X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/dcf924a345ea8ffbc1cf6b40b5f75c6005e504c0..a4a6984d5e2d8ae424cc8f31c223304b260d7670:/src/common/string.cpp?ds=inline diff --git a/src/common/string.cpp b/src/common/string.cpp index 1b42fd7893..9c01501994 100644 --- a/src/common/string.cpp +++ b/src/common/string.cpp @@ -35,11 +35,11 @@ #include "wx/defs.h" #include "wx/string.h" #include "wx/intl.h" -#if wxUSE_THREADS - #include -#endif + #include "wx/thread.h" #endif +#include "wx/regex.h" // for wxString::Matches() + #include #include #include @@ -85,10 +85,16 @@ static const struct { wxStringData data; wxChar dummy; -} g_strEmpty = { {-1, 0, 0}, _T('\0') }; +} g_strEmpty = { {-1, 0, 0}, wxT('\0') }; + +#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 // empty C style string: points to 'string data' byte of g_strEmpty -extern const wxChar WXDLLEXPORT *g_szNul = &g_strEmpty.dummy; +extern const wxChar WXDLLEXPORT *wxEmptyString = &g_strEmpty.dummy; // ---------------------------------------------------------------------------- // conditional compilation @@ -104,28 +110,32 @@ extern const wxChar WXDLLEXPORT *g_szNul = &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. +// +// FIXME currently, this is only for ANSI (!Unicode) strings, so we call this +// function wxVsnprintfA (A for ANSI), should also find one for Unicode +// strings in Unicode build #ifdef __WXMSW__ - #ifdef __VISUALC__ - #define wxVsnprintf _vsnprintf + #if defined(__VISUALC__) || (defined(__MINGW32__) && wxUSE_NORLANDER_HEADERS) + #define wxVsnprintfA _vsnprintf #endif +#elif defined(__WXMAC__) + #define wxVsnprintfA vsnprintf #else // !Windows #ifdef HAVE_VSNPRINTF - #define wxVsnprintf vsnprintf + #define wxVsnprintfA vsnprintf #endif #endif // Windows/!Windows -#ifndef wxVsnprintf +#ifndef wxVsnprintfA // 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) + #define wxVsnprintfA(buf, len, format, arg) vsprintf(buf, format, arg) #if defined(__VISUALC__) #pragma message("Using sprintf() because no snprintf()-like function defined") - #elif defined(__GNUG__) && !defined(__UNIX__) + #elif defined(__GNUG__) #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 @@ -138,14 +148,14 @@ extern const wxChar WXDLLEXPORT *g_szNul = &g_strEmpty.dummy; // global functions // ---------------------------------------------------------------------------- -#ifdef wxSTD_STRING_COMPATIBILITY +#if defined(wxSTD_STRING_COMPATIBILITY) && wxUSE_STD_IOSTREAM // MS Visual C++ version 5.0 provides the new STL headers as well as the old // iostream ones. // // ATTN: you can _not_ use both of these in the same program! -istream& operator>>(istream& is, wxString& WXUNUSED(str)) +wxSTD istream& operator>>(wxSTD istream& is, wxString& WXUNUSED(str)) { #if 0 int w = is.width(0); @@ -176,8 +186,54 @@ istream& operator>>(istream& is, wxString& WXUNUSED(str)) return is; } +wxSTD ostream& operator<<(wxSTD ostream& os, const wxString& str) +{ + os << str.c_str(); + return os; +} + #endif //std::string compatibility +extern int WXDLLEXPORT wxVsnprintf(wxChar *buf, size_t len, + const wxChar *format, va_list argptr) +{ +#if wxUSE_UNICODE + // FIXME should use wvsnprintf() or whatever if it's available + wxString s; + int iLen = s.PrintfV(format, argptr); + if ( iLen != -1 ) + { + wxStrncpy(buf, s.c_str(), len); + buf[len-1] = wxT('\0'); + } + + return iLen; +#else // ANSI + // vsnprintf() will not terminate the string with '\0' if there is not + // enough place, but we want the string to always be NUL terminated + int rc = wxVsnprintfA(buf, len - 1, format, argptr); + if ( rc == -1 ) + { + buf[len] = 0; + } + + return rc; +#endif // Unicode/ANSI +} + +extern int WXDLLEXPORT wxSnprintf(wxChar *buf, size_t len, + const wxChar *format, ...) +{ + va_list argptr; + va_start(argptr, format); + + int iLen = wxVsnprintf(buf, len, format, argptr); + + va_end(argptr); + + return