X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/49dcf8bfb80ab5571f81b57f86051a6f4fcad677..068b764a5a7cc2929b1fd8ea5643f43b5cbc8762:/src/common/string.cpp diff --git a/src/common/string.cpp b/src/common/string.cpp index b006012042..6a49e9d42a 100644 --- a/src/common/string.cpp +++ b/src/common/string.cpp @@ -1,18 +1,15 @@ ///////////////////////////////////////////////////////////////////////////// // Name: string.cpp // Purpose: wxString class -// Author: Vadim Zeitlin +// Author: Vadim Zeitlin, Ryan Norton // Modified by: // Created: 29/01/98 // RCS-ID: $Id$ // Copyright: (c) 1998 Vadim Zeitlin +// (c) 2004 Ryan Norton // Licence: wxWindows licence ///////////////////////////////////////////////////////////////////////////// -#if defined(__GNUG__) && !defined(NO_GCC_PRAGMA) - #pragma implementation "string.h" -#endif - /* * About ref counting: * 1) all empty strings use g_strEmpty, nRefs = -1 (set in Init()) @@ -42,6 +39,10 @@ #include #include +#ifndef __WXMSW__ +#include +#endif + #ifdef __SALFORDC__ #include #endif @@ -55,14 +56,9 @@ // static class variables definition // --------------------------------------------------------------------------- -#if defined(__VISAGECPP__) && __IBMCPP__ >= 400 -// must define this static for VA or else you get multiply defined symbols -// everywhere -const unsigned int wxSTRING_MAXLEN = UINT_MAX - 100; -#endif // Visual Age - #if !wxUSE_STL - const size_t wxStringBase::npos = wxSTRING_MAXLEN; + //According to STL _must_ be a -1 size_t + const size_t wxStringBase::npos = (size_t) -1; #endif // ---------------------------------------------------------------------------- @@ -208,7 +204,7 @@ void wxStringBase::InitWith(const wxChar *psz, size_t nPos, size_t nLength) wxFAIL_MSG( _T("out of memory in wxStringBase::InitWith") ); return; } - wxMemcpy(m_pchData, psz + nPos, nLength); + wxTmemcpy(m_pchData, psz + nPos, nLength); } } @@ -273,7 +269,7 @@ bool wxStringBase::CopyBeforeWrite() // allocation failures are handled by the caller return false; } - memcpy(m_pchData, pData->data(), nLen*sizeof(wxChar)); + wxTmemcpy(m_pchData, pData->data(), nLen); } wxASSERT( !GetStringData()->IsShared() ); // we must be the only owner @@ -316,13 +312,15 @@ bool wxStringBase::AllocBeforeWrite(size_t nLen) 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 + // it doesn't really matter what the string length is as it's going to be + // overwritten later but, for extra safety, set it to 0 for now as we may + // have some junk in m_pchData + GetStringData()->nDataLength = 0; + return true; } @@ -330,7 +328,7 @@ wxStringBase& wxStringBase::append(size_t n, wxChar ch) { size_type len = length(); - if ( !CopyBeforeWrite() || !Alloc(len + n) ) { + if ( !Alloc(len + n) || !CopyBeforeWrite() ) { wxFAIL_MSG( _T("out of memory in wxStringBase::append") ); } GetStringData()->nDataLength = len + n; @@ -384,7 +382,9 @@ bool wxStringBase::Alloc(size_t nLen) // allocation failure handled by caller return false; } - memcpy(m_pchData, pData->data(), nOldLen*sizeof(wxChar)); + // +1 to copy the terminator, too + memcpy(m_pchData, pData->data(), (nOldLen+1)*sizeof(wxChar)); + GetStringData()->nDataLength = nOldLen; } else { nLen += EXTRA_ALLOC; @@ -449,7 +449,7 @@ wxStringBase& wxStringBase::insert(size_t nPos, const wxChar *sz, size_t n) if ( n == npos ) n = wxStrlen(sz); if ( n == 0 ) return *this; - if ( !CopyBeforeWrite() || !Alloc(length() + n) ) { + if ( !Alloc(length() + n) || !CopyBeforeWrite() ) { wxFAIL_MSG( _T("out of memory in wxStringBase::insert") ); } @@ -475,25 +475,31 @@ size_t wxStringBase::find(const wxStringBase& str, size_t nStart) const wxASSERT( nStart <= length() ); //anchor - const wxChar* p = (const wxChar*)wxMemchr(c_str() + nStart, - str.c_str()[0], + const wxChar* p = (const wxChar*)wxTmemchr(c_str() + nStart, + str.c_str()[0], length() - nStart); - + if(!p) return npos; while(p - c_str() + str.length() <= length() && - wxMemcmp(p, str.c_str(), str.length()) ) + wxTmemcmp(p, str.c_str(), str.length()) ) { + //Previosly passed as the first argument to wxTmemchr, + //but C/C++ standard does not specify evaluation order + //of arguments to functions - + //http://embedded.com/showArticle.jhtml?articleID=9900607 + ++p; + //anchor again - p = (const wxChar*)wxMemchr(++p, - str.c_str()[0], + p = (const wxChar*)wxTmemchr(p, + str.c_str()[0], length() - (p - c_str())); if(!p) return npos; } - + return (p - c_str() + str.length() <= length()) ? p - c_str() : npos; } @@ -506,7 +512,7 @@ size_t wxStringBase::find(wxChar ch, size_t nStart) const { wxASSERT( nStart <= length() ); - const wxChar *p = (const wxChar*)wxMemchr(c_str() + nStart, ch, length() - nStart); + const wxChar *p = (const wxChar*)wxTmemchr(c_str() + nStart, ch, length() - nStart); return p == NULL ? npos : p - c_str(); } @@ -533,7 +539,7 @@ size_t wxStringBase::rfind(const wxStringBase& str, size_t nStart) const const wxChar *cursor = c_str() + top; do { - if ( wxMemcmp(cursor, str.c_str(), + if ( wxTmemcmp(cursor, str.c_str(), str.length()) == 0 ) { return cursor - c_str(); @@ -580,7 +586,7 @@ size_t wxStringBase::find_first_of(const wxChar* sz, size_t nStart) const size_t i; for(i = nStart; i < this->length(); ++i) { - if (wxMemchr(sz, *(c_str() + i), len)) + if (wxTmemchr(sz, *(c_str() + i), len)) break; } @@ -609,10 +615,10 @@ size_t wxStringBase::find_last_of(const wxChar* sz, size_t nStart) const } size_t len = wxStrlen(sz); - + for ( const wxChar *p = c_str() + nStart; p >= c_str(); --p ) { - if ( wxMemchr(sz, *p, len) ) + if ( wxTmemchr(sz, *p, len) ) return p - c_str(); } @@ -641,7 +647,7 @@ size_t wxStringBase::find_first_not_of(const wxChar* sz, size_t nStart) const size_t i; for(i = nStart; i < this->length(); ++i) { - if (!wxMemchr(sz, *(c_str() + i), len)) + if (!wxTmemchr(sz, *(c_str() + i), len)) break; } @@ -685,7 +691,7 @@ size_t wxStringBase::find_last_not_of(const wxChar* sz, size_t nStart) const for ( const wxChar *p = c_str() + nStart; p >= c_str(); --p ) { - if ( !wxMemchr(sz, *p,len) ) + if ( !wxTmemchr(sz, *p,len) ) return p - c_str(); } @@ -741,7 +747,7 @@ wxStringBase& wxStringBase::replace(size_t nStart, size_t nLen, wxStringBase& wxStringBase::replace(size_t nStart, size_t nLen, size_t nCount, wxChar ch) { - return replace(nStart, nLen, wxStringBase(ch, nCount).c_str()); + return replace(nStart, nLen, wxStringBase(nCount, ch).c_str()); } wxStringBase& wxStringBase::replace(size_t nStart, size_t nLen, @@ -915,20 +921,17 @@ static inline int wxDoCmp(const wxChar* s1, size_t l1, const wxChar* s2, size_t l2) { if( l1 == l2 ) - return wxMemcmp(s1, s2, l1); + return wxTmemcmp(s1, s2, l1); else if( l1 < l2 ) { - int ret = wxMemcmp(s1, s2, l1); + int ret = wxTmemcmp(s1, s2, l1); return ret == 0 ? -1 : ret; } - else if( l1 > l2 ) + else { - int ret = wxMemcmp(s1, s2, l2); + int ret = wxTmemcmp(s1, s2, l2); return ret == 0 ? +1 : ret; } - - wxFAIL; // must never get there - return 0; // quiet compilers } int STRINGCLASS::compare(const wxStringBase& str) const @@ -984,106 +987,6 @@ int STRINGCLASS::compare(size_t nStart, size_t nLen, // wxString class core // =========================================================================== -// --------------------------------------------------------------------------- -// common conversion routines -// --------------------------------------------------------------------------- - -#if wxUSE_WCHAR_T - -//Convert a wide character string of a specified length -//to a multi-byte character string, ignoring intermittent null characters -//returns the actual length of the string -inline size_t wxMbstr(wxCharBuffer& buffer, const wchar_t* szString, - size_t nStringLen, wxMBConv& conv) -{ - const wchar_t* szEnd = szString + nStringLen + 1; - const wchar_t* szPos = szString; - const wchar_t* szStart = szPos; - - size_t nActualLength = 0; - - //Convert the string until the length() is reached, continuing the - //loop every time a null character is reached - while(szPos != szEnd) - { - wxASSERT(szPos < szEnd); //something is _really_ screwed up if this rings true - - //Get the length of the current (sub)string - size_t nLen = conv.WC2MB(NULL, szPos, 0); - - wxASSERT(nLen != (size_t)-1); //should not be true! If it is system wctomb could be bad - - nActualLength += nLen + 1; - - wxASSERT(nActualLength <= (nStringLen<<1) + 1); //If this is true it means buffer overflow - - //Convert the current (sub)string - if ( conv.WC2MB(&buffer.data()[szPos - szStart], szPos, nLen + 1) == (size_t)-1 ) - { - //error - return empty buffer - wxFAIL_MSG(wxT("Error converting wide-character string to a multi-byte string")); - buffer.data()[0] = '\0'; - return 0; - } - - //Increment to next (sub)string - //Note that we have to use wxWcslen here instead of nLen - //here because XX2XX gives us the size of the output buffer, - //not neccessarly the length of the string - szPos += wxWcslen(szPos) + 1; - } - - return nActualLength - 1; //success - return actual length -} - -//Convert a multi-byte character string of a specified length -//to a wide character string, ignoring intermittent null characters -//returns the actual length -inline size_t wxWcstr( wxWCharBuffer& buffer, const char* szString, - size_t nStringLen, wxMBConv& conv) -{ - const char* szEnd = szString + nStringLen + 1; - const char* szPos = szString; - const char* szStart = szPos; - - size_t nActualLength = 0; - - //Convert the string until the length() is reached, continuing the - //loop every time a null character is reached - while(szPos != szEnd) - { - wxASSERT(szPos < szEnd); //something is _really_ screwed up if this rings true - - //Get the length of the current (sub)string - size_t nLen = conv.MB2WC(NULL, szPos, 0); - - wxASSERT(nLen != (size_t)-1); //should not be true! If it is system mbtowc could be bad - - nActualLength += nLen + 1; - - wxASSERT(nActualLength <= nStringLen + 1); //If this is true it means buffer overflow - - //Convert the current (sub)string - if ( conv.MB2WC(&buffer.data()[szPos - szStart], szPos, nLen + 1) == (size_t)-1 ) - { - //error - return empty buffer - wxFAIL_MSG(wxT("Error converting multi-byte string to a wide-character string")); - buffer.data()[0] = '\0'; - return 0; - } - - //Increment to next (sub)string - //Note that we have to use strlen here instead of nLen - //here because XX2XX gives us the size of the output buffer, - //not neccessarly the length of the string - szPos += strlen(szPos) + 1; - } - - return nActualLength - 1; //success - return actual length -} - -#endif //wxUSE_WCHAR_T - // --------------------------------------------------------------------------- // construction and conversion // --------------------------------------------------------------------------- @@ -1123,37 +1026,25 @@ wxString::wxString(const char *psz, wxMBConv& conv, size_t nLength) nLen = 0; } + // anything to do? if ( (nLen != 0) && (nLen != (size_t)-1) ) { - //When converting mb->wc it never inflates to more characters than the length - wxWCharBuffer buffer(nLen + 