X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/25dd56b1d32f01cb770fbbeb50851c2814fd09a6..4c663122e48dfd9eca0170b0e56541efc211c996:/src/common/string.cpp diff --git a/src/common/string.cpp b/src/common/string.cpp index c866ca4589..f0cb6d2910 100644 --- a/src/common/string.cpp +++ b/src/common/string.cpp @@ -6,10 +6,10 @@ // Created: 29/01/98 // RCS-ID: $Id$ // Copyright: (c) 1998 Vadim Zeitlin -// Licence: wxWindows license +// Licence: wxWindows licence ///////////////////////////////////////////////////////////////////////////// -#ifdef __GNUG__ +#if defined(__GNUG__) && !defined(NO_GCC_PRAGMA) #pragma implementation "string.h" #endif @@ -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! @@ -65,14 +55,26 @@ // static class variables definition // --------------------------------------------------------------------------- -#ifdef wxSTD_STRING_COMPATIBILITY - const size_t wxString::npos = wxSTRING_MAXLEN; -#endif // wxSTD_STRING_COMPATIBILITY +#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 + +#if !wxUSE_STL + const size_t wxStringBase::npos = wxSTRING_MAXLEN; +#endif // ---------------------------------------------------------------------------- // static data // ---------------------------------------------------------------------------- +#if wxUSE_STL + +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) @@ -80,60 +82,25 @@ 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 -// ---------------------------------------------------------------------------- +extern const wxChar WXDLLIMPEXP_BASE *wxEmptyString = &g_strEmpty.dummy; -// 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 // ---------------------------------------------------------------------------- // global functions // ---------------------------------------------------------------------------- -#ifdef wxSTD_STRING_COMPATIBILITY +#if 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); @@ -164,7 +131,13 @@ istream& operator>>(istream& is, wxString& WXUNUSED(str)) return is; } -#endif //std::string compatibility +wxSTD ostream& operator<<(wxSTD ostream& os, const wxString& str) +{ + os << str.c_str(); + return os; +} + +#endif // wxUSE_STD_IOSTREAM // ---------------------------------------------------------------------------- // private classes @@ -176,15 +149,15 @@ 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"), @@ -195,114 +168,75 @@ istream& operator>>(istream& is, wxString& WXUNUSED(str)) #define STATISTICS_ADD(av, val) #endif // WXSTRING_STATISTICS +#if !wxUSE_STL + // =========================================================================== -// wxString class core +// wxStringData class deallocation // =========================================================================== -// --------------------------------------------------------------------------- -// construction -// --------------------------------------------------------------------------- - -// constructs string of copies of character -wxString::wxString(wxChar ch, size_t nLength) +#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() { - Init(); - - if ( nLength > 0 ) { - AllocBuffer(nLength); - -#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 wxStringBase::InitWith") ); + return; + } memcpy(m_pchData, psz + nPos, nLength*sizeof(wxChar)); } } -#ifdef wxSTD_STRING_COMPATIBILITY - // poor man's iterators are "void *" pointers -wxString::wxString(const void *pStart, const void *pEnd) +wxStringBase::wxStringBase(const void *pStart, const void *pEnd) { InitWith((const wxChar *)pStart, 0, (const wxChar *)pEnd - (const wxChar *)pStart); } -#endif //std::string compatibility - -#if wxUSE_UNICODE - -// 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; - - // nLength is number of *Unicode* characters here! - if ((nLen != (size_t)-1) && (nLen > nLength)) - nLen = nLength; - - // empty? - if ( (nLen != 0) && (nLen != (size_t)-1) ) { - AllocBuffer(nLen); - conv.MB2WC(m_pchData, psz, nLen); - } - else { - Init(); - } -} - -#else - -#if wxUSE_WCHAR_T -// from wide string -wxString::wxString(const wchar_t *pwz) +wxStringBase::wxStringBase(size_type n, wxChar ch) { - // first get necessary size - size_t nLen = pwz ? wxWC2MB((char *) NULL, pwz, 0) : 0; - - // empty? - if ( (nLen != 0) && (nLen != (size_t)-1) ) { - AllocBuffer(nLen); - wxWC2MB(m_pchData, pwz, nLen); - } - else { - Init(); - } + Init(); + append(n, ch); } -#endif - -#endif // --------------------------------------------------------------------------- // memory allocation // --------------------------------------------------------------------------- // allocates memory needed to store a C string of length nLen -void wxString::AllocBuffer(size_t nLen) +bool wxStringBase::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); @@ -311,50 +245,116 @@ 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 wxStringBase::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 wxStringBase::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; +} + +wxStringBase& wxStringBase::append(size_t n, wxChar ch) +{ + size_type len = length(); + + if ( !CopyBeforeWrite() || !Alloc(len + n) ) { + wxFAIL_MSG( _T("out of memory in wxStringBase::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 wxStringBase::resize(size_t nSize, wxChar 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 -void wxString::Alloc(size_t nLen) +bool wxStringBase::Alloc(size_t nLen) { wxStringData *pData = GetStringData(); if ( pData->nAllocLength <= nLen ) { @@ -362,291 +362,887 @@ 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() +wxStringBase::iterator wxStringBase::erase(iterator it) { - 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)); - - wxASSERT( p != NULL ); // can't free memory? - wxASSERT( p == pData ); // we're decrementing the size - block shouldn't move! + size_type idx = it - begin(); + erase(idx, 1); + return begin() + idx; } -// get the pointer to writable buffer of (at least) nLen bytes -wxChar *wxString::GetWriteBuf(size_t nLen) +wxStringBase& wxStringBase::erase(size_t nStart, size_t nLen) { - AllocBeforeWrite(nLen); - - wxASSERT( GetStringData()->nRefs == 1 ); - GetStringData()->Validate(FALSE); + wxASSERT(nStart <= length()); + size_t strLen = length() - nStart; + // delete nLen or up to the end of the string characters + nLen = strLen < nLen ? strLen : nLen; + wxString