X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/68919656bbaa4b292162e790b272e205aa41175c..7af61b2e180b95f0009e41a5c2c09006d5077e9c:/src/common/string.cpp?ds=sidebyside diff --git a/src/common/string.cpp b/src/common/string.cpp index 2b95e23f12..7759ed8766 100644 --- a/src/common/string.cpp +++ b/src/common/string.cpp @@ -6,11 +6,11 @@ // Created: 29/01/98 // RCS-ID: $Id$ // Copyright: (c) 1998 Vadim Zeitlin -// Licence: wxWindows license +// Licence: wxWindows license ///////////////////////////////////////////////////////////////////////////// #ifdef __GNUG__ -#pragma implementation "string.h" + #pragma implementation "string.h" #endif /* @@ -28,80 +28,129 @@ #include "wx/wxprec.h" #ifdef __BORLANDC__ -#pragma hdrstop + #pragma hdrstop #endif #ifndef WX_PRECOMP -#include "wx/defs.h" -#include "wx/string.h" + #include "wx/defs.h" + #include "wx/string.h" + #include "wx/intl.h" +#if wxUSE_THREADS + #include +#endif #endif #include #include #include +#ifdef __SALFORDC__ + #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! +#define EXTRA_ALLOC (19 - nLen % 16) + // --------------------------------------------------------------------------- // static class variables definition // --------------------------------------------------------------------------- -#ifdef STD_STRING_COMPATIBILITY - const size_t wxString::npos = STRING_MAXLEN; -#endif +#ifdef wxSTD_STRING_COMPATIBILITY + const size_t wxString::npos = wxSTRING_MAXLEN; +#endif // wxSTD_STRING_COMPATIBILITY -// =========================================================================== -// static class data, special inlines -// =========================================================================== +// ---------------------------------------------------------------------------- +// static data +// ---------------------------------------------------------------------------- + +// for an empty string, GetStringData() will return this address: this +// structure has the same layout as wxStringData and it's data() method will +// return the empty string (dummy pointer) +static const struct +{ + wxStringData data; + wxChar dummy; +} g_strEmpty = { {-1, 0, 0}, _T('\0') }; -// for an empty string, GetStringData() will return this address -static int g_strEmpty[] = { -1, // ref count (locked) - 0, // current length - 0, // allocated memory - 0 }; // string data -// empty string shares memory with g_strEmpty -static wxStringData *g_strNul = (wxStringData*)&g_strEmpty; // empty C style string: points to 'string data' byte of g_strEmpty -extern const char *g_szNul = (const char *)(&g_strEmpty[3]); +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 -// =========================================================================== +// ---------------------------------------------------------------------------- // global functions -// =========================================================================== +// ---------------------------------------------------------------------------- -#ifdef STD_STRING_COMPATIBILITY +#ifdef wxSTD_STRING_COMPATIBILITY // 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! -#if 0 // def _MSC_VER - #include - #define NAMESPACE std:: -#else - #include - #define NAMESPACE -#endif //Visual C++ -NAMESPACE istream& operator>>(NAMESPACE istream& is, wxString& WXUNUSED(str)) +istream& operator>>(istream& is, wxString& WXUNUSED(str)) { #if 0 int w = is.width(0); if ( is.ipfx(0) ) { - NAMESPACE streambuf *sb = is.rdbuf(); + streambuf *sb = is.rdbuf(); str.erase(); while ( true ) { int ch = sb->sbumpc (); if ( ch == EOF ) { - is.setstate(NAMESPACE ios::eofbit); + is.setstate(ios::eofbit); break; } else if ( isspace(ch) ) { sb->sungetc(); break; } - + str += ch; if ( --w == 1 ) break; @@ -110,13 +159,42 @@ NAMESPACE istream& operator>>(NAMESPACE istream& is, wxString& WXUNUSED(str)) is.isfx(); if ( str.length() == 0 ) - is.setstate(NAMESPACE ios::failbit); + is.setstate(ios::failbit); #endif return is; } #endif //std::string compatibility +// ---------------------------------------------------------------------------- +// private classes +// ---------------------------------------------------------------------------- + +// this small class is used to gather statistics for performance tuning +//#define WXSTRING_STATISTICS +#ifdef WXSTRING_STATISTICS + class Averager + { + public: + Averager(const char *sz) { m_sz = sz; m_nTotal = m_nCount = 0; } + ~Averager() + { printf("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; + } g_averageLength("allocation size"), + g_averageSummandLength("summand length"), + g_averageConcatHit("hit probability in concat"), + g_averageInitialLength("initial string length"); + + #define STATISTICS_ADD(av, val) g_average##av.Add(val) +#else + #define STATISTICS_ADD(av, val) +#endif // WXSTRING_STATISTICS + // =========================================================================== // wxString class core // =========================================================================== @@ -125,104 +203,97 @@ NAMESPACE istream& operator>>(NAMESPACE istream& is, wxString& WXUNUSED(str)) // construction // --------------------------------------------------------------------------- -// construct an empty string -wxString::wxString() -{ - Init(); -} - -// copy constructor -wxString::wxString(const wxString& stringSrc) -{ - wxASSERT( stringSrc.GetStringData()->IsValid() ); - - if ( stringSrc.IsEmpty() ) { - // nothing to do for an empty string - Init(); - } - else { - m_pchData = stringSrc.m_pchData; // share same data - GetStringData()->Lock(); // => one more copy - } -} - // constructs string of copies of character -wxString::wxString(char ch, size_t nLength) +wxString::wxString(wxChar ch, size_t nLength) { Init(); if ( nLength > 0 ) { AllocBuffer(nLength); - wxASSERT( sizeof(char) == 1 ); // can't use memset if not - +#if wxUSE_UNICODE + // memset only works on char + for (size_t n=0; n 0 ) { // trailing '\0' is written in AllocBuffer() AllocBuffer(nLength); - memcpy(m_pchData, psz + nPos, nLength*sizeof(char)); + memcpy(m_pchData, psz + nPos, nLength*sizeof(wxChar)); } } - -// take first nLength characters of C string psz -// (default value of STRING_MAXLEN means take all the string) -wxString::wxString(const char *psz, size_t nLength) -{ - InitWith(psz, 0, nLength); -} - -// the same as previous constructor, but for compilers using unsigned char -wxString::wxString(const unsigned char* psz, size_t nLength) -{ - InitWith((const char *)psz, 0, nLength); -} - -#ifdef STD_STRING_COMPATIBILITY -// ctor from a substring -wxString::wxString(const wxString& s, size_t nPos, size_t nLen) -{ - InitWith(s.c_str(), nPos, nLen == npos ? 0 : nLen); -} +#ifdef wxSTD_STRING_COMPATIBILITY // poor man's iterators are "void *" pointers wxString::wxString(const void *pStart, const void *pEnd) { - InitWith((const char *)pStart, 0, - (const char *)pEnd - (const char *)pStart); + 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 = conv.MB2WC((wchar_t *) NULL, psz, 0); + + // nLength is number of *Unicode* characters here! + if (nLen > nLength) + nLen = nLength; + + // empty? + if ( nLen != 0 ) { + AllocBuffer(nLen); + conv.MB2WC(m_pchData, psz, nLen); + } + else { + Init(); + } +} + +#else + // from wide string wxString::wxString(const wchar_t *pwz) { // first get necessary size - size_t nLen = wcstombs(NULL, pwz, 0); + + size_t nLen = wxWC2MB((char *) NULL, pwz, 0); // empty? if ( nLen != 0 ) { AllocBuffer(nLen); - wcstombs(m_pchData, pwz, nLen); + wxWC2MB(m_pchData, pwz, nLen); } else { Init(); } } +#endif + // --------------------------------------------------------------------------- // memory allocation // --------------------------------------------------------------------------- @@ -233,33 +304,18 @@ void wxString::AllocBuffer(size_t nLen) wxASSERT( nLen > 0 ); // wxASSERT( nLen <= INT_MAX-1 ); // max size (enough room for 1 extra) + STATISTICS_ADD(Length, nLen); + // allocate memory: // 1) one extra character for '\0' termination // 2) sizeof(wxStringData) for housekeeping info - wxStringData* pData = (wxStringData*)new char[sizeof(wxStringData) + - (nLen + 1)*sizeof(char)]; + wxStringData* pData = (wxStringData*) + malloc(sizeof(wxStringData) + (nLen + EXTRA_ALLOC + 1)*sizeof(wxChar)); pData->nRefs = 1; - pData->data()[nLen] = '\0'; pData->nDataLength = nLen; - pData->nAllocLength = nLen; + pData->nAllocLength = nLen + EXTRA_ALLOC; m_pchData = pData->data(); // data starts after wxStringData -} - -// releases the string memory and reinits it -void wxString::Reinit() -{ - GetStringData()->Unlock(); - Init(); -} - -// wrapper around wxString::Reinit -void wxString::Empty() -{ - if ( GetStringData()->nDataLength != 0 ) - Reinit(); - - wxASSERT( GetStringData()->nDataLength == 0 ); - wxASSERT( GetStringData()->nAllocLength == 0 ); + m_pchData[nLen] = _T('\0'); } // must be called before changing this string @@ -269,11 +325,12 @@ void wxString::CopyBeforeWrite() if ( pData->IsShared() ) { pData->Unlock(); // memory not freed because shared - AllocBuffer(pData->nDataLength); - memcpy(m_pchData, pData->data(), (pData->nDataLength + 1)*sizeof(char)); + size_t nLen = pData->nDataLength; + AllocBuffer(nLen); + memcpy(m_pchData, pData->data(), nLen*sizeof(wxChar)); } - wxASSERT( !pData->IsShared() ); // we must be the only owner + wxASSERT( !GetStringData()->IsShared() ); // we must be the only owner } // must be called before replacing contents of this string @@ -282,27 +339,94 @@ void wxString::AllocBeforeWrite(size_t nLen) wxASSERT( nLen != 0 ); // doesn't make any sense // must not share string and must have enough space - register wxStringData* pData = GetStringData(); + wxStringData* pData = GetStringData(); if ( pData->IsShared() || (nLen > pData->nAllocLength) ) { // can't work with old buffer, get new one pData->Unlock(); AllocBuffer(nLen); } + else { + // update the string length + pData->nDataLength = nLen; + } wxASSERT( !GetStringData()->IsShared() ); // we must be the only owner } +// allocate enough memory for nLen characters +void wxString::Alloc(size_t nLen) +{ + wxStringData *pData = GetStringData(); + if ( pData->nAllocLength <= nLen ) { + if ( pData->IsEmpty() ) { + nLen += EXTRA_ALLOC; + + wxStringData* pData = (wxStringData*) + malloc(sizeof(wxStringData) + (nLen + 1)*sizeof(wxChar)); + pData->nRefs = 1; + pData->nDataLength = 0; + pData->nAllocLength = nLen; + m_pchData = pData->data(); // data starts after wxStringData + m_pchData[0u] = _T('\0'); + } + else if ( pData->IsShared() ) { + pData->Unlock(); // memory not freed because shared + size_t nOldLen = pData->nDataLength; + AllocBuffer(nLen); + memcpy(m_pchData, pData->data(), nOldLen*sizeof(wxChar)); + } + else { + nLen += EXTRA_ALLOC; + + wxStringData *p = (wxStringData *) + realloc(pData, sizeof(wxStringData) + (nLen + 1)*sizeof(wxChar)); + + if ( p == NULL ) { + // @@@ what to do on memory error? + return; + } + + // 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(); + } + } + //else: we've already got enough +} + +// shrink to minimal size (releasing extra memory) +void 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)); + + wxASSERT( p != NULL ); // can't free memory? + wxASSERT( p == pData ); // we're decrementing the size - block shouldn't move! +} + // get the pointer to writable buffer of (at least) nLen bytes -char *wxString::GetWriteBuf(size_t nLen) +wxChar *wxString::GetWriteBuf(size_t nLen) { AllocBeforeWrite(nLen); + + wxASSERT( GetStringData()->nRefs == 1 ); + GetStringData()->Validate(FALSE); + return m_pchData; } -// dtor frees memory if no other strings use it -wxString::~wxString() +// put string back in a reasonable state after GetWriteBuf +void wxString::UngetWriteBuf() { - GetStringData()->Unlock(); + GetStringData()->nDataLength = wxStrlen(m_pchData); + GetStringData()->Validate(TRUE); } // --------------------------------------------------------------------------- @@ -315,23 +439,25 @@ wxString::~wxString() // assignment operators // --------------------------------------------------------------------------- -// helper function: does real copy -void wxString::AssignCopy(size_t nSrcLen, const char *pszSrcData) +// helper function: does real copy +void wxString::AssignCopy(size_t nSrcLen, const wxChar *pszSrcData) { if ( nSrcLen == 0 ) { Reinit(); } else { AllocBeforeWrite(nSrcLen); - memcpy(m_pchData, pszSrcData, nSrcLen*sizeof(char)); + memcpy(m_pchData, pszSrcData, nSrcLen*sizeof(wxChar)); GetStringData()->nDataLength = nSrcLen; - m_pchData[nSrcLen] = '\0'; + m_pchData[nSrcLen] = _T('\0'); } } // assigns one string to another wxString& wxString::operator=(const wxString& stringSrc) { + wxASSERT( stringSrc.GetStringData()->IsValid() ); + // don't copy string over itself if ( m_pchData != stringSrc.m_pchData ) { if ( stringSrc.GetStringData()->IsEmpty() ) { @@ -349,19 +475,21 @@ wxString& wxString::operator=(const wxString& stringSrc) } // assigns a single character -wxString& wxString::operator=(char ch) +wxString& wxString::operator=(wxChar ch) { AssignCopy(1, &ch); return *this; } // assigns C string -wxString& wxString::operator=(const char *psz) +wxString& wxString::operator=(const wxChar *psz) { - AssignCopy(Strlen(psz), psz); + AssignCopy(wxStrlen(psz), psz); return *this; } +#if !wxUSE_UNICODE + // same as 'signed char' variant wxString& wxString::operator=(const unsigned char* psz) { @@ -376,95 +504,58 @@ wxString& wxString::operator=(const wchar_t *pwz) return *this; } +#endif + // --------------------------------------------------------------------------- // string concatenation // --------------------------------------------------------------------------- -// concatenate two sources -// NB: assume that 'this' is a new wxString object -void wxString::ConcatCopy(int nSrc1Len, const char *pszSrc1Data, - int nSrc2Len, const char *pszSrc2Data) -{ - int nNewLen = nSrc1Len + nSrc2Len; - if ( nNewLen != 0 ) - { - AllocBuffer(nNewLen); - memcpy(m_pchData, pszSrc1Data, nSrc1Len*sizeof(char)); - memcpy(m_pchData + nSrc1Len, pszSrc2Data, nSrc2Len*sizeof(char)); - } -} - // add something to this string -void wxString::ConcatSelf(int nSrcLen, const char *pszSrcData) +void wxString::ConcatSelf(int nSrcLen, const wxChar *pszSrcData) { + STATISTICS_ADD(SummandLength, nSrcLen); + // concatenating an empty string is a NOP - if ( nSrcLen != 0 ) { - register wxStringData *pData = GetStringData(); + 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() || - pData->nDataLength + nSrcLen > pData->nAllocLength ) { - // we have to grow the buffer, use the ConcatCopy routine - // (which will allocate memory) + if ( pData->IsShared() ) { + STATISTICS_ADD(ConcatHit, 0); + + // we have to allocate another buffer wxStringData* pOldData = GetStringData(); - ConcatCopy(pOldData->nDataLength, m_pchData, nSrcLen, pszSrcData); + AllocBuffer(nNewLen); + memcpy(m_pchData, pOldData->data(), nLen*sizeof(wxChar)); pOldData->Unlock(); } - else { - // fast concatenation when buffer big enough - memcpy(m_pchData + pData->nDataLength, pszSrcData, nSrcLen*sizeof(char)); - pData->nDataLength += nSrcLen; + else if ( nNewLen > pData->nAllocLength ) { + STATISTICS_ADD(ConcatHit, 0); - // should be enough space - wxASSERT( pData->nDataLength <= pData->nAllocLength ); - - m_pchData[pData->nDataLength] = '\0'; // put terminating '\0' + // we have to grow the buffer + Alloc(nNewLen); } - } -} + else { + STATISTICS_ADD(ConcatHit, 1); -/* - * string may be concatenated with other string, C string or a character - */ + // the buffer is already big enough + } -void wxString::operator+=(const wxString& string) -{ - ConcatSelf(string.Len(), string); -} + // should be enough space + wxASSERT( nNewLen <= GetStringData()->nAllocLength ); -void wxString::operator+=(const char *psz) -{ - ConcatSelf(Strlen(psz), psz); -} + // fast concatenation - all is done in our buffer + memcpy(m_pchData + nLen, pszSrcData, nSrcLen*sizeof(wxChar)); -void wxString::operator+=(char ch) -{ - ConcatSelf(1, &ch); + m_pchData[nNewLen] = _T('\0'); // put terminating '\0' + GetStringData()->nDataLength = nNewLen; // and fix the length + } + //else: the string to append was empty } /* - * Same as above but return the result - */ - -wxString& wxString::operator<<(const wxString& string) -{ - ConcatSelf(string.Len(), string); - return *this; -} - -wxString& wxString::operator<<(const char *psz) -{ - ConcatSelf(Strlen(psz), psz); - return *this; -} - -wxString& wxString::operator<<(char ch) -{ - ConcatSelf(1, &ch); - return *this; -} - -/* * concatenation functions come in 5 flavours: * string + string * char + string and string + char @@ -473,39 +564,56 @@ wxString& wxString::operator<<(char ch) wxString operator+(const wxString& string1, const wxString& string2) { - wxString s; - s.ConcatCopy(string1.GetStringData()->nDataLength, string1.m_pchData, - string2.GetStringData()->nDataLength, string2.m_pchData); + wxASSERT( string1.GetStringData()->IsValid() ); + wxASSERT( string2.GetStringData()->IsValid() ); + + wxString s = string1; + s += string2; + return s; } -wxString operator+(const wxString& string1, char ch) +wxString operator+(const wxString& string, wxChar ch) { - wxString s; - s.ConcatCopy(string1.GetStringData()->nDataLength, string1.m_pchData, 1, &ch); + wxASSERT( string.GetStringData()->IsValid() ); + + wxString s = string; + s += ch; + return s; } -wxString operator+(char ch, const wxString& string) +wxString operator+(wxChar ch, const wxString& string) { - wxString s; - s.ConcatCopy(1, &ch, string.GetStringData()->nDataLength, string.m_pchData); + wxASSERT( string.GetStringData()->IsValid() ); + + wxString s = ch; + s += string; + return s; } -wxString operator+(const wxString& string, const char *psz) +wxString operator+(const wxString& string, const wxChar *psz) { + wxASSERT( string.GetStringData()->IsValid() ); + wxString s; - s.ConcatCopy(string.GetStringData()->nDataLength, string.m_pchData, - Strlen(psz), psz); + s.Alloc(wxStrlen(psz) + string.Len()); + s = string; + s += psz; + return s; } -wxString operator+(const char *psz, const wxString& string) +wxString operator+(const wxChar *psz, const wxString& string) { + wxASSERT( string.GetStringData()->IsValid() ); + wxString s; - s.ConcatCopy(Strlen(psz), psz, - string.GetStringData()->nDataLength, string.m_pchData); + s.Alloc(wxStrlen(psz) + string.Len()); + s = psz; + s += string; + return s; } @@ -520,32 +628,42 @@ wxString operator+(const char *psz, const wxString& string) // helper function: clone the data attached to this string void wxString::AllocCopy(wxString& dest, int nCopyLen, int nCopyIndex) const { - if ( nCopyLen == 0 ) - { + if ( nCopyLen == 0 ) { dest.Init(); } - else - { + else { dest.AllocBuffer(nCopyLen); - memcpy(dest.m_pchData, m_pchData + nCopyIndex, nCopyLen*sizeof(char)); + memcpy(dest.m_pchData, m_pchData + nCopyIndex, nCopyLen*sizeof(wxChar)); } } // extract string of length nCount starting at nFirst -// default value of nCount is 0 and means "till the end" wxString wxString::Mid(size_t nFirst, size_t nCount) const { + wxStringData *pData = GetStringData(); + size_t nLen = pData->nDataLength; + + // default value of nCount is wxSTRING_MAXLEN and means "till the end" + if ( nCount == wxSTRING_MAXLEN ) + { + nCount = nLen - nFirst; + } + // out-of-bounds requests return sensible things - if ( nCount == 0 ) - nCount = GetStringData()->nDataLength - nFirst; + if ( nFirst + nCount > nLen ) + { + nCount = nLen - nFirst; + } - if ( nFirst + nCount > (size_t)GetStringData()->nDataLength ) - nCount = GetStringData()->nDataLength - nFirst; - if ( nFirst > (size_t)GetStringData()->nDataLength ) + if ( nFirst > nLen ) + { + // AllocCopy() will return empty string nCount = 0; + } wxString dest; AllocCopy(dest, nCount, nFirst); + return dest; } @@ -562,11 +680,11 @@ wxString wxString::Right(size_t nCount) const // get all characters after the last occurence of ch // (returns the whole string if ch not found) -wxString wxString::Right(char ch) const +wxString wxString::AfterLast(wxChar ch) const { wxString str; int iPos = Find(ch, TRUE); - if ( iPos == NOT_FOUND ) + if ( iPos == wxNOT_FOUND ) str = *this; else str = c_str() + iPos + 1; @@ -587,10 +705,10 @@ wxString wxString::Left(size_t nCount) const // get all characters before the first occurence of ch // (returns the whole string if ch not found) -wxString wxString::Left(char ch) const +wxString wxString::BeforeFirst(wxChar ch) const { wxString str; - for ( const char *pc = m_pchData; *pc != '\0' && *pc != ch; pc++ ) + for ( const wxChar *pc = m_pchData; *pc != _T('\0') && *pc != ch; pc++ ) str += *pc; return str; @@ -598,11 +716,11 @@ wxString wxString::Left(char ch) const /// get all characters before the last occurence of ch /// (returns empty string if ch not found) -wxString wxString::Before(char ch) const +wxString wxString::BeforeLast(wxChar ch) const { wxString str; int iPos = Find(ch, TRUE); - if ( iPos != NOT_FOUND && iPos != 0 ) + if ( iPos != wxNOT_FOUND && iPos != 0 ) str = wxString(c_str(), iPos); return str; @@ -610,28 +728,28 @@ wxString wxString::Before(char ch) const /// get all characters after the first occurence of ch /// (returns empty string if ch not found) -wxString wxString::After(char ch) const +wxString wxString::AfterFirst(wxChar ch) const { wxString str; int iPos = Find(ch); - if ( iPos != NOT_FOUND ) + if ( iPos != wxNOT_FOUND ) str = c_str() + iPos + 1; return str; } // replace first (or all) occurences of some substring with another one -uint wxString::Replace(const char *szOld, const char *szNew, bool bReplaceAll) +size_t wxString::Replace(const wxChar *szOld, const wxChar *szNew, bool bReplaceAll) { - uint uiCount = 0; // count of replacements made + size_t uiCount = 0; // count of replacements made - uint uiOldLen = Strlen(szOld); + size_t uiOldLen = wxStrlen(szOld); wxString strTemp; - const char *pCurrent = m_pchData; - const char *pSubstr; - while ( *pCurrent != '\0' ) { - pSubstr = strstr(pCurrent, szOld); + const wxChar *pCurrent = m_pchData; + const wxChar *pSubstr; + while ( *pCurrent != _T('\0') ) { + pSubstr = wxStrstr(pCurrent, szOld); if ( pSubstr == NULL ) { // strTemp is unused if no replacements were made, so avoid the copy if ( uiCount == 0 ) @@ -664,38 +782,34 @@ uint wxString::Replace(const char *szOld, const char *szNew, bool bReplaceAll) bool wxString::IsAscii() const { - const char *s = (const char*) *this; + const wxChar *s = (const wxChar*) *this; while(*s){ if(!isascii(*s)) return(FALSE); s++; } return(TRUE); } - + bool wxString::IsWord() const { - const char *s = (const char*) *this; + const wxChar *s = (const wxChar*) *this; while(*s){ - if(!isalpha(*s)) return(FALSE); + if(!wxIsalpha(*s)) return(FALSE); s++; } return(TRUE); } - + bool wxString::IsNumber() const { - const char *s = (const char*) *this; + const wxChar *s = (const wxChar*) *this; while(*s){ - if(!isdigit(*s)) return(FALSE); + if(!wxIsdigit(*s)) return(FALSE); s++; } return(TRUE); } -// kludge: we don't have declaraton of wxStringData here, so we add offsets -// manually to get to the "length" field of wxStringData structure -bool wxString::IsEmpty() const { return Len() == 0; } - wxString wxString::Strip(stripType w) const { wxString s = *this; @@ -704,33 +818,6 @@ wxString wxString::Strip(stripType w) const return s; } -/// case-insensitive strcmp() (platform independent) -int Stricmp(const char *psz1, const char *psz2) -{ -#if defined(_MSC_VER) - return _stricmp(psz1, psz2); -#elif defined(__BORLANDC__) - return stricmp(psz1, psz2); -#elif defined(__UNIX__) || defined(__GNUWIN32__) - return strcasecmp(psz1, psz2); -#else - // almost all compilers/libraries provide this function (unfortunately under - // different names), that's why we don't implement our own which will surely - // be more efficient than this code (uncomment to use): - /* - register char c1, c2; - do { - c1 = tolower(*psz1++); - c2 = tolower(*psz2++); - } while ( c1 && (c1 == c2) ); - - return c1 - c2; - */ - - #error "Please define string case-insensitive compare for your OS/compiler" -#endif // OS/compiler -} - // --------------------------------------------------------------------------- // case conversion // --------------------------------------------------------------------------- @@ -739,8 +826,8 @@ wxString& wxString::MakeUpper() { CopyBeforeWrite(); - for ( char *p = m_pchData; *p; p++ ) - *p = (char)toupper(*p); + for ( wxChar *p = m_pchData; *p; p++ ) + *p = (wxChar)wxToupper(*p); return *this; } @@ -748,9 +835,9 @@ wxString& wxString::MakeUpper() wxString& wxString::MakeLower() { CopyBeforeWrite(); - - for ( char *p = m_pchData; *p; p++ ) - *p = (char)tolower(*p); + + for ( wxChar *p = m_pchData; *p; p++ ) + *p = (wxChar)wxTolower(*p); return *this; } @@ -762,37 +849,47 @@ wxString& wxString::MakeLower() // trims spaces (in the sense of isspace) from left or right side wxString& wxString::Trim(bool bFromRight) { - CopyBeforeWrite(); - - if ( bFromRight ) + // first check if we're going to modify the string at all + if ( !IsEmpty() && + ( + (bFromRight && wxIsspace(GetChar(Len() - 1))) || + (!bFromRight && wxIsspace(GetChar(0u))) + ) + ) { - // find last non-space character - char *psz = m_pchData + GetStringData()->nDataLength - 1; - while ( isspace(*psz) && (psz >= m_pchData) ) - psz--; - - // truncate at trailing space start - *++psz = '\0'; - GetStringData()->nDataLength = psz - m_pchData; - } - else - { - // find first non-space character - const char *psz = m_pchData; - while ( isspace(*psz) ) - psz++; - - // fix up data and length - int nDataLength = GetStringData()->nDataLength - (psz - m_pchData); - memmove(m_pchData, psz, (nDataLength + 1)*sizeof(char)); - GetStringData()->nDataLength = nDataLength; + // 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) ) + psz--; + + // truncate at trailing space start + *++psz = _T('\0'); + GetStringData()->nDataLength = psz - m_pchData; + } + else + { + // find first non-space character + const wxChar *psz = m_pchData; + while ( wxIsspace(*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; + } } return *this; } // adds nCount characters chPad to the string from either side -wxString& wxString::Pad(size_t nCount, char chPad, bool bFromRight) +wxString& wxString::Pad(size_t nCount, wxChar chPad, bool bFromRight) { wxString s(chPad, nCount); @@ -810,36 +907,68 @@ wxString& wxString::Pad(size_t nCount, char chPad, bool bFromRight) // truncate the string wxString& wxString::Truncate(size_t uiLen) { - *(m_pchData + uiLen) = '\0'; - GetStringData()->nDataLength = uiLen; + if ( uiLen < Len() ) { + CopyBeforeWrite(); + + *(m_pchData + uiLen) = _T('\0'); + GetStringData()->nDataLength = uiLen; + } + //else: nothing to do, string is already short enough return *this; } // --------------------------------------------------------------------------- -// finding (return NOT_FOUND if not found and index otherwise) +// finding (return wxNOT_FOUND if not found and index otherwise) // --------------------------------------------------------------------------- // find a character -int wxString::Find(char ch, bool bFromEnd) const +int wxString::Find(wxChar ch, bool bFromEnd) const { - const char *psz = bFromEnd ? strrchr(m_pchData, ch) : strchr(m_pchData, ch); + const wxChar *psz = bFromEnd ? wxStrrchr(m_pchData, ch) : wxStrchr(m_pchData, ch); - return (psz == NULL) ? NOT_FOUND : psz - m_pchData; + return (psz == NULL) ? wxNOT_FOUND : psz - (const wxChar*) m_pchData; } // find a sub-string (like strstr) -int wxString::Find(const char *pszSub) const +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) { - const char *psz = strstr(m_pchData, pszSub); + wxString res; + res.Printf(_T("%f"), f); - return (psz == NULL) ? NOT_FOUND : psz - m_pchData; + return (*this) << res; +} + +wxString& wxString::operator<<(double d) +{ + wxString res; + res.Printf(_T("%g"), d); + + return (*this) << res; } // --------------------------------------------------------------------------- // formatted output // --------------------------------------------------------------------------- -int wxString::Printf(const char *pszFormat, ...) +int wxString::Printf(const wxChar *pszFormat, ...) { va_list argptr; va_start(argptr, pszFormat); @@ -851,117 +980,461 @@ int wxString::Printf(const char *pszFormat, ...) return iLen; } -int wxString::PrintfV(const char* pszFormat, va_list argptr) +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 iLen = vsprintf(s_szScratch, pszFormat, argptr); - AllocBeforeWrite(iLen); - strcpy(m_pchData, s_szScratch); +#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]; - return iLen; +#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; + } + + // still not enough, double it again + buffer = (char *)realloc(buffer, size *= 2); + } + + if ( !buffer ) { + // out of memory + return -1; + } + } + + wxString s(buffer); + *this = s; + + if ( buffer != s_szScratch ) + free(buffer); +#endif + + return Len(); } -#if 0 -int wxString::Scanf(const char *pszFormat, ...) const +// ---------------------------------------------------------------------------- +// misc other operations +// ---------------------------------------------------------------------------- +bool wxString::Matches(const wxChar *pszMask) const { - va_list argptr; - va_start(argptr, pszFormat); + // 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; - int iLen = ScanfV(pszFormat, argptr); + pszTxt++; + pszMask++; + break; - va_end(argptr); + 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; + + // -1 to compensate "++" in the loop + pszTxt = pMatch + uiLenMask - 1; + pszMask += uiLenMask - 1; + } + break; - return iLen; + default: + if ( *pszMask != *pszTxt ) + return FALSE; + break; + } + } + + // match only if nothing left + return *pszTxt == _T('\0'); } -int wxString::ScanfV(const char *pszFormat, va_list argptr) const +// Count the number of chars +int wxString::Freq(wxChar ch) const { -#ifdef __WINDOWS__ - wxMessageBox("ScanfV not implemented"); - return 0; -#else - return vsscanf(c_str(), pszFormat, argptr); -#endif + int count = 0; + int len = Len(); + for (int i = 0; i < len; i++) + { + if (GetChar(i) == ch) + count ++; + } + return count; } -#endif + +// 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 // --------------------------------------------------------------------------- -#ifdef STD_STRING_COMPATIBILITY +#ifdef wxSTD_STRING_COMPATIBILITY wxString& wxString::insert(size_t nPos, const wxString& str) { + wxASSERT( str.GetStringData()->IsValid() ); wxASSERT( nPos <= Len() ); - wxString strTmp; - char *pc = strTmp.GetWriteBuf(Len() + str.Len()); - strncpy(pc, c_str(), nPos); - strcpy(pc + nPos, str); - strcpy(pc + nPos + str.Len(), c_str() + nPos); - *this = strTmp; - - return *this; + 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; } size_t wxString::find(const wxString& str, size_t nStart) const { + wxASSERT( str.GetStringData()->IsValid() ); wxASSERT( nStart <= Len() ); - const char *p = strstr(c_str() + nStart, str); - + const wxChar *p = wxStrstr(c_str() + nStart, str); + return p == NULL ? npos : p - c_str(); } // VC++ 1.5 can't cope with the default argument in the header. -#if ! (defined(_MSC_VER) && !defined(__WIN32__)) -size_t wxString::find(const char* sz, size_t nStart, size_t n) 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); } -#endif - -size_t wxString::find(char ch, size_t nStart) const +#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 { wxASSERT( nStart <= Len() ); - const char *p = strchr(c_str() + nStart, ch); - + const wxChar *p = wxStrchr(c_str() + nStart, ch); + return p == NULL ? npos : p - c_str(); } +#endif size_t wxString::rfind(const wxString& str, size_t nStart) const { + wxASSERT( str.GetStringData()->IsValid() ); wxASSERT( nStart <= Len() ); // # could be quicker than that - const char *p = c_str() + (nStart == npos ? Len() : nStart); + const wxChar *p = c_str() + (nStart == npos ? Len() : nStart); while ( p >= c_str() + str.Len() ) { - if ( strncmp(p - str.Len(), str, str.Len()) == 0 ) + if ( wxStrncmp(p - str.Len(), str, str.Len()) == 0 ) return p - str.Len() - c_str(); p--; } - + return npos; } - + // VC++ 1.5 can't cope with the default argument in the header. -#if ! (defined(_MSC_VER) && !defined(__WIN32__)) -size_t wxString::rfind(const char* sz, size_t nStart, size_t n) 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); } -size_t wxString::rfind(char ch, size_t nStart) const +size_t wxString::rfind(wxChar ch, size_t nStart) const { wxASSERT( nStart <= Len() ); - const char *p = strrchr(c_str() + nStart, ch); - + const wxChar *p = wxStrrchr(c_str() + nStart, ch); + return p == NULL ? npos : p - c_str(); } -#endif +#endif // VC++ 1.5 wxString wxString::substr(size_t nStart, size_t nLen) const { @@ -987,33 +1460,33 @@ wxString& wxString::erase(size_t nStart, size_t nLen) return *this; } -wxString& wxString::replace(size_t nStart, size_t nLen, const char *sz) +wxString& wxString::replace(size_t nStart, size_t nLen, const wxChar *sz) { - wxASSERT( nStart + nLen <= Strlen(sz) ); + wxASSERT( nStart + nLen <= wxStrlen(sz) ); wxString strTmp; if ( nStart != 0 ) strTmp.append(c_str(), nStart); strTmp += sz; strTmp.append(c_str() + nStart + nLen); - + *this = strTmp; return *this; } -wxString& wxString::replace(size_t nStart, size_t nLen, size_t nCount, char ch) +wxString& wxString::replace(size_t nStart, size_t nLen, size_t nCount, wxChar ch) { return replace(nStart, nLen, wxString(ch, nCount)); } -wxString& wxString::replace(size_t nStart, size_t nLen, - const wxString& str, size_t nStart2, size_t nLen2) +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)); } -wxString& wxString::replace(size_t nStart, size_t nLen, - const char* sz, size_t nCount) +wxString& wxString::replace(size_t nStart, size_t nLen, + const wxChar* sz, size_t nCount) { return replace(nStart, nLen, wxString(sz, nCount)); } @@ -1037,39 +1510,35 @@ wxArrayString::wxArrayString() { m_nSize = m_nCount = 0; - m_pItems = NULL; + m_pItems = (wxChar **) NULL; } // copy ctor wxArrayString::wxArrayString(const wxArrayString& src) { - m_nSize = src.m_nSize; - m_nCount = src.m_nCount; - - if ( m_nSize != 0 ) - m_pItems = new char *[m_nSize]; - else - m_pItems = NULL; + m_nSize = + m_nCount = 0; + m_pItems = (wxChar **) NULL; - if ( m_nCount != 0 ) - memcpy(m_pItems, src.m_pItems, m_nCount*sizeof(char *)); + *this = src; } -// copy operator +// assignment operator wxArrayString& wxArrayString::operator=(const wxArrayString& src) { - DELETEA(m_pItems); + if ( m_nSize > 0 ) + Clear(); - m_nSize = src.m_nSize; - m_nCount = src.m_nCount; + if ( src.m_nCount > ARRAY_DEFAULT_INITIAL_SIZE ) + Alloc(src.m_nCount); - if ( m_nSize != 0 ) - m_pItems = new char *[m_nSize]; - else - m_pItems = NULL; + // 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(char *)); + memcpy(m_pItems, src.m_pItems, m_nCount*sizeof(wxChar *)); return *this; } @@ -1082,21 +1551,26 @@ void wxArrayString::Grow() if( m_nSize == 0 ) { // was empty, alloc some memory m_nSize = ARRAY_DEFAULT_INITIAL_SIZE; - m_pItems = new char *[m_nSize]; + 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 >> 1; + size_t nIncrement = m_nSize < ARRAY_DEFAULT_INITIAL_SIZE + ? ARRAY_DEFAULT_INITIAL_SIZE : m_nSize >> 1; if ( nIncrement > ARRAY_MAXSIZE_INCREMENT ) nIncrement = ARRAY_MAXSIZE_INCREMENT; m_nSize += nIncrement; - char **pNew = new char *[m_nSize]; + wxChar **pNew = new wxChar *[m_nSize]; // copy data to new location - memcpy(pNew, m_pItems, m_nCount*sizeof(char *)); + memcpy(pNew, m_pItems, m_nCount*sizeof(wxChar *)); // delete old memory (but do not release the strings!) - DELETEA(m_pItems); + wxDELETEA(m_pItems); m_pItems = pNew; } @@ -1123,11 +1597,10 @@ void wxArrayString::Clear() { Free(); - m_nSize = + m_nSize = m_nCount = 0; - DELETEA(m_pItems); - m_pItems = NULL; + wxDELETEA(m_pItems); } // dtor @@ -1135,7 +1608,7 @@ wxArrayString::~wxArrayString() { Free(); - DELETEA(m_pItems); + wxDELETEA(m_pItems); } // pre-allocates memory (frees the previous data!) @@ -1146,23 +1619,35 @@ void wxArrayString::Alloc(size_t nSize) // only if old buffer was not big enough if ( nSize > m_nSize ) { Free(); - DELETEA(m_pItems); - m_pItems = new char *[nSize]; + wxDELETEA(m_pItems); + m_pItems = new wxChar *[nSize]; m_nSize = nSize; } m_nCount = 0; } -// searches the array for an item (forward or backwards) +// minimizes the memory usage by freeing unused memory +void wxArrayString::Shrink() +{ + // only do it if we have some memory to free + if( m_nCount < m_nSize ) { + // allocates exactly as much memory as we need + wxChar **pNew = new wxChar *[m_nCount]; -// Robert Roebling (changed to bool from bool) + // copy data to new location + memcpy(pNew, m_pItems, m_nCount*sizeof(wxChar *)); + delete [] m_pItems; + m_pItems = pNew; + } +} -int wxArrayString::Index(const char *sz, bool bCase, bool bFromEnd) const +// 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 ) { - uint ui = m_nCount; + size_t ui = m_nCount; do { if ( STRING(m_pItems[--ui])->IsSameAs(sz, bCase) ) return ui; @@ -1171,37 +1656,41 @@ int wxArrayString::Index(const char *sz, bool bCase, bool bFromEnd) const } } else { - for( uint ui = 0; ui < m_nCount; ui++ ) { + for( size_t ui = 0; ui < m_nCount; ui++ ) { if( STRING(m_pItems[ui])->IsSameAs(sz, bCase) ) return ui; } } - return NOT_FOUND; + return wxNOT_FOUND; } // add item at the end -void wxArrayString::Add(const wxString& src) +void wxArrayString::Add(const