1); + //Convert string + size_t nRealSize; + wxWCharBuffer theBuffer = conv.cMB2WC(psz, nLen, &nRealSize); - //Convert the string - size_t nActualLength = wxWcstr(buffer, psz, nLen, conv); - - if ( !Alloc(nActualLength + 1) ) - { - wxFAIL_MSG(wxT("Out of memory in wxString")); - } - else - { - //Copy the data - assign(buffer.data(), nActualLength); - } + //Copy + if (nRealSize) + assign( theBuffer.data() , nRealSize - 1 ); } -} +} //Convert wxString in Unicode mode to a multi-byte string const wxCharBuffer wxString::mb_str(wxMBConv& conv) const { - //*2 is the worst case - probably for UTF8 - wxCharBuffer buffer((length() << 1) + 1); - - //Do the actual conversion (will return a blank string on error) - wxMbstr(buffer, (*this).c_str(), length(), conv); - - return buffer; + size_t dwOutSize; + return conv.cWC2MB(c_str(), length(), &dwOutSize); } #else // ANSI @@ -1195,37 +1086,24 @@ wxString::wxString(const wchar_t *pwz, wxMBConv& conv, size_t nLength) // anything to do? if ( (nLen != 0) && (nLen != (size_t)-1) ) { - //*2 is the worst case - probably for UTF8 - wxCharBuffer buffer((nLen << 1) + 1); - - //do the actual conversion (if it fails we get an empty string) - size_t nActualLength = wxMbstr(buffer, pwz, nLen, conv); - - if ( !Alloc(nActualLength + 1) ) - { - wxFAIL_MSG(wxT("Out of memory in wxString")); - } - else - { - //copy the data - assign(buffer.data(), nActualLength); - } + //Convert string + size_t nRealSize; + wxCharBuffer theBuffer = conv.cWC2MB(pwz, nLen, &nRealSize); + + //Copy + if (nRealSize) + assign( theBuffer.data() , nRealSize - 1 ); } } -//Converts this string to a wide character string if unicode +//Converts this string to a wide character string if unicode //mode is not enabled and wxUSE_WCHAR_T is enabled const wxWCharBuffer wxString::wc_str(wxMBConv& conv) const { - //mb->wc never inflates to more than the length - wxWCharBuffer buffer(length() + 1); - - //Do the actual conversion (will return a blank string on error) - wxWcstr(buffer, (*this).c_str(), length(), conv); - - return buffer; + size_t dwOutSize; + return conv.cMB2WC(c_str(), length(), &dwOutSize); } - + #endif // wxUSE_WCHAR_T #endif // Unicode/ANSI @@ -1351,7 +1229,7 @@ wxString operator+(const wxString& str, const wxChar *psz) if ( !s.Alloc(wxStrlen(psz) + str.Len()) ) { wxFAIL_MSG( _T("out of memory in wxString::operator+") ); } - s = str; + s += str; s += psz; return s; @@ -1399,7 +1277,7 @@ static inline int wxDoCmpNoCase(const wxChar* s1, size_t l1, if(wxTolower(s1[i]) != wxTolower(s2[i])) break; } - return i == l1 ? 0 : s1[i] < s2[i] ? -1 : 1; + return i == l1 ? 0 : wxTolower(s1[i]) < wxTolower(s2[i]) ? -1 : 1; } else if( l1 < l2 ) { @@ -1408,20 +1286,17 @@ static inline int wxDoCmpNoCase(const wxChar* s1, size_t l1, if(wxTolower(s1[i]) != wxTolower(s2[i])) break; } - return i == l1 ? -1 : s1[i] < s2[i] ? -1 : 1; + return i == l1 ? -1 : wxTolower(s1[i]) < wxTolower(s2[i]) ? -1 : 1; } - else if( l1 > l2 ) + else { for(i = 0; i < l2; ++i) { if(wxTolower(s1[i]) != wxTolower(s2[i])) break; } - return i == l2 ? 1 : s1[i] < s2[i] ? -1 : 1; + return i == l2 ? 1 : wxTolower(s1[i]) < wxTolower(s2[i]) ? -1 : 1; } - - wxFAIL; // must never get there - return 0; // quiet compilers } int wxString::CmpNoCase(const wxString& s) const @@ -1484,12 +1359,13 @@ const wxCharBuffer wxString::ToAscii() const // this will allocate enough space for the terminating NUL too wxCharBuffer buffer(length()); - signed char *dest = (signed char *)buffer.data(); + + char *dest = buffer.data(); const wchar_t *pwc = c_str(); for ( ;; ) { - *dest++ = *pwc > SCHAR_MAX ? '_' : *pwc; + *dest++ = (char)(*pwc > SCHAR_MAX ? wxT('_') : *pwc); // the