X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/7aa88ac4af764fa92c781056bc23103cadea21a4..84c5b38d579e140df28cfbb649587e8862148c89:/src/common/string.cpp diff --git a/src/common/string.cpp b/src/common/string.cpp index e9e906d49b..beb618873d 100644 --- a/src/common/string.cpp +++ b/src/common/string.cpp @@ -10,7 +10,7 @@ ///////////////////////////////////////////////////////////////////////////// #ifdef __GNUG__ -#pragma implementation "string.h" + #pragma implementation "string.h" #endif /* @@ -28,20 +28,27 @@ #include "wx/wxprec.h" #ifdef __BORLANDC__ -#pragma hdrstop + #pragma hdrstop #endif #ifndef WX_PRECOMP #include "wx/defs.h" #include "wx/string.h" #include "wx/intl.h" + #include "wx/thread.h" #endif +#include "wx/regex.h" // for wxString::Matches() + #include #include #include -#ifdef wxUSE_WCSRTOMBS +#ifdef __SALFORDC__ + #include +#endif + +#if wxUSE_WCSRTOMBS #include // for wcsrtombs(), see comments where it's used #endif // GNU @@ -49,6 +56,11 @@ IMPLEMENT_DYNAMIC_CLASS(wxString, wxObject) #endif //WXSTRING_IS_WXOBJECT +#if wxUSE_UNICODE +#undef wxUSE_EXPERIMENTAL_PRINTF +#define wxUSE_EXPERIMENTAL_PRINTF 1 +#endif + // allocating extra space for each string consumes more memory but speeds up // the concatenation operations (nLen is the current string's length) // NB: EXTRA_ALLOC must be >= 0! @@ -58,9 +70,9 @@ // 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 data @@ -72,62 +84,88 @@ static const struct { wxStringData data; - char dummy; -} g_strEmpty = { {-1, 0, 0}, '\0' }; + wxChar dummy; +} 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 char *g_szNul = &g_strEmpty.dummy; +extern const wxChar WXDLLEXPORT *wxEmptyString = &g_strEmpty.dummy; // ---------------------------------------------------------------------------- -// global functions +// conditional compilation // ---------------------------------------------------------------------------- -#ifdef STD_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 wxUSE_IOSTREAMH -#include -#define NAMESPACE -#else -#include -# ifdef _MSC_VER - using namespace std; -# endif -// for msvc (bcc50+ also) you don't need these NAMESPACE defines, -// using namespace std; takes care of that. -#define NAMESPACE std:: +#if !defined(__WXSW__) && wxUSE_UNICODE + #ifdef wxUSE_EXPERIMENTAL_PRINTF + #undef wxUSE_EXPERIMENTAL_PRINTF + #endif + #define wxUSE_EXPERIMENTAL_PRINTF 1 #endif +// 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 _MSC_VER - #define wxVsprintf _vsnprintf + #if defined(__VISUALC__) || (defined(__MINGW32__) && wxUSE_NORLANDER_HEADERS) + #define wxVsnprintfA _vsnprintf #endif -#else - #if defined ( HAVE_VSNPRINTF ) - #define wxVsprintf vsnprintf +#elif defined(__WXMAC__) + #define wxVsnprintfA vsnprintf +#else // !Windows + #ifdef HAVE_VSNPRINTF + #define wxVsnprintfA vsnprintf #endif +#endif // Windows/!Windows + +#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 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__) + #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 -#ifndef wxVsprintf - // vsprintf() is ANSI so we can always use it, but it's unsafe! - #define wxVsprintf(buffer,len,format,argptr) vsprintf(buffer,format, argptr) - #pragma message("Using sprintf() because no snprintf()-like function defined") -#endif +// ---------------------------------------------------------------------------- +// global functions +// ---------------------------------------------------------------------------- -NAMESPACE istream& operator>>(NAMESPACE istream& is, wxString& WXUNUSED(str)) +#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! + +wxSTD istream& operator>>(wxSTD 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) ) { @@ -143,13 +181,59 @@ 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; } +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 // ---------------------------------------------------------------------------- @@ -188,80 +272,97 @@ NAMESPACE istream& operator>>(NAMESPACE istream& is, wxString& WXUNUSED(str)) // --------------------------------------------------------------------------- // 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)); } } -// 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 +#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 -// from wide string -wxString::wxString(const wchar_t *pwz) +#if