X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/d85cfb3784a10430cc17618645fa66ff3df08040..4200d1f4164a82739e3e54f5a85f25910c8a8d1d:/src/common/string.cpp diff --git a/src/common/string.cpp b/src/common/string.cpp index 07a532d180..c199f98b4c 100644 --- a/src/common/string.cpp +++ b/src/common/string.cpp @@ -208,7 +208,7 @@ void wxStringBase::InitWith(const wxChar *psz, size_t nPos, size_t nLength) wxFAIL_MSG( _T("out of memory in wxStringBase::InitWith") ); return; } - memcpy(m_pchData, psz + nPos, nLength*sizeof(wxChar)); + wxMemcpy(m_pchData, psz + nPos, nLength); } } @@ -474,9 +474,27 @@ size_t wxStringBase::find(const wxStringBase& str, size_t nStart) const wxASSERT( str.GetStringData()->IsValid() ); wxASSERT( nStart <= length() ); - const wxChar *p = wxStrstr(c_str() + nStart, str.c_str()); + //anchor + const wxChar* p = (const wxChar*)wxMemchr(c_str() + nStart, + str.c_str()[0], + length() - nStart); - return p == NULL ? npos : p - c_str(); + if(!p) + return npos; + + while(p - c_str() + str.length() <= length() && + wxMemcmp(p, str.c_str(), str.length()) ) + { + //anchor again + 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; } size_t wxStringBase::find(const wxChar* sz, size_t nStart, size_t n) const @@ -488,7 +506,7 @@ size_t wxStringBase::find(wxChar ch, size_t nStart) const { wxASSERT( nStart <= length() ); - const wxChar *p = wxStrchr(c_str() + nStart, ch); + const wxChar *p = (const wxChar*)wxMemchr(c_str() + nStart, ch, length() - nStart); return p == NULL ? npos : p - c_str(); } @@ -515,8 +533,8 @@ size_t wxStringBase::rfind(const wxStringBase& str, size_t nStart) const const wxChar *cursor = c_str() + top; do { - if ( memcmp(cursor, str.c_str(), - str.length() * sizeof(wxChar)) == 0 ) + if ( wxMemcmp(cursor, str.c_str(), + str.length()) == 0 ) { return cursor - c_str(); } @@ -555,12 +573,21 @@ size_t wxStringBase::rfind(wxChar ch, size_t nStart) const size_t wxStringBase::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 - c_str(); - else + wxASSERT(nStart <= length()); + + size_t len = wxStrlen(sz); + + size_t i; + for(i = nStart; i < this->length(); ++i) + { + if (wxMemchr(sz, *(c_str() + i), len)) + break; + } + + if(i == this->length()) return npos; + else + return i; } size_t wxStringBase::find_first_of(const wxChar* sz, size_t nStart, @@ -581,9 +608,11 @@ size_t wxStringBase::find_last_of(const wxChar* sz, size_t nStart) const _T("invalid index in find_last_of()") ); } + size_t len = wxStrlen(sz); + for ( const wxChar *p = c_str() + nStart; p >= c_str(); --p ) { - if ( wxStrchr(sz, *p) ) + if ( wxMemchr(sz, *p, len) ) return p - c_str(); } @@ -607,11 +636,19 @@ size_t wxStringBase::find_first_not_of(const wxChar* sz, size_t nStart) const wxASSERT( nStart <= length() ); } - size_t nAccept = wxStrspn(c_str() + nStart, sz); - if ( nAccept >= length() - nStart ) - return npos; - else - return nStart + nAccept; + size_t len = wxStrlen(sz); + + size_t i; + for(i = nStart; i < this->length(); ++i) + { + if (!wxMemchr(sz, *(c_str() + i), len)) + break; + } + + if(i == this->length()) + return npos; + else + return i; } size_t wxStringBase::find_first_not_of(const wxChar* sz, size_t nStart, @@ -644,10 +681,12 @@ size_t wxStringBase::find_last_not_of(const wxChar* sz, size_t nStart) const wxASSERT( nStart <= length() ); } + size_t