iLen; +} + // ---------------------------------------------------------------------------- // private classes // ---------------------------------------------------------------------------- @@ -237,10 +293,12 @@ void wxString::InitWith(const wxChar *psz, size_t nPos, size_t nLength) { Init(); - wxASSERT( nPos <= wxStrlen(psz) ); + // if the length is not given, assume the string to be NUL terminated + if ( nLength == wxSTRING_MAXLEN ) { + wxASSERT_MSG( nPos <= wxStrlen(psz), _T("index out of bounds") ); - if ( nLength == wxSTRING_MAXLEN ) nLength = wxStrlen(psz + nPos); + } STATISTICS_ADD(InitialLength, nLength); @@ -284,27 +342,27 @@ wxString::wxString(const char *psz, wxMBConv& conv, size_t nLength) } } -#else +#else // ANSI #if wxUSE_WCHAR_T // from wide string -wxString::wxString(const wchar_t *pwz) +wxString::wxString(const wchar_t *pwz, wxMBConv& conv) { // first get necessary size - size_t nLen = pwz ? wxWC2MB((char *) NULL, pwz, 0) : 0; + size_t nLen = pwz ? conv.WC2MB((char *) NULL, pwz, 0) : 0; // empty? if ( (nLen != 0) && (nLen != (size_t)-1) ) { AllocBuffer(nLen); - wxWC2MB(m_pchData, pwz, nLen); + conv.WC2MB(m_pchData, pwz, nLen); } else { Init(); } } -#endif +#endif // wxUSE_WCHAR_T -#endif +#endif // Unicode/ANSI // --------------------------------------------------------------------------- // memory allocation @@ -313,8 +371,13 @@ wxString::wxString(const wchar_t *pwz) // allocates memory needed to store a C string of length nLen void wxString::AllocBuffer(size_t nLen) { - wxASSERT( nLen > 0 ); // - wxASSERT( nLen <= INT_MAX-1 ); // max size (enough room for 1 extra) + // allocating 0 sized buffer doesn't make sense, all empty strings should + // reuse g_strEmpty + wxASSERT( nLen > 0 ); + + // make sure that we don't overflow + wxASSERT( nLen < (INT_MAX / sizeof(wxChar)) - + (sizeof(wxStringData) + EXTRA_ALLOC + 1) ); STATISTICS_ADD(Length, nLen); @@ -327,7 +390,7 @@ void wxString::AllocBuffer(size_t nLen) 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'); } // must be called before changing this string @@ -352,13 +415,35 @@ void wxString::AllocBeforeWrite(size_t nLen) // 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); } 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; + + wxStringData *pDataOld = pData; + pData = (wxStringData*) + realloc(pData, sizeof(wxStringData) + (nLen + 1)*sizeof(wxChar)); + if ( !pData ) { + // out of memory + free(pDataOld); + + // FIXME we're going to crash... + return; + } + + pData->nAllocLength = nLen; + m_pchData = pData->data(); + } + + // now we have enough space, just update the string length pData->nDataLength = nLen; } @@ -379,7 +464,7 @@ void wxString::Alloc(size_t nLen) 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 @@ -390,11 +475,15 @@ void wxString::Alloc(size_t nLen) else { nLen += EXTRA_ALLOC; + wxStringData *pDataOld = pData; wxStringData *p = (wxStringData *) realloc(pData, sizeof(wxStringData) + (nLen + 1)*sizeof(wxChar)); if ( p == NULL ) { - // @@@ what to do on memory error? + // don't leak memory + free(pDataOld); + + // FIXME what to do on memory error? return; } @@ -412,13 +501,16 @@ void wxString::Shrink() { wxStringData *pData = GetStringData(); - // this variable is unused in release build, so avoid the compiler warning by - // just not declaring it + // 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)); + // we rely on a reasonable realloc() implementation here - so far I haven't + // seen any which wouldn't behave like this + wxASSERT( p != NULL ); // can't free memory? wxASSERT( p == pData ); // we're decrementing the size - block shouldn't move! } @@ -441,6 +533,12 @@ void wxString::UngetWriteBuf() GetStringData()->Validate(TRUE); } +void wxString::UngetWriteBuf(size_t nLen) +{ + GetStringData()->nDataLength = nLen; + GetStringData()->Validate(TRUE); +} + // --------------------------------------------------------------------------- // data