()[i2]); swap(strTmp); return *this; } wxStringImpl& wxStringImpl::replace(size_t nStart, size_t nLen, size_t nCount, wxStringCharType ch) { return replace(nStart, nLen, wxStringImpl(nCount, ch).c_str()); } wxStringImpl& wxStringImpl::replace(size_t nStart, size_t nLen, const wxStringImpl& str, size_t nStart2, size_t nLen2) { return replace(nStart, nLen, str.substr(nStart2, nLen2)); } wxStringImpl& wxStringImpl::replace(size_t nStart, size_t nLen, const wxChar* sz, size_t nCount) { return replace(nStart, nLen, wxStringImpl(sz, nCount).c_str()); } wxStringImpl wxStringImpl::substr(size_t nStart, size_t nLen) const { if ( nLen == npos ) nLen = length() - nStart; return wxStringImpl(*this, nStart, nLen); } // assigns one string to another wxStringImpl& wxStringImpl::operator=(const wxStringImpl& stringSrc) { wxASSERT( stringSrc.GetStringData()->IsValid() ); // don't copy string over itself if ( m_pchData != stringSrc.m_pchData ) { if ( stringSrc.GetStringData()->IsEmpty() ) { Reinit(); } else { // adjust references GetStringData()->Unlock(); m_pchData = stringSrc.m_pchData; GetStringData()->Lock(); } } return *this; } // assigns a single character wxStringImpl& wxStringImpl::operator=(wxStringCharType ch) { wxChar c(ch); if ( !AssignCopy(1, &c) ) { wxFAIL_MSG( _T("out of memory in wxStringImpl::operator=(wxChar)") ); } return *this; } // assigns C string wxStringImpl& wxStringImpl::operator=(const wxChar *psz) { if ( !AssignCopy(wxStrlen(psz), psz) ) { wxFAIL_MSG( _T("out of memory in wxStringImpl::operator=(const wxChar *)") ); } return *this; } // helper function: does real copy bool wxStringImpl::AssignCopy(size_t nSrcLen, const wxChar *pszSrcData) { if ( nSrcLen == 0 ) { Reinit(); } else { if ( !AllocBeforeWrite(nSrcLen) ) { // allocation failure handled by caller return false; } memcpy(m_pchData, pszSrcData, nSrcLen*sizeof(wxChar)); GetStringData()->nDataLength = nSrcLen; m_pchData[nSrcLen] = wxT('\0'); } return true; } // --------------------------------------------------------------------------- // string concatenation // --------------------------------------------------------------------------- // add something to this string bool wxStringImpl::ConcatSelf(size_t nSrcLen, const wxChar *pszSrcData, size_t nMaxLen) { STATISTICS_ADD(SummandLength, nSrcLen); nSrcLen = nSrcLen < nMaxLen ? nSrcLen : nMaxLen; // concatenating an empty string is a NOP if ( nSrcLen > 0 ) { wxStringData *pData = GetStringData(); size_t nLen = pData->nDataLength; size_t nNewLen = nLen + nSrcLen; // alloc new buffer if current is too small if ( pData->IsShared() ) { STATISTICS_ADD(ConcatHit, 0); // we have to allocate another buffer wxStringData* pOldData = GetStringData(); if ( !AllocBuffer(nNewLen) ) { // allocation failure handled by caller return false; } memcpy(m_pchData, pOldData->data(), nLen*sizeof(wxChar)); pOldData->Unlock(); } else if ( nNewLen > pData->nAllocLength ) { STATISTICS_ADD(ConcatHit, 0); reserve(nNewLen); // we have to grow the buffer if ( capacity() < nNewLen ) { // allocation failure handled by caller return false; } } else { STATISTICS_ADD(ConcatHit, 1); // the buffer is already big enough } // should be enough space wxASSERT( nNewLen <= GetStringData()->nAllocLength ); // fast concatenation - all is done in our buffer memcpy(m_pchData + nLen, pszSrcData, nSrcLen*sizeof(wxChar)); m_pchData[nNewLen] = wxT('\0'); // put terminating '\0' GetStringData()->nDataLength = nNewLen; // and fix the length } //else: the string to append was empty return true; } #endif // !wxUSE_STL_BASED_WXSTRING #ifdef HAVE_STD_STRING_COMPARE // NB: Comparison code (both if HAVE_STD_STRING_COMPARE and if not) works with // UTF-8 encoded strings too, thanks to UTF-8's design which allows us to // sort strings in characters code point order by sorting the byte sequence // in byte values order (i.e. what strcmp() and memcmp() do). int wxString::compare(const wxString& str) const { return m_impl.compare(str.m_impl); } int wxString::compare(size_t nStart, size_t nLen, const wxString& str) const { size_t pos, len; PosLenToImpl(nStart, nLen, &pos, &len); return m_impl.compare(pos, len, str.m_impl); } int wxString::compare(size_t nStart, size_t nLen, const wxString& str, size_t nStart2, size_t nLen2) const { size_t pos, len; PosLenToImpl(nStart, nLen, &pos, &len); size_t pos2, len2; str.PosLenToImpl(nStart2, nLen2, &pos2, &len2); return m_impl.compare(pos, len, str.m_impl, pos2, len2); } int wxString::compare(const char* sz) const { return m_impl.compare(ImplStr(sz)); } int wxString::compare(const wchar_t* sz) const { return m_impl.compare(ImplStr(sz)); } int wxString::compare(size_t nStart, size_t nLen, const char* sz, size_t nCount) const { size_t pos, len; PosLenToImpl(nStart, nLen, &pos, &len); SubstrBufFromMB str(ImplStr(sz, nCount)); return m_impl.compare(pos, len, str.data, str.len); } int wxString::compare(size_t nStart, size_t nLen, const wchar_t* sz, size_t nCount) const { size_t pos, len; PosLenToImpl(nStart, nLen, &pos, &len); SubstrBufFromWC str(ImplStr(sz, nCount)); return m_impl.compare(pos, len, str.data, str.len); } #else // !HAVE_STD_STRING_COMPARE static inline int wxDoCmp(const wxStringCharType* s1, size_t l1, const wxStringCharType* s2, size_t l2) { if( l1 == l2 ) return wxStringMemcmp(s1, s2, l1); else if( l1 < l2 ) { int ret = wxStringMemcmp(s1, s2, l1); return ret == 0 ? -1 : ret; } else { int ret = wxStringMemcmp(s1, s2, l2); return ret == 0 ? +1 : ret; } } int wxString::compare(const wxString& str) const { return ::wxDoCmp(m_impl.data(), m_impl.length(), str.m_impl.data(), str.m_impl.length()); } int wxString::compare(size_t nStart, size_t nLen, const wxString& str) const { wxASSERT(nStart <= length()); size_type strLen = length() - nStart; nLen = strLen < nLen ? strLen : nLen; size_t pos, len; PosLenToImpl(nStart, nLen, &pos, &len); return ::wxDoCmp(m_impl.data() + pos, len, str.m_impl.data(), str.m_impl.length()); } int wxString::compare(size_t nStart, size_t nLen, const wxString& str, size_t nStart2, size_t