output string can't have embedded NULs anyhow, so we can safely // stop at first of them even if we do have any @@ -1711,7 +1587,7 @@ bool wxString::IsNumber() const { const wxChar *s = (const wxChar*) *this; if (wxStrlen(s)) - if ((s[0] == '-') || (s[0] == '+')) s++; + if ((s[0] == wxT('-')) || (s[0] == wxT('+'))) s++; while(*s){ if(!wxIsdigit(*s)) return(false); s++; @@ -1763,7 +1639,7 @@ inline int wxSafeIsspace(wxChar ch) { return (ch < 127) && wxIsspace(ch); } wxString& wxString::Trim(bool bFromRight) { // first check if we're going to modify the string at all - if ( !IsEmpty() && + if ( !empty() && ( (bFromRight && wxSafeIsspace(GetChar(Len() - 1))) || (!bFromRight && wxSafeIsspace(GetChar(0u))) @@ -1916,50 +1792,55 @@ wxString wxString::FormatV(const wxChar *pszFormat, va_list argptr) int wxString::Printf(const wxChar *pszFormat, ...) { - va_list argptr; - va_start(argptr, pszFormat); + va_list argptr; + va_start(argptr, pszFormat); - int iLen = PrintfV(pszFormat, argptr); + int iLen = PrintfV(pszFormat, argptr); - va_end(argptr); + va_end(argptr); - return iLen; + return iLen; } int wxString::PrintfV(const wxChar* pszFormat, va_list argptr) { int size = 1024; - int len; for ( ;; ) { + wxStringBuffer tmp(*this, size + 1); + wxChar* buf = tmp; + + if ( !buf ) { - wxStringBuffer tmp(*this, size + 1); - wxChar* buf = tmp; + // out of memory + return -1; + } - if ( !buf ) - { - // out of memory - return -1; - } + // wxVsnprintf() may modify the original arg pointer, so pass it + // only a copy + va_list argptrcopy; + wxVaCopy(argptrcopy, argptr); + int len = wxVsnprintf(buf, size, pszFormat, argptrcopy); + va_end(argptrcopy); - // wxVsnprintf() may modify the original arg pointer, so pass it - // only a copy - va_list argptrcopy; - wxVaCopy(argptrcopy, argptr); - len = wxVsnprintf(buf, size, pszFormat, argptrcopy); - va_end(argptrcopy); - - // some implementations of vsnprintf() don't NUL terminate - // the string if there is not enough space for it so - // always do it manually - buf[size] = _T('\0'); - } + // some implementations of vsnprintf() don't NUL terminate + // the string if there is not enough space for it so + // always do it manually + buf[size] = _T('\0'); // vsnprintf() may return either -1 (traditional Unix behaviour) or the // total number of characters which would have been written if the // buffer were large enough - if ( len >= 0 && len <= size ) + // also, it may return an errno may be something like EILSEQ, + // in which case we need to break out + if ( (len >= 0 && len <= size) + // No EOVERFLOW on Windows nor Palm 6.0 nor OpenVMS nor MacOS (not X) + // not OS/2 (not Innotek libc). +#if !defined(__WXMSW__) && !defined(__WXPALMOS__) && !defined( __VMS ) && !(defined(__WXMAC__) && !defined(__WXMAC_OSX__)) && !(defined(__EMX__) && !defined(__INNOTEK_LIBC__)) + || errno != EOVERFLOW +#endif + ) { // ok, there was enough space break; @@ -2152,6 +2033,24 @@ int wxString::sprintf(const wxChar *pszFormat, ...) #include "wx/arrstr.h" +wxArrayString::wxArrayString(size_t sz, const wxChar** a) +{ +#if !wxUSE_STL + Init(false); +#endif + for (size_t i=0; i < sz; i++) + Add(a[i]); +} + +wxArrayString::wxArrayString(size_t sz, const wxString* a) +{ +#if !wxUSE_STL + Init(false); +#endif + for (size_t i=0; i < sz; i++) + Add(a[i]); +} + #if !wxUSE_STL // size increment = min(50% of current size, ARRAY_MAXSIZE_INCREMENT) @@ -2329,6 +2228,11 @@ wxString* wxArrayString::GetStringArray() const return array; } +void wxArrayString::Remove(size_t nIndex, size_t nRemove) +{ + RemoveAt(nIndex, nRemove); +} + #endif // WXWIN_COMPATIBILITY_2_4 // searches the array for an item (forward or backwards)