wxUSE_UNICODE + +// from multibyte string +wxString::wxString(const char *psz, wxMBConv& conv, size_t nLength) { // first get necessary size + size_t nLen = psz ? conv.MB2WC((wchar_t *) NULL, psz, 0) : 0; - // NB: GNU libc5 wcstombs() is completely broken, don't use it (it doesn't - // honor the 3rd parameter, thus it will happily crash here). -#ifdef wxUSE_WCSRTOMBS - // don't know if it's really needed (or if we can pass NULL), but better safe - // than quick - mbstate_t mbstate; - size_t nLen = wcsrtombs((char *) NULL, &pwz, 0, &mbstate); -#else // !GNU libc - size_t nLen = wcstombs((char *) NULL, pwz, 0); -#endif // GNU + // nLength is number of *Unicode* characters here! + if ((nLen != (size_t)-1) && (nLen > nLength)) + nLen = nLength; + + // empty? + if ( (nLen != 0) && (nLen != (size_t)-1) ) { + AllocBuffer(nLen); + conv.MB2WC(m_pchData, psz, nLen); + } + else { + Init(); + } +} + +#else // ANSI + +#if wxUSE_WCHAR_T +// from wide string +wxString::wxString(const wchar_t *pwz, wxMBConv& conv) +{ + // first get necessary size + size_t nLen = pwz ? conv.WC2MB((char *) NULL, pwz, 0) : 0; // empty? - if ( nLen != 0 ) { + if ( (nLen != 0) && (nLen != (size_t)-1) ) { AllocBuffer(nLen); - wcstombs(m_pchData, pwz, nLen); + conv.WC2MB(m_pchData, pwz, nLen); } else { Init(); } } +#endif // wxUSE_WCHAR_T + +#endif // Unicode/ANSI // --------------------------------------------------------------------------- // memory allocation @@ -270,8 +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); @@ -279,12 +385,12 @@ void wxString::AllocBuffer(size_t nLen) // 1) one extra character for '\0' termination // 2) sizeof(wxStringData) for housekeeping info wxStringData* pData = (wxStringData*) - malloc(sizeof(wxStringData) + (nLen + EXTRA_ALLOC + 1)*sizeof(char)); + malloc(sizeof(wxStringData) + (nLen + EXTRA_ALLOC + 1)*sizeof(wxChar)); pData->nRefs = 1; pData->nDataLength = nLen; pData->nAllocLength = nLen + EXTRA_ALLOC; m_pchData = pData->data(); // data starts after wxStringData - m_pchData[nLen] = '\0'; + m_pchData[nLen] = wxT('\0'); } // must be called before changing this string @@ -296,7 +402,7 @@ void wxString::CopyBeforeWrite() pData->Unlock(); // memory not freed because shared size_t nLen = pData->nDataLength; AllocBuffer(nLen); - memcpy(m_pchData, pData->data(), nLen*sizeof(char)); + memcpy(m_pchData, pData->data(), nLen*sizeof(wxChar)); } wxASSERT( !GetStringData()->IsShared() ); // we must be the only owner @@ -309,11 +415,37 @@ 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 { + 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; + } wxASSERT( !GetStringData()->IsShared() ); // we must be the only owner } @@ -327,27 +459,31 @@ void wxString::Alloc(size_t nLen) nLen += EXTRA_ALLOC; wxStringData* pData = (wxStringData*) - malloc(sizeof(wxStringData) + (nLen + 1)*sizeof(char)); + 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] = '\0'; + m_pchData[0u] = wxT('\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(char)); + memcpy(m_pchData, pData->data(), nOldLen*sizeof(wxChar)); } else { nLen += EXTRA_ALLOC; + wxStringData *pDataOld = pData; wxStringData *p = (wxStringData *) - realloc(pData, sizeof(wxStringData) + (nLen + 1)*sizeof(char)); + 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; } @@ -365,19 +501,22 @@ 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(char)); + 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! } // 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); @@ -390,7 +529,13 @@ char *wxString::GetWriteBuf(size_t nLen) // put string back in a reasonable state after GetWriteBuf void wxString::UngetWriteBuf() { - GetStringData()->nDataLength = strlen(m_pchData); + GetStringData()->nDataLength = wxStrlen(m_pchData); + GetStringData()->Validate(TRUE); +} + +void wxString::UngetWriteBuf(size_t nLen) +{ + GetStringData()->nDataLength = nLen; GetStringData()->Validate(TRUE); } @@ -405,16 +550,16 @@ void wxString::UngetWriteBuf() // --------------------------------------------------------------------------- // helper function: does real copy -void wxString::AssignCopy(size_t nSrcLen, const char *pszSrcData) +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] = wxT('\0'); } } @@ -440,19 +585,21 @@ 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) { @@ -460,19 +607,23 @@ wxString& wxString::operator=(const