nLen2) const { wxASSERT(nStart <= length()); wxASSERT(nStart2 <= str.length()); size_type strLen = length() - nStart, strLen2 = str.length() - nStart2; nLen = strLen < nLen ? strLen : nLen; nLen2 = strLen2 < nLen2 ? strLen2 : nLen2; size_t pos, len; PosLenToImpl(nStart, nLen, &pos, &len); size_t pos2, len2; str.PosLenToImpl(nStart2, nLen2, &pos2, &len2); return ::wxDoCmp(m_impl.data() + pos, len, str.m_impl.data() + pos2, len2); } int wxString::compare(const char* sz) const { SubstrBufFromMB str(ImplStr(sz, npos)); if ( str.len == npos ) str.len = wxStringStrlen(str.data); return ::wxDoCmp(m_impl.data(), m_impl.length(), str.data, str.len); } int wxString::compare(const wchar_t* sz) const { SubstrBufFromWC str(ImplStr(sz, npos)); if ( str.len == npos ) str.len = wxStringStrlen(str.data); return ::wxDoCmp(m_impl.data(), m_impl.length(), str.data, str.len); } int wxString::compare(size_t nStart, size_t nLen, const char* sz, size_t nCount) const { wxASSERT(nStart <= length()); size_type strLen = length() - nStart; nLen = strLen < nLen ? strLen : nLen; size_t pos, len; PosLenToImpl(nStart, nLen, &pos, &len); SubstrBufFromMB str(ImplStr(sz, nCount)); if ( str.len == npos ) str.len = wxStringStrlen(str.data); return ::wxDoCmp(m_impl.data() + pos, len, str.data, str.len); } int wxString::compare(size_t nStart, size_t nLen, const wchar_t* sz, size_t nCount) const { wxASSERT(nStart <= length()); size_type strLen = length() - nStart; nLen = strLen < nLen ? strLen : nLen; size_t pos, len; PosLenToImpl(nStart, nLen, &pos, &len); SubstrBufFromWC str(ImplStr(sz, nCount)); if ( str.len == npos ) str.len = wxStringStrlen(str.data); return ::wxDoCmp(m_impl.data() + pos, len, str.data, str.len); } #endif // HAVE_STD_STRING_COMPARE/!HAVE_STD_STRING_COMPARE // =========================================================================== // wxString class core // =========================================================================== // --------------------------------------------------------------------------- // construction and conversion // --------------------------------------------------------------------------- #if wxUSE_UNICODE /* static */ wxString::SubstrBufFromMB wxString::ConvertStr(const char *psz, size_t nLength, const wxMBConv& conv) { // anything to do? if ( !psz || nLength == 0 ) return SubstrBufFromMB(); if ( nLength == npos ) nLength = wxNO_LEN; size_t wcLen; wxWCharBuffer wcBuf(conv.cMB2WC(psz, nLength, &wcLen)); if ( !wcLen ) return SubstrBufFromMB(); else return SubstrBufFromMB(wcBuf, wcLen); } #else /* static */ wxString::SubstrBufFromWC wxString::ConvertStr(const wchar_t *pwz, size_t nLength, const wxMBConv& conv) { // anything to do? if ( !pwz || nLength == 0 ) return SubstrBufFromWC(); if ( nLength == npos ) nLength = wxNO_LEN; size_t mbLen; wxCharBuffer mbBuf(conv.cWC2MB(pwz, nLength, &mbLen)); if ( !mbLen ) return SubstrBufFromWC(); else return SubstrBufFromWC(mbBuf, mbLen); } #endif #if wxUSE_UNICODE // from multibyte string wxString::wxString(const char *psz, const wxMBConv& conv, size_t nLength) { // FIXME-UTF8: this will need changes // anything to do? if ( psz && nLength != 0 ) { if ( nLength == npos ) { nLength = wxNO_LEN; } size_t nLenWide; wxWCharBuffer wbuf = conv.cMB2WC(psz, nLength, &nLenWide); if ( nLenWide ) assign(wbuf, nLenWide); } } wxString::wxString(const char *psz, size_t nLength) { assign(psz, nLength); } //Convert wxString in Unicode mode to a multi-byte string const wxCharBuffer wxString::mb_str(const wxMBConv& conv) const { return conv.cWC2MB(c_str(), length() + 1 /* size, not length */, NULL); } #else // ANSI #if wxUSE_WCHAR_T // from wide string wxString::wxString(const wchar_t *pwz, const wxMBConv& conv, size_t nLength) { // FIXME-UTF8: this will need changes // anything to do? if ( pwz && nLength != 0 ) { if ( nLength == npos ) { nLength = wxNO_LEN; } size_t nLenMB; wxCharBuffer buf = conv.cWC2MB(pwz, nLength, &nLenMB); if ( nLenMB ) assign(buf, nLenMB); } } wxString::wxString(const wchar_t *pwz, size_t nLength) { assign(pwz, nLength); } //Converts this string to a wide character string if unicode //mode is not enabled and wxUSE_WCHAR_T is enabled const wxWCharBuffer wxString::wc_str(const wxMBConv& conv) const { return conv.cMB2WC(c_str(), length() + 1 /* size, not length */, NULL); } #endif // wxUSE_WCHAR_T #endif // Unicode/ANSI // shrink to minimal size (releasing extra memory) bool wxString::Shrink() { wxString tmp(begin(), end()); swap(tmp); return tmp.length() == length(); } #if !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8 // get the pointer to writable buffer of (at least) nLen bytes wxChar *wxStringImpl::DoGetWriteBuf(size_t nLen) { if ( !AllocBeforeWrite(nLen) ) { // allocation failure handled by caller return NULL; } wxASSERT( GetStringData()->nRefs == 1 ); GetStringData()->Validate(false); return m_pchData; } // put string back in a reasonable state after GetWriteBuf void wxStringImpl::DoUngetWriteBuf() { DoUngetWriteBuf(wxStrlen(m_pchData)); } void wxStringImpl::DoUngetWriteBuf(size_t nLen) { wxStringData * const pData = GetStringData(); wxASSERT_MSG( nLen < pData->nAllocLength, _T("buffer overrun") ); // the strings we store are always NUL-terminated pData->data()[nLen] = _T('\0'); pData->nDataLength = nLen; pData->Validate(true); } // deprecated compatibility code: #if WXWIN_COMPATIBILITY_2_8 wxChar *wxString::GetWriteBuf(size_t nLen) { return DoGetWriteBuf(nLen); } void wxString::UngetWriteBuf() { DoUngetWriteBuf(); } void wxString::UngetWriteBuf(size_t nLen) { DoUngetWriteBuf(nLen); } #endif // WXWIN_COMPATIBILITY_2_8 #endif // !