strTmp(c_str(), nStart); + strTmp.append(c_str() + nStart + nLen, length() - nStart - nLen); - return m_pchData; + swap(strTmp); + return *this; } -// put string back in a reasonable state after GetWriteBuf -void wxString::UngetWriteBuf() +wxStringBase& wxStringBase::insert(size_t nPos, const wxChar *sz, size_t n) { - GetStringData()->nDataLength = wxStrlen(m_pchData); - GetStringData()->Validate(TRUE); -} + wxASSERT( nPos <= length() ); -// --------------------------------------------------------------------------- -// data access -// --------------------------------------------------------------------------- + if ( n == npos ) n = wxStrlen(sz); + if ( n == 0 ) return *this; -// all functions are inline in string.h + if ( !CopyBeforeWrite() || !Alloc(length() + n) ) { + wxFAIL_MSG( _T("out of memory in wxStringBase::insert") ); + } -// --------------------------------------------------------------------------- -// assignment operators -// --------------------------------------------------------------------------- + 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'; -// helper function: does real copy -void wxString::AssignCopy(size_t nSrcLen, const wxChar *pszSrcData) + return *this; +} + +void wxStringBase::swap(wxStringBase& str) { - if ( nSrcLen == 0 ) { - Reinit(); - } - else { - AllocBeforeWrite(nSrcLen); - memcpy(m_pchData, pszSrcData, nSrcLen*sizeof(wxChar)); - GetStringData()->nDataLength = nSrcLen; - m_pchData[nSrcLen] = _T('\0'); - } + wxChar* tmp = str.m_pchData; + str.m_pchData = m_pchData; + m_pchData = tmp; } -// assigns one string to another -wxString& wxString::operator=(const wxString& stringSrc) +size_t wxStringBase::find(const wxStringBase& str, size_t nStart) const { - wxASSERT( stringSrc.GetStringData()->IsValid() ); + wxASSERT( str.GetStringData()->IsValid() ); + wxASSERT( nStart <= length() ); - // 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(); - } - } + const wxChar *p = wxStrstr(c_str() + nStart, str.c_str()); - return *this; + return p == NULL ? npos : p - c_str(); } -// assigns a single character -wxString& wxString::operator=(wxChar ch) +size_t wxStringBase::find(const wxChar* sz, size_t nStart, size_t n) const { - AssignCopy(1, &ch); - return *this; + return find(wxStringBase(sz, n), nStart); } -// assigns C string -wxString& wxString::operator=(const wxChar *psz) +size_t wxStringBase::find(wxChar ch, size_t nStart) const { - AssignCopy(wxStrlen(psz), psz); - return *this; -} + wxASSERT( nStart <= length() ); -#if !wxUSE_UNICODE + const wxChar *p = wxStrchr(c_str() + nStart, ch); -// same as 'signed char' variant -wxString& wxString::operator=(const unsigned char* psz) -{ - *this = (const char *)psz; - return *this; + return p == NULL ? npos : p - c_str(); } -#if wxUSE_WCHAR_T -wxString& wxString::operator=(const wchar_t *pwz) +size_t wxStringBase::rfind(const wxStringBase& str, size_t nStart) const { - wxString str(pwz); - *this = str; - return *this; -} -#endif + wxASSERT( str.GetStringData()->IsValid() ); + wxASSERT( nStart == npos || nStart <= length() ); -#endif + if ( length() >= str.length() ) + { + // avoids a corner case later + if ( length() == 0 && str.length() == 0 ) + return 0; -// --------------------------------------------------------------------------- -// string concatenation -// --------------------------------------------------------------------------- + // "top" is the point where search starts from + size_t top = length() - str.length(); -// add something to this string -void wxString::ConcatSelf(int nSrcLen, const wxChar *pszSrcData) -{ - STATISTICS_ADD(SummandLength, nSrcLen); + if ( nStart == npos ) + nStart = length() - 1; + if ( nStart < top ) + top = nStart; - // concatenating an empty string is a NOP - if ( nSrcLen > 0 ) { - wxStringData *pData = GetStringData(); - size_t nLen = pData->nDataLength; - size_t nNewLen = nLen + nSrcLen; + const wxChar *cursor = c_str() + top; + do + { + if ( memcmp(cursor, str.c_str(), + str.length() * sizeof(wxChar)) == 0 ) + { + return cursor - c_str(); + } + } while ( cursor-- > c_str() ); + } + + return npos; +} - // alloc new buffer if current is too small - if ( pData->IsShared() ) { - STATISTICS_ADD(ConcatHit, 0); +size_t wxStringBase::rfind(const wxChar* sz, size_t nStart, size_t n) const +{ + return rfind(wxStringBase(sz, n), nStart); +} - // we have to allocate another buffer - wxStringData* pOldData = GetStringData(); - AllocBuffer(nNewLen); - memcpy(m_pchData, pOldData->data(), nLen*sizeof(wxChar)); - pOldData->Unlock(); +size_t wxStringBase::rfind(wxChar ch, size_t nStart) const +{ + if ( nStart == npos ) + { + nStart = length(); } - else if ( nNewLen > pData->nAllocLength ) { - STATISTICS_ADD(ConcatHit, 0); - - // we have to grow the buffer - Alloc(nNewLen); + else + { + wxASSERT( nStart <= length() ); } - else { - STATISTICS_ADD(ConcatHit, 1); - // the buffer is already big enough + const wxChar *actual; + for ( actual = c_str() + ( nStart == npos ? length() : nStart + 1 ); + actual > c_str(); --actual ) + { + if ( *(actual - 1) == ch ) + return (actual - 1) - c_str(); } - // 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] = _T('\0'); // put terminating '\0' - GetStringData()->nDataLength = nNewLen; // and fix the length - } - //else: the string to append was empty + return npos; } -/* - * concatenation functions come in 5 flavours: - * string + string - * char + string and string + char - * C str + string and string + C str - */ - -wxString operator+(const wxString& string1, const wxString& string2) +size_t wxStringBase::find_first_of(const wxChar* sz, size_t nStart) const { - wxASSERT( string1.GetStringData()->IsValid() ); - wxASSERT( string2.GetStringData()->IsValid() ); - - wxString s = string1; - s += string2; - - return s; + const wxChar *start = c_str() + nStart; + const wxChar *firstOf = wxStrpbrk(start, sz); + if ( firstOf ) + return firstOf - c_str(); + else + return npos; } -wxString operator+(const wxString& string, wxChar ch) +size_t wxStringBase::find_first_of(const wxChar* sz, size_t nStart, + size_t n) const { - wxASSERT( string.GetStringData()->IsValid() ); - - wxString s = string; - s += ch; - - return s; + return find_first_of(wxStringBase(sz, n), nStart); } -wxString operator+(wxChar ch, const wxString& string) +size_t wxStringBase::find_last_of(const