wxString& str) { + wxASSERT( str.GetStringData()->IsValid() ); + Grow(); // the string data must not be deleted! - src.GetStringData()->Lock(); - m_pItems[m_nCount++] = (char *)src.c_str(); + str.GetStringData()->Lock(); + m_pItems[m_nCount++] = (wxChar *)str.c_str(); } // add item at the given position -void wxArrayString::Insert(const wxString& src, size_t nIndex) +void wxArrayString::Insert(const wxString& str, size_t nIndex) { - wxCHECK_RET( nIndex <= m_nCount, "bad index in wxArrayString::Insert" ); + wxASSERT( str.GetStringData()->IsValid() ); + + wxCHECK_RET( nIndex <= m_nCount, _("bad index in wxArrayString::Insert") ); Grow(); - memmove(&m_pItems[nIndex + 1], &m_pItems[nIndex], - (m_nCount - nIndex)*sizeof(char *)); + memmove(&m_pItems[nIndex + 1], &m_pItems[nIndex], + (m_nCount - nIndex)*sizeof(wxChar *)); - src.GetStringData()->Lock(); - m_pItems[nIndex] = (char *)src.c_str(); + str.GetStringData()->Lock(); + m_pItems[nIndex] = (wxChar *)str.c_str(); m_nCount++; } @@ -1209,31 +1698,419 @@ void wxArrayString::Insert(const wxString& src, size_t nIndex) // removes item from array (by index) void wxArrayString::Remove(size_t nIndex) { - wxCHECK_RET( nIndex <= m_nCount, "bad index in wxArrayString::Remove" ); + wxCHECK_RET( nIndex <= m_nCount, _("bad index in wxArrayString::Remove") ); // release our lock Item(nIndex).GetStringData()->Unlock(); - memmove(&m_pItems[nIndex], &m_pItems[nIndex + 1], - (m_nCount - nIndex - 1)*sizeof(char *)); + memmove(&m_pItems[nIndex], &m_pItems[nIndex + 1], + (m_nCount - nIndex - 1)*sizeof(wxChar *)); m_nCount--; } // removes item from array (by value) -void wxArrayString::Remove(const char *sz) +void wxArrayString::Remove(const wxChar *sz) { int iIndex = Index(sz); - wxCHECK_RET( iIndex != NOT_FOUND, - "removing inexistent element in wxArrayString::Remove" ); + wxCHECK_RET( iIndex != wxNOT_FOUND, + _("removing inexistent element in wxArrayString::Remove") ); + + Remove(iIndex); +} + +// ---------------------------------------------------------------------------- +// sorting +// ---------------------------------------------------------------------------- + +// we can only sort one array at a time with the quick-sort based +// implementation +#if wxUSE_THREADS + // need a critical section to protect access to gs_compareFunction and + // gs_sortAscending variables + static wxCriticalSection *gs_critsectStringSort = NULL; + + // call this before the value of the global sort vars is changed/after + // you're finished with them + #define START_SORT() wxASSERT( !gs_critsectStringSort ); \ + gs_critsectStringSort = new wxCriticalSection; \ + gs_critsectStringSort->Enter() + #define END_SORT() gs_critsectStringSort->Leave(); \ + delete gs_critsectStringSort; \ + gs_critsectStringSort = NULL +#else // !threads + #define START_SORT() + #define END_SORT() +#endif // wxUSE_THREADS + +// function to use for string comparaison +static wxArrayString::CompareFunction gs_compareFunction = NULL; + +// if we don't use the compare function, this flag tells us if we sort the +// array in ascending or descending order +static bool gs_sortAscending = TRUE; + +// function which is called by quick sort +static int wxStringCompareFunction(const void *first, const void *second) +{ + wxString *strFirst = (wxString *)first; + wxString *strSecond = (wxString *)second; + + if ( gs_compareFunction ) { + return gs_compareFunction(*strFirst, *strSecond); + } + else { + // maybe we should use wxStrcoll + int result = wxStrcmp(strFirst->c_str(), strSecond->c_str()); - Remove((size_t)iIndex); + return gs_sortAscending ? result : -result; + } } // sort array elements using passed comparaison function +void wxArrayString::Sort(CompareFunction compareFunction) +{ + START_SORT(); + + wxASSERT( !gs_compareFunction ); // must have been reset to NULL + gs_compareFunction = compareFunction; + + DoSort(); + + END_SORT(); +} -void wxArrayString::Sort(bool bCase, bool bReverse) +void wxArrayString::Sort(bool reverseOrder) { - //@@@@ TO DO - //qsort(m_pItems, m_nCount, sizeof(char *), fCmp); + START_SORT(); + + wxASSERT( !gs_compareFunction ); // must have been reset to NULL + gs_sortAscending = !reverseOrder; + + DoSort(); + + 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 *) 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 +{ + 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; +} + +// ---------------------------------------------------------------------------- +// 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) +{ + 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 + } +} + +wxCSConv::~wxCSConv() +{ + free(m_name); +} + +void wxCSConv::LoadNow() +{ +// wxPrintf(_T("Conversion request\n")); + if (m_deferred) { + 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; c<=n; c++) + buf[c] = m_cset->data[psz[c]]; + } else { + // latin-1 (direct) + for (size_t c=0; c<=n; c++) + buf[c] = psz[c]; + } + } + return n; +} + +size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const +{ + ((wxCSConv *)this)->LoadNow(); // discard constness + if (buf) { + if (m_cset) { + for (size_t c=0; c<=n; c++) { + size_t n; + for (n=0; (n<256) && (m_cset->data[n] != psz[c]); n++); + buf[c] = (n>0xff) ? '?' : n; + } + } else { + // latin-1 (direct) + for (size_t c=0; c<=n; c++) + buf[c] = (psz[c]>0xff) ? '?' : psz[c]; + } + } + return n; +} +