X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/c565abe1c595587ff98fcc6c29537883df0defa4..245ff47f7c0715d95aac7844604b0a74633e8dc9:/src/common/string.cpp diff --git a/src/common/string.cpp b/src/common/string.cpp index 573e5073c2..762410411d 100644 --- a/src/common/string.cpp +++ b/src/common/string.cpp @@ -24,6 +24,7 @@ #ifndef WX_PRECOMP #include "wx/string.h" #include "wx/wxcrtvararg.h" + #include "wx/log.h" #endif #include @@ -36,6 +37,12 @@ #include #include "wx/hashmap.h" +#include "wx/vector.h" +#include "wx/xlocale.h" + +#ifdef __WXMSW__ + #include "wx/msw/wrapwin.h" +#endif // __WXMSW__ // string handling functions used by wxString: #if wxUSE_UNICODE_UTF8 @@ -50,6 +57,18 @@ #define wxStringStrlen wxStrlen #endif +// ---------------------------------------------------------------------------- +// global variables +// ---------------------------------------------------------------------------- + +namespace wxPrivate +{ + +static UntypedBufferData s_untypedNullData(NULL, 0); + +UntypedBufferData * const untypedNullDataPtr = &s_untypedNullData; + +} // namespace wxPrivate // --------------------------------------------------------------------------- // static class variables definition @@ -91,7 +110,7 @@ static wxStrCacheInitializer gs_stringCacheInit; // gdb seems to be unable to display thread-local variables correctly, at least // not my 6.4.98 version under amd64, so provide this debugging helper to do it -#ifdef __WXDEBUG__ +#if wxDEBUG_LEVEL >= 2 struct wxStrCacheDumper { @@ -116,7 +135,7 @@ struct wxStrCacheDumper void wxDumpStrCache() { wxStrCacheDumper::ShowAll(); } -#endif // __WXDEBUG__ +#endif // wxDEBUG_LEVEL >= 2 #ifdef wxPROFILE_STRING_CACHE @@ -168,7 +187,13 @@ static wxStrCacheStatsDumper s_showCacheStats; wxSTD ostream& operator<<(wxSTD ostream& os, const wxCStrData& str) { #if wxUSE_UNICODE && !wxUSE_UNICODE_UTF8 - return os << (const char *)str.AsCharBuf(); + const wxScopedCharBuffer buf(str.AsCharBuf()); + if ( !buf ) + os.clear(wxSTD ios_base::failbit); + else + os << buf.data(); + + return os; #else return os << str.AsInternal(); #endif @@ -179,13 +204,13 @@ wxSTD ostream& operator<<(wxSTD ostream& os, const wxString& str) return os << str.c_str(); } -wxSTD ostream& operator<<(wxSTD ostream& os, const wxCharBuffer& str) +wxSTD ostream& operator<<(wxSTD ostream& os, const wxScopedCharBuffer& str) { return os << str.data(); } #ifndef __BORLANDC__ -wxSTD ostream& operator<<(wxSTD ostream& os, const wxWCharBuffer& str) +wxSTD ostream& operator<<(wxSTD ostream& os, const wxScopedWCharBuffer& str) { return os << str.data(); } @@ -203,7 +228,7 @@ wxSTD wostream& operator<<(wxSTD wostream& wos, const wxCStrData& str) return wos << str.AsWChar(); } -wxSTD wostream& operator<<(wxSTD wostream& wos, const wxWCharBuffer& str) +wxSTD wostream& operator<<(wxSTD wostream& wos, const wxScopedWCharBuffer& str) { return wos << str.data(); } @@ -375,7 +400,7 @@ const char* wxCStrData::AsChar() const // adding more fields to wxString and require profiling results // to be sure that we really gain enough from them to justify // doing it. - wxCharBuffer buf(str->mb_str()); + wxScopedCharBuffer buf(str->mb_str()); // if it failed, return empty string and not NULL to avoid crashes in code // written with either wxWidgets 2 wxString or std::string behaviour in @@ -406,7 +431,7 @@ const wchar_t* wxCStrData::AsWChar() const wxString *str = wxConstCast(m_str, wxString); // convert the string: - wxWCharBuffer buf(str->wc_str()); + wxScopedWCharBuffer buf(str->wc_str()); // notice that here, unlike above in AsChar(), conversion can't fail as our // internal UTF-8 is always well-formed -- or the string was corrupted and @@ -445,15 +470,15 @@ wxString::SubstrBufFromMB wxString::ConvertStr(const char *psz, size_t nLength, { // anything to do? if ( !psz || nLength == 0 ) - return SubstrBufFromMB(L"", 0); + return SubstrBufFromMB(wxWCharBuffer(L""), 0); if ( nLength == npos ) nLength = wxNO_LEN; size_t wcLen; - wxWCharBuffer wcBuf(conv.cMB2WC(psz, nLength, &wcLen)); + wxScopedWCharBuffer wcBuf(conv.cMB2WC(psz, nLength, &wcLen)); if ( !wcLen ) - return SubstrBufFromMB(_T(""), 0); + return SubstrBufFromMB(wxWCharBuffer(L""), 0); else return SubstrBufFromMB(wcBuf, wcLen); } @@ -466,7 +491,7 @@ wxString::SubstrBufFromMB wxString::ConvertStr(const char *psz, size_t nLength, { // anything to do? if ( !psz || nLength == 0 ) - return SubstrBufFromMB("", 0); + return SubstrBufFromMB(wxCharBuffer(""), 0); // if psz is already in UTF-8, we don't have to do the roundtrip to // wchar_t* and back: @@ -479,7 +504,8 @@ wxString::SubstrBufFromMB wxString::ConvertStr(const char *psz, size_t nLength, // we must pass the real string length to SubstrBufFromMB ctor if ( nLength == npos ) nLength = psz ? strlen(psz) : 0; - return SubstrBufFromMB(wxCharBuffer::CreateNonOwned(psz), nLength); + return SubstrBufFromMB(wxScopedCharBuffer::CreateNonOwned(psz, nLength), + nLength); } // else: do the roundtrip through wchar_t* } @@ -489,9 +515,9 @@ wxString::SubstrBufFromMB wxString::ConvertStr(const char *psz, size_t nLength, // first convert to wide string: size_t wcLen; - wxWCharBuffer wcBuf(conv.cMB2WC(psz, nLength, &wcLen)); + wxScopedWCharBuffer wcBuf(conv.cMB2WC(psz, nLength, &wcLen)); if ( !wcLen ) - return SubstrBufFromMB("", 0); + return SubstrBufFromMB(wxCharBuffer(""), 0); // and then to UTF-8: SubstrBufFromMB buf(ConvertStr(wcBuf, wcLen, wxMBConvStrictUTF8())); @@ -509,15 +535,15 @@ wxString::SubstrBufFromWC wxString::ConvertStr(const wchar_t *pwz, size_t nLengt { // anything to do? if ( !pwz || nLength == 0 ) - return SubstrBufFromWC("", 0); + return SubstrBufFromWC(wxCharBuffer(""), 0); if ( nLength == npos ) nLength = wxNO_LEN; size_t mbLen; - wxCharBuffer mbBuf(conv.cWC2MB(pwz, nLength, &mbLen)); + wxScopedCharBuffer mbBuf(conv.cWC2MB(pwz, nLength, &mbLen)); if ( !mbLen ) - return SubstrBufFromWC("", 0); + return SubstrBufFromWC(wxCharBuffer(""), 0); else return SubstrBufFromWC(mbBuf, mbLen); } @@ -527,50 +553,54 @@ wxString::SubstrBufFromWC wxString::ConvertStr(const wchar_t *pwz, size_t nLengt #if wxUSE_UNICODE_WCHAR //Convert wxString in Unicode mode to a multi-byte string -const wxCharBuffer wxString::mb_str(const wxMBConv& conv) const +const wxScopedCharBuffer wxString::mb_str(const wxMBConv& conv) const { - return conv.cWC2MB(wx_str(), length() + 1 /* size, not length */, NULL); + // NB: Length passed to cWC2MB() doesn't include terminating NUL, it's + // added by it automatically. If we passed length()+1 