wxChar* sz, size_t nStart) const { - wxASSERT( string.GetStringData()->IsValid() ); + if ( nStart == npos ) + { + nStart = length() - 1; + } + else + { + wxASSERT_MSG( nStart <= length(), + _T("invalid index in find_last_of()") ); + } - wxString s = ch; - s += string; + for ( const wxChar *p = c_str() + nStart; p >= c_str(); --p ) + { + if ( wxStrchr(sz, *p) ) + return p - c_str(); + } - return s; + return npos; } -wxString operator+(const wxString& string, const wxChar *psz) +size_t wxStringBase::find_last_of(const wxChar* sz, size_t nStart, + size_t n) const { - wxASSERT( string.GetStringData()->IsValid() ); + return find_last_of(wxStringBase(sz, n), nStart); +} - wxString s; - s.Alloc(wxStrlen(psz) + string.Len()); - s = string; - s += psz; +size_t wxStringBase::find_first_not_of(const wxChar* sz, size_t nStart) const +{ + if ( nStart == npos ) + { + nStart = length(); + } + else + { + wxASSERT( nStart <= length() ); + } - return s; + size_t nAccept = wxStrspn(c_str() + nStart, sz); + if ( nAccept >= length() - nStart ) + return npos; + else + return nStart + nAccept; } -wxString operator+(const wxChar *psz, const wxString& string) +size_t wxStringBase::find_first_not_of(const wxChar* sz, size_t nStart, + size_t n) const { - wxASSERT( string.GetStringData()->IsValid() ); + return find_first_not_of(wxStringBase(sz, n), nStart); +} - wxString s; - s.Alloc(wxStrlen(psz) + string.Len()); - s = psz; - s += string; +size_t wxStringBase::find_first_not_of(wxChar ch, size_t nStart) const +{ + wxASSERT( nStart <= length() ); - return s; + for ( const wxChar *p = c_str() + nStart; *p; p++ ) + { + if ( *p != ch ) + return p - c_str(); + } + + return npos; } -// =========================================================================== -// other common string functions -// =========================================================================== +size_t wxStringBase::find_last_not_of(const wxChar* sz, size_t nStart) const +{ + if ( nStart == npos ) + { + nStart = length() - 1; + } + else + { + wxASSERT( nStart <= length() ); + } + + for ( const wxChar *p = c_str() + nStart; p >= c_str(); --p ) + { + if ( !wxStrchr(sz, *p) ) + return p - c_str(); + } + + return npos; +} + +size_t wxStringBase::find_last_not_of(const wxChar* sz, size_t nStart, + size_t n) const +{ + return find_last_not_of(wxStringBase(sz, n), nStart); +} + +size_t wxStringBase::find_last_not_of(wxChar ch, size_t nStart) const +{ + if ( nStart == npos ) + { + nStart = length() - 1; + } + else + { + wxASSERT( nStart <= length() ); + } + + for ( const wxChar *p = c_str() + nStart; p >= c_str(); --p ) + { + if ( *p != ch ) + return p - c_str(); + } + + return npos; +} + +wxStringBase& wxStringBase::replace(size_t nStart, size_t nLen, + const wxChar *sz) +{ + wxASSERT_MSG( nStart <= length(), + _T("index out of bounds in wxStringBase::replace") ); + size_t strLen = length() - nStart; + nLen = strLen < nLen ? strLen : nLen; + + wxStringBase strTmp; + strTmp.reserve(length()); // micro optimisation to avoid multiple mem allocs + + if ( nStart != 0 ) + strTmp.append(c_str(), nStart); + strTmp.append(sz); + strTmp.append(c_str() + nStart + nLen); + + swap(strTmp); + return *this; +} + +wxStringBase& wxStringBase::replace(size_t nStart, size_t nLen, + size_t nCount, wxChar ch) +{ + return replace(nStart, nLen, wxStringBase(ch, nCount).c_str()); +} + +wxStringBase& wxStringBase::replace(size_t nStart, size_t nLen, + const wxStringBase& str, + size_t nStart2, size_t nLen2) +{ + return replace(nStart, nLen, str.substr(nStart2, nLen2)); +} + +wxStringBase& wxStringBase::replace(size_t nStart, size_t nLen, + const wxChar* sz, size_t nCount) +{ + return replace(nStart, nLen, wxStringBase(sz, nCount).c_str()); +} + +wxStringBase wxStringBase::substr(size_t nStart, size_t nLen) const +{ + if ( nLen == npos ) + nLen = length() - nStart; + return wxStringBase(*this, nStart, nLen); +} + +// assigns one string to another +wxStringBase& wxStringBase::operator=(const wxStringBase& 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 +wxStringBase& wxStringBase::operator=(wxChar ch) +{ + if ( !AssignCopy(1, &ch) ) { + wxFAIL_MSG( _T("out of memory in wxStringBase::operator=(wxChar)") ); + } + return *this; +} + +// assigns C string +wxStringBase& wxStringBase::operator=(const wxChar *psz) +{ + if ( !AssignCopy(wxStrlen(psz), psz) ) { + wxFAIL_MSG( _T("out of memory in wxStringBase::operator=(const wxChar *)") ); + } + return *this; +} + +// helper function: does real copy +bool wxStringBase::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 wxStringBase::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; +} // --------------------------------------------------------------------------- // simple sub-string extraction // --------------------------------------------------------------------------- // helper function: clone the data attached to this string -void wxString::AllocCopy(wxString& dest, int nCopyLen, int nCopyIndex) const +bool wxStringBase::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; +} + +#endif // !wxUSE_STL + +#if !wxUSE_STL || !defined(HAVE_STD_STRING_COMPARE) + +#if !wxUSE_STL + #define STRINGCLASS wxStringBase +#else + #define STRINGCLASS wxString +#endif + +static inline int wxDoCmp(const wxChar* s1, size_t l1, + const wxChar* s2, size_t l2) +{ + if( l1 == l2 ) + return wxStrncmp(s1, s2, l1); + else if( l1 < l2 ) + { + int ret = wxStrncmp(s1, s2, l1); + return ret == 0 ? -1 : ret; + } + else if( l1 > l2 ) + { + int ret = wxStrncmp(s1, s2, l2); + return ret == 0 ? +1 : ret; + } + + wxFAIL; // must never get there + return 0; // quiet compilers +} + +#if wxUSE_STL + +int STRINGCLASS::compare(const wxStringBase& str) const +{ + return ::wxDoCmp(data(), length(), str.data(), str.length()); +} + +#endif + +int STRINGCLASS::compare(size_t nStart, size_t nLen, + const wxStringBase& str) const +{ + wxASSERT(nStart <= length()); + size_type strLen = length() - nStart; + nLen = strLen < nLen ? strLen : nLen; + return ::wxDoCmp(data() + nStart, nLen, str.data(), str.length()); +} + +int STRINGCLASS::compare(size_t nStart, size_t nLen, + const wxStringBase& 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; + return ::wxDoCmp(data() + nStart, nLen, str.data() + nStart2, nLen2); +} + +#if wxUSE_STL + +int STRINGCLASS::compare(const