X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/38444daf663a7e4dfdc4f7ba8387ee5df16789f0..3faedeb065c67cb17ee8e271c9506e0477af5e3d:/src/common/string.cpp diff --git a/src/common/string.cpp b/src/common/string.cpp index 2d371b9fa4..791c5ce049 100644 --- a/src/common/string.cpp +++ b/src/common/string.cpp @@ -6,7 +6,7 @@ // Created: 29/01/98 // RCS-ID: $Id$ // Copyright: (c) 1998 Vadim Zeitlin -// Licence: wxWindows license +// Licence: wxWindows licence ///////////////////////////////////////////////////////////////////////////// #ifdef __GNUG__ @@ -35,9 +35,7 @@ #include "wx/defs.h" #include "wx/string.h" #include "wx/intl.h" -#if wxUSE_THREADS - #include -#endif + #include "wx/thread.h" #endif #include @@ -48,14 +46,6 @@ #include #endif -#if wxUSE_WCSRTOMBS - #include // for wcsrtombs(), see comments where it's used -#endif // GNU - -#ifdef WXSTRING_IS_WXOBJECT - IMPLEMENT_DYNAMIC_CLASS(wxString, wxObject) -#endif //WXSTRING_IS_WXOBJECT - // 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! @@ -65,6 +55,12 @@ // static class variables definition // --------------------------------------------------------------------------- +#if defined(__VISAGECPP__) && __IBMCPP__ >= 400 +// must define this static for VA or else you get multiply defined symbols +// everywhere +const unsigned int wxSTRING_MAXLEN = UINT_MAX - 100; +#endif // Visual Age + #ifdef wxSTD_STRING_COMPATIBILITY const size_t wxString::npos = wxSTRING_MAXLEN; #endif // wxSTD_STRING_COMPATIBILITY @@ -80,60 +76,23 @@ static const struct { wxStringData data; wxChar dummy; -} g_strEmpty = { {-1, 0, 0}, _T('\0') }; +} g_strEmpty = { {-1, 0, 0}, wxT('\0') }; // empty C style string: points to 'string data' byte of g_strEmpty -extern const wxChar WXDLLEXPORT *g_szNul = &g_strEmpty.dummy; - -// ---------------------------------------------------------------------------- -// conditional compilation -// ---------------------------------------------------------------------------- - -// we want to find out if the current platform supports vsnprintf()-like -// function: for Unix this is done with configure, for Windows we test the -// compiler explicitly. -#ifdef __WXMSW__ - #ifdef __VISUALC__ - #define wxVsnprintf _vsnprintf - #endif -#else // !Windows - #ifdef HAVE_VSNPRINTF - #define wxVsnprintf vsnprintf - #endif -#endif // Windows/!Windows - -#ifndef wxVsnprintf - // in this case we'll use vsprintf() (which is ANSI and thus should be - // always available), but it's unsafe because it doesn't check for buffer - // size - so give a warning - #define wxVsnprintf(buffer,len,format,argptr) vsprintf(buffer,format, argptr) - - #if defined(__VISUALC__) - #pragma message("Using sprintf() because no snprintf()-like function defined") - #elif defined(__GNUG__) && !defined(__UNIX__) - #warning "Using sprintf() because no snprintf()-like function defined" - #elif defined(__MWERKS__) - #warning "Using sprintf() because no snprintf()-like function defined" - #endif //compiler -#endif // no vsnprintf - -#ifdef _AIX - // AIX has vsnprintf, but there's no prototype in the system headers. - extern "C" int vsnprintf(char* str, size_t n, const char* format, va_list ap); -#endif +extern const wxChar WXDLLEXPORT *wxEmptyString = &g_strEmpty.dummy; // ---------------------------------------------------------------------------- // global functions // ---------------------------------------------------------------------------- -#ifdef wxSTD_STRING_COMPATIBILITY +#if defined(wxSTD_STRING_COMPATIBILITY) && wxUSE_STD_IOSTREAM // MS Visual C++ version 5.0 provides the new STL headers as well as the old // iostream ones. // // ATTN: you can _not_ use both of these in the same program! -istream& operator>>(istream& is, wxString& WXUNUSED(str)) +wxSTD istream& operator>>(wxSTD istream& is, wxString& WXUNUSED(str)) { #if 0 int w = is.width(0); @@ -164,6 +123,12 @@ istream& operator>>(istream& is, wxString& WXUNUSED(str)) return is; } +wxSTD ostream& operator<<(wxSTD ostream& os, const wxString& str) +{ + os << str.c_str(); + return os; +} + #endif //std::string compatibility // ---------------------------------------------------------------------------- @@ -176,15 +141,15 @@ istream& operator>>(istream& is, wxString& WXUNUSED(str)) class Averager { public: - Averager(const char *sz) { m_sz = sz; m_nTotal = m_nCount = 0; } + Averager(const wxChar *sz) { m_sz = sz; m_nTotal = m_nCount = 0; } ~Averager() - { printf("wxString: average %s = %f\n", m_sz, ((float)m_nTotal)/m_nCount); } + { 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 char *m_sz; + const wxChar *m_sz; } g_averageLength("allocation size"), g_averageSummandLength("summand length"), g_averageConcatHit("hit probability in concat"), @@ -209,11 +174,15 @@ wxString::wxString(wxChar ch, size_t nLength) Init(); if ( nLength > 0 ) { - AllocBuffer(nLength); + if ( !AllocBuffer(nLength) ) { + wxFAIL_MSG( _T("out of memory in wxString::wxString") ); + return; + } #if wxUSE_UNICODE - // memset only works on char - for (size_t n=0; n 0 ) { // trailing '\0' is written in AllocBuffer() - AllocBuffer(nLength); + if ( !AllocBuffer(nLength) ) { + wxFAIL_MSG( _T("out of memory in wxString::InitWith") ); + return; + } memcpy(m_pchData, psz + nPos, nLength*sizeof(wxChar)); } } @@ -255,52 +229,97 @@ wxString::wxString(const void *pStart, const void *pEnd) // from multibyte string wxString::wxString(const char *psz, wxMBConv& conv, size_t nLength) { - // first get necessary size - size_t nLen = psz ? conv.MB2WC((wchar_t *) NULL, psz, 0) : 0; + // first get the size of the buffer we need + size_t nLen; + if ( psz ) + { + // calculate the needed size ourselves or use a provide one + nLen = nLength == wxSTRING_MAXLEN ? conv.MB2WC(NULL, psz, 0) : nLength; + } + else + { + // nothing to convert + nLen = 0; + } - // nLength is number of *Unicode* characters here! - if (nLen > nLength) - nLen = nLength; + // anything to do? + if ( (nLen != 0) && (nLen != (size_t)-1) ) + { + if ( !AllocBuffer(nLen) ) + { + wxFAIL_MSG( _T("out of memory in wxString::wxString") ); + return; + } + + // MB2WC wants the buffer size, not the string length + if ( conv.MB2WC(m_pchData, psz, nLen + 1) != (size_t)-1 ) + { + // initialized ok + return; + } + //else: the conversion failed -- leave the string empty (what else?) + } - // empty? - if ( nLen != 0 ) { - AllocBuffer(nLen); - conv.MB2WC(m_pchData, psz, nLen); - } - else { Init(); - } } -#else +#else // ANSI +#if wxUSE_WCHAR_T // from wide string -wxString::wxString(const wchar_t *pwz) +wxString::wxString(const wchar_t *pwz, wxMBConv& conv, size_t nLength) { - // first get necessary size - size_t nLen = pwz ? wxWC2MB((char *) NULL, pwz, 