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8 // --------------------------------------------------------------------------- // data access // --------------------------------------------------------------------------- // all functions are inline in string.h // --------------------------------------------------------------------------- // assignment operators // --------------------------------------------------------------------------- #if !wxUSE_UNICODE // same as 'signed char' variant wxString& wxString::operator=(const unsigned char* psz) { *this = (const char *)psz; return *this; } #if wxUSE_WCHAR_T wxString& wxString::operator=(const wchar_t *pwz) { wxString str(pwz); swap(str); return *this; } #endif #endif /* * concatenation functions come in 5 flavours: * string + string * char + string and string + char * C str + string and string + C str */ wxString operator+(const wxString& str1, const wxString& str2) { #if !wxUSE_STL_BASED_WXSTRING wxASSERT( str1.IsValid() ); wxASSERT( str2.IsValid() ); #endif wxString s = str1; s += str2; return s; } wxString operator+(const wxString& str, wxUniChar ch) { #if !wxUSE_STL_BASED_WXSTRING wxASSERT( str.IsValid() ); #endif wxString s = str; s += ch; return s; } wxString operator+(wxUniChar ch, const wxString& str) { #if !wxUSE_STL_BASED_WXSTRING wxASSERT( str.IsValid() ); #endif wxString s = ch; s += str; return s; } wxString operator+(const wxString& str, const char *psz) { #if !wxUSE_STL_BASED_WXSTRING wxASSERT( str.IsValid() ); #endif wxString s; if ( !s.Alloc(strlen(psz) + str.length()) ) { wxFAIL_MSG( _T("out of memory in wxString::operator+") ); } s += str; s += psz; return s; } wxString operator+(const wxString& str, const wchar_t *pwz) { #if !wxUSE_STL_BASED_WXSTRING wxASSERT( str.IsValid() ); #endif wxString s; if ( !s.Alloc(wxWcslen(pwz) + str.length()) ) { wxFAIL_MSG( _T("out of memory in wxString::operator+") ); } s += str; s += pwz; return s; } wxString operator+(const char *psz, const wxString& str) { #if !wxUSE_STL_BASED_WXSTRING wxASSERT( str.IsValid() ); #endif wxString s; if ( !s.Alloc(strlen(psz) + str.length()) ) { wxFAIL_MSG( _T("out of memory in wxString::operator+") ); } s = psz; s += str; return s; } wxString operator+(const wchar_t *pwz, const wxString& str) { #if !wxUSE_STL_BASED_WXSTRING wxASSERT( str.IsValid() ); #endif wxString s; if ( !s.Alloc(wxWcslen(pwz) + str.length()) ) { wxFAIL_MSG( _T("out of memory in wxString::operator+") ); } s = pwz; s += str; return s; } // --------------------------------------------------------------------------- // find_{first,last}_[not]_of functions // --------------------------------------------------------------------------- #if !wxUSE_STL_BASED_WXSTRING || wxUSE_UNICODE_UTF8 // NB: All these functions are implemented with the argument being wxChar*, // i.e. widechar string in any Unicode build, even though native string // representation is char* in the UTF-8 build. This is because we couldn't // use memchr() to determine if a character is in a set encoded as UTF-8. size_t wxString::find_first_of(const wxChar* sz, size_t nStart) const { return find_first_of(sz, nStart, wxStrlen(sz)); } size_t wxString::find_first_not_of(const wxChar* sz, size_t nStart) const { return find_first_not_of(sz, nStart, wxStrlen(sz)); } size_t wxString::find_first_of(const wxChar* sz, size_t nStart, size_t n) const { wxASSERT_MSG( nStart <= length(), _T("invalid index") ); size_t idx = nStart; for ( const_iterator i = begin() + nStart; i != end(); ++idx, ++i ) { if ( wxTmemchr(sz, *i, n) ) return idx; } return npos; } size_t wxString::find_first_not_of(const wxChar* sz, size_t nStart, size_t n) const { wxASSERT_MSG( nStart <= length(), _T("invalid index") ); size_t idx = nStart; for ( const_iterator i = begin() + nStart; i != end(); ++idx, ++i ) { if ( !wxTmemchr(sz, *i, n) ) return idx; } return npos; } size_t wxString::find_last_of(const wxChar* sz, size_t nStart) const { return find_last_of(sz, nStart, wxStrlen(sz)); } size_t wxString::find_last_not_of(const wxChar* sz, size_t nStart) const { return find_last_not_of(sz, nStart, wxStrlen(sz)); } size_t wxString::find_last_of(const wxChar* sz, size_t nStart, size_t n) const { size_t len = length(); if ( nStart == npos ) { nStart = len - 1; } else { wxASSERT_MSG( nStart <= len, _T("invalid index") ); } size_t idx = nStart; for ( const_reverse_iterator i = rbegin() + (len - nStart - 1); i != rend(); --idx, ++i ) { if ( wxTmemchr(sz, *i, n) ) return idx; } return npos; } size_t wxString::find_last_not_of(const wxChar* sz, size_t nStart, size_t n) const { size_t len = length(); if ( nStart == npos ) { nStart = len - 1; } else { wxASSERT_MSG( nStart <= len, _T("invalid index") ); } size_t idx = nStart; for ( const_reverse_iterator i = rbegin() + (len - nStart - 1); i != rend(); --idx, ++i ) { if ( !wxTmemchr(sz, *i, n) ) return idx; } return npos; } size_t wxString::find_first_not_of(wxUniChar ch, size_t nStart) const { wxASSERT_MSG( nStart <= length(), _T("invalid index") ); size_t idx = nStart; for ( const_iterator i = begin() + nStart; i != end(); ++idx, ++i ) { if ( *i != ch ) return idx; } return npos; } size_t wxString::find_last_not_of(wxUniChar ch, size_t nStart) const { size_t len = length(); if ( nStart == npos ) { nStart = len - 1; } else { wxASSERT_MSG( nStart <= len, _T("invalid index") ); } size_t idx = nStart; for ( const_reverse_iterator i = rbegin() + (len - nStart - 1); i != rend(); --idx, ++i ) { if ( *i != ch ) return idx; } return npos; } // the functions above were implemented for wchar_t* arguments in Unicode // build and char* in ANSI build; below are implementations for the other // version: #if wxUSE_UNICODE #define wxOtherCharType char #define STRCONV (const wxChar*)wxConvLibc.cMB2WC #else #define wxOtherCharType wchar_t #define STRCONV (const wxChar*)wxConvLibc.cWC2MB #endif size_t wxString::find_first_of(const wxOtherCharType* sz, size_t nStart) const { return find_first_of(STRCONV(sz), nStart); } size_t wxString::find_first_of(const wxOtherCharType* sz, size_t nStart, size_t n) const { return find_first_of(STRCONV(sz, n, NULL), nStart, n); } size_t wxString::find_last_of(const wxOtherCharType* sz, size_t nStart) const { return find_last_of(STRCONV(sz), nStart); } size_t wxString::find_last_of(const wxOtherCharType* sz, size_t nStart, size_t n) const { return find_last_of(STRCONV(sz, n, NULL), nStart, n); } size_t wxString::find_first_not_of(const