wxChar* sz) const +{ + size_t nLen = wxStrlen(sz); + return ::wxDoCmp(data(), length(), sz, nLen); +} + +#endif + +int STRINGCLASS::compare(size_t nStart, size_t nLen, + const wxChar* sz, size_t nCount) const +{ + wxASSERT(nStart <= length()); + size_type strLen = length() - nStart; + nLen = strLen < nLen ? strLen : nLen; + if( nCount == npos ) + nCount = wxStrlen(sz); + + return ::wxDoCmp(data() + nStart, nLen, sz, nCount); +} + +#undef STRINGCLASS + +#endif // !wxUSE_STL || !defined(HAVE_STD_STRING_COMPARE) + +// =========================================================================== +// wxString class core +// =========================================================================== + +// --------------------------------------------------------------------------- +// construction +// --------------------------------------------------------------------------- + +#if wxUSE_UNICODE + +// from multibyte string +wxString::wxString(const char *psz, wxMBConv& conv, size_t nLength) +{ + // first get the size of the buffer we need + size_t nLen; + if ( psz ) + { + // calculate the needed size ourselves or use the provided one + nLen = nLength == npos ? conv.MB2WC(NULL, psz, 0) : nLength; + } + else + { + // nothing to convert + nLen = 0; + } + + // anything to do? + if ( (nLen != 0) && (nLen != (size_t)-1) ) + { + if ( !Alloc(nLen) ) + { + wxFAIL_MSG( _T("out of memory in wxString::wxString") ); + } + else + { + wxWCharBuffer buf(nLen + 1); + // MB2WC wants the buffer size, not the string length hence +1 + nLen = conv.MB2WC(buf.data(), psz, nLen + 1); + + if ( nLen != (size_t)-1 ) + { + // initialized ok, set the real length as nLength specified by + // the caller could be greater than the real string length + assign(buf.data(), nLen); + return; + } + //else: the conversion failed -- leave the string empty (what else?) + } + } +} + +#else // ANSI + +#if wxUSE_WCHAR_T +// from wide string +wxString::wxString(const wchar_t *pwz, wxMBConv& conv, size_t nLength) +{ + // first get the size of the buffer we need + size_t nLen; + if ( pwz ) + { + // calculate the needed size ourselves or use the provided one + nLen = nLength == npos ? conv.WC2MB(NULL, pwz, 0) : nLength; + } + else + { + // nothing to convert + nLen = 0; + } + + // anything to do? + if ( (nLen != 0) && (nLen != (size_t)-1) ) + { + if ( !Alloc(nLen) ) + { + wxFAIL_MSG( _T("out of memory in wxString::wxString") ); + } + else + { + wxCharBuffer buf(nLen); + // WC2MB wants the buffer size, not the string length + if ( conv.WC2MB(buf.data(), pwz, nLen + 1) != (size_t)-1 ) + { + // initialized ok + assign(buf.data(), nLen); + return; + } + //else: the conversion failed -- leave the string empty (what else?) + } + } + + // leave empty +} +#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 +// get the pointer to writable buffer of (at least) nLen bytes +wxChar *wxString::GetWriteBuf(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 wxString::UngetWriteBuf() +{ + GetStringData()->nDataLength = wxStrlen(m_pchData); + GetStringData()->Validate(TRUE); +} + +void wxString::UngetWriteBuf(size_t nLen) +{ + GetStringData()->nDataLength = nLen; + GetStringData()->Validate(TRUE); +} +#endif + +// --------------------------------------------------------------------------- +// 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 + wxASSERT( str1.GetStringData()->IsValid() ); + wxASSERT( str2.GetStringData()->IsValid() ); +#endif + + wxString s = str1; + s += str2; + + return s; +} + +wxString operator+(const wxString& str, wxChar ch) +{ +#if !wxUSE_STL + wxASSERT( str.GetStringData()->IsValid() ); +#endif + + wxString s = str; + s += ch; + + return s; +} + +wxString operator+(wxChar ch, const wxString& str) +{ +#if !wxUSE_STL + wxASSERT( str.GetStringData()->IsValid() ); +#endif + + wxString s = ch; + s += str; + + return s; +} + +wxString operator+(const wxString& str, const wxChar *psz) +{ +#if !wxUSE_STL + wxASSERT( str.GetStringData()->IsValid() ); +#endif + + wxString s; + 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& str) +{ +#if !wxUSE_STL + wxASSERT( str.GetStringData()->IsValid() ); +#endif + + wxString s; + if ( !s.Alloc(wxStrlen(psz) + str.Len()) ) { + wxFAIL_MSG( _T("out of memory in wxString::operator+") ); + } + s = psz; + s += str; + + return s; +} + +// =========================================================================== +// 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 + // extract string of length nCount starting at nFirst wxString wxString::Mid(size_t nFirst, size_t nCount) const { - wxStringData *pData = GetStringData(); - size_t nLen = pData->nDataLength; + size_t nLen = length(); - // default value of nCount is wxSTRING_MAXLEN and means "till the end" - if ( nCount == wxSTRING_MAXLEN ) + // default value of nCount is npos and means "till the end" + if ( nCount == npos ) { nCount = nLen - nFirst; } @@ -663,20 +1259,53 @@ wxString wxString::Mid(size_t nFirst, size_t nCount) const nCount = 0; } - wxString dest; - AllocCopy(dest, nCount, nFirst); + 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; +} + // extract nCount last (rightmost) characters wxString wxString::Right(size_t nCount) const { - if ( nCount > (size_t)GetStringData()->nDataLength ) - nCount = GetStringData()->nDataLength; + if ( nCount > length() ) + nCount = length(); - wxString dest; - AllocCopy(dest, nCount, GetStringData()->nDataLength - nCount); + wxString dest(*this, length() - nCount, nCount); + if ( dest.length() != nCount ) { + wxFAIL_MSG( _T("out of memory in wxString::Right") ); + } return dest; } @@ -697,11 +1326,13 @@ wxString wxString::AfterLast(wxChar ch) const // extract nCount first (leftmost) characters wxString wxString::Left(size_t nCount) const { - if ( nCount > (size_t)GetStringData()->nDataLength ) - nCount = GetStringData()->nDataLength; + if ( nCount > length() ) + nCount = length(); - wxString dest; - AllocCopy(dest, nCount, 0); + wxString dest(*this, 0, nCount); + if ( dest.length() != nCount ) { + wxFAIL_MSG( _T("out of memory in wxString::Left") ); + } return dest; } @@ -709,11 +1340,9 @@ wxString wxString::Left(size_t nCount) const // (returns the whole string if ch not found) wxString wxString::BeforeFirst(wxChar ch) const { - wxString str; - for ( const wxChar *pc = m_pchData; *pc != _T('\0') && *pc != ch; pc++ ) - str += *pc; - - return str; + int iPos = Find(ch); + if ( iPos == wxNOT_FOUND ) iPos = length(); + return wxString(*this, 0, iPos); } /// get all characters before the last occurence of ch @@ -741,16 +1370,21 @@ wxString wxString::AfterFirst(wxChar ch) const } // replace first (or all) occurences of some substring with another one -size_t wxString::Replace(const wxChar *szOld, const wxChar *szNew, bool bReplaceAll) +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); wxString strTemp; - const wxChar *pCurrent = m_pchData; + const wxChar *pCurrent = c_str(); 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 @@ -762,7 +1396,12 @@ size_t wxString::Replace(const wxChar *szOld, const wxChar *szNew, bool bReplace } else { // take chars before match - strTemp.ConcatSelf(pSubstr - pCurrent, pCurrent); + size_type len = strTemp.length(); + strTemp.append(pCurrent, pSubstr - pCurrent); + if ( strTemp.length() != (size_t)(len + pSubstr - pCurrent) ) { + wxFAIL_MSG( _T("out of memory in wxString::Replace") ); + return 0; + } strTemp += szNew; pCurrent = pSubstr + uiOldLen; // restart after match @@ -777,7 +1416,7 @@ size_t wxString::Replace(const wxChar *szOld, const wxChar *szNew, bool bReplace } // only done if there were replacements, otherwise would have returned above - *this = strTemp; + swap(strTemp); return uiCount; } @@ -805,6 +1444,8 @@ 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++; @@ -826,20 +1467,16 @@ wxString wxString::Strip(stripType w) const wxString& wxString::MakeUpper() { - CopyBeforeWrite(); - - for ( wxChar *p = m_pchData; *p; p++ ) - *p = (wxChar)wxToupper(*p); + for ( iterator it = begin(), en = end(); it != en; ++it ) + *it = (wxChar)wxToupper(*it); return *this; } wxString& wxString::MakeLower() { - CopyBeforeWrite(); - - for ( wxChar *p = m_pchData; *p; p++ ) - *p = (wxChar)wxTolower(*p); + for ( iterator it = begin(), en = end(); it != en; ++it ) + *it = (wxChar)wxTolower(*it); return *this; } @@ -848,42 +1485,45 @@ 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 ( bFromRight ) { // find last non-space character - wxChar *psz = m_pchData + GetStringData()->nDataLength - 1; - while ( wxIsspace(*psz) && (psz >= m_pchData) ) + iterator psz = begin() + length() - 1; + while ( wxSafeIsspace(*psz) && (psz >= begin()) ) psz--; // truncate at trailing space start - *++psz = _T('\0'); - GetStringData()->nDataLength = psz - m_pchData; + *++psz = wxT('\0'); + erase(psz, end()); } else { // find first non-space character - const wxChar *psz = m_pchData; - while ( wxIsspace(*psz) ) + iterator psz = begin(); + while ( wxSafeIsspace(*psz) ) psz++; // fix up data and length - int nDataLength = GetStringData()->nDataLength - (psz - (const wxChar*) m_pchData); - memmove(m_pchData, psz, (nDataLength + 1)*sizeof(wxChar)); - GetStringData()->nDataLength = nDataLength; + erase(begin(), psz); } } @@ -900,7 +1540,7 @@ wxString& wxString::Pad(size_t nCount, wxChar chPad, bool bFromRight) else { s += *this; - *this = s; + swap(s); } return *this; @@ -910,10 +1550,7 @@ wxString& wxString::Pad(size_t nCount, wxChar chPad, bool bFromRight) wxString& wxString::Truncate(size_t uiLen) { if ( uiLen < Len() ) { - CopyBeforeWrite(); - - *(m_pchData + uiLen) = _T('\0'); - GetStringData()->nDataLength = uiLen; + erase(begin() + uiLen, end()); } //else: nothing to do, string is already short enough @@ -927,712 +1564,347 @@ wxString& wxString::Truncate(size_t uiLen) // find a character int wxString::Find(wxChar ch, bool bFromEnd) const { - const wxChar *psz = bFromEnd ? wxStrrchr(m_pchData, ch) : wxStrchr(m_pchData, ch); + size_type idx = bFromEnd ? find_last_of(ch) : find_first_of(ch); - return (psz == NULL) ? wxNOT_FOUND : psz - (const wxChar*) m_pchData; + return (idx == npos) ? wxNOT_FOUND : (int)idx; } // find a sub-string (like strstr) int wxString::Find(const wxChar *pszSub) const { - const wxChar *psz = wxStrstr(m_pchData, pszSub); - - return (psz == NULL) ? wxNOT_FOUND : psz - (const wxChar*) m_pchData; -} - -// --------------------------------------------------------------------------- -// stream-like operators -// --------------------------------------------------------------------------- -wxString& wxString::operator<<(int i) -{ - wxString res; - res.Printf(_T("%d"), i); - - return (*this) << res; -} - -wxString& wxString::operator<<(float f) -{ - wxString res; - res.Printf(_T("%f"), f); + size_type idx = find(pszSub); - return (*this) << res; + return (idx == npos) ? wxNOT_FOUND : (int)idx; } -wxString& wxString::operator<<(double d) -{ - wxString res; - res.Printf(_T("%g"), d); - - return (*this) << res; -} +// ---------------------------------------------------------------------------- +// conversion to numbers +// ---------------------------------------------------------------------------- -// --------------------------------------------------------------------------- -// formatted output -// --------------------------------------------------------------------------- -int wxString::Printf(const wxChar *pszFormat, ...) +bool wxString::ToLong(long *val, int base) const { - va_list argptr; - va_start(argptr, pszFormat); - - int iLen = PrintfV(pszFormat, argptr); + wxCHECK_MSG( val, FALSE, _T("NULL pointer in wxString::ToLong") ); + wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") ); - va_end(argptr); + const wxChar *start = c_str(); + wxChar *end; + *val = wxStrtol(start, &end, base); - return iLen; + // 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); } -int wxString::PrintfV(const wxChar* pszFormat, va_list argptr) +bool wxString::ToULong(unsigned long *val, int base) const { - // 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 - -#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, wxConvLibc, 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]; - -#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; - } + wxCHECK_MSG( val, FALSE, _T("NULL pointer in wxString::ToULong") ); + wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") ); - // still not enough, double it again - buffer = (char *)realloc(buffer, size *= 2); - } - - if ( !buffer ) { - // out of memory - return -1; - } - } + const wxChar *start = c_str(); + wxChar *end; + *val = wxStrtoul(start, &end, base); - wxString s(buffer); - *this = s; - - if ( buffer != s_szScratch ) - free(buffer); -#endif - - return Len(); + // 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); } -// ---------------------------------------------------------------------------- -// misc other operations -// ---------------------------------------------------------------------------- -bool wxString::Matches(const wxChar *pszMask) const +bool wxString::ToDouble(double *val) const { - // check char by char - const wxChar *pszTxt; - for ( pszTxt = c_str(); *pszMask != _T('\0'); pszMask++, pszTxt++ ) { - switch ( *pszMask ) { - case _T('?'): - if ( *pszTxt == _T('\0') ) - return FALSE; - - pszTxt++; - pszMask++; - break; - - case _T('*'): - { - // ignore special chars immediately following this one - while ( *pszMask == _T('*') || *pszMask == _T('?') ) - pszMask++; - - // if there is nothing more, match - if ( *pszMask == _T('\0') ) - return TRUE; - - // are there any other metacharacters in the mask? - size_t uiLenMask; - const wxChar *pEndMask = wxStrpbrk(pszMask, _T("*?")); - - 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; + wxCHECK_MSG( val, FALSE, _T("NULL pointer in wxString::ToDouble") ); - // -1 to compensate "++" in the loop - pszTxt = pMatch + uiLenMask - 1; - pszMask += uiLenMask - 1; - } - break; - - default: - if ( *pszMask != *pszTxt ) - return FALSE; - break; - } - } + const wxChar *start = c_str(); + wxChar *end; + *val = wxStrtod(start, &end); - // match only if nothing left - return *pszTxt == _T('\0'); -} - -// Count the number of chars -int wxString::Freq(wxChar ch) const -{ - int count = 0; - int len = Len(); - for (int i = 0; i < len; i++) - { - if (GetChar(i) == ch) - count ++; - } - return count; + // 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); } -// 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(); } - -int wxString::sprintf(const wxChar *pszFormat, ...) - { - va_list argptr; - va_start(argptr, pszFormat); - int iLen = PrintfV(pszFormat, argptr); - va_end(argptr); - return iLen; - } - // --------------------------------------------------------------------------- -// standard C++ library string functions +// formatted output // --------------------------------------------------------------------------- -#ifdef wxSTD_STRING_COMPATIBILITY -wxString& wxString::insert(size_t nPos, const wxString& str) +/* static */ +wxString wxString::Format(const wxChar *pszFormat, ...) { - wxASSERT( str.GetStringData()->IsValid() ); - wxASSERT( nPos <= Len() ); - - if ( !str.IsEmpty() ) { - wxString strTmp; - wxChar *pc = strTmp.GetWriteBuf(Len() + str.Len()); - wxStrncpy(pc, c_str(), nPos); - wxStrcpy(pc + nPos, str); - wxStrcpy(pc + nPos + str.Len(), c_str() + nPos); - strTmp.UngetWriteBuf(); - *this = strTmp; - } - - return *this; -} + va_list argptr; + va_start(argptr, pszFormat); -size_t wxString::find(const wxString& str, size_t nStart) const -{ - wxASSERT( str.GetStringData()->IsValid() ); - wxASSERT( nStart <= Len() ); + wxString s; + s.PrintfV(pszFormat, argptr); - const wxChar *p = wxStrstr(c_str() + nStart, str); + va_end(argptr); - return p == NULL ? npos : p - c_str(); + return s; } -// VC++ 1.5 can't cope with the default argument in the header. -#if !defined(__VISUALC__) || defined(__WIN32__) -size_t wxString::find(const wxChar* sz, size_t nStart, size_t n) const +/* static */ +wxString wxString::FormatV(const wxChar *pszFormat, va_list argptr) { - return find(wxString(sz, n == npos ? 0 : n), nStart); + wxString s; + s.PrintfV(pszFormat, argptr); + return s; } -#endif // VC++ 1.5 -// Gives a duplicate symbol (presumably a case-insensitivity problem) -#if !defined(__BORLANDC__) -size_t wxString::find(wxChar ch, size_t nStart) const +int wxString::Printf(const wxChar *pszFormat, ...) { - wxASSERT( nStart <= Len() ); - - const wxChar *p = wxStrchr(c_str() + nStart, ch); + va_list argptr; + va_start(argptr, pszFormat); - return p == NULL ? npos : p - c_str(); -} -#endif + int iLen = PrintfV(pszFormat, argptr); -size_t wxString::rfind(const wxString& str, size_t nStart) const -{ - wxASSERT( str.GetStringData()->IsValid() ); - wxASSERT( nStart <= Len() ); - - // 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 ) - return p - str.Len() - c_str(); - p--; - } + va_end(argptr); - return npos; + return iLen; } -// VC++ 1.5 can't cope with the default argument in the header. -#if !defined(__VISUALC__) || defined(__WIN32__) -size_t wxString::rfind(const wxChar* sz, size_t nStart, size_t n) const +int wxString::PrintfV(const wxChar* pszFormat, va_list argptr) { - return rfind(wxString(sz, n == npos ? 0 : n), nStart); -} + int size = 1024; + int len; -size_t wxString::rfind(wxChar ch, size_t nStart) const -{ - if ( nStart == npos ) - { - nStart = Len(); - } - else + for ( ;; ) { - wxASSERT( nStart <= Len() ); - } + { + wxStringBuffer tmp(*this, size + 1); + wxChar* buf = tmp; - const wxChar *p = wxStrrchr(c_str(), ch); + if ( !buf ) + { + // out of memory + return -1; + } - if ( p == NULL ) - return npos; + len = wxVsnprintf(buf, size, pszFormat, argptr); - size_t result = p - c_str(); - return ( result > nStart ) ? npos : result; -} -#endif // VC++ 1.5 + // 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'); + } -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 - start; - else - return npos; -} + // 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 + if ( len >= 0 && len <= size ) + { + // ok, there was enough space + break; + } -size_t wxString::find_last_of(const wxChar* sz, size_t nStart) const -{ - if ( nStart == npos ) - { - nStart = Len(); - } - else - { - wxASSERT( nStart <= Len() ); + // still not enough, double it again + size *= 2; } - for ( const char *p = c_str() + length() - 1; p >= c_str(); p-- ) - { - if ( wxStrchr(sz, *p) ) - return p - c_str(); - } + // we could have overshot + Shrink(); - return npos; + return Len(); } -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 = strspn(c_str() + nStart, sz); - if ( nAccept >= length() - nStart ) - return npos; - else - return nAccept; -} +// ---------------------------------------------------------------------------- +// misc other operations +// ---------------------------------------------------------------------------- -size_t wxString::find_first_not_of(wxChar ch, size_t nStart) const +// 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 { - wxASSERT( nStart <= Len() ); - - for ( const char *p = c_str() + nStart; *p; p++ ) + // 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 ) { - if ( *p != ch ) - return p - c_str(); + switch ( *pszMask ) + { + case _T('?'): + pattern += _T('.'); + break; + + case _T('*'): + pattern += _T(".