access // --------------------------------------------------------------------------- @@ -461,7 +559,7 @@ void wxString::AssignCopy(size_t nSrcLen, const wxChar *pszSrcData) AllocBeforeWrite(nSrcLen); memcpy(m_pchData, pszSrcData, nSrcLen*sizeof(wxChar)); GetStringData()->nDataLength = nSrcLen; - m_pchData[nSrcLen] = _T('\0'); + m_pchData[nSrcLen] = wxT('\0'); } } @@ -493,6 +591,7 @@ wxString& wxString::operator=(wxChar ch) return *this; } + // assigns C string wxString& wxString::operator=(const wxChar *psz) { @@ -563,7 +662,7 @@ void wxString::ConcatSelf(int nSrcLen, const wxChar *pszSrcData) // fast concatenation - all is done in our buffer memcpy(m_pchData + nLen, pszSrcData, nSrcLen*sizeof(wxChar)); - m_pchData[nNewLen] = _T('\0'); // put terminating '\0' + m_pchData[nNewLen] = wxT('\0'); // put terminating '\0' GetStringData()->nDataLength = nNewLen; // and fix the length } //else: the string to append was empty @@ -681,6 +780,35 @@ wxString wxString::Mid(size_t nFirst, size_t nCount) const return dest; } +// check that the tring starts with prefix and return the rest of the string +// in the provided pointer if it is not NULL, otherwise return FALSE +bool wxString::StartsWith(const wxChar *prefix, wxString *rest) const +{ + wxASSERT_MSG( prefix, _T("invalid parameter in wxString::StartsWith") ); + + // first check if the beginning of the string matches the prefix: note + // that we don't have to check that we don't run out of this string as + // when we reach the terminating NUL, either prefix string ends too (and + // then it's ok) or we break out of the loop because there is no match + const wxChar *p = c_str(); + while ( *prefix ) + { + if ( *prefix++ != *p++ ) + { + // no match + return FALSE; + } + } + + if ( rest ) + { + // put the rest of the string into provided pointer + *rest = p; + } + + return TRUE; +} + // extract nCount last (rightmost) characters wxString wxString::Right(size_t nCount) const { @@ -722,7 +850,7 @@ wxString wxString::Left(size_t nCount) const wxString wxString::BeforeFirst(wxChar ch) const { wxString str; - for ( const wxChar *pc = m_pchData; *pc != _T('\0') && *pc != ch; pc++ ) + for ( const wxChar *pc = m_pchData; *pc != wxT('\0') && *pc != ch; pc++ ) str += *pc; return str; @@ -762,7 +890,7 @@ size_t wxString::Replace(const wxChar *szOld, const wxChar *szNew, bool bReplace wxString strTemp; const wxChar *pCurrent = m_pchData; const wxChar *pSubstr; - while ( *pCurrent != _T('\0') ) { + while ( *pCurrent != wxT('\0') ) { pSubstr = wxStrstr(pCurrent, szOld); if ( pSubstr == NULL ) { // strTemp is unused if no replacements were made, so avoid the copy @@ -817,6 +945,8 @@ 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++; @@ -860,14 +990,22 @@ 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))) ) ) { @@ -878,18 +1016,18 @@ wxString& wxString::Trim(bool bFromRight) { // find last non-space character wxChar *psz = m_pchData + GetStringData()->nDataLength - 1; - while ( wxIsspace(*psz) && (psz >= m_pchData) ) + while ( wxSafeIsspace(*psz) && (psz >= m_pchData) ) psz--; // truncate at trailing space start - *++psz = _T('\0'); + *++psz = wxT('\0'); GetStringData()->nDataLength = psz - m_pchData; } else { // find first non-space character const wxChar *psz = m_pchData; - while ( wxIsspace(*psz) ) + while ( wxSafeIsspace(*psz) ) psz++; // fix up data and length @@ -924,7 +1062,7 @@ wxString& wxString::Truncate(size_t uiLen) if ( uiLen < Len() ) { CopyBeforeWrite(); - *(m_pchData + uiLen) = _T('\0'); + *(m_pchData + uiLen) = wxT('\0'); GetStringData()->nDataLength = uiLen; } //else: nothing to do, string is already short enough @@ -952,36 +1090,75 @@ int wxString::Find(const wxChar *pszSub) const return (psz == NULL) ? wxNOT_FOUND : psz - (const wxChar*) m_pchData; } -// --------------------------------------------------------------------------- -// stream-like