wxOtherCharType* sz, size_t nStart) const { return find_first_not_of(STRCONV(sz), nStart); } size_t wxString::find_first_not_of(const wxOtherCharType* sz, size_t nStart, size_t n) const { return find_first_not_of(STRCONV(sz, n, NULL), nStart, n); } size_t wxString::find_last_not_of(const wxOtherCharType* sz, size_t nStart) const { return find_last_not_of(STRCONV(sz), nStart); } size_t wxString::find_last_not_of(const wxOtherCharType* sz, size_t nStart, size_t n) const { return find_last_not_of(STRCONV(sz, n, NULL), nStart, n); } #undef wxOtherCharType #undef STRCONV #endif // !wxUSE_STL_BASED_WXSTRING || wxUSE_UNICODE_UTF8 // =========================================================================== // other common string functions // =========================================================================== int wxString::CmpNoCase(const wxString& s) const { // FIXME-UTF8: use wxUniChar::ToLower/ToUpper once added size_t idx = 0; const_iterator i1 = begin(); const_iterator end1 = end(); const_iterator i2 = s.begin(); const_iterator end2 = s.end(); for ( ; i1 != end1 && i2 != end2; ++idx, ++i1, ++i2 ) { wxUniChar lower1 = (wxChar)wxTolower(*i1); wxUniChar lower2 = (wxChar)wxTolower(*i2); if ( lower1 != lower2 ) return lower1 < lower2 ? -1 : 1; } size_t len1 = length(); size_t len2 = s.length(); if ( len1 < len2 ) return -1; else if ( len1 > len2 ) return 1; return 0; } #if wxUSE_UNICODE #ifdef __MWERKS__ #ifndef __SCHAR_MAX__ #define __SCHAR_MAX__ 127 #endif #endif wxString wxString::FromAscii(const char *ascii) { if (!ascii) return wxEmptyString; size_t len = strlen( ascii ); wxString res; if ( len ) { wxStringBuffer buf(res, len); wchar_t *dest = buf; for ( ;; ) { if ( (*dest++ = (wchar_t)(unsigned char)*ascii++) == L'\0' ) break; } } return res; } wxString wxString::FromAscii(const char ascii) { // What do we do with '\0' ? wxString res; res += (wchar_t)(unsigned char) ascii; return res; } const wxCharBuffer wxString::ToAscii() const { // this will allocate enough space for the terminating NUL too wxCharBuffer buffer(length()); char *dest = buffer.data(); const wchar_t *pwc = c_str(); for ( ;; ) { *dest++ = (char)(*pwc > SCHAR_MAX ? wxT('_') : *pwc); // the output string can't have embedded NULs anyhow, so we can safely // stop at first of them even if we do have any if ( !*pwc++ ) break; } return buffer; } #endif // Unicode // extract string of length nCount starting at nFirst wxString wxString::Mid(size_t nFirst, size_t nCount) const { size_t nLen = length(); // default value of nCount is npos and means "till the end" if ( nCount == npos ) { nCount = nLen - nFirst; } // out-of-bounds requests return sensible things if ( nFirst + nCount > nLen ) { nCount = nLen - nFirst; } if ( nFirst > nLen ) { // AllocCopy() will return empty string return wxEmptyString; } wxString dest(*this, nFirst, nCount); if ( dest.length() != nCount ) { wxFAIL_MSG( _T("out of memory in wxString::Mid") ); } return dest; } // check that the string starts with prefix and return the rest of the string // in the provided pointer if it is not NULL, otherwise return false bool wxString::StartsWith(const wxChar *prefix, wxString *rest) const { wxASSERT_MSG( prefix, _T("invalid parameter in wxString::StartsWith") ); // first check if the beginning of the string matches the prefix: note // that we don't have to check that we don't run out of this string as // when we reach the terminating NUL, either prefix string ends too (and // then it's ok) or we break out of the loop because there is no match const wxChar *p = c_str(); while ( *prefix ) { if ( *prefix++ != *p++ ) { // no match return false; } } if ( rest ) { // put the rest of the string into provided pointer *rest = p; } return true; } // check that the string ends with suffix and return the rest of it in the // provided pointer if it is not NULL, otherwise return false bool wxString::EndsWith(const wxChar *suffix, wxString *rest) const { wxASSERT_MSG( suffix, _T("invalid parameter in wxString::EndssWith") ); int start = length() - wxStrlen(suffix); if ( start < 0 || wxStrcmp(wx_str() + start, suffix) != 0 ) return false; if ( rest ) { // put the rest of the string into provided pointer rest->assign(*this, 0, start); } return true; } // extract nCount last (rightmost) characters wxString wxString::Right(size_t nCount) const { if ( nCount > length() ) nCount = length(); wxString dest(*this, length() - nCount, nCount); if ( dest.length() != nCount ) { wxFAIL_MSG( _T("out of memory in wxString::Right") ); } return dest; } // get all characters after the last occurence of ch // (returns the whole string if ch not found) wxString wxString::AfterLast(wxUniChar ch) const { wxString str; int iPos = Find(ch, true); if ( iPos == wxNOT_FOUND ) str = *this; else str = wx_str() + iPos + 1; return str; } // extract nCount first (leftmost) characters wxString wxString::Left(size_t nCount) const { if ( nCount > length() ) nCount = length(); wxString dest(*this, 0, nCount); if ( dest.length() != nCount ) { wxFAIL_MSG( _T("out of memory in wxString::Left") ); } return dest; } // get all characters before the first occurence of ch // (returns the whole string if ch not found) wxString wxString::BeforeFirst(wxUniChar ch) const { int iPos = Find(ch); if ( iPos == wxNOT_FOUND ) iPos = length(); return wxString(*this, 0, iPos); } /// get all characters before the last occurence of ch /// (returns empty string if ch not found) wxString wxString::BeforeLast(wxUniChar ch) const { wxString str; int iPos = Find(ch, true); if ( iPos != wxNOT_FOUND && iPos != 0 ) str = wxString(c_str(), iPos); return str; } /// get all characters after the first occurence of ch /// (returns empty string if ch not found) wxString wxString::AfterFirst(wxUniChar ch) const { wxString str; int iPos = Find(ch); if ( iPos != wxNOT_FOUND ) str = wx_str() + iPos + 1; return str; } // replace first (or all) occurences of some substring with another one size_t wxString::Replace(const wxChar *szOld, const wxChar *szNew, bool bReplaceAll) { // if