*"); + break; + + case _T('^'): + case _T('.'): + case _T('$'): + case _T('('): + case _T(')'): + case _T('|'): + case _T('+'): + case _T('\\'): + // these characters are special in a RE, quote them + // (however note that we don't quote '[' and ']' to allow + // using them for Unix shell like matching) + pattern += _T('\\'); + // fall through + + default: + pattern += *pszMask; + } + + pszMask++; } + pattern += _T('$'); - return npos; -} + // and now use it + return wxRegEx(pattern, wxRE_NOSUB | wxRE_EXTENDED).Matches(c_str()); +#else // !wxUSE_REGEX + // TODO: this is, of course, awfully inefficient... -size_t wxString::find_last_not_of(const wxChar* sz, size_t nStart) const -{ - if ( nStart == npos ) - { - nStart = Len(); - } - else - { - wxASSERT( nStart <= Len() ); - } + // the char currently being checked + const wxChar *pszTxt = c_str(); - for ( const char *p = c_str() + nStart - 1; p >= c_str(); p-- ) - { - if ( !wxStrchr(sz, *p) ) - return p - c_str(); - } + // the last location where '*' matched + const wxChar *pszLastStarInText = NULL; + const wxChar *pszLastStarInMask = NULL; - return npos; -} +match: + for ( ; *pszMask != wxT('\0'); pszMask++, pszTxt++ ) { + switch ( *pszMask ) { + case wxT('?'): + if ( *pszTxt == wxT('\0') ) + return FALSE; -size_t wxString::find_last_not_of(wxChar ch, size_t nStart) const -{ - if ( nStart == npos ) - { - nStart = Len(); - } - else - { - wxASSERT( nStart <= Len() ); - } + // pszTxt and pszMask will be incremented in the loop statement - for ( const char *p = c_str() + nStart - 1; p >= c_str(); p-- ) - { - if ( *p != ch ) - return p - c_str(); - } + break; - return npos; -} + 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++; -wxString wxString::substr(size_t nStart, size_t nLen) const -{ - // npos means 'take all' - if ( nLen == npos ) - nLen = 0; + // if there is nothing more, match + if ( *pszMask == wxT('\0') ) + return TRUE; - wxASSERT( nStart + nLen <= Len() ); + // are there any other metacharacters in the mask? + size_t uiLenMask; + const wxChar *pEndMask = wxStrpbrk(pszMask, wxT("*?")); - return wxString(c_str() + nStart, nLen == npos ? 0 : nLen); -} + 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& wxString::erase(size_t nStart, size_t nLen) -{ - wxString strTmp(c_str(), nStart); - if ( nLen != npos ) { - wxASSERT( nStart + nLen <= Len() ); + 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; - strTmp.append(c_str() + nStart + nLen); + default: + if ( *pszMask != *pszTxt ) + return FALSE; + break; + } } - *this = strTmp; - return *this; -} + // match only if nothing left + if ( *pszTxt == wxT('\0') ) + return TRUE; -wxString& wxString::replace(size_t nStart, size_t nLen, const wxChar *sz) -{ - wxASSERT( nStart + nLen <= wxStrlen(sz) ); + // if we failed to match, backtrack if we can + if ( pszLastStarInText ) { + pszTxt = pszLastStarInText + 1; + pszMask = pszLastStarInMask; - wxString strTmp; - if ( nStart != 0 ) - strTmp.append(c_str(), nStart); - strTmp += sz; - strTmp.append(c_str() + nStart + nLen); + pszLastStarInText = NULL; - *this = strTmp; - return *this; -} + // don't bother resetting pszLastStarInMask, it's unnecessary -wxString& wxString::replace(size_t nStart, size_t nLen, size_t nCount, wxChar ch) -{ - return replace(nStart, nLen, wxString(ch, nCount)); -} + goto match; + } -wxString& wxString::replace(size_t nStart, size_t nLen, - const wxString& str, size_t nStart2, size_t nLen2) -{ - return replace(nStart, nLen, str.substr(nStart2, nLen2)); + return FALSE; +#endif // wxUSE_REGEX/!wxUSE_REGEX } -wxString& wxString::replace(size_t nStart, size_t nLen, - const wxChar* sz, size_t nCount) +// Count the number of chars +int wxString::Freq(wxChar ch) const { - return replace(nStart, nLen, wxString(sz, nCount)); + int count = 0; + int len = Len(); + for (int i = 0; i < len; i++) + { + if (GetChar(i) == ch) + count ++; + } + return count; } -#endif //std::string compatibility +// 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(); } + +int wxString::sprintf(const wxChar *pszFormat, ...) + { + va_list argptr; + va_start(argptr, pszFormat); + int iLen = PrintfV(pszFormat, argptr); + va_end(argptr); + return iLen; + } // ============================================================================ // ArrayString // ============================================================================ -// size increment = max(50% of current size, ARRAY_MAXSIZE_INCREMENT) +#include "wx/arrstr.h" + +#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) +#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; } @@ -1643,40 +1915,50 @@ 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]; @@ -1725,11 +2007,14 @@ 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) { - wxASSERT( nSize > 0 ); - // only if old buffer was not big enough if ( nSize > m_nSize ) { Free(); @@ -1747,7 +2032,7 @@ 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 *)); @@ -1756,23 +2041,68 @@ void wxArrayString::Shrink() } } +#if WXWIN_COMPATIBILITY_2_4 + +// 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; +} + +#endif // WXWIN_COMPATIBILITY_2_4 + // 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; + } } } @@ -1780,46 +2110,99 @@ 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") ); + + Insert(str, lo, nInsert); + + return (size_t)lo; + } + else { + wxASSERT( str.GetStringData()->IsValid() ); + + Grow(nInsert); - Grow(); + for (size_t i = 0; i < nInsert; i++) + { + // the string data must not be deleted! + str.GetStringData()->Lock(); - // the string data must not be deleted! - str.GetStringData()->Lock(); - m_pItems[m_nCount++] = (wxChar *)str.c_str(); + // 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; +} + +// expand the array +void wxArrayString::SetCount(size_t count) +{ + Alloc(count); - m_nCount++; + 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::RemoveAt(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) @@ -1828,9 +2211,16 @@ 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") ); + + RemoveAt(iIndex); +} - Remove(iIndex); +void