operators -// --------------------------------------------------------------------------- -wxString& wxString::operator<<(int i) +// ---------------------------------------------------------------------------- +// conversion to numbers +// ---------------------------------------------------------------------------- + +bool wxString::ToLong(long *val) const { - wxString res; - res.Printf(_T("%d"), i); + wxCHECK_MSG( val, FALSE, _T("NULL pointer in wxString::ToLong") ); + + const wxChar *start = c_str(); + wxChar *end; + *val = wxStrtol(start, &end, 10); - return (*this) << res; + // return TRUE only if scan was stopped by the terminating NUL and if the + // string was not empty to start with + return !*end && (end != start); } -wxString& wxString::operator<<(float f) +bool wxString::ToULong(unsigned long *val) const { - wxString res; - res.Printf(_T("%f"), f); + wxCHECK_MSG( val, FALSE, _T("NULL pointer in wxString::ToULong") ); + + const wxChar *start = c_str(); + wxChar *end; + *val = wxStrtoul(start, &end, 10); - return (*this) << res; + // return TRUE only if scan was stopped by the terminating NUL and if the + // string was not empty to start with + return !*end && (end != start); } -wxString& wxString::operator<<(double d) +bool wxString::ToDouble(double *val) const { - wxString res; - res.Printf(_T("%g"), d); + wxCHECK_MSG( val, FALSE, _T("NULL pointer in wxString::ToDouble") ); - return (*this) << res; + const wxChar *start = c_str(); + wxChar *end; + *val = wxStrtod(start, &end); + + // return TRUE only if scan was stopped by the terminating NUL and if the + // string was not empty to start with + return !*end && (end != start); } // --------------------------------------------------------------------------- // formatted output // --------------------------------------------------------------------------- + +/* static */ +wxString wxString::Format(const wxChar *pszFormat, ...) +{ + va_list argptr; + va_start(argptr, pszFormat); + + wxString s; + s.PrintfV(pszFormat, argptr); + + va_end(argptr); + + return s; +} + +/* static */ +wxString wxString::FormatV(const wxChar *pszFormat, va_list argptr) +{ + wxString s; + s.PrintfV(pszFormat, argptr); + return s; +} + int wxString::Printf(const wxChar *pszFormat, ...) { va_list argptr; @@ -996,282 +1173,279 @@ int wxString::Printf(const wxChar *pszFormat, ...) int wxString::PrintfV(const wxChar* pszFormat, va_list argptr) { - // static buffer to avoid dynamic memory allocation each time - char s_szScratch[1024]; // using static buffer causes internal compiler err -#if 0 -#if wxUSE_THREADS - // protect the static buffer - static wxCriticalSection critsect; - wxCriticalSectionLocker lock(critsect); -#endif -#endif - #if wxUSE_EXPERIMENTAL_PRINTF -// the new implementation + // the new implementation + + // buffer to avoid dynamic memory allocation each time for small strings + char szScratch[1024]; Reinit(); for (size_t n = 0; pszFormat[n]; n++) - if (pszFormat[n] == _T('%')) { + if (pszFormat[n] == wxT('%')) { static char s_szFlags[256] = "%"; size_t flagofs = 1; bool adj_left = FALSE, in_prec = FALSE, - prec_dot = FALSE, done = 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) { + switch (pszFormat[++n]) { + case wxT('\0'): + done = TRUE; + break; + case wxT('%'): + *this += wxT('%'); + done = TRUE; + break; + case wxT('#'): + case wxT('0'): + case wxT(' '): + case wxT('+'): + case wxT('\''): + CHECK_PREC + s_szFlags[flagofs++] = pszFormat[n]; + break; + case wxT('-'): + CHECK_PREC + adj_left = TRUE; + s_szFlags[flagofs++] = pszFormat[n]; + break; + case wxT('.'): + 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 wxT('h'): + ilen = -1; + CHECK_PREC + s_szFlags[flagofs++] = pszFormat[n]; + break; + case wxT('l'): + ilen = 1; + CHECK_PREC + s_szFlags[flagofs++] = pszFormat[n]; + break; + case