we tried to replace an empty string we'd enter an infinite loop below wxCHECK_MSG( szOld && *szOld && szNew, 0, _T("wxString::Replace(): invalid parameter") ); size_t uiCount = 0; // count of replacements made size_t uiOldLen = wxStrlen(szOld); size_t uiNewLen = wxStrlen(szNew); size_t dwPos = 0; while ( this->c_str()[dwPos] != wxT('\0') ) { //DO NOT USE STRSTR HERE //this string can contain embedded null characters, //so strstr will function incorrectly dwPos = find(szOld, dwPos); if ( dwPos == npos ) break; // exit the loop else { //replace this occurance of the old string with the new one replace(dwPos, uiOldLen, szNew, uiNewLen); //move up pos past the string that was replaced dwPos += uiNewLen; //increase replace count ++uiCount; // stop now? if ( !bReplaceAll ) break; // exit the loop } } return uiCount; } bool wxString::IsAscii() const { const wxChar *s = (const wxChar*) *this; while(*s){ if(!isascii(*s)) return(false); s++; } return(true); } bool wxString::IsWord() const { const wxChar *s = (const wxChar*) *this; while(*s){ if(!wxIsalpha(*s)) return(false); s++; } return(true); } bool wxString::IsNumber() const { const wxChar *s = (const wxChar*) *this; if (wxStrlen(s)) if ((s[0] == wxT('-')) || (s[0] == wxT('+'))) s++; while(*s){ if(!wxIsdigit(*s)) return(false); s++; } return(true); } wxString wxString::Strip(stripType w) const { wxString s = *this; if ( w & leading ) s.Trim(false); if ( w & trailing ) s.Trim(true); return s; } // --------------------------------------------------------------------------- // case conversion // --------------------------------------------------------------------------- wxString& wxString::MakeUpper() { for ( iterator it = begin(), en = end(); it != en; ++it ) *it = (wxChar)wxToupper(*it); return *this; } wxString& wxString::MakeLower() { for ( iterator it = begin(), en = end(); it != en; ++it ) *it = (wxChar)wxTolower(*it); return *this; } // --------------------------------------------------------------------------- // trimming and padding // --------------------------------------------------------------------------- // some compilers (VC++ 6.0 not to name them) return true for a call to // isspace('ê') in the C locale which seems to be broken to me, but we have to // live with this by checking that the character is a 7 bit one - even if this // may fail to detect some spaces (I don't know if Unicode doesn't have // space-like symbols somewhere except in the first 128 chars), it is arguably // still better than trimming away accented letters inline int wxSafeIsspace(wxChar ch) { return (ch < 127) && wxIsspace(ch); } // trims spaces (in the sense of isspace) from left or right side wxString& wxString::Trim(bool bFromRight) { // first check if we're going to modify the string at all if ( !empty() && ( (bFromRight && wxSafeIsspace(GetChar(length() - 1))) || (!bFromRight && wxSafeIsspace(GetChar(0u))) ) ) { if ( bFromRight ) { // find last non-space character reverse_iterator psz = rbegin(); while ( (psz != rend()) && wxSafeIsspace(*psz) ) psz++; // truncate at trailing space start erase(psz.base(), end()); } else { // find first non-space character iterator psz = begin(); while ( (psz != end()) && wxSafeIsspace(*psz) ) psz++; // fix up data and length erase(begin(), psz); } } return *this; } // adds nCount characters chPad to the string from either side wxString& wxString::Pad(size_t nCount, wxUniChar chPad, bool bFromRight) { wxString s(chPad, nCount); if ( bFromRight ) *this += s; else { s += *this; swap(s); } return *this; } // truncate the string wxString& wxString::Truncate(size_t uiLen) { if ( uiLen < length() ) { erase(begin() + uiLen, end()); } //else: nothing to do, string is already short enough return *this; } // --------------------------------------------------------------------------- // finding (return wxNOT_FOUND if not found and index otherwise) // --------------------------------------------------------------------------- // find a character int wxString::Find(wxUniChar ch, bool bFromEnd) const { size_type idx = bFromEnd ? find_last_of(ch) : find_first_of(ch); return (idx == npos) ? wxNOT_FOUND : (int)idx; } // find a sub-string (like strstr) int wxString::Find(const wxChar *pszSub) const { size_type idx = find(pszSub); return (idx == npos) ? wxNOT_FOUND : (int)idx; } // ---------------------------------------------------------------------------- // conversion to numbers // ---------------------------------------------------------------------------- // the implementation of all the functions below is exactly the same so factor // it out template bool wxStringToIntType(const wxChar *start, T *val, int base, F func) { wxCHECK_MSG( val, false, _T("NULL output pointer") ); wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") ); #ifndef __WXWINCE__ errno = 0; #endif wxChar *end; *val = (*func)(start, &end, base); // return true only if scan was stopped by the terminating NUL and if the // string was not empty to start with and no under/overflow occurred return !*end && (end != start) #ifndef __WXWINCE__ && (errno != ERANGE) #endif ; } bool wxString::ToLong(long *val, int base) const { return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtol); } bool wxString::ToULong(unsigned long *val, int base) const { return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtoul); } bool wxString::ToLongLong(wxLongLong_t *val, int base) const { #ifdef wxHAS_STRTOLL return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtoll); #else // TODO: implement this ourselves wxUnusedVar(val); wxUnusedVar(base); return false; #endif // wxHAS_STRTOLL } bool wxString::ToULongLong(wxULongLong_t *val, int base) const { #ifdef wxHAS_STRTOLL return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtoull); #else // TODO: implement this ourselves wxUnusedVar(val); wxUnusedVar(base); return false; #endif } bool wxString::ToDouble(double *val) const { wxCHECK_MSG( val, false, _T("NULL pointer in wxString::ToDouble") ); #ifndef __WXWINCE__ errno = 0; #endif const wxChar *start = c_str(); wxChar *end; *val = wxStrtod(start, &end); // return true only if scan was stopped by the terminating NUL and if the // string was not empty to start with and no under/overflow occurred return !