0) : 0; + // first get the size of the buffer we need + size_t nLen; + if ( pwz ) + { + // calculate the needed size ourselves or use a provide one + nLen = nLength == wxSTRING_MAXLEN ? conv.WC2MB(NULL, pwz, 0) : nLength; + } + else + { + // nothing to convert + nLen = 0; + } + + // anything to do? + if ( (nLen != 0) && (nLen != (size_t)-1) ) + { + if ( !AllocBuffer(nLen) ) + { + wxFAIL_MSG( _T("out of memory in wxString::wxString") ); + return; + } + + // WC2MB wants the buffer size, not the string length + if ( conv.WC2MB(m_pchData, pwz, nLen + 1) != (size_t)-1 ) + { + // initialized ok + return; + } + //else: the conversion failed -- leave the string empty (what else?) + } - // empty? - if ( nLen != 0 ) { - AllocBuffer(nLen); - wxWC2MB(m_pchData, pwz, nLen); - } - else { Init(); - } } +#endif // wxUSE_WCHAR_T -#endif +#endif // Unicode/ANSI // --------------------------------------------------------------------------- // memory allocation // --------------------------------------------------------------------------- // allocates memory needed to store a C string of length nLen -void wxString::AllocBuffer(size_t nLen) +bool wxString::AllocBuffer(size_t nLen) { - wxASSERT( nLen > 0 ); // - wxASSERT( nLen <= INT_MAX-1 ); // max size (enough room for 1 extra) + // 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); @@ -309,50 +328,87 @@ void wxString::AllocBuffer(size_t nLen) // 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] = _T('\0'); + m_pchData[nLen] = wxT('\0'); + return TRUE; } // must be called before changing this string -void wxString::CopyBeforeWrite() +bool wxString::CopyBeforeWrite() { wxStringData* pData = GetStringData(); if ( pData->IsShared() ) { pData->Unlock(); // memory not freed because shared size_t nLen = pData->nDataLength; - AllocBuffer(nLen); + if ( !AllocBuffer(nLen) ) { + // allocation failures are handled by the caller + return FALSE; + } memcpy(m_pchData, pData->data(), nLen*sizeof(wxChar)); } wxASSERT( !GetStringData()->IsShared() ); // we must be the only owner + + return TRUE; } // must be called before replacing contents of this string -void wxString::AllocBeforeWrite(size_t nLen) +bool wxString::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() || (nLen > pData->nAllocLength) ) { + if ( pData->IsShared() || pData->IsEmpty() ) { // can't work with old buffer, get new one pData->Unlock(); - AllocBuffer(nLen); + if ( !AllocBuffer(nLen) ) { + // allocation failures are handled by the caller + return FALSE; + } } else { - // update the string length + 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(); + } + + // now we have enough space, just update the string length pData->nDataLength = nLen; } wxASSERT( !GetStringData()->IsShared() ); // we must be the only owner + + return TRUE; } // allocate enough memory for nLen characters -void wxString::Alloc(size_t nLen) +bool wxString::Alloc(size_t nLen) { wxStringData *pData = GetStringData(); if ( pData->nAllocLength <= nLen ) { @@ -360,59 +416,85 @@ void wxString::Alloc(size_t nLen) nLen += EXTRA_ALLOC; wxStringData* pData = (wxStringData*) - malloc(sizeof(wxStringData) + (nLen + 1)*sizeof(wxChar)); + 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] = _T('\0'); + m_pchData[0u] = wxT('\0'); } else if ( pData->IsShared() ) { pData->Unlock(); // memory not freed because shared size_t nOldLen = pData->nDataLength; - AllocBuffer(nLen); + if ( !AllocBuffer(nLen) ) { + // allocation failure handled by caller + return FALSE; + } memcpy(m_pchData, pData->data(), nOldLen*sizeof(wxChar)); } else { nLen += EXTRA_ALLOC; - wxStringData *p = (wxStringData *) + pData = (wxStringData *) realloc(pData, sizeof(wxStringData) + (nLen + 1)*sizeof(wxChar)); - if ( p == NULL ) { - // @@@ what to do on memory error? - return; + 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) - p->nAllocLength = nLen; - m_pchData = p->data(); + pData->nAllocLength = nLen; + m_pchData = pData->data(); } } //else: we've already got enough + return TRUE; } // shrink to minimal size (releasing extra memory) -void wxString::Shrink() +bool wxString::Shrink() { wxStringData *pData = GetStringData(); - // this variable is unused in release build, so avoid the compiler warning by - // just not declaring it -#ifdef __WXDEBUG__ - void *p = -#endif - realloc(pData, sizeof(wxStringData) + (pData->nDataLength + 1)*sizeof(wxChar)); + size_t nLen = pData->nDataLength; + void *p = realloc(pData, sizeof(wxStringData) + (nLen + 1)*sizeof(wxChar)); + + if ( p == NULL) { + wxFAIL_MSG( _T("out of memory reallocating wxString data") ); + // keep previous data since reallocation failed + return FALSE; + } + + if ( p != pData ) + { + // contrary to what one might believe, some realloc() implementation do + // move the memory block even when its size is reduced + pData = (wxStringData *)p; + + m_pchData = pData->data(); + } + + pData->nAllocLength = nLen; - wxASSERT( p != NULL ); // can't free memory? - wxASSERT( p == pData ); // we're decrementing the size - block shouldn't move! + return TRUE; } // get the pointer to writable buffer of (at least) nLen bytes wxChar *wxString::GetWriteBuf(size_t nLen) { - AllocBeforeWrite(nLen); + if ( !AllocBeforeWrite(nLen) ) { + // allocation failure handled by caller + return NULL; + } wxASSERT( GetStringData()->nRefs == 1 ); GetStringData()->Validate(FALSE); @@ -427,6 +509,12 @@ void wxString::UngetWriteBuf() GetStringData()->Validate(TRUE); } +void wxString::UngetWriteBuf(size_t nLen) +{ + GetStringData()->nDataLength = nLen; + GetStringData()->Validate(TRUE); +} + // --------------------------------------------------------------------------- // data access // --------------------------------------------------------------------------- @@ -438,17 +526,21 @@ void wxString::UngetWriteBuf() // --------------------------------------------------------------------------- // helper function: does real copy -void wxString::AssignCopy(size_t nSrcLen, const wxChar *pszSrcData) +bool wxString::AssignCopy(size_t nSrcLen, const wxChar *pszSrcData) { if ( nSrcLen == 0 ) { Reinit(); } else { - AllocBeforeWrite(nSrcLen); + if ( !AllocBeforeWrite(nSrcLen) ) { + // allocation failure handled by caller + return FALSE; + } memcpy(m_pchData, pszSrcData, nSrcLen*sizeof(wxChar)); GetStringData()->nDataLength = nSrcLen; - m_pchData[nSrcLen] = _T('\0'); + m_pchData[nSrcLen] = wxT('\0'); } + return TRUE; } // assigns one string to another @@ -475,14 +567,19 @@ wxString& wxString::operator=(const wxString& stringSrc) // assigns a single character wxString& wxString::operator=(wxChar ch) { - AssignCopy(1, &ch); + if ( !AssignCopy(1, &ch) ) { + wxFAIL_MSG( _T("out of