wxT('q'): + case wxT('L'): + ilen = 2; + CHECK_PREC + s_szFlags[flagofs++] = pszFormat[n]; + break; + case wxT('Z'): + ilen = 3; + CHECK_PREC + s_szFlags[flagofs++] = pszFormat[n]; + break; + case wxT('*'): + { + 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 wxT('1'): case wxT('2'): case wxT('3'): + case wxT('4'): case wxT('5'): case wxT('6'): + case wxT('7'): case wxT('8'): case wxT('9'): + { + int len = 0; + CHECK_PREC + while ((pszFormat[n]>=wxT('0')) && (pszFormat[n]<=wxT('9'))) { + s_szFlags[flagofs++] = pszFormat[n]; + len = len*10 + (pszFormat[n] - wxT('0')); + n++; + } + if (in_prec) max_width = len; + else min_width = len; + n--; // the main loop pre-increments n again + } + break; + case wxT('d'): + case wxT('i'): + case wxT('o'): + case wxT('u'): + case wxT('x'): + case wxT('X'): + CHECK_PREC + s_szFlags[flagofs++] = pszFormat[n]; + s_szFlags[flagofs] = '\0'; + if (ilen == 0 ) { + int val = va_arg(argptr, int); + ::sprintf(szScratch, s_szFlags, val); + } + else if (ilen == -1) { + short int val = va_arg(argptr, short int); + ::sprintf(szScratch, s_szFlags, val); + } + else if (ilen == 1) { + long int val = va_arg(argptr, long int); + ::sprintf(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); + long long int val = va_arg(argptr, long long int); + ::sprintf(szScratch, s_szFlags, val); #else - long int val = va_arg(argptr, long int); - ::sprintf(s_szScratch, s_szFlags, val); + long int val = va_arg(argptr, long int); + ::sprintf(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 *); + } + else if (ilen == 3) { + size_t val = va_arg(argptr, size_t); + ::sprintf(szScratch, s_szFlags, val); + } + *this += wxString(szScratch); + done = TRUE; + break; + case wxT('e'): + case wxT('E'): + case wxT('f'): + case wxT('g'): + case wxT('G'): + CHECK_PREC + s_szFlags[flagofs++] = pszFormat[n]; + s_szFlags[flagofs] = '\0'; + if (ilen == 2) { + long double val = va_arg(argptr, long double); + ::sprintf(szScratch, s_szFlags, val); + } else { + double val = va_arg(argptr, double); + ::sprintf(szScratch, s_szFlags, val); + } + *this += wxString(szScratch); + done = TRUE; + break; + case wxT('p'): + { + void *val = va_arg(argptr, void *); + CHECK_PREC + s_szFlags[flagofs++] = pszFormat[n]; + s_szFlags[flagofs] = '\0'; + ::sprintf(szScratch, s_szFlags, val); + *this += wxString(szScratch); + done = TRUE; + } + break; + case wxT('c'): + { + wxChar val = va_arg(argptr, int); + // we don't need to honor padding here, do we? + *this += val; + done = TRUE; + } + break; + case wxT('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); + // 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; - } + if (s.Len() < min_width) + s.Pad(min_width - s.Len(), wxT(' '), adj_left); + *this += s; + } else { + wxChar *val = va_arg(argptr, wxChar *); + 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 += wxT('%'); // 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; + // buffer to avoid dynamic memory allocation each time for small strings + char szScratch[1024]; + + // NB: wxVsnprintf() may return either less than the buffer size or -1 if + // there is not enough place depending on implementation + int iLen = wxVsnprintfA(szScratch, WXSIZEOF(szScratch), (char *)pszFormat, argptr); + if ( iLen != -1 ) { + // the whole string is in szScratch + *this = 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 ) { + bool outOfMemory = FALSE; + int size = 2*WXSIZEOF(szScratch); + while ( !outOfMemory ) { + char *buf = GetWriteBuf(size); + if ( buf ) + iLen = wxVsnprintfA(buf, size, pszFormat, argptr); + else + outOfMemory = TRUE; + + UngetWriteBuf(); + + if ( iLen != -1 ) { // ok, there was enough space break; } // still not enough, double it again - buffer = (char *)realloc(buffer, size *= 2); + size *= 2; } - if ( !buffer ) { + if ( outOfMemory ) { // out of memory return -1; } } - - wxString s(buffer); - *this = s; - - if ( buffer != s_szScratch ) - free(buffer); -#endif +#endif // wxUSE_EXPERIMENTAL_PRINTF/!wxUSE_EXPERIMENTAL_PRINTF return Len(); } @@ -1279,33 +1453,93 @@ int wxString::PrintfV(const wxChar* pszFormat, va_list argptr) // ---------------------------------------------------------------------------- // misc other operations // ---------------------------------------------------------------------------- + +// returns TRUE if the string matches the pattern which may contain '*' and +// '?' metacharacters (as usual, '?' matches any character and '*' any number +// of them) bool wxString::Matches(const wxChar *pszMask) const { - // check char by char - const wxChar *pszTxt; - for ( pszTxt = c_str(); *pszMask != _T('\0'); pszMask++, pszTxt++ ) { +#if wxUSE_REGEX + // first translate the shell-like mask into a regex + wxString pattern; + pattern.reserve(wxStrlen(pszMask)); + + pattern += _T('^'); + while ( *pszMask ) + { + switch ( *pszMask ) + { + case _T('?'): + pattern += _T('.'); + break; + + case _T('*'): + pattern += _T(".*"); + break; + + case _T('^'): + case _T('.'): + case _T('$'): + case _T('('): + case _T(')'): + case _T('|'): + case _T('+'): + case _T('\\'): + // these characters are special in a RE, quote them + // (however note that we don't quote '[' and ']' to allow + // using them for Unix shell like matching) + pattern += _T('\\'); + // fall through + + default: + pattern += *pszMask; + } + + pszMask++; + } + pattern += _T('$'); + + // and now use it + return wxRegEx(pattern, wxRE_NOSUB | wxRE_EXTENDED).Matches(c_str()); +#else // !wxUSE_REGEX + // TODO: this is, of course, awfully inefficient... + + // the char currently being checked + const wxChar *pszTxt = c_str(); + + // the last location where '*' matched + const wxChar *pszLastStarInText = NULL; + const wxChar *pszLastStarInMask = NULL; + +match: + for ( ; *pszMask != wxT('\0'); pszMask++, pszTxt++ ) { switch ( *pszMask ) { - case _T('?'): - if ( *pszTxt == _T('\0') ) + case wxT('?'): + if ( *pszTxt == wxT('\0') ) return FALSE; - pszTxt++; - pszMask++; + // pszTxt and pszMask will be incremented in the loop statement + break; - case _T('*'): + case wxT('*'): { + // remember where we started to be able to backtrack later + pszLastStarInText = pszTxt; + pszLastStarInMask = pszMask; + // ignore special chars immediately following this one - while ( *pszMask == _T('*') || *pszMask == _T('?') ) + // (should this be an error?) + while ( *pszMask == wxT('*') || *pszMask == wxT('?') ) pszMask++; // if there is nothing more, match - if ( *pszMask == _T('\0') ) + if ( *pszMask == wxT('\0') ) return TRUE; // are there any other metacharacters in the mask? size_t uiLenMask; - const wxChar *pEndMask = wxStrpbrk(pszMask, _T("*?")); + const wxChar *pEndMask = wxStrpbrk(pszMask, wxT("*?")); if ( pEndMask != NULL ) { // we have to match the string between two metachars @@ -1335,7 +1569,23 @@ bool wxString::Matches(const wxChar *pszMask) const } // match only if nothing left - return *pszTxt == _T('\0'); + if ( *pszTxt == wxT('\0') ) + return TRUE; + + // if we failed to match, backtrack if we can + if ( pszLastStarInText ) { + pszTxt = pszLastStarInText + 1; + pszMask = pszLastStarInMask; + + pszLastStarInText = NULL; + + // don't bother resetting pszLastStarInMask, it's unnecessary + + goto match; + } + + return FALSE; +#endif // wxUSE_REGEX/!wxUSE_REGEX } // Count the number of chars @@ -1370,8 +1620,34 @@ int wxString::sprintf(const wxChar *pszFormat, ...) // --------------------------------------------------------------------------- // standard C++ library string functions // --------------------------------------------------------------------------- + #ifdef wxSTD_STRING_COMPATIBILITY +void wxString::resize(size_t nSize, wxChar ch) +{ + size_t len = length(); + + if ( nSize < len ) + { + Truncate(nSize); + } + else if ( nSize > len ) + { + *this += wxString(ch, len - nSize); + } + //else: we have exactly the specified length, nothing to do +} + +void wxString::swap(wxString& str) +{ + // this is slightly less efficient than fiddling with m_pchData directly, + // but it is still quite efficient as we don't copy the string here because + // ref count always stays positive + wxString tmp = str; + str = *this; + *this = str; +} + wxString& wxString::insert(size_t nPos, const wxString& str) { wxASSERT( str.GetStringData()->IsValid() ); @@ -1404,7 +1680,7 @@ size_t wxString::find(const wxString& str, size_t nStart) const #if !defined(__VISUALC__) || defined(__WIN32__) size_t wxString::find(const wxChar* sz, size_t nStart, size_t n) const { - return find(wxString(sz, n == npos ? 0 : n), nStart); + return find(wxString(sz, n), nStart); } #endif // VC++ 1.5 @@ -1469,7 +1745,7 @@ 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; + return firstOf - c_str(); else return npos; } @@ -1565,17 +1841,6 @@ size_t wxString::find_last_not_of(wxChar ch, size_t nStart) const return npos; } -wxString wxString::substr(size_t nStart, size_t nLen) const -{ - // npos means 'take all' - if ( nLen == npos ) - nLen = 0; - - wxASSERT( nStart + nLen <= Len() ); - - return wxString(c_str() + nStart, nLen == npos ? 0 : nLen); -} - wxString& wxString::erase(size_t nStart, size_t nLen) { wxString strTmp(c_str(), nStart); @@ -1591,13 +1856,15 @@ wxString& wxString::erase(size_t nStart, size_t nLen) wxString& wxString::replace(size_t nStart, size_t nLen, const wxChar *sz) { - wxASSERT( nStart + nLen <= wxStrlen(sz) ); + wxASSERT_MSG( nStart + nLen <= Len(), + _T("index out of bounds in wxString::replace") ); wxString strTmp; + strTmp.Alloc(Len()); // micro optimisation to avoid multiple mem allocs + if ( nStart != 0 ) strTmp.append(c_str(), nStart); - strTmp += sz; - strTmp.append(c_str() + nStart + nLen); + strTmp << sz << c_str() + nStart + nLen; *this = strTmp; return *this; @@ -1635,11 +1902,12 @@ wxString& wxString::replace(size_t nStart, size_t nLen, #define STRING(p) ((wxString *)(&(p))) // ctor -wxArrayString::wxArrayString() +wxArrayString::wxArrayString(bool autoSort) { m_nSize = m_nCount = 0; m_pItems = (wxChar **) NULL; + m_autoSort = autoSort; } // copy ctor @@ -1648,6 +1916,7 @@ wxArrayString::wxArrayString(const wxArrayString& src) m_nSize = m_nCount = 0; m_pItems = (wxChar **) NULL; + m_autoSort = src.m_autoSort; *this = src; } @@ -1658,26 +1927,34 @@ 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() { // only do it if no more place - if( m_nCount == m_nSize ) { - if( m_nSize == 0 ) { + if ( m_nCount == m_nSize ) { + // 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; m_pItems = new wxChar *[m_nSize]; @@ -1685,8 +1962,6 @@ void wxArrayString::Grow() 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 ? ARRAY_DEFAULT_INITIAL_SIZE : m_nSize >> 1; @@ -1762,7 +2037,7 @@ 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 *)); @@ -1774,20 +2049,46 @@ void wxArrayString::Shrink() // 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; + } } } @@ -1795,15 +2096,47 @@ 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) +{ + 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; + } + } - Grow(); + wxASSERT_MSG( lo == hi, wxT("binary search broken") ); - // the string data must not be deleted! - str.GetStringData()->Lock(); - m_pItems[m_nCount++] = (wxChar *)str.c_str(); + Insert(str, lo); + + return (size_t)lo; + } + else { + wxASSERT( str.GetStringData()->IsValid() ); + + Grow(); + + // the string data must not be deleted! + str.GetStringData()->Lock(); + + // just append + m_pItems[m_nCount] = (wxChar *)str.c_str(); // const_cast + + return m_nCount++; + } } // add item at the given position @@ -1811,7 +2144,7 @@ void wxArrayString::Insert(const wxString& str, size_t nIndex) { 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") ); Grow(); @@ -1827,7 +2160,7 @@ void wxArrayString::Insert(const wxString& str, 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, wxT("bad index in wxArrayString::Remove") ); // release our lock Item(nIndex).GetStringData()->Unlock(); @@ -1843,7 +2176,7 @@ void wxArrayString::Remove(const wxChar *sz) int iIndex = Index(sz); wxCHECK_RET( iIndex != wxNOT_FOUND, - _("removing inexistent element in wxArrayString::Remove") ); + wxT("removing inexistent element in wxArrayString::Remove") ); Remove(iIndex); } @@ -1880,7 +2213,7 @@ 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) +static int LINKAGEMODE wxStringCompareFunction(const void *first, const void *second) { wxString *strFirst = (wxString *)first; wxString *strSecond = (wxString *)second; @@ -1906,6 +2239,9 @@ void wxArrayString::Sort(CompareFunction compareFunction) DoSort(); + // reset it to NULL so that Sort(bool) will work the next time + gs_compareFunction = NULL; + END_SORT(); } @@ -1923,428 +2259,24 @@ void wxArrayString::Sort(bool reverseOrder) void wxArrayString::DoSort() { + wxCHECK_RET( !m_autoSort, wxT("can't use this method with sorted arrays") ); + // just sort the pointers using qsort() - of course it only works because // wxString() *is* a pointer to its data qsort(m_pItems, m_nCount, sizeof(wxChar *), wxStringCompareFunction); } -// ============================================================================ -// MBConv -// ============================================================================ - -WXDLLEXPORT_DATA(wxMBConv *) 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 +bool wxArrayString::operator==(const wxArrayString& a) 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; - } -} - -size_t wxMBConvGdk::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 -// ---------------------------------------------------------------------------- + if ( m_nCount != a.m_nCount ) + return FALSE; -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 -{ - return 0; -} - -size_t wxMBConvUTF7::WC2MB(char * WXUNUSED(buf), - const wchar_t * WXUNUSED(psz), - size_t WXUNUSED(n)) const -{ - return 0; -} - -// ---------------------------------------------------------------------------- -// 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) && (wxString(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(); -} - -static wxCharacterSet *wxFindCharacterSet(const wxChar *charset) -{ - if (!charset) return (wxCharacterSet *)NULL; - wxLoadCharacterSets(); - for (size_t n=0; n4) { - if (wxString(charset,4) == _T("8859")) { - codeset << _T("8859-"); - if (*charset == _T('-')) charset++; - } - } - } - codeset << charset; - codeset.MakeUpper(); - m_name = wxStrdup(codeset.c_str()); - m_deferred = TRUE; -#endif - } -} - -void wxCSConv::LoadNow() -{ -// wxPrintf(_T("Conversion request\n")); - if (m_deferred) { - if (!m_name) { -#ifdef __UNIX__ - wxChar *lang = wxGetenv(_T("LANG")); - wxChar *dot = lang ? wxStrchr(lang, _T('.')) : (wxChar *)NULL; - if (dot) SetName(dot+1); -#endif - } - m_cset = wxFindCharacterSet(m_name); - m_deferred = FALSE; - } -} - -size_t wxCSConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const -{ - ((wxCSConv *)this)->LoadNow(); // discard constness - if (buf) { - if (m_cset) { - for (size_t c=0; cdata[(unsigned char)(psz[c])]; - } else { - // latin-1 (direct) - for (size_t c=0; cLoadNow(); // discard constness - if (buf) { - if (m_cset) { - for (size_t c=0; cdata[n] != psz[c]); n++); - buf[c] = (n>0xff) ? '?' : n; - } - } else { - // latin-1 (direct) - for (size_t c=0; c0xff) ? '?' : psz[c]; - } - return n; - } - return wcslen(psz); -} - -#endif//wxUSE_WCHAR_T - -#if wxUSE_WCHAR_T -const wxWCharBuffer wxMBConv::cMB2WC(const char *psz) const + for ( size_t n = 0; n < m_nCount; n++ ) { - 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); + if ( Item(n) != a[n] ) + return FALSE; } -#endif//wxUSE_WCHAR_T + return TRUE; +}