unsigned char* psz) return *this; } +#if wxUSE_WCHAR_T wxString& wxString::operator=(const wchar_t *pwz) { wxString str(pwz); *this = str; return *this; } +#endif + +#endif // --------------------------------------------------------------------------- // string concatenation // --------------------------------------------------------------------------- // 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); @@ -489,7 +640,7 @@ void wxString::ConcatSelf(int nSrcLen, const char *pszSrcData) // we have to allocate another buffer wxStringData* pOldData = GetStringData(); AllocBuffer(nNewLen); - memcpy(m_pchData, pOldData->data(), nLen*sizeof(char)); + memcpy(m_pchData, pOldData->data(), nLen*sizeof(wxChar)); pOldData->Unlock(); } else if ( nNewLen > pData->nAllocLength ) { @@ -508,9 +659,9 @@ void wxString::ConcatSelf(int nSrcLen, const char *pszSrcData) wxASSERT( nNewLen <= GetStringData()->nAllocLength ); // fast concatenation - all is done in our buffer - memcpy(m_pchData + nLen, pszSrcData, nSrcLen*sizeof(char)); + memcpy(m_pchData + nLen, pszSrcData, nSrcLen*sizeof(wxChar)); - m_pchData[nNewLen] = '\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 @@ -534,7 +685,7 @@ wxString operator+(const wxString& string1, const wxString& string2) return s; } -wxString operator+(const wxString& string, char ch) +wxString operator+(const wxString& string, wxChar ch) { wxASSERT( string.GetStringData()->IsValid() ); @@ -544,7 +695,7 @@ wxString operator+(const wxString& string, char ch) return s; } -wxString operator+(char ch, const wxString& string) +wxString operator+(wxChar ch, const wxString& string) { wxASSERT( string.GetStringData()->IsValid() ); @@ -554,24 +705,24 @@ wxString operator+(char ch, const wxString& 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.Alloc(Strlen(psz) + string.Len()); + 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.Alloc(Strlen(psz) + string.Len()); + s.Alloc(wxStrlen(psz) + string.Len()); s = psz; s += string; @@ -594,7 +745,7 @@ void wxString::AllocCopy(wxString& dest, int nCopyLen, int nCopyIndex) const } else { dest.AllocBuffer(nCopyLen); - memcpy(dest.m_pchData, m_pchData + nCopyIndex, nCopyLen*sizeof(char)); + memcpy(dest.m_pchData, m_pchData + nCopyIndex, nCopyLen*sizeof(wxChar)); } } @@ -604,8 +755,8 @@ wxString wxString::Mid(size_t nFirst, size_t nCount) const wxStringData *pData = GetStringData(); size_t nLen = pData->nDataLength; - // default value of nCount is STRING_MAXLEN and means "till the end" - if ( nCount == STRING_MAXLEN ) + // default value of nCount is wxSTRING_MAXLEN and means "till the end" + if ( nCount == wxSTRING_MAXLEN ) { nCount = nLen - nFirst; } @@ -628,6 +779,35 @@ 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 { @@ -641,11 +821,11 @@ 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; @@ -666,10 +846,10 @@ 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 != wxT('\0') && *pc != ch; pc++ ) str += *pc; return str; @@ -677,11 +857,11 @@ 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; @@ -689,28 +869,28 @@ 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 -size_t wxString::Replace(const char *szOld, const char *szNew, bool bReplaceAll) +size_t wxString::Replace(const wxChar *szOld, const wxChar *szNew, bool bReplaceAll) { size_t uiCount = 0; // count of replacements made - size_t 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 != wxT('\0') ) { + pSubstr = wxStrstr(pCurrent, szOld); if ( pSubstr == NULL ) { // strTemp is unused if no replacements were made, so avoid the copy if ( uiCount == 0 ) @@ -743,7 +923,7 @@ size_t 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++; @@ -753,9 +933,9 @@ bool wxString::IsAscii() const 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); @@ -763,9 +943,11 @@ bool wxString::IsWord() const bool wxString::IsNumber() const { - const char *s = (const char*) *this; + const wxChar *s = (const wxChar*) *this; + if (wxStrlen(s)) + if ((s[0] == '-') || (s[0] == '+')) s++; while(*s){ - if(!isdigit(*s)) return(FALSE); + if(!wxIsdigit(*s)) return(FALSE); s++; } return(TRUE); @@ -787,8 +969,8 @@ 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; } @@ -797,8 +979,8 @@ 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; } @@ -807,14 +989,22 @@ 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 && isspace(GetChar(Len() - 