*end && (end != start) #ifndef __WXWINCE__ && (errno != ERANGE) #endif ; } // --------------------------------------------------------------------------- // formatted output // --------------------------------------------------------------------------- /* static */ #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN wxString wxStringPrintfMixinBase::DoFormat(const wxChar *format, ...) #else wxString wxString::DoFormat(const wxChar *format, ...) #endif { va_list argptr; va_start(argptr, format); wxString s; s.PrintfV(format, argptr); va_end(argptr); return s; } /* static */ wxString wxString::FormatV(const wxString& format, va_list argptr) { wxString s; s.PrintfV(format, argptr); return s; } #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN int wxStringPrintfMixinBase::DoPrintf(const wxChar *format, ...) #else int wxString::DoPrintf(const wxChar *format, ...) #endif { va_list argptr; va_start(argptr, format); #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN // get a pointer to the wxString instance; we have to use dynamic_cast<> // because it's the only cast that works safely for downcasting when // multiple inheritance is used: wxString *str = static_cast(this); #else wxString *str = this; #endif int iLen = str->PrintfV(format, argptr); va_end(argptr); return iLen; } int wxString::PrintfV(const wxString& format, va_list argptr) { int size = 1024; for ( ;; ) { wxStringBuffer tmp(*this, size + 1); wxChar *buf = tmp; if ( !buf ) { // out of memory return -1; } // wxVsnprintf() may modify the original arg pointer, so pass it // only a copy va_list argptrcopy; wxVaCopy(argptrcopy, argptr); int len = wxVsnprintf(buf, size, format, argptrcopy); va_end(argptrcopy); // some implementations of vsnprintf() don't NUL terminate // the string if there is not enough space for it so // always do it manually buf[size] = _T('\0'); // vsnprintf() may return either -1 (traditional Unix behaviour) or the // total number of characters which would have been written if the // buffer were large enough (newer standards such as Unix98) if ( len < 0 ) { #if wxUSE_WXVSNPRINTF // we know that our own implementation of wxVsnprintf() returns -1 // only for a format error - thus there's something wrong with // the user's format string return -1; #else // assume that system version only returns error if not enough space // still not enough, as we don't know how much we need, double the // current size of the buffer size *= 2; #endif // wxUSE_WXVSNPRINTF/!wxUSE_WXVSNPRINTF } else if ( len >= size ) { #if wxUSE_WXVSNPRINTF // we know that our own implementation of wxVsnprintf() returns // size+1 when there's not enough space but that's not the size // of the required buffer! size *= 2; // so we just double the current size of the buffer #else // some vsnprintf() implementations NUL-terminate the buffer and // some don't in len == size case, to be safe always add 1 size = len + 1; #endif } else // ok, there was enough space { break; } } // we could have overshot Shrink(); return length(); } // ---------------------------------------------------------------------------- // misc other operations // ---------------------------------------------------------------------------- // returns true if the string matches the pattern which may contain '*' and // '?' metacharacters (as usual, '?' matches any character and '*' any number // of them) bool wxString::Matches(const wxChar *pszMask) const { // I disable this code as it doesn't seem to be faster (in fact, it seems // to be much slower) than the old, hand-written code below and using it // here requires always linking with libregex even if the user code doesn't // use it #if 0 // wxUSE_REGEX // first translate the shell-like mask into a regex wxString pattern; pattern.reserve(wxStrlen(pszMask)); pattern += _T('^'); while ( *pszMask ) { switch ( *pszMask ) { case _T('?'): pattern += _T('.'); break; case _T('*'): pattern += _T(".*"); break; case _T('^'): case _T('.'): case _T('$'): case _T('('): case _T(')'): case _T('|'): case _T('+'): case _T('\\'): // these characters are special in a RE, quote them // (however note that we don't quote '[' and ']' to allow // using them for Unix shell like matching) pattern += _T('\\'); // fall through default: pattern += *pszMask; } pszMask++; } pattern += _T('$'); // and now use it return wxRegEx(pattern, wxRE_NOSUB | wxRE_EXTENDED).Matches(c_str()); #else // !wxUSE_REGEX // TODO: this is, of course, awfully inefficient... // the char currently being checked const wxChar *pszTxt = c_str(); // the last location where '*' matched const wxChar *pszLastStarInText = NULL; const wxChar *pszLastStarInMask = NULL; match: for ( ; *pszMask != wxT('\0'); pszMask++, pszTxt++ ) { switch ( *pszMask ) { case wxT('?'): if ( *pszTxt == wxT('\0') ) return false; // pszTxt and pszMask will be incremented in the loop statement break; case wxT('*'): { // remember where we started to be able to backtrack later pszLastStarInText = pszTxt; pszLastStarInMask = pszMask; // ignore special chars immediately following this one // (should this be an error?) while ( *pszMask == wxT('*') || *pszMask == wxT('?') ) pszMask++; // if there is nothing more, match if ( *pszMask == wxT('\0') ) return true; // are there any other metacharacters in the mask? size_t uiLenMask; const wxChar *pEndMask = wxStrpbrk(pszMask, wxT("*?")); if ( pEndMask != NULL ) { // we have to match the string between two metachars uiLenMask = pEndMask - pszMask; } else { // we have to match the remainder of the string uiLenMask = wxStrlen(pszMask); } wxString strToMatch(pszMask, uiLenMask); const wxChar* pMatch = wxStrstr(pszTxt, strToMatch); if ( pMatch == NULL ) return false; // -1 to compensate "++" in the loop pszTxt = pMatch + uiLenMask - 1; pszMask += uiLenMask - 1; } break; default: if ( *pszMask != *pszTxt ) return false; break; } } // match only if nothing left if ( *pszTxt == wxT('\0') ) return true; // if we failed to match, backtrack if we can if ( pszLastStarInText ) { pszTxt = pszLastStarInText + 1; pszMask = pszLastStarInMask; pszLastStarInText = NULL; // don't bother resetting pszLastStarInMask, it's unnecessary goto match; } return false; #endif // wxUSE_REGEX/!wxUSE_REGEX } // Count the number of chars int wxString::Freq(wxUniChar ch) const { int count = 0; for ( const_iterator i = begin(); i != end(); ++i ) { if ( *i == ch ) count ++; } return count; } // convert to upper case, return the copy of the string wxString wxString::Upper() const { wxString s(*this); return s.MakeUpper(); } // convert to lower case, return the copy of the string wxString wxString::Lower() const { wxString s(*this); return s.MakeLower(); } // ============================================================================ // ArrayString // ============================================================================ #include "wx/arrstr.h" wxArrayString::wxArrayString(size_t sz, const wxChar** a) { #if !wxUSE_STL Init(false); #endif for (size_t i=0; i < sz; i++) Add(a[i]); } wxArrayString::wxArrayString(size_t sz, const wxString* a) { #if !wxUSE_STL Init(false); #endif for (size_t i=0; i < sz; i++) Add(a[i]); } #if !wxUSE_STL // size increment = min(50% of current size, ARRAY_MAXSIZE_INCREMENT) #define ARRAY_MAXSIZE_INCREMENT 4096 #ifndef ARRAY_DEFAULT_INITIAL_SIZE // also defined in dynarray.h #define ARRAY_DEFAULT_INITIAL_SIZE (16) #endif // ctor void wxArrayString::Init(bool autoSort) { m_nSize = m_nCount = 0; m_pItems = NULL; m_autoSort = autoSort; } // copy ctor wxArrayString::wxArrayString(const wxArrayString& src) { Init(src.m_autoSort); *this = src; } // assignment operator wxArrayString& wxArrayString::operator=(const wxArrayString& src) { if ( m_nSize > 0 ) Clear(); Copy(src); m_autoSort = src.m_autoSort; return *this; } void wxArrayString::Copy(const wxArrayString& src) { if ( src.m_nCount > ARRAY_DEFAULT_INITIAL_SIZE ) Alloc(src.m_nCount); for ( size_t n = 0; n < src.m_nCount; n++ ) Add(src[n]); } // grow the array void wxArrayString::Grow(size_t nIncrement) { // only do it if no more place if ( (m_nSize - m_nCount) < nIncrement ) { // if ARRAY_DEFAULT_INITIAL_SIZE were set to 0, the initially empty would // be never resized! #if ARRAY_DEFAULT_INITIAL_SIZE == 0 #error "ARRAY_DEFAULT_INITIAL_SIZE must be > 0!" #endif if ( m_nSize == 0 ) { // was empty, alloc some memory m_nSize = ARRAY_DEFAULT_INITIAL_SIZE; if (m_nSize < nIncrement) m_nSize = nIncrement; m_pItems = new wxString[m_nSize]; } else { // otherwise when it's called for the first time, nIncrement would be 0 // and the array would never be expanded // add 50% but not too much size_t ndefIncrement = m_nSize < ARRAY_DEFAULT_INITIAL_SIZE ? ARRAY_DEFAULT_INITIAL_SIZE : m_nSize >> 1; if ( ndefIncrement > ARRAY_MAXSIZE_INCREMENT ) ndefIncrement = ARRAY_MAXSIZE_INCREMENT; if ( nIncrement < ndefIncrement ) nIncrement = ndefIncrement; m_nSize += nIncrement; wxString *pNew = new wxString[m_nSize]; // copy data to new location for ( size_t j = 0; j < m_nCount; j++ ) pNew[j] = m_pItems[j]; // delete old memory (but do not release the strings!) wxDELETEA(m_pItems); m_pItems = pNew; } } } // deletes all the strings from the list void wxArrayString::Empty() { m_nCount = 0; } // as Empty, but also frees memory void wxArrayString::Clear() { m_nSize = m_nCount = 0; wxDELETEA(m_pItems); } // dtor wxArrayString::~wxArrayString() { wxDELETEA(m_pItems); } void wxArrayString::reserve(size_t nSize) { Alloc(nSize); } // pre-allocates memory (frees the previous data!) void wxArrayString::Alloc(size_t nSize) { // only if old buffer was not big enough if ( nSize > m_nSize ) { wxString *pNew = new wxString[nSize]; if ( !pNew ) return; for ( size_t j = 0; j < m_nCount; j++ ) pNew[j] = m_pItems[j]; delete [] m_pItems; m_pItems = pNew; m_nSize = nSize; } } // minimizes the memory usage by freeing unused memory void wxArrayString::Shrink() { // only do it if we have some memory to free if( m_nCount < m_nSize ) { // allocates exactly as much memory as we need wxString *pNew = new wxString[m_nCount]; // copy data to new location for ( size_t j = 0; j < m_nCount; j++ ) pNew[j] = m_pItems[j]; delete [] m_pItems; m_pItems = pNew; } } // searches the array for an item (forward or backwards) int wxArrayString::Index(const wxChar *sz, bool bCase, bool bFromEnd) const { if ( m_autoSort ) { // use binary search in the sorted array wxASSERT_MSG( bCase && !bFromEnd, wxT("search parameters ignored for auto sorted array") ); size_t i, lo = 0, hi = m_nCount; int res; while ( lo < hi ) { i = (lo + hi)/2; res = wxStrcmp(sz, m_pItems[i]); if ( res < 0 ) hi = i; else if ( res > 0 ) lo = i + 1; else return i; } return wxNOT_FOUND; } else { // use linear search in unsorted array if ( bFromEnd ) { if ( m_nCount > 0 ) { size_t ui = m_nCount; do { if ( m_pItems[--ui].IsSameAs(sz, bCase) ) return ui; } while ( ui != 0 ); } } else { for( size_t ui = 0; ui < m_nCount; ui++ ) { if( m_pItems[ui].IsSameAs(sz, bCase) ) return ui; } } } return wxNOT_FOUND; } // add item at the end size_t wxArrayString::Add(const wxString& str, size_t nInsert) { if ( m_autoSort ) { // insert the string at the correct position to keep the array sorted size_t i, lo = 0, hi = m_nCount; int res; while ( lo < hi ) { i = (lo + hi)/2; res = str.Cmp(m_pItems[i]); if ( res < 0 ) hi = i; else if ( res > 0 ) lo = i + 1; else { lo = hi = i; break; } } wxASSERT_MSG( lo == hi, wxT("binary search broken") ); Insert(str, lo, nInsert); return (size_t)lo; } else { Grow(nInsert); for (size_t i = 0; i < nInsert; i++) { // just append m_pItems[m_nCount + i] = str; } size_t ret = m_nCount; m_nCount += nInsert; return ret; } } // add item at the given position void wxArrayString::Insert(const wxString& str, size_t nIndex, size_t nInsert) { wxCHECK_RET( nIndex <= m_nCount, wxT("bad index in wxArrayString::Insert") ); wxCHECK_RET( m_nCount <= m_nCount + nInsert, wxT("array size overflow in wxArrayString::Insert") ); Grow(nInsert); for (int j = m_nCount - nIndex - 1; j >= 0; j--) m_pItems[nIndex + nInsert + j] = m_pItems[nIndex + j]; for (size_t i = 0; i < nInsert; i++) { m_pItems[nIndex + i] = str; } m_nCount += nInsert; } // range insert (STL 23.2.4.3) void wxArrayString::insert(iterator it, const_iterator first, const_iterator last) { const int idx = it - begin(); // grow it once Grow(last - first); // reset "it" since it can change inside Grow() it = begin() + idx; while ( first != last ) { it = insert(it, *first); // insert returns an iterator to the last element inserted but we need // insert the next after this one, that is before the next one ++it; ++first; } } // expand the array void wxArrayString::SetCount(size_t count) { Alloc(count); wxString s; while ( m_nCount < count ) m_pItems[m_nCount++] = s; } // removes item from array (by index) void wxArrayString::RemoveAt(size_t nIndex, size_t nRemove) { wxCHECK_RET( nIndex < m_nCount, wxT("bad index in wxArrayString::Remove") ); wxCHECK_RET( nIndex + nRemove <= m_nCount, wxT("removing too many elements in wxArrayString::Remove") ); for ( size_t j = 0; j < m_nCount - nIndex -nRemove; j++) m_pItems[nIndex + j] = m_pItems[nIndex + nRemove + j]; m_nCount -= nRemove; } // removes item from array (by value) void wxArrayString::Remove(const wxChar *sz) { int iIndex = Index(sz); wxCHECK_RET( iIndex != wxNOT_FOUND, wxT("removing inexistent element in wxArrayString::Remove") ); RemoveAt(iIndex); } void wxArrayString::assign(const_iterator first, const_iterator last) { reserve(last - first); for(; first != last; ++first) push_back(*first); } // ---------------------------------------------------------------------------- // sorting // ---------------------------------------------------------------------------- // we can only sort one array at a time with the quick-sort based // implementation #if wxUSE_THREADS // need a critical section to protect access to gs_compareFunction and // gs_sortAscending variables static wxCriticalSection gs_critsectStringSort; #endif // wxUSE_THREADS // function to use for string comparaison static wxArrayString::CompareFunction gs_compareFunction = NULL; // if we don't use the compare function, this flag tells us if we sort the // array in ascending or descending order static bool gs_sortAscending = true; // function which is called by quick sort extern "C" int wxC_CALLING_CONV // LINKAGEMODE wxStringCompareFunction(const void *first, const void *second) { wxString *strFirst = (wxString *)first; wxString *strSecond = (wxString *)second; if ( gs_compareFunction ) { return gs_compareFunction(*strFirst, *strSecond); } else { // maybe we should use wxStrcoll int result = strFirst->Cmp(*strSecond); return gs_sortAscending ? result : -result; } } // sort array elements using passed comparaison function void wxArrayString::Sort(CompareFunction compareFunction) { wxCRIT_SECT_LOCKER(lockCmpFunc, gs_critsectStringSort); wxASSERT( !gs_compareFunction ); // must have been reset to NULL gs_compareFunction = compareFunction; DoSort(); // reset it to NULL so that Sort(bool) will work the next time gs_compareFunction = NULL; } extern "C" { typedef int (wxC_CALLING_CONV * wxStringCompareFn)(const void *first, const void *second); } void wxArrayString::Sort(CompareFunction2 compareFunction) { qsort(m_pItems, m_nCount, sizeof(wxString), (wxStringCompareFn)compareFunction); } void wxArrayString::Sort(bool reverseOrder) { Sort(reverseOrder ? wxStringSortDescending : wxStringSortAscending); } void wxArrayString::DoSort() { wxCHECK_RET( !m_autoSort, wxT("can't use this method with sorted arrays") ); qsort(m_pItems, m_nCount, sizeof(wxString), wxStringCompareFunction); } bool wxArrayString::operator==(const wxArrayString& a) const { if ( m_nCount != a.m_nCount ) return false; for ( size_t n = 0; n < m_nCount; n++ ) { if ( Item(n) != a[n] ) return false; } return true; } #endif // !wxUSE_STL int wxCMPFUNC_CONV wxStringSortAscending(wxString* s1, wxString* s2) { return s1->Cmp(*s2); } int wxCMPFUNC_CONV wxStringSortDescending(wxString* s1, wxString* s2) { return -s1->Cmp(*s2); } // =========================================================================== // wxJoin and wxSplit // =========================================================================== #include "wx/tokenzr.h" wxString wxJoin(const wxArrayString& arr, const wxChar sep, const wxChar escape) { size_t count = arr.size(); if ( count == 0 ) return wxEmptyString; wxString str; // pre-allocate memory using the estimation of the average length of the // strings in the given array: this is very imprecise, of course, but // better than nothing str.reserve(count*(arr[0].length() + arr[count-1].length()) / 2); if ( escape == wxT('\0') ) { // escaping is disabled: for ( size_t i = 0; i < count; i++ ) { if ( i ) str += sep; str += arr[i]; } } else // use escape character { for ( size_t n = 0; n < count; n++ ) { if ( n ) str += sep; for ( wxString::const_iterator i = arr[n].begin(), end = arr[n].end(); i != end; ++i ) { const wxChar ch = *i; if ( ch == sep ) str += escape; // escape this separator str += ch; } } } str.Shrink(); // release extra memory if we allocated too much return str; } wxArrayString wxSplit(const wxString& str, const wxChar sep, const wxChar escape) { if ( escape == wxT('\0') ) { // simple case: we don't need to honour the escape character return wxStringTokenize(str, sep, wxTOKEN_RET_EMPTY_ALL); } wxArrayString ret; wxString curr; wxChar prev = wxT('\0'); for ( wxString::const_iterator i = str.begin(), end = str.end(); i != end; ++i ) { const wxChar ch = *i; if ( ch == sep ) { if ( prev == escape ) { // remove the escape character and don't consider this // occurrence of 'sep' as a real separator *curr.rbegin() = sep; } else // real separator { ret.push_back(curr); curr.clear(); } } else // normal character { curr += ch; } prev = ch; } // add the last token if ( !curr.empty() || prev == sep ) ret.Add(curr); return ret; }