memory in wxString::operator=(wxChar)") ); + } return *this; } + // assigns C string wxString& wxString::operator=(const wxChar *psz) { - AssignCopy(wxStrlen(psz), psz); + if ( !AssignCopy(wxStrlen(psz), psz) ) { + wxFAIL_MSG( _T("out of memory in wxString::operator=(const wxChar *)") ); + } return *this; } @@ -495,12 +592,14 @@ wxString& wxString::operator=(const unsigned char* psz) return *this; } +#if wxUSE_WCHAR_T wxString& wxString::operator=(const wchar_t *pwz) { wxString str(pwz); *this = str; return *this; } +#endif #endif @@ -509,7 +608,7 @@ wxString& wxString::operator=(const wchar_t *pwz) // --------------------------------------------------------------------------- // add something to this string -void wxString::ConcatSelf(int nSrcLen, const wxChar *pszSrcData) +bool wxString::ConcatSelf(size_t nSrcLen, const wxChar *pszSrcData) { STATISTICS_ADD(SummandLength, nSrcLen); @@ -525,7 +624,10 @@ void wxString::ConcatSelf(int nSrcLen, const wxChar *pszSrcData) // we have to allocate another buffer wxStringData* pOldData = GetStringData(); - AllocBuffer(nNewLen); + if ( !AllocBuffer(nNewLen) ) { + // allocation failure handled by caller + return FALSE; + } memcpy(m_pchData, pOldData->data(), nLen*sizeof(wxChar)); pOldData->Unlock(); } @@ -533,7 +635,10 @@ void wxString::ConcatSelf(int nSrcLen, const wxChar *pszSrcData) STATISTICS_ADD(ConcatHit, 0); // we have to grow the buffer - Alloc(nNewLen); + if ( !Alloc(nNewLen) ) { + // allocation failure handled by caller + return FALSE; + } } else { STATISTICS_ADD(ConcatHit, 1); @@ -547,10 +652,11 @@ void wxString::ConcatSelf(int nSrcLen, const wxChar *pszSrcData) // fast concatenation - all is done in our buffer memcpy(m_pchData + nLen, pszSrcData, nSrcLen*sizeof(wxChar)); - m_pchData[nNewLen] = _T('\0'); // put terminating '\0' + m_pchData[nNewLen] = wxT('\0'); // put terminating '\0' GetStringData()->nDataLength = nNewLen; // and fix the length } //else: the string to append was empty + return TRUE; } /* @@ -560,57 +666,61 @@ void wxString::ConcatSelf(int nSrcLen, const wxChar *pszSrcData) * C str + string and string + C str */ -wxString operator+(const wxString& string1, const wxString& string2) +wxString operator+(const wxString& str1, const wxString& str2) { - wxASSERT( string1.GetStringData()->IsValid() ); - wxASSERT( string2.GetStringData()->IsValid() ); + wxASSERT( str1.GetStringData()->IsValid() ); + wxASSERT( str2.GetStringData()->IsValid() ); - wxString s = string1; - s += string2; + wxString s = str1; + s += str2; return s; } -wxString operator+(const wxString& string, wxChar ch) +wxString operator+(const wxString& str, wxChar ch) { - wxASSERT( string.GetStringData()->IsValid() ); + wxASSERT( str.GetStringData()->IsValid() ); - wxString s = string; + wxString s = str; s += ch; return s; } -wxString operator+(wxChar ch, const wxString& string) +wxString operator+(wxChar ch, const wxString& str) { - wxASSERT( string.GetStringData()->IsValid() ); + wxASSERT( str.GetStringData()->IsValid() ); wxString s = ch; - s += string; + s += str; return s; } -wxString operator+(const wxString& string, const wxChar *psz) +wxString operator+(const wxString& str, const wxChar *psz) { - wxASSERT( string.GetStringData()->IsValid() ); + wxASSERT( str.GetStringData()->IsValid() ); wxString s; - s.Alloc(wxStrlen(psz) + string.Len()); - s = string; + if ( !s.Alloc(wxStrlen(psz) + str.Len()) ) { + wxFAIL_MSG( _T("out of memory in wxString::operator+") ); + } + s = str; s += psz; return s; } -wxString operator+(const wxChar *psz, const wxString& string) +wxString operator+(const wxChar *psz, const wxString& str) { - wxASSERT( string.GetStringData()->IsValid() ); + wxASSERT( str.GetStringData()->IsValid() ); wxString s; - s.Alloc(wxStrlen(psz) + string.Len()); + if ( !s.Alloc(wxStrlen(psz) + str.Len()) ) { + wxFAIL_MSG( _T("out of memory in wxString::operator+") ); + } s = psz; - s += string; + s += str; return s; } @@ -619,20 +729,83 @@ wxString operator+(const wxChar *psz, const wxString& string) // other common string functions // =========================================================================== +#if wxUSE_UNICODE + +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()); + + signed char *dest = (signed char *)buffer.data(); + + const wchar_t *pwc = c_str(); + for ( ;; ) + { + *dest++ = *pwc > SCHAR_MAX ? '_' : *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 + // --------------------------------------------------------------------------- // simple sub-string extraction // --------------------------------------------------------------------------- // helper function: clone the data attached to this string -void wxString::AllocCopy(wxString& dest, int nCopyLen, int nCopyIndex) const +bool wxString::AllocCopy(wxString& dest, int nCopyLen, int nCopyIndex) const { if ( nCopyLen == 0 ) { dest.Init(); } else { - dest.AllocBuffer(nCopyLen); + if ( !dest.AllocBuffer(nCopyLen) ) { + // allocation failure handled by caller + return FALSE; + } memcpy(dest.m_pchData, m_pchData + nCopyIndex, nCopyLen*sizeof(wxChar)); } + return TRUE; } // extract string of length nCount starting at nFirst @@ -660,11 +833,42 @@ wxString wxString::Mid(size_t nFirst, size_t nCount) const } wxString dest; - AllocCopy(dest, nCount, nFirst); + if ( !AllocCopy(dest, nCount, nFirst) ) { + wxFAIL_MSG( _T("out of memory in wxString::Mid") ); + } return dest; } +// check that the tring 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; +} + // extract nCount last (rightmost) characters wxString wxString::Right(size_t nCount) const { @@ -672,7 +876,9 @@ wxString wxString::Right(size_t nCount) const nCount = GetStringData()->nDataLength; wxString dest; - AllocCopy(dest, nCount, GetStringData()->nDataLength - nCount); + if ( !AllocCopy(dest, nCount, GetStringData()->nDataLength - nCount) ) { + wxFAIL_MSG( _T("out of memory in wxString::Right") ); + } return dest; } @@ -697,7 +903,9 @@ wxString wxString::Left(size_t nCount) const nCount = GetStringData()->nDataLength; wxString dest; - AllocCopy(dest, nCount, 0); + if ( !AllocCopy(dest, nCount, 0) ) { + wxFAIL_MSG( _T("out of memory in wxString::Left") ); + } return dest; } @@ -706,7 +914,7 @@ wxString wxString::Left(size_t nCount) const wxString wxString::BeforeFirst(wxChar ch) const { wxString str; - for ( const wxChar *pc = m_pchData; *pc != _T('\0') && *pc != ch; pc++ ) + for ( const wxChar *pc = m_pchData; *pc != wxT('\0') && *pc != ch; pc++ ) str += *pc; return str; @@ -746,7 +954,7 @@ size_t wxString::Replace(const wxChar *szOld, const wxChar *szNew, bool bReplace wxString strTemp; const wxChar *pCurrent = m_pchData; const wxChar *pSubstr; - while ( *pCurrent != _T('\0') ) { + while ( *pCurrent != wxT('\0') ) { pSubstr = wxStrstr(pCurrent, szOld); if ( pSubstr == NULL ) { // strTemp is unused if no replacements were made, so avoid the copy @@ -758,7 +966,10 @@ size_t wxString::Replace(const wxChar *szOld, const wxChar *szNew, bool bReplace } else { // take chars before match - strTemp.ConcatSelf(pSubstr - pCurrent, pCurrent); + if ( !strTemp.ConcatSelf(pSubstr - pCurrent, pCurrent) ) { + wxFAIL_MSG( _T("out of memory in wxString::Replace") ); + return 0; + } strTemp += szNew; pCurrent = pSubstr + uiOldLen; // restart after match @@ -801,6 +1012,8 @@ bool wxString::IsWord() const bool wxString::IsNumber() const { const wxChar *s = (const wxChar*) *this; + if (wxStrlen(s)) + if ((s[0] == '-') || (s[0] == '+')) s++; while(*s){ if(!wxIsdigit(*s)) return(FALSE); s++; @@ -822,7 +1035,10 @@ wxString wxString::Strip(stripType w) const wxString& wxString::MakeUpper() { - CopyBeforeWrite(); + if ( !CopyBeforeWrite() ) { + wxFAIL_MSG( _T("out of memory in wxString::MakeUpper") ); + return *this; + } for ( wxChar *p = m_pchData; *p; p++ ) *p = (wxChar)wxToupper(*p); @@ -832,7 +1048,10 @@ wxString& wxString::MakeUpper() wxString& wxString::MakeLower() { - CopyBeforeWrite(); + if ( !CopyBeforeWrite() ) { + wxFAIL_MSG( _T("out of memory in wxString::MakeLower") ); + return *this; + } for ( wxChar *p = m_pchData; *p; p++ ) *p = (wxChar)wxTolower(*p); @@ -844,36 +1063,47 @@ wxString& wxString::MakeLower() // 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 ( !IsEmpty() && - ( - (bFromRight && wxIsspace(GetChar(Len() - 1))) || - (!bFromRight && wxIsspace(GetChar(0u))) + if ( !IsEmpty() && + ( + (bFromRight && wxSafeIsspace(GetChar(Len() - 1))) || + (!bFromRight && wxSafeIsspace(GetChar(0u))) ) ) { // ok, there is at least one space to trim - CopyBeforeWrite(); + if ( !CopyBeforeWrite() ) { + wxFAIL_MSG( _T("out of memory in wxString::Trim") ); + return *this; + } if ( bFromRight ) { // find last non-space character wxChar *psz = m_pchData + GetStringData()->nDataLength - 1; - while ( wxIsspace(*psz) && (psz >= m_pchData) ) + while ( wxSafeIsspace(*psz) && (psz >= m_pchData) ) psz--; // truncate at trailing space start - *++psz = _T('\0'); + *++psz = wxT('\0'); GetStringData()->nDataLength = psz - m_pchData; } else { // find first non-space character const wxChar *psz = m_pchData; - while ( wxIsspace(*psz) ) + while ( wxSafeIsspace(*psz) ) psz++; // fix up data and length @@ -906,9 +1136,12 @@ wxString& wxString::Pad(size_t nCount, wxChar chPad, bool bFromRight) wxString& wxString::Truncate(size_t uiLen) { if ( uiLen < Len() ) { - CopyBeforeWrite(); + if ( !CopyBeforeWrite() ) { + wxFAIL_MSG( _T("out of memory in wxString::Truncate") ); + return *this; + } - *(m_pchData + uiLen) = _T('\0'); + *(m_pchData + uiLen) = wxT('\0'); GetStringData()->nDataLength = uiLen; } //else: nothing to do, string is already short enough @@ -936,36 +1169,77 @@ int wxString::Find(const wxChar *pszSub) const return (psz == NULL) ? wxNOT_FOUND : psz - (const wxChar*) m_pchData; } -// --------------------------------------------------------------------------- -// stream-like operators -// --------------------------------------------------------------------------- -wxString& wxString::operator<<(int i) +// ---------------------------------------------------------------------------- +// conversion to numbers +// ---------------------------------------------------------------------------- + +bool wxString::ToLong(long *val, int base) const { - wxString res; - res.Printf(_T("%d"), i); + wxCHECK_MSG( val, FALSE, _T("NULL pointer in wxString::ToLong") ); + wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") ); + + const wxChar *start = c_str(); + wxChar *end; + *val = wxStrtol(start, &end, base); - return (*this) << res; + // return TRUE only if scan was stopped by the terminating NUL and if the + // string was not empty to start with + return !*end && (end != start); } -wxString& wxString::operator<<(float f) +bool wxString::ToULong(unsigned long *val, int base) const { - wxString res; - res.Printf(_T("%f"), f); + wxCHECK_MSG( val, FALSE, _T("NULL pointer in wxString::ToULong") ); + wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") ); + + const wxChar *start = c_str(); + wxChar *end; + *val = wxStrtoul(start, &end, base); - return (*this) << res; + // return TRUE only if scan was stopped by the terminating NUL and if the + // string was not empty to start with + return !*end && (end != start); } -wxString& wxString::operator<<(double d) +bool wxString::ToDouble(double *val) const { - wxString res; - res.Printf(_T("%g"), d); + wxCHECK_MSG( val, FALSE, _T("NULL pointer in wxString::ToDouble") ); + + const wxChar *start = c_str(); + wxChar *end; + *val = wxStrtod(start, &end); - return (*this) << res; + // return TRUE only if scan was stopped by the terminating NUL and if the + // string was not empty to start with + return !*end && (end != start); } // --------------------------------------------------------------------------- // formatted output // --------------------------------------------------------------------------- + +/* static */ +wxString wxString::Format(const wxChar *pszFormat, ...) +{ + va_list argptr; + va_start(argptr, pszFormat); + + wxString s; + s.PrintfV(pszFormat, argptr); + + va_end(argptr); + + return s; +} + +/* static */ +wxString wxString::FormatV(const wxChar *pszFormat, va_list argptr) +{ + wxString s; + s.PrintfV(pszFormat, argptr); + return s; +} + int wxString::Printf(const wxChar *pszFormat, ...) { va_list argptr; @@ -980,313 +1254,134 @@ int wxString::Printf(const wxChar *pszFormat, ...) int wxString::PrintfV(const wxChar* pszFormat, va_list argptr) { - // static buffer to avoid dynamic memory allocation each time - static char s_szScratch[1024]; -#if wxUSE_THREADS - // protect the static buffer - static wxCriticalSection critsect; - wxCriticalSectionLocker lock(critsect); -#endif + int size = 1024; + for ( ;; ) + { + wxChar *buf = GetWriteBuf(size + 1); + if ( !buf ) + { + // out of memory + return -1; + } -#if 1 // the new implementation - - Reinit(); - for (size_t n = 0; pszFormat[n]; n++) - if (pszFormat[n] == _T('%')) { - static char s_szFlags[256] = "%"; - size_t flagofs = 1; - bool adj_left = FALSE, in_prec = FALSE, - prec_dot = FALSE, done = FALSE; - int ilen = 0; - size_t min_width = 0, max_width = wxSTRING_MAXLEN; - do { -#define CHECK_PREC if (in_prec && !prec_dot) { s_szFlags[flagofs++] = '.'; prec_dot = TRUE; } - switch (pszFormat[++n]) { - case _T('\0'): - done = TRUE; - break; - case _T('%'): - *this += _T('%'); - done = TRUE; - break; - case _T('#'): - case _T('0'): - case _T(' '): - case _T('+'): - case _T('\''): - CHECK_PREC - s_szFlags[flagofs++] = pszFormat[n]; - break; - case _T('-'): - CHECK_PREC - adj_left = TRUE; - s_szFlags[flagofs++] = pszFormat[n]; - break; - case _T('.'): - CHECK_PREC - in_prec = TRUE; - prec_dot = FALSE; - max_width = 0; - // dot will be auto-added to s_szFlags if non-negative number follows - break; - case _T('h'): - ilen = -1; - CHECK_PREC - s_szFlags[flagofs++] = pszFormat[n]; - break; - case _T('l'): - ilen = 1; - CHECK_PREC - s_szFlags[flagofs++] = pszFormat[n]; - break; - case _T('q'): - case _T('L'): - ilen = 2; - CHECK_PREC - s_szFlags[flagofs++] = pszFormat[n]; - break; - case _T('Z'): - ilen = 3; - CHECK_PREC - s_szFlags[flagofs++] = pszFormat[n]; - break; - case _T('*'): - { - int len = va_arg(argptr, int); - if (in_prec) { - if (len<0) break; - CHECK_PREC - max_width = len; - } else { - if (len<0) { - adj_left = !adj_left; - s_szFlags[flagofs++] = '-'; - len = -len; - } - min_width = len; - } - flagofs += ::sprintf(s_szFlags+flagofs,"%d",len); - } - break; - case _T('1'): case _T('2'): case _T('3'): - case _T('4'): case _T('5'): case _T('6'): - case _T('7'): case _T('8'): case _T('9'): - { - int len = 0; - CHECK_PREC - while ((pszFormat[n]>=_T('0')) && (pszFormat[n]<=_T('9'))) { - s_szFlags[flagofs++] = pszFormat[n]; - len = len*10 + (pszFormat[n] - _T('0')); - n++; - } - if (in_prec) max_width = len; - else min_width = len; - n--; // the main loop pre-increments n again - } - break; - case _T('d'): - case _T('i'): - case _T('o'): - case _T('u'): - case _T('x'): - case _T('X'): - CHECK_PREC - s_szFlags[flagofs++] = pszFormat[n]; - s_szFlags[flagofs] = '\0'; - if (ilen == 0 ) { - int val = va_arg(argptr, int); - ::sprintf(s_szScratch, s_szFlags, val); - } - else if (ilen == -1) { - short int val = va_arg(argptr, short int); - ::sprintf(s_szScratch, s_szFlags, val); - } - else if (ilen == 1) { - long int val = va_arg(argptr, long int); - ::sprintf(s_szScratch, s_szFlags, val); - } - else if (ilen == 2) { -#if SIZEOF_LONG_LONG - long long int val = va_arg(argptr, long long int); - ::sprintf(s_szScratch, s_szFlags, val); -#else - long int val = va_arg(argptr, long int); - ::sprintf(s_szScratch, s_szFlags, val); -#endif - } - else if (ilen == 3) { - size_t val = va_arg(argptr, size_t); - ::sprintf(s_szScratch, s_szFlags, val); - } - *this += wxString(s_szScratch); - done = TRUE; - break; - case _T('e'): - case _T('E'): - case _T('f'): - case _T('g'): - case _T('G'): - CHECK_PREC - s_szFlags[flagofs++] = pszFormat[n]; - s_szFlags[flagofs] = '\0'; - if (ilen == 2) { - long double val = va_arg(argptr, long double); - ::sprintf(s_szScratch, s_szFlags, val); - } else { - double val = va_arg(argptr, double); - ::sprintf(s_szScratch, s_szFlags, val); - } - *this += wxString(s_szScratch); - done = TRUE; - break; - case _T('p'): - { - void *val = va_arg(argptr, void *); - CHECK_PREC - s_szFlags[flagofs++] = pszFormat[n]; - s_szFlags[flagofs] = '\0'; - ::sprintf(s_szScratch, s_szFlags, val); - *this += wxString(s_szScratch); - done = TRUE; - } - break; - case _T('c'): - { - wxChar val = va_arg(argptr, int); - // we don't need to honor padding here, do we? - *this += val; - done = TRUE; - } - break; - case _T('s'): - if (ilen == -1) { - // wx extension: we'll let %hs mean non-Unicode strings - char *val = va_arg(argptr, char *); -#if wxUSE_UNICODE - // ASCII->Unicode constructor handles max_width right - wxString s(val, wxConv_libc, max_width); -#else - size_t len = wxSTRING_MAXLEN; - if (val) { - for (len = 0; val[len] && (len= 1) { - long int *val = va_arg(argptr, long int *); - *val = Len(); - } - done = TRUE; - break; - default: - if (wxIsalpha(pszFormat[n])) - // probably some flag not taken care of here yet - s_szFlags[flagofs++] = pszFormat[n]; - else { - // bad format - *this += _T('%'); // just to pass the glibc tst-printf.c - n--; - done = TRUE; - } - break; - } -#undef CHECK_PREC - } while (!done); - } else *this += pszFormat[n]; + int len = wxVsnprintf(buf, size, pszFormat, argptr); -#else - // NB: wxVsnprintf() may return either less than the buffer size or -1 if there - // is not enough place depending on implementation - int iLen = wxVsnprintf(s_szScratch, WXSIZEOF(s_szScratch), pszFormat, argptr); - char *buffer; - if ( iLen < (int)WXSIZEOF(s_szScratch) ) { - buffer = s_szScratch; - } - else { - int size = WXSIZEOF(s_szScratch) * 2; - buffer = (char *)malloc(size); - while ( buffer != NULL ) { - iLen = wxVsnprintf(buffer, WXSIZEOF(s_szScratch), pszFormat, argptr); - if ( iLen < size ) { - // ok, there was enough space - break; - } + // 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'); - // still not enough, double it again - buffer = (char *)realloc(buffer, size *= 2); - } + UngetWriteBuf(); - if ( !buffer ) { - // out of memory - return -1; - } - } + if ( len >= 0 ) + { + // ok, there was enough space + break; + } - wxString s(buffer); - *this = s; + // still not enough, double it again + size *= 2; + } - if ( buffer != s_szScratch ) - free(buffer); -#endif + // we could have overshot + Shrink(); - return Len(); + return Len(); } // ---------------------------------------------------------------------------- // 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 { - // check char by char - const wxChar *pszTxt; - for ( pszTxt = c_str(); *pszMask != _T('\0'); pszMask++, pszTxt++ ) { + // 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 _T('?'): - if ( *pszTxt == _T('\0') ) + case wxT('?'): + if ( *pszTxt == wxT('\0') ) return FALSE; - pszTxt++; - pszMask++; + // pszTxt and pszMask will be incremented in the loop statement + break; - case _T('*'): + case wxT('*'): { + // remember where we started to be able to backtrack later + pszLastStarInText = pszTxt; + pszLastStarInMask = pszMask; + // ignore special chars immediately following this one - while ( *pszMask == _T('*') || *pszMask == _T('?') ) + // (should this be an error?) + while ( *pszMask == wxT('*') || *pszMask == wxT('?') ) pszMask++; // if there is nothing more, match - if ( *pszMask == _T('\0') ) + if ( *pszMask == wxT('\0') ) return TRUE; // are there any other metacharacters in the mask? size_t uiLenMask; - const wxChar *pEndMask = wxStrpbrk(pszMask, _T("*?")); + const wxChar *pEndMask = wxStrpbrk(pszMask, wxT("*?")); if ( pEndMask != NULL ) { // we have to match the string between two metachars @@ -1316,7 +1411,23 @@ bool wxString::Matches(const wxChar *pszMask) const } // match only if nothing left - return *pszTxt == _T('\0'); + 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 @@ -1351,8 +1462,34 @@ int wxString::sprintf(const wxChar *pszFormat, ...) // --------------------------------------------------------------------------- // standard C++ library string functions // --------------------------------------------------------------------------- + #ifdef wxSTD_STRING_COMPATIBILITY +void wxString::resize(size_t nSize, wxChar ch) +{ + size_t len = length(); + + if ( nSize < len ) + { + Truncate(nSize); + } + else if ( nSize > len ) + { + *this += wxString(ch, nSize - len); + } + //else: we have exactly the specified length, nothing to do +} + +void wxString::swap(wxString& str) +{ + // this is slightly less efficient than fiddling with m_pchData directly, + // but it is still quite efficient as we don't copy the string here because + // ref count always stays positive + wxString tmp = str; + str = *this; + *this = tmp; +} + wxString& wxString::insert(size_t nPos, const wxString& str) { wxASSERT( str.GetStringData()->IsValid() ); @@ -1385,7 +1522,7 @@ size_t wxString::find(const wxString& str, size_t nStart) const #if !defined(__VISUALC__) || defined(__WIN32__) size_t wxString::find(const wxChar* sz, size_t nStart, size_t n) const { - return find(wxString(sz, n == npos ? 0 : n), nStart); + return find(wxString(sz, n), nStart); } #endif // VC++ 1.5 @@ -1404,9 +1541,9 @@ size_t wxString::find(wxChar ch, size_t nStart) const size_t wxString::rfind(const wxString& str, size_t nStart) const { wxASSERT( str.GetStringData()->IsValid() ); - wxASSERT( nStart <= Len() ); + wxASSERT( nStart == npos || nStart <= Len() ); - // # could be quicker than that + // TODO could be made much quicker than that const wxChar *p = c_str() + (nStart == npos ? Len() : nStart); while ( p >= c_str() + str.Len() ) { if ( wxStrncmp(p - str.Len(), str, str.Len()) == 0 ) @@ -1421,28 +1558,129 @@ size_t wxString::rfind(const wxString& str, size_t nStart) const #if !defined(__VISUALC__) || defined(__WIN32__) size_t wxString::rfind(const wxChar* sz, size_t nStart, size_t n) const { - return rfind(wxString(sz, n == npos ? 0 : n), nStart); + return rfind(wxString(sz, n == npos ? wxSTRING_MAXLEN : n), nStart); } size_t wxString::rfind(wxChar ch, size_t nStart) const { - wxASSERT( nStart <= Len() ); + if ( nStart == npos ) + { + nStart = Len(); + } + else + { + wxASSERT( nStart <= Len() ); + } - const wxChar *p = wxStrrchr(c_str() + nStart, ch); + const wxChar *p = wxStrrchr(c_str(), ch); - return p == NULL ? npos : p - c_str(); + if ( p == NULL ) + return npos; + + size_t result = p - c_str(); + return ( result > nStart ) ? npos : result; } #endif // VC++ 1.5 -wxString wxString::substr(size_t nStart, size_t nLen) const +size_t wxString::find_first_of(const wxChar* sz, size_t nStart) const +{ + const wxChar *start = c_str() + nStart; + const wxChar *firstOf = wxStrpbrk(start, sz); + if ( firstOf ) + return firstOf - c_str(); + else + return npos; +} + +size_t wxString::find_last_of(const wxChar* sz, size_t nStart) const +{ + if ( nStart == npos ) + { + nStart = Len(); + } + else + { + wxASSERT( nStart <= Len() ); + } + + for ( const wxChar *p = c_str() + length() - 1; p >= c_str(); p-- ) + { + if ( wxStrchr(sz, *p) ) + return p - c_str(); + } + + return npos; +} + +size_t wxString::find_first_not_of(const wxChar* sz, size_t nStart) const +{ + if ( nStart == npos ) + { + nStart = Len(); + } + else + { + wxASSERT( nStart <= Len() ); + } + + size_t nAccept = wxStrspn(c_str() + nStart, sz); + if ( nAccept >= length() - nStart ) + return npos; + else + return nAccept; +} + +size_t wxString::find_first_not_of(wxChar ch, size_t nStart) const +{ + wxASSERT( nStart <= Len() ); + + for ( const wxChar *p = c_str() + nStart; *p; p++ ) + { + if ( *p != ch ) + return p - c_str(); + } + + return npos; +} + +size_t wxString::find_last_not_of(const wxChar* sz, size_t nStart) const +{ + if ( nStart == npos ) + { + nStart = Len(); + } + else + { + wxASSERT( nStart <= Len() ); + } + + for ( const wxChar *p = c_str() + nStart - 1; p >= c_str(); p-- ) + { + if ( !wxStrchr(sz, *p) ) + return p - c_str(); + } + + return npos; +} + +size_t wxString::find_last_not_of(wxChar ch, size_t nStart) const { - // npos means 'take all' - if ( nLen == npos ) - nLen = 0; + if ( nStart == npos ) + { + nStart = Len(); + } + else + { + wxASSERT( nStart <= Len() ); + } - wxASSERT( nStart + nLen <= Len() ); + for ( const wxChar *p = c_str() + nStart - 1; p >= c_str(); p-- ) + { + if ( *p != ch ) + return p - c_str(); + } - return wxString(c_str() + nStart, nLen == npos ? 0 : nLen); + return npos; } wxString& wxString::erase(size_t nStart, size_t nLen) @@ -1460,13 +1698,15 @@ wxString& wxString::erase(size_t nStart, size_t nLen) wxString& wxString::replace(size_t nStart, size_t nLen, const wxChar *sz) { - wxASSERT( nStart + nLen <= wxStrlen(sz) ); + wxASSERT_MSG( nStart + nLen <= Len(), + _T("index out of bounds in wxString::replace") ); wxString strTmp; + strTmp.Alloc(Len()); // micro optimisation to avoid multiple mem allocs + if ( nStart != 0 ) strTmp.append(c_str(), nStart); - strTmp += sz; - strTmp.append(c_str() + nStart + nLen); + strTmp << sz << c_str() + nStart + nLen; *this = strTmp; return *this; @@ -1495,28 +1735,28 @@ wxString& wxString::replace(size_t nStart, size_t nLen, // ArrayString // ============================================================================ -// size increment = max(50% of current size, ARRAY_MAXSIZE_INCREMENT) +// 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) +#define ARRAY_DEFAULT_INITIAL_SIZE (16) #endif #define STRING(p) ((wxString *)(&(p))) // ctor -wxArrayString::wxArrayString() +void wxArrayString::Init(bool autoSort) { m_nSize = m_nCount = 0; m_pItems = (wxChar **) NULL; + m_autoSort = autoSort; } // copy ctor wxArrayString::wxArrayString(const wxArrayString& src) { - m_nSize = - m_nCount = 0; - m_pItems = (wxChar **) NULL; + Init(src.m_autoSort); *this = src; } @@ -1527,40 +1767,50 @@ 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); - // we can't just copy the pointers here because otherwise we would share - // the strings with another array for ( size_t n = 0; n < src.m_nCount; n++ ) Add(src[n]); - - if ( m_nCount != 0 ) - memcpy(m_pItems, src.m_pItems, m_nCount*sizeof(wxChar *)); - - return *this; } // grow the array -void wxArrayString::Grow() +void wxArrayString::Grow(size_t nIncrement) { // only do it if no more place - if( m_nCount == m_nSize ) { - if( m_nSize == 0 ) { + 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 wxChar *[m_nSize]; } else { // otherwise when it's called for the first time, nIncrement would be 0 // and the array would never be expanded - wxASSERT( ARRAY_DEFAULT_INITIAL_SIZE != 0 ); - // add 50% but not too much - size_t nIncrement = m_nSize < ARRAY_DEFAULT_INITIAL_SIZE + size_t ndefIncrement = m_nSize < ARRAY_DEFAULT_INITIAL_SIZE ? ARRAY_DEFAULT_INITIAL_SIZE : m_nSize >> 1; - if ( nIncrement > ARRAY_MAXSIZE_INCREMENT ) - nIncrement = ARRAY_MAXSIZE_INCREMENT; + if ( ndefIncrement > ARRAY_MAXSIZE_INCREMENT ) + ndefIncrement = ARRAY_MAXSIZE_INCREMENT; + if ( nIncrement < ndefIncrement ) + nIncrement = ndefIncrement; m_nSize += nIncrement; wxChar **pNew = new wxChar *[m_nSize]; @@ -1612,8 +1862,6 @@ wxArrayString::~wxArrayString() // pre-allocates memory (frees the previous data!) void wxArrayString::Alloc(size_t nSize) { - wxASSERT( nSize > 0 ); - // only if old buffer was not big enough if ( nSize > m_nSize ) { Free(); @@ -1631,7 +1879,7 @@ 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 - wxChar **pNew = new wxChar *[m_nCount]; + wxChar **pNew = new wxChar *[m_nCount]; // copy data to new location memcpy(pNew, m_pItems, m_nCount*sizeof(wxChar *)); @@ -1640,23 +1888,64 @@ void wxArrayString::Shrink() } } +// return a wxString[] as required for some control ctors. +wxString* wxArrayString::GetStringArray() const +{ + wxString *array = 0; + + if( m_nCount > 0 ) + { + array = new wxString[m_nCount]; + for( size_t i = 0; i < m_nCount; i++ ) + array[i] = m_pItems[i]; + } + + return array; +} + // searches the array for an item (forward or backwards) int wxArrayString::Index(const wxChar *sz, bool bCase, bool bFromEnd) const { - if ( bFromEnd ) { - if ( m_nCount > 0 ) { - size_t ui = m_nCount; - do { - if ( STRING(m_pItems[--ui])->IsSameAs(sz, bCase) ) - return ui; - } - while ( ui != 0 ); + 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 { - for( size_t ui = 0; ui < m_nCount; ui++ ) { - if( STRING(m_pItems[ui])->IsSameAs(sz, bCase) ) - return ui; + // use linear search in unsorted array + if ( bFromEnd ) { + if ( m_nCount > 0 ) { + size_t ui = m_nCount; + do { + if ( STRING(m_pItems[--ui])->IsSameAs(sz, bCase) ) + return ui; + } + while ( ui != 0 ); + } + } + else { + for( size_t ui = 0; ui < m_nCount; ui++ ) { + if( STRING(m_pItems[ui])->IsSameAs(sz, bCase) ) + return ui; + } } } @@ -1664,46 +1953,99 @@ int wxArrayString::Index(const wxChar *sz, bool bCase, bool bFromEnd) const } // add item at the end -void wxArrayString::Add(const wxString& str) -{ - wxASSERT( str.GetStringData()->IsValid() ); +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 = wxStrcmp(str, 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") ); - Grow(); + Insert(str, lo, nInsert); + + return (size_t)lo; + } + else { + wxASSERT( str.GetStringData()->IsValid() ); - // the string data must not be deleted! - str.GetStringData()->Lock(); - m_pItems[m_nCount++] = (wxChar *)str.c_str(); + Grow(nInsert); + + for (size_t i = 0; i < nInsert; i++) + { + // the string data must not be deleted! + str.GetStringData()->Lock(); + + // just append + m_pItems[m_nCount + i] = (wxChar *)str.c_str(); // const_cast + } + 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) +void wxArrayString::Insert(const wxString& str, size_t nIndex, size_t nInsert) { wxASSERT( str.GetStringData()->IsValid() ); - wxCHECK_RET( nIndex <= m_nCount, _("bad index in wxArrayString::Insert") ); + 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(); + Grow(nInsert); - memmove(&m_pItems[nIndex + 1], &m_pItems[nIndex], + memmove(&m_pItems[nIndex + nInsert], &m_pItems[nIndex], (m_nCount - nIndex)*sizeof(wxChar *)); - str.GetStringData()->Lock(); - m_pItems[nIndex] = (wxChar *)str.c_str(); + for (size_t i = 0; i < nInsert; i++) + { + str.GetStringData()->Lock(); + m_pItems[nIndex + i] = (wxChar *)str.c_str(); + } + m_nCount += nInsert; +} - m_nCount++; +// expand the array +void wxArrayString::SetCount(size_t count) +{ + Alloc(count); + + wxString s; + while ( m_nCount < count ) + m_pItems[m_nCount++] = (wxChar *)s.c_str(); } // removes item from array (by index) -void wxArrayString::Remove(size_t nIndex) +void wxArrayString::Remove(size_t nIndex, size_t nRemove) { - wxCHECK_RET( nIndex <= m_nCount, _("bad index in wxArrayString::Remove") ); + 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") ); // release our lock - Item(nIndex).GetStringData()->Unlock(); + for (size_t i = 0; i < nRemove; i++) + Item(nIndex + i).GetStringData()->Unlock(); - memmove(&m_pItems[nIndex], &m_pItems[nIndex + 1], - (m_nCount - nIndex - 1)*sizeof(wxChar *)); - m_nCount--; + memmove(&m_pItems[nIndex], &m_pItems[nIndex + nRemove], + (m_nCount - nIndex - nRemove)*sizeof(wxChar *)); + m_nCount -= nRemove; } // removes item from array (by value) @@ -1712,7 +2054,7 @@ void wxArrayString::Remove(const wxChar *sz) int iIndex = Index(sz); wxCHECK_RET( iIndex != wxNOT_FOUND, - _("removing inexistent element in wxArrayString::Remove") ); + wxT("removing inexistent element in wxArrayString::Remove") ); Remove(iIndex); } @@ -1749,7 +2091,8 @@ static wxArrayString::CompareFunction gs_compareFunction = NULL; static bool gs_sortAscending = TRUE; // function which is called by quick sort -static int wxStringCompareFunction(const void *first, const void *second) +extern "C" int wxC_CALLING_CONV // LINKAGEMODE +wxStringCompareFunction(const void *first, const void *second) { wxString *strFirst = (wxString *)first; wxString *strSecond = (wxString *)second; @@ -1775,6 +2118,9 @@ void wxArrayString::Sort(CompareFunction compareFunction) DoSort(); + // reset it to NULL so that Sort(bool) will work the next time + gs_compareFunction = NULL; + END_SORT(); } @@ -1792,331 +2138,24 @@ void wxArrayString::Sort(bool reverseOrder) void wxArrayString::DoSort() { + wxCHECK_RET( !m_autoSort, wxT("can't use this method with sorted arrays") ); + // just sort the pointers using qsort() - of course it only works because // wxString() *is* a pointer to its data qsort(m_pItems, m_nCount, sizeof(wxChar *), wxStringCompareFunction); } -// ============================================================================ -// MBConv -// ============================================================================ - -WXDLLEXPORT_DATA(wxMBConv *) wxConv_current = &wxConv_libc; - -// ---------------------------------------------------------------------------- -// standard libc conversion -// ---------------------------------------------------------------------------- - -WXDLLEXPORT_DATA(wxMBConv) wxConv_libc; - -size_t wxMBConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const -{ - return wxMB2WC(buf, psz, n); -} - -size_t wxMBConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const +bool wxArrayString::operator==(const wxArrayString& a) const { - return wxWC2MB(buf, psz, n); -} - -// ---------------------------------------------------------------------------- -// standard file conversion -// ---------------------------------------------------------------------------- - -WXDLLEXPORT_DATA(wxMBConv_file) wxConv_file; - -// just use the libc conversion for now -size_t wxMBConv_file::MB2WC(wchar_t *buf, const char *psz, size_t n) const -{ - return wxMB2WC(buf, psz, n); -} - -size_t wxMBConv_file::WC2MB(char *buf, const wchar_t *psz, size_t n) const -{ - return wxWC2MB(buf, psz, n); -} - -// ---------------------------------------------------------------------------- -// standard gdk conversion -// ---------------------------------------------------------------------------- - -#if defined(__WXGTK__) && (GTK_MINOR_VERSION > 0) -WXDLLEXPORT_DATA(wxMBConv_gdk) wxConv_gdk; - -#include - -size_t wxMBConv_gdk::MB2WC(wchar_t *buf, const char *psz, size_t n) const -{ - if (buf) { - return gdk_mbstowcs((GdkWChar *)buf, psz, n); - } else { - GdkWChar *nbuf = new GdkWChar[n=strlen(psz)]; - size_t len = gdk_mbstowcs(nbuf, psz, n); - delete [] nbuf; - return len; - } -} - -size_t wxMBConv_gdk::WC2MB(char *buf, const wchar_t *psz, size_t n) const -{ - char *mbstr = gdk_wcstombs((GdkWChar *)psz); - size_t len = mbstr ? strlen(mbstr) : 0; - if (buf) { - if (len > n) len = n; - memcpy(buf, psz, len); - if (len < n) buf[len] = 0; - } - return len; -} -#endif // GTK > 1.0 - -// ---------------------------------------------------------------------------- -// UTF-7 -// ---------------------------------------------------------------------------- - -WXDLLEXPORT_DATA(wxMBConv_UTF7) wxConv_UTF7; - -// TODO: write actual implementations of UTF-7 here -size_t wxMBConv_UTF7::MB2WC(wchar_t *buf, const char *psz, size_t n) const -{ - return 0; -} - -size_t wxMBConv_UTF7::WC2MB(char *buf, const wchar_t *psz, size_t n) const -{ - return 0; -} + if ( m_nCount != a.m_nCount ) + return FALSE; -// ---------------------------------------------------------------------------- -// UTF-8 -// ---------------------------------------------------------------------------- - -WXDLLEXPORT_DATA(wxMBConv_UTF8) wxConv_UTF8; - -// TODO: write actual implementations of UTF-8 here -size_t wxMBConv_UTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const -{ - return wxMB2WC(buf, psz, n); -} - -size_t wxMBConv_UTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const -{ - return wxWC2MB(buf, psz, n); -} - -// ---------------------------------------------------------------------------- -// specified character set -// ---------------------------------------------------------------------------- - -class wxCharacterSet -{ -public: - wxArrayString names; - wchar_t *data; -}; - -#ifndef WX_PRECOMP - #include "wx/dynarray.h" - #include "wx/filefn.h" - #include "wx/textfile.h" - #include "wx/tokenzr.h" - #include "wx/utils.h" -#endif - -WX_DECLARE_OBJARRAY(wxCharacterSet, wxCSArray); -#include "wx/arrimpl.cpp" -WX_DEFINE_OBJARRAY(wxCSArray); - -static wxCSArray wxCharsets; - -static void wxLoadCharacterSets(void) -{ - static bool already_loaded = FALSE; - - if (already_loaded) return; - -#if defined(__UNIX__) - // search through files in /usr/share/i18n/charmaps - wxString fname; - for (fname = ::wxFindFirstFile(_T("/usr/share/i18n/charmaps/*")); - !fname.IsEmpty(); - fname = ::wxFindNextFile()) { - wxTextFile cmap(fname); - if (cmap.Open()) { - wxCharacterSet *cset = new wxCharacterSet; - wxString comchar,escchar; - bool in_charset = FALSE; - - // wxFprintf(stderr,_T("Loaded: %s\n"),fname.c_str()); - - wxString line; - for (line = cmap.GetFirstLine(); - !cmap.Eof(); - line = cmap.GetNextLine()) { - // wxFprintf(stderr,_T("line contents: %s\n"),line.c_str()); - wxStringTokenizer token(line); - wxString cmd = token.GetNextToken(); - if (cmd == comchar) { - if (token.GetNextToken() == _T("alias")) { - wxStringTokenizer names(token.GetNextToken(),_T("/")); - wxString name; - while (!(name = names.GetNextToken()).IsEmpty()) - cset->names.Add(name); - } - } - else if (cmd == _T("")) - cset->names.Add(token.GetNextToken()); - else if (cmd == _T("")) - comchar = token.GetNextToken(); - else if (cmd == _T("")) - escchar = token.GetNextToken(); - else if (cmd == _T("")) { - delete cset; - cset = (wxCharacterSet *) NULL; - break; // we don't support multibyte charsets ourselves (yet) - } - else if (cmd == _T("CHARMAP")) { - cset->data = (wchar_t *)calloc(256, sizeof(wchar_t)); - in_charset = TRUE; - } - else if (cmd == _T("END")) { - if (token.GetNextToken() == _T("CHARMAP")) - in_charset = FALSE; - } - else if (in_charset) { - // format: /x00 NULL (NUL) - // /x41 LATIN CAPITAL LETTER A - wxString hex = token.GetNextToken(); - // skip whitespace (why doesn't wxStringTokenizer do this?) - while (wxIsEmpty(hex) && token.HasMoreTokens()) hex = token.GetNextToken(); - wxString uni = token.GetNextToken(); - // skip whitespace again - while (wxIsEmpty(uni) && token.HasMoreTokens()) uni = token.GetNextToken(); - - if ((hex.GetChar(0) == escchar) && (hex.GetChar(1) == _T('x')) && - (uni.Left(2) == _T("=0) { - unsigned long uni1 = ::wxHexToDec(uni.Mid(2,2)); - unsigned long uni2 = ::wxHexToDec(uni.Mid(4,2)); - cset->data[pos] = (uni1 << 16) | uni2; - // wxFprintf(stderr,_T("char %02x mapped to %04x (%c)\n"),pos,cset->data[pos],cset->data[pos]); - } - } - } - } - if (cset) { - cset->names.Shrink(); - wxCharsets.Add(cset); - } - } - } -#endif - wxCharsets.Shrink(); - already_loaded = TRUE; -} - -static wxCharacterSet *wxFindCharacterSet(const wxChar *charset) -{ - if (!charset) return (wxCharacterSet *)NULL; - wxLoadCharacterSets(); - for (size_t n=0; n4) { - if (wxString(charset,4) == _T("8859")) { - codeset << _T("8859-"); - if (*charset == _T('-')) charset++; - } - } - } - codeset << charset; - codeset.MakeUpper(); - m_name = wxStrdup(codeset.c_str()); - m_deferred = TRUE; -#endif - } -} - -void wxCSConv::LoadNow() -{ -// wxPrintf(_T("Conversion request\n")); - if (m_deferred) { - if (!m_name) { -#ifdef __UNIX__ - wxChar *lang = wxGetenv(_T("LANG")); - wxChar *dot = lang ? wxStrchr(lang, _T('.')) : (wxChar *)NULL; - if (dot) SetName(dot+1); -#endif + for ( size_t n = 0; n < m_nCount; n++ ) + { + if ( Item(n) != a[n] ) + return FALSE; } - m_cset = wxFindCharacterSet(m_name); - m_deferred = FALSE; - } -} -size_t wxCSConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const -{ - ((wxCSConv *)this)->LoadNow(); // discard constness - if (buf) { - if (m_cset) { - for (size_t c=0; cdata[psz[c]]; - } else { - // latin-1 (direct) - for (size_t c=0; cLoadNow(); // discard constness - if (buf) { - if (m_cset) { - for (size_t c=0; cdata[n] != psz[c]); n++); - buf[c] = (n>0xff) ? '?' : n; - } - } else { - // latin-1 (direct) - for (size_t c=0; c0xff) ? '?' : psz[c]; - } - return n; - } - return wcslen(psz); -}