1))) || - (!bFromRight && isspace(GetChar(0u))) + if ( !IsEmpty() && + ( + (bFromRight && wxSafeIsspace(GetChar(Len() - 1))) || + (!bFromRight && wxSafeIsspace(GetChar(0u))) ) ) { @@ -824,24 +1014,24 @@ wxString& wxString::Trim(bool bFromRight) if ( bFromRight ) { // find last non-space character - char *psz = m_pchData + GetStringData()->nDataLength - 1; - while ( isspace(*psz) && (psz >= m_pchData) ) + wxChar *psz = m_pchData + GetStringData()->nDataLength - 1; + while ( wxSafeIsspace(*psz) && (psz >= m_pchData) ) psz--; // truncate at trailing space start - *++psz = '\0'; + *++psz = wxT('\0'); GetStringData()->nDataLength = psz - m_pchData; } else { // find first non-space character - const char *psz = m_pchData; - while ( isspace(*psz) ) + const wxChar *psz = m_pchData; + while ( wxSafeIsspace(*psz) ) psz++; // fix up data and length - int nDataLength = GetStringData()->nDataLength - (psz - m_pchData); - memmove(m_pchData, psz, (nDataLength + 1)*sizeof(char)); + int nDataLength = GetStringData()->nDataLength - (psz - (const wxChar*) m_pchData); + memmove(m_pchData, psz, (nDataLength + 1)*sizeof(wxChar)); GetStringData()->nDataLength = nDataLength; } } @@ -850,7 +1040,7 @@ wxString& wxString::Trim(bool bFromRight) } // 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); @@ -868,63 +1058,107 @@ 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) = wxT('\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 char *psz = strstr(m_pchData, pszSub); + const wxChar *psz = wxStrstr(m_pchData, pszSub); - return (psz == NULL) ? NOT_FOUND : psz - m_pchData; + 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("%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("%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("%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 // --------------------------------------------------------------------------- -int wxString::Printf(const char *pszFormat, ...) + +/* 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; va_start(argptr, pszFormat); @@ -936,75 +1170,375 @@ 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_EXPERIMENTAL_PRINTF + // 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] == wxT('%')) { + 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 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(szScratch, s_szFlags, val); +#else + 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(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); +#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 += wxT('%'); // just to pass the glibc tst-printf.c + n--; + done = TRUE; + } + break; + } +#undef CHECK_PREC + } while (!done); + } else *this += pszFormat[n]; - int iLen = wxVsprintf(s_szScratch, WXSIZEOF(s_szScratch), pszFormat, argptr); - char *buffer; - if ( (size_t)iLen < WXSIZEOF(s_szScratch) ) { - buffer = s_szScratch; +#else + // 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 = wxVsprintf(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; } } +#endif // wxUSE_EXPERIMENTAL_PRINTF/!wxUSE_EXPERIMENTAL_PRINTF - AllocBeforeWrite(iLen); - strcpy(m_pchData, buffer); - - if ( buffer != s_szScratch ) - free(buffer); - - return iLen; + return Len(); } // ---------------------------------------------------------------------------- // misc other operations // ---------------------------------------------------------------------------- -bool wxString::Matches(const char *pszMask) const + +// returns TRUE if the string matches the pattern which may contain '*' and +// '?' metacharacters (as usual, '?' matches any character and '*' any number +// of them) +bool wxString::Matches(const wxChar *pszMask) const { - // check char by char - const char *pszTxt; - for ( pszTxt = c_str(); *pszMask != '\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 '?': - if ( *pszTxt == '\0' ) + case wxT('?'): + if ( *pszTxt == wxT('\0') ) return FALSE; - pszTxt++; - pszMask++; + // pszTxt and pszMask will be incremented in the loop statement + break; - case '*': + 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 == '*' || *pszMask == '?' ) + // (should this be an error?) + while ( *pszMask == wxT('*') || *pszMask == wxT('?') ) pszMask++; // if there is nothing more, match - if ( *pszMask == '\0' ) + if ( *pszMask == wxT('\0') ) return TRUE; // are there any other metacharacters in the