X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/19db911ad22775daac04871c7d31a878b75e83dd..3b01b1fe4cb656f39290bf7f3891667855abe601:/src/common/string.cpp?ds=sidebyside diff --git a/src/common/string.cpp b/src/common/string.cpp index f601c39f5f..f6b388e45b 100644 --- a/src/common/string.cpp +++ b/src/common/string.cpp @@ -273,7 +273,7 @@ bool wxStringBase::CopyBeforeWrite() // allocation failures are handled by the caller return false; } - memcpy(m_pchData, pData->data(), nLen*sizeof(wxChar)); + wxMemcpy(m_pchData, pData->data(), nLen); } wxASSERT( !GetStringData()->IsShared() ); // we must be the only owner @@ -475,10 +475,10 @@ 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*)wxMemchr(c_str() + nStart, + str.c_str()[0], length() - nStart); - + if(!p) return npos; @@ -486,14 +486,14 @@ size_t wxStringBase::find(const wxStringBase& str, size_t nStart) const wxMemcmp(p, str.c_str(), str.length()) ) { //anchor again - p = (const wxChar*)wxMemchr(++p, - str.c_str()[0], + p = (const wxChar*)wxMemchr(++p, + str.c_str()[0], length() - (p - c_str())); if(!p) return npos; } - + return (p - c_str() + str.length() <= length()) ? p - c_str() : npos; } @@ -609,7 +609,7 @@ 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) ) @@ -741,7 +741,7 @@ 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, @@ -988,19 +988,26 @@ int STRINGCLASS::compare(size_t nStart, size_t nLen, // common conversion routines // --------------------------------------------------------------------------- +size_t wxString::WorstEncodingCase(size_t len, const wxMBConv& WXUNUSED(conv)) +{ + //Worst case for UTF7 + return len * 5; +} + #if wxUSE_WCHAR_T //Convert a wide character string of a specified length //to a multi-byte character string, ignoring intermittent null characters -inline wxCharBuffer wxMbstr(const wchar_t* szString, size_t nStringLen, wxMBConv& conv) +//returns the actual length of the string +inline size_t wxMbstr(char* szBuffer, 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; - - //Create the buffer we'll return on success - wxCharBuffer buffer(nStringLen + 1); - + + 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) @@ -1009,38 +1016,51 @@ inline wxCharBuffer wxMbstr(const wchar_t* szString, size_t nStringLen, wxMBConv //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 - + + //Invalid conversion? + if( nLen == (size_t)-1 ) + { + szBuffer[0] = '\0'; + return 0; + } + + //Increase the actual length (+1 for current null character) + nActualLength += nLen + 1; + + //If this is true it means buffer overflow + wxASSERT( nActualLength <= wxString::WorstEncodingCase(nStringLen, conv) + 1 ); + //Convert the current (sub)string - if ( conv.WC2MB(&buffer.data()[szPos - szStart], szPos, nLen + 1) == (size_t)-1 ) + if(conv.WC2MB(&szBuffer[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 buffer; - } - + szBuffer[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 buffer; //success - return converted string + + 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 -inline wxWCharBuffer wxWcstr(const char* szString, size_t nStringLen, wxMBConv& conv) +//returns the actual length +inline size_t wxWcstr( wchar_t* szBuffer, const char* szString, + size_t nStringLen, wxMBConv& conv) { const char* szEnd = szString + nStringLen + 1; const char* szPos = szString; const char* szStart = szPos; - - wxWCharBuffer buffer(nStringLen + 1); - + + 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) @@ -1049,26 +1069,37 @@ inline wxWCharBuffer wxWcstr(const char* szString, size_t nStringLen, wxMBConv& //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 - + + //Invalid conversion? + if( nLen == (size_t)-1 ) + { + szBuffer[0] = '\0'; + return 0; + } + + //Increase the actual length (+1 for current null character) + nActualLength += nLen + 1; + + //If this is true it means buffer overflow + wxASSERT(nActualLength <= nStringLen + 1); + //Convert the current (sub)string - if ( conv.MB2WC(&buffer.data()[szPos - szStart], szPos, nLen + 1) == (size_t)-1 ) + if ( conv.MB2WC(&szBuffer[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 buffer; - } - + szBuffer[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 buffer; //success - return converted string + + return nActualLength - 1; //success - return actual length } #endif //wxUSE_WCHAR_T @@ -1115,15 +1146,26 @@ wxString::wxString(const char *psz, wxMBConv& conv, size_t nLength) // anything to do? if ( (nLen != 0) && (nLen != (size_t)-1) ) { - //do the actual conversion (if it fails we get an empty string) - (*this) = wxWcstr(psz, nLen, conv); + //When converting mb->wc it never inflates to more characters than the length + wxStringBufferLength internalBuffer(*this, nLen + 1); + + //Do the actual conversion & Set the length of the buffer + internalBuffer.SetLength( + wxWcstr(internalBuffer, psz, nLen, conv) + ); } -} +} //Convert wxString in Unicode mode to a multi-byte string const wxCharBuffer wxString::mb_str(wxMBConv& conv) const { - return wxMbstr((*this).c_str(), length(), conv); + //Create the buffer + wxCharBuffer buffer( wxString::WorstEncodingCase(length(), conv) + 1); + + //Do the actual conversion (will return a blank string on error) + wxMbstr(buffer.data(), (*this).c_str(), length(), conv); + + return buffer; } #else // ANSI @@ -1165,19 +1207,30 @@ wxString::wxString(const wchar_t *pwz, wxMBConv& conv, size_t nLength) // anything to do? if ( (nLen != 0) && (nLen != (size_t)-1) ) { - //do the actual conversion (if it fails we get an empty string) - (*this) = wxMbstr(pwz, nLen, conv); + //Create a wxStringBufferLength which will access the internal + //C char pointer in non-stl mode + wxStringBufferLength internalBuffer(*this, wxString::WorstEncodingCase(nLen, conv) + 1); + + //Do the actual conversion & Set the length of the buffer + internalBuffer.SetLength( + wxMbstr(internalBuffer, pwz, nLen, conv) + ); } } -//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) - return wxWcstr((*this).c_str(), length(), conv); + wxWcstr(buffer.data(), (*this).c_str(), length(), conv); + + return buffer; } - + #endif // wxUSE_WCHAR_T #endif // Unicode/ANSI @@ -1441,7 +1494,7 @@ const wxCharBuffer wxString::ToAscii() const const wchar_t *pwc = c_str(); for ( ;; ) { - *dest++ = *pwc > SCHAR_MAX ? '_' : *pwc; + *dest++ = *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 @@ -1663,7 +1716,7 @@ 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++;