wxArrayString::assign(const_iterator first, const_iterator last) +{ + reserve(last - first); + for(; first != last; ++first) + push_back(*first); } // ---------------------------------------------------------------------------- @@ -1865,7 +2255,8 @@ 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; @@ -1891,446 +2282,55 @@ void wxArrayString::Sort(CompareFunction compareFunction) DoSort(); - END_SORT(); -} - -void wxArrayString::Sort(bool reverseOrder) -{ - START_SORT(); - - wxASSERT( !gs_compareFunction ); // must have been reset to NULL - gs_sortAscending = !reverseOrder; - - DoSort(); + // reset it to NULL so that Sort(bool) will work the next time + gs_compareFunction = NULL; END_SORT(); } -void wxArrayString::DoSort() -{ - // 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 *) wxConvCurrent = &wxConvLibc; -#if !wxUSE_WCHAR_T -WXDLLEXPORT_DATA(wxMBConv) wxConvLibc, wxConvFile; -#endif - -#if wxUSE_WCHAR_T - -// ---------------------------------------------------------------------------- -// standard libc conversion -// ---------------------------------------------------------------------------- - -WXDLLEXPORT_DATA(wxMBConv) wxConvLibc; - -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 -{ - return wxWC2MB(buf, psz, n); -} - -// ---------------------------------------------------------------------------- -// standard file conversion -// ---------------------------------------------------------------------------- - -WXDLLEXPORT_DATA(wxMBConvFile) wxConvFile; - -// just use the libc conversion for now -size_t wxMBConvFile::MB2WC(wchar_t *buf, const char *psz, size_t n) const -{ - return wxMB2WC(buf, psz, n); -} - -size_t wxMBConvFile::WC2MB(char *buf, const wchar_t *psz, size_t n) const -{ - return wxWC2MB(buf, psz, n); -} - -// ---------------------------------------------------------------------------- -// standard gdk conversion -// ---------------------------------------------------------------------------- - -#ifdef __WXGTK12__ -WXDLLEXPORT_DATA(wxMBConvGdk) wxConvGdk; - -#include - -size_t wxMBConvGdk::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; - } -} +typedef int (wxC_CALLING_CONV * wxStringCompareFn)(const void *first, const void *second); -size_t wxMBConvGdk::WC2MB(char *buf, const wchar_t *psz, size_t n) const +void wxArrayString::Sort(CompareFunction2 compareFunction) { - 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; + qsort(m_pItems, m_nCount, sizeof(wxChar *), (wxStringCompareFn)compareFunction); } -#endif // GTK > 1.0 - -// ---------------------------------------------------------------------------- -// UTF-7 -// ---------------------------------------------------------------------------- - -WXDLLEXPORT_DATA(wxMBConvUTF7) wxConvUTF7; - -#if 0 -static char utf7_setD[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ" - "abcdefghijklmnopqrstuvwxyz" - "0123456789'(),-./:?"; -static char utf7_setO[]="!\"#$%&*;<=>@[]^_`{|}"; -static char utf7_setB[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ" - "abcdefghijklmnopqrstuvwxyz" - "0123456789+/"; -#endif -// TODO: write actual implementations of UTF-7 here -size_t wxMBConvUTF7::MB2WC(wchar_t * WXUNUSED(buf), - const char * WXUNUSED(psz), - size_t WXUNUSED(n)) const +void wxArrayString::Sort(bool reverseOrder) { - return 0; + Sort(reverseOrder ? wxStringSortDescending : wxStringSortAscending); } -size_t wxMBConvUTF7::WC2MB(char * WXUNUSED(buf), - const wchar_t * WXUNUSED(psz), - size_t WXUNUSED(n)) const +void wxArrayString::DoSort() { - return 0; -} + wxCHECK_RET( !m_autoSort, wxT("can't use this method with sorted arrays") ); -// ---------------------------------------------------------------------------- -// UTF-8 -// ---------------------------------------------------------------------------- - -WXDLLEXPORT_DATA(wxMBConvUTF8) wxConvUTF8; - -static unsigned long utf8_max[]={0x7f,0x7ff,0xffff,0x1fffff,0x3ffffff,0x7fffffff,0xffffffff}; - -size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const -{ - size_t len = 0; - - while (*psz && ((!buf) || (len>cnt); - while (cnt--) { - cc = *psz++; - if ((cc&0xC0)!=0x80) { - // invalid UTF-8 sequence - return (size_t)-1; - } - res=(res<<6)|(cc&0x3f); - } - if (res<=utf8_max[ocnt]) { - // illegal UTF-8 encoding - return (size_t)-1; - } - if (buf) *buf++=res; - len++; - } - } - } - if (buf && (lenutf8_max[cnt]; cnt++); - if (!cnt) { - // plain ASCII char - if (buf) *buf++=cc; - len++; - } else { - len+=cnt+1; - if (buf) { - *buf++=(-128>>cnt)|((cc>>(cnt*6))&(0x3f>>cnt)); - while (cnt--) - *buf++=0x80|((cc>>(cnt*6))&0x3f); - } - } - } - if (buf && (lennames.Add(token.GetNextToken()); - } - 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.Len() > 2) && (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); - } + for ( size_t n = 0; n < m_nCount; n++ ) + { + if ( Item(n) != a[n] ) + return FALSE; } - } -#endif - wxCharsets.Shrink(); -} - -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 - } + return TRUE; } -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 - } - m_cset = wxFindCharacterSet(m_name); - m_deferred = FALSE; - } -} +#endif // !wxUSE_STL -size_t wxCSConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const +int wxStringSortAscending(wxString* s1, wxString* s2) { - ((wxCSConv *)this)->LoadNow(); // discard constness - if (buf) { - if (m_cset) { - for (size_t c=0; cdata[(unsigned char)(psz[c])]; - } else { - // latin-1 (direct) - for (size_t c=0; cc_str(), s2->c_str()); } -size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const +int wxStringSortDescending(wxString* s1, wxString* s2) { - ((wxCSConv *)this)->LoadNow(); // 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); + return -wxStrcmp(s1->c_str(), s2->c_str()); } - -#endif//wxUSE_WCHAR_T - -#if wxUSE_WCHAR_T -const wxWCharBuffer wxMBConv::cMB2WC(const char *psz) const - { - if (psz) { - size_t nLen = MB2WC((wchar_t *) NULL, psz, 0); - wxWCharBuffer buf(nLen); - MB2WC(WCSTRINGCAST buf, psz, nLen); - return buf; - } else return wxWCharBuffer((wchar_t *) NULL); - } - -const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *psz) const - { - if (psz) { - size_t nLen = WC2MB((char *) NULL, psz, 0); - wxCharBuffer buf(nLen); - WC2MB(MBSTRINGCAST buf, psz, nLen); - return buf; - } else return wxCharBuffer((char *) NULL); - } - -#endif//wxUSE_WCHAR_T -