mask? size_t uiLenMask; - const char *pEndMask = strpbrk(pszMask, "*?"); + const wxChar *pEndMask = wxStrpbrk(pszMask, wxT("*?")); if ( pEndMask != NULL ) { // we have to match the string between two metachars @@ -1012,11 +1546,11 @@ bool wxString::Matches(const char *pszMask) const } else { // we have to match the remainder of the string - uiLenMask = strlen(pszMask); + uiLenMask = wxStrlen(pszMask); } wxString strToMatch(pszMask, uiLenMask); - const char* pMatch = strstr(pszTxt, strToMatch); + const wxChar* pMatch = wxStrstr(pszTxt, strToMatch); if ( pMatch == NULL ) return FALSE; @@ -1034,13 +1568,84 @@ bool wxString::Matches(const char *pszMask) const } // match only if nothing left - return *pszTxt == '\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 +int wxString::Freq(wxChar ch) const +{ + int count = 0; + int len = Len(); + for (int i = 0; i < len; i++) + { + if (GetChar(i) == ch) + count ++; + } + return count; } +// 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 + +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) { @@ -1049,10 +1654,10 @@ wxString& wxString::insert(size_t nPos, const wxString& str) if ( !str.IsEmpty() ) { 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); + 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; } @@ -1065,37 +1670,40 @@ 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); + return find(wxString(sz, n), nStart); } -#endif +#endif // VC++ 1.5 -size_t wxString::find(char ch, size_t nStart) const +// 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); + // TODO could be made much quicker than that + 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--; } @@ -1104,31 +1712,132 @@ size_t wxString::rfind(const wxString& str, size_t nStart) const } // 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); + 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() ); + if ( nStart == npos ) + { + nStart = Len(); + } + else + { + wxASSERT( nStart <= Len() ); + } - const char *p = strrchr(c_str() + nStart, ch); + const wxChar *p = wxStrrchr(c_str(), ch); - return p == NULL ? npos : p - c_str(); + if ( p == NULL ) + return npos; + + size_t result = p - c_str(); + return ( result > nStart ) ? npos : result; } -#endif +#endif // VC++ 1.5 -wxString wxString::substr(size_t nStart, size_t nLen) const +size_t wxString::find_first_of(const wxChar* sz, size_t nStart) const { - // npos means 'take all' - if ( nLen == npos ) - nLen = 0; + const wxChar *start = c_str() + nStart; + const wxChar *firstOf = wxStrpbrk(start, sz); + if ( firstOf ) + return firstOf - c_str(); + else + return npos; +} + +size_t wxString::find_last_of(const wxChar* sz, size_t nStart) const +{ + if ( nStart == npos ) + { + nStart = Len(); + } + else + { + wxASSERT( nStart <= Len() ); + } + + for ( const wxChar *p = c_str() + length() - 1; p >= c_str(); p-- ) + { + if ( wxStrchr(sz, *p) ) + return p - c_str(); + } + + return npos; +} + +size_t wxString::find_first_not_of(const wxChar* sz, size_t nStart) const +{ + if ( nStart == npos ) + { + nStart = Len(); + } + else + { + wxASSERT( nStart <= Len() ); + } + + size_t nAccept = wxStrspn(c_str() + nStart, sz); + if ( nAccept >= length() - nStart ) + return npos; + else + return nAccept; +} + +size_t wxString::find_first_not_of(wxChar ch, size_t nStart) const +{ + wxASSERT( nStart <= Len() ); - wxASSERT( nStart + nLen <= Len() ); + for ( const wxChar *p = c_str() + nStart; *p; p++ ) + { + if ( *p != ch ) + return p - c_str(); + } - return wxString(c_str() + nStart, nLen == npos ? 0 : nLen); + return npos; +} + +size_t wxString::find_last_not_of(const wxChar* sz, size_t nStart) const +{ + if ( nStart == npos ) + { + nStart = Len(); + } + else + { + wxASSERT( nStart <= Len() ); + } + + for ( const wxChar *p = c_str() + nStart - 1; p >= c_str(); p-- ) + { + if ( !wxStrchr(sz, *p) ) + return p - c_str(); + } + + return npos; +} + +size_t wxString::find_last_not_of(wxChar ch, size_t nStart) const +{ + if ( nStart == npos ) + { + nStart = Len(); + } + else + { + wxASSERT( nStart <= Len() ); + } + + for ( const wxChar *p = c_str() + nStart - 1; p >= c_str(); p-- ) + { + if ( *p != ch ) + return p - c_str(); + } + + return npos; } wxString& wxString::erase(size_t nStart, size_t nLen) @@ -1144,21 +1853,23 @@ 