len = wxStrlen(sz); + for ( const wxChar *p = c_str() + nStart; p >= c_str(); --p ) { - if ( !wxStrchr(sz, *p) ) - return p - c_str(); + if ( !wxMemchr(sz, *p,len) ) + return p - c_str(); } return npos; @@ -702,7 +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, @@ -876,15 +915,15 @@ static inline int wxDoCmp(const wxChar* s1, size_t l1, const wxChar* s2, size_t l2) { if( l1 == l2 ) - return wxStrncmp(s1, s2, l1); + return wxMemcmp(s1, s2, l1); else if( l1 < l2 ) { - int ret = wxStrncmp(s1, s2, l1); + int ret = wxMemcmp(s1, s2, l1); return ret == 0 ? -1 : ret; } else if( l1 > l2 ) { - int ret = wxStrncmp(s1, s2, l2); + int ret = wxMemcmp(s1, s2, l2); return ret == 0 ? +1 : ret; } @@ -892,15 +931,11 @@ static inline int wxDoCmp(const wxChar* s1, size_t l1, return 0; // quiet compilers } -#if wxUSE_STL - int STRINGCLASS::compare(const wxStringBase& str) const { return ::wxDoCmp(data(), length(), str.data(), str.length()); } -#endif - int STRINGCLASS::compare(size_t nStart, size_t nLen, const wxStringBase& str) const { @@ -923,16 +958,12 @@ int STRINGCLASS::compare(size_t nStart, size_t nLen, return ::wxDoCmp(data() + nStart, nLen, str.data() + nStart2, nLen2); } -#if wxUSE_STL - int STRINGCLASS::compare(const wxChar* sz) const { size_t nLen = wxStrlen(sz); return ::wxDoCmp(data(), length(), sz, nLen); } -#endif - int STRINGCLASS::compare(size_t nStart, size_t nLen, const wxChar* sz, size_t nCount) const { @@ -954,7 +985,107 @@ int STRINGCLASS::compare(size_t nStart, size_t nLen, // =========================================================================== // --------------------------------------------------------------------------- -// construction +// 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 // --------------------------------------------------------------------------- #if wxUSE_UNICODE @@ -981,7 +1112,10 @@ wxString::wxString(const char *psz, wxMBConv& conv, size_t nLength) if ( psz ) { // calculate the needed size ourselves or use the provided one - nLen = conv.MB2WC(NULL, psz, 0); + if (nLength == npos) + nLen = strlen(psz); + else + nLen = nLength; } else { @@ -992,28 +1126,36 @@ wxString::wxString(const char *psz, wxMBConv& conv, size_t nLength) // anything to do? if ( (nLen != 0) && (nLen != (size_t)-1) ) { - if ( !Alloc(nLen) ) + //When converting mb->wc it never inflates to more characters than the length + wxWCharBuffer buffer(nLen + 1); + + //Convert the string + size_t nActualLength = wxWcstr(buffer, psz, nLen, conv); + + if ( !Alloc(nActualLength + 1) ) { - wxFAIL_MSG( _T("out of memory in wxString::wxString") ); + wxFAIL_MSG(wxT("Out of memory in wxString")); } else { - wxWCharBuffer buf(nLen); - // MB2WC wants the buffer size, not the string length hence +1 - nLen = conv.MB2WC(buf.data(), psz, nLen + 1); - - if ( nLen != (size_t)-1 ) - { - // initialized ok, set the real length as nLength specified by - // the caller could be greater than the real string length - assign(buf.data(), nLen); - return; - } - //else: the conversion failed -- leave the string empty (what else?) + //Copy the data + assign(buffer.data(), nActualLength); } } } +//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; +} + #else // ANSI #if wxUSE_WCHAR_T @@ -1039,7 +1181,10 @@ wxString::wxString(const wchar_t *pwz, wxMBConv& conv, size_t nLength) if ( pwz ) { // calculate the needed size ourselves or use the provided one - nLen = conv.WC2MB(NULL, pwz, 0); + if (nLength == npos) + nLen = wxWcslen(pwz); + else + nLen = nLength; } else { @@ -1050,26 +1195,37 @@ wxString::wxString(const wchar_t *pwz, wxMBConv& conv, size_t nLength) // anything to do? if ( (nLen != 0) && (nLen != (size_t)-1) ) { - if ( !Alloc(nLen) ) + //*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( _T("out of memory in wxString::wxString") ); + wxFAIL_MSG(wxT("Out of memory in wxString")); } else { - wxCharBuffer buf(nLen); - // WC2MB wants the buffer size, not the string length - if ( conv.WC2MB(buf.data(), pwz, nLen + 1) != (size_t)-1 ) - { - // initialized ok - assign(buf.data(), nLen); - return; - } - //else: the conversion failed -- leave the string empty (what else?) + //copy the data + assign(buffer.data(), nActualLength); } } +} + +//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); - // leave empty + //Do the actual conversion (will return a blank string on error) + wxWcstr(buffer, (*this).c_str(), length(), conv); + + return buffer; } + #endif // wxUSE_WCHAR_T #endif // Unicode/ANSI @@ -1221,6 +1377,66 @@ wxString operator+(const wxChar *psz, const wxString& str) // other common string functions // =========================================================================== +int wxString::Cmp(const wxString& s) const +{ + return compare(s); +} + +int wxString::Cmp(const wxChar* psz) const +{ + return compare(psz); +} + +static inline int wxDoCmpNoCase(const wxChar* s1, size_t l1, + const wxChar* s2, size_t l2) +{ + size_t i; + + if( l1 == l2 ) + { + for(i = 0; i < l1; ++i) + { + if(wxTolower(s1[i]) != wxTolower(s2[i])) + break; + } + return i == l1 ? 0 : s1[i] < s2[i] ? -1 : 1; + } + else if( l1 < l2 ) + { + for(i = 0; i < l1; ++i) + { + if(wxTolower(s1[i]) != wxTolower(s2[i])) + break; + } + return i == l1 ? -1 : s1[i] < s2[i] ? -1 : 1; + } + else if( l1 > l2 ) + { + for(i = 0; i < l2; ++i) + { + if(wxTolower(s1[i]) != wxTolower(s2[i])) + break; + } + return i == l2 ? 1 : s1[i] < s2[i] ? -1 : 1; + } + + wxFAIL; // must never get there + return 0; // quiet compilers +} + +int wxString::CmpNoCase(const wxString& s) const +{ + return wxDoCmpNoCase(data(), length(), s.data(), s.length()); +} + +int wxString::CmpNoCase(const wxChar* psz) const +{ + int nLen = wxStrlen(psz); + + return wxDoCmpNoCase(data(), length(), psz, nLen); +} + + #if wxUSE_UNICODE #ifdef __MWERKS__ @@ -2176,7 +2392,7 @@ size_t wxArrayString::Add(const wxString& str, size_t nInsert) while ( lo < hi ) { i = (lo + hi)/2; - res = wxStrcmp(str, m_pItems[i]); + res = str.Cmp(m_pItems[i]); if ( res < 0 ) hi = i; else if ( res > 0 ) @@ -2345,7 +2561,7 @@ wxStringCompareFunction(const void *first, const void *second) } else { // maybe we should use wxStrcoll - int result = wxStrcmp(strFirst->c_str(), strSecond->c_str()); + int result = strFirst->Cmp(*strSecond); return gs_sortAscending ? result : -result; } @@ -2406,10 +2622,10 @@ bool wxArrayString::operator==(const wxArrayString& a) const int wxCMPFUNC_CONV wxStringSortAscending(wxString* s1, wxString* s2) { - return wxStrcmp(s1->c_str(), s2->c_str()); + return s1->Cmp(*s2); } int wxCMPFUNC_CONV wxStringSortDescending(wxString* s1, wxString* s2) { - return -wxStrcmp(s1->c_str(), s2->c_str()); + return -s1->Cmp(*s2); }