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_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; } -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)); } @@ -1170,7 +1881,7 @@ wxString& wxString::replace(size_t nStart, size_t nLen, } wxString& wxString::replace(size_t nStart, size_t nLen, - const char* sz, size_t nCount) + const wxChar* sz, size_t nCount) { return replace(nStart, nLen, wxString(sz, nCount)); } @@ -1190,11 +1901,12 @@ 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 = (char **) NULL; + m_pItems = (wxChar **) NULL; + m_autoSort = autoSort; } // copy ctor @@ -1202,7 +1914,8 @@ wxArrayString::wxArrayString(const wxArrayString& src) { m_nSize = m_nCount = 0; - m_pItems = (char **) NULL; + m_pItems = (wxChar **) NULL; + m_autoSort = src.m_autoSort; *this = src; } @@ -1213,45 +1926,51 @@ 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(char *)); - - 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 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 < 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!) wxDELETEA(m_pItems); @@ -1304,46 +2023,119 @@ void wxArrayString::Alloc(size_t nSize) if ( nSize > m_nSize ) { Free(); wxDELETEA(m_pItems); - m_pItems = new char *[nSize]; + m_pItems = new wxChar *[nSize]; m_nSize = nSize; } m_nCount = 0; } -// searches the array for an item (forward or backwards) -int wxArrayString::Index(const char *sz, bool bCase, bool bFromEnd) const +// minimizes the memory usage by freeing unused memory +void wxArrayString::Shrink() { - 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 ); + // 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]; + + // copy data to new location + memcpy(pNew, m_pItems, m_nCount*sizeof(wxChar *)); + delete [] m_pItems; + m_pItems = pNew; + } +} + +// searches the array for an item (forward or backwards) +int wxArrayString::Index(const wxChar *sz, bool bCase, bool bFromEnd) const +{ + if ( m_autoSort ) { + // use binary search in the sorted array + wxASSERT_MSG( bCase && !bFromEnd, + wxT("search parameters ignored for auto sorted array") ); + + size_t i, + lo = 0, + hi = m_nCount; + int res; + while ( lo < hi ) { + i = (lo + hi)/2; + + res = wxStrcmp(sz, m_pItems[i]); + if ( res < 0 ) + hi = i; + else if ( res > 0 ) + lo = i + 1; + else + return i; } + + return wxNOT_FOUND; } else { - 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; + } } } - return NOT_FOUND; + return wxNOT_FOUND; } // 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++] = (char *)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 @@ -1351,15 +2143,15 @@ 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(); memmove(&m_pItems[nIndex + 1], &m_pItems[nIndex], - (m_nCount - nIndex)*sizeof(char *)); + (m_nCount - nIndex)*sizeof(wxChar *)); str.GetStringData()->Lock(); - m_pItems[nIndex] = (char *)str.c_str(); + m_pItems[nIndex] = (wxChar *)str.c_str(); m_nCount++; } @@ -1367,31 +2159,123 @@ 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(); memmove(&m_pItems[nIndex], &m_pItems[nIndex + 1], - (m_nCount - nIndex - 1)*sizeof(char *)); + (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, + wxT("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 LINKAGEMODE wxStringCompareFunction(const void *first, const void *second) +{ + wxString *strFirst = (wxString *)first; + wxString *strSecond = (wxString *)second; + + if ( gs_compareFunction ) { + return gs_compareFunction(*strFirst, *strSecond); + } + else { + // maybe we should use wxStrcoll + int result = wxStrcmp(strFirst->c_str(), strSecond->c_str()); + + return gs_sortAscending ? result : -result; + } +} + // sort array elements using passed comparaison function +void wxArrayString::Sort(CompareFunction compareFunction) +{ + START_SORT(); -void wxArrayString::Sort(bool WXUNUSED(bCase), bool WXUNUSED(bReverse) ) + wxASSERT( !gs_compareFunction ); // must have been reset to NULL + gs_compareFunction = compareFunction; + + DoSort(); + + // reset it to NULL so that Sort(bool) will work the next time + gs_compareFunction = NULL; + + END_SORT(); +} + +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() +{ + 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); +} + +bool wxArrayString::operator==(const wxArrayString& a) const +{ + if ( m_nCount != a.m_nCount ) + return FALSE; + + for ( size_t n = 0; n < m_nCount; n++ ) + { + if ( Item(n) != a[n] ) + return FALSE; + } + + return TRUE; +} +