here, it would + // create a buffer with 2 trailing NULs of length one greater than + // expected. + return conv.cWC2MB(wx_str(), length(), NULL); } #elif wxUSE_UNICODE_UTF8 -const wxWCharBuffer wxString::wc_str() const +const wxScopedWCharBuffer wxString::wc_str() const { + // NB: Length passed to cMB2WC() doesn't include terminating NUL, it's + // added by it automatically. If we passed length()+1 here, it would + // create a buffer with 2 trailing NULs of length one greater than + // expected. return wxMBConvStrictUTF8().cMB2WC ( m_impl.c_str(), - m_impl.length() + 1, // size, not length + m_impl.length(), NULL ); } -const wxCharBuffer wxString::mb_str(const wxMBConv& conv) const +const wxScopedCharBuffer wxString::mb_str(const wxMBConv& conv) const { if ( conv.IsUTF8() ) - return wxCharBuffer::CreateNonOwned(m_impl.c_str()); - - // FIXME-UTF8: use wc_str() here once we have buffers with length + return wxScopedCharBuffer::CreateNonOwned(m_impl.c_str(), m_impl.length()); - size_t wcLen; - wxWCharBuffer wcBuf(wxMBConvStrictUTF8().cMB2WC - ( - m_impl.c_str(), - m_impl.length() + 1, // size - &wcLen - )); - if ( !wcLen ) + wxScopedWCharBuffer wcBuf(wc_str()); + if ( !wcBuf.length() ) return wxCharBuffer(""); - return conv.cWC2MB(wcBuf, wcLen+1, NULL); + return conv.cWC2MB(wcBuf.data(), wcBuf.length(), NULL); } #else // ANSI //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(const wxMBConv& conv) const +const wxScopedWCharBuffer wxString::wc_str(const wxMBConv& conv) const { - return conv.cMB2WC(wx_str(), length() + 1 /* size, not length */, NULL); + // NB: Length passed to cMB2WC() doesn't include terminating NUL, it's + // added by it automatically. If we passed length()+1 here, it would + // create a buffer with 2 trailing NULs of length one greater than + // expected. + return conv.cMB2WC(wx_str(), length(), NULL); } #endif // Unicode/ANSI @@ -1097,8 +1127,36 @@ size_t wxString::find_last_not_of(const wxOtherCharType* sz, size_t nStart, int wxString::CmpNoCase(const wxString& s) const { - // FIXME-UTF8: use wxUniChar::ToLower/ToUpper once added +#if defined(__WXMSW__) && !wxUSE_UNICODE_UTF8 + // prefer to use CompareString() if available as it's more efficient than + // doing it manual or even using wxStricmp() (see #10375) + switch ( ::CompareString(LOCALE_USER_DEFAULT, NORM_IGNORECASE, + m_impl.c_str(), m_impl.length(), + s.m_impl.c_str(), s.m_impl.length()) ) + { + case CSTR_LESS_THAN: + return -1; + + case CSTR_EQUAL: + return 0; + + case CSTR_GREATER_THAN: + return 1; + + default: + wxFAIL_MSG( "unexpected CompareString() return value" ); + // fall through + + case 0: + wxLogLastError("CompareString"); + // use generic code below + } +#endif // __WXMSW__ && !wxUSE_UNICODE_UTF8 + + // do the comparison manually: notice that we can't use wxStricmp() as it + // doesn't handle embedded NULs + // FIXME-UTF8: use wxUniChar::ToLower/ToUpper once added const_iterator i1 = begin(); const_iterator end1 = end(); const_iterator i2 = s.begin(); @@ -1172,7 +1230,7 @@ wxString wxString::FromAscii(char ascii) return wxString(wxUniChar((wchar_t)c)); } -const wxCharBuffer wxString::ToAscii() const +const wxScopedCharBuffer wxString::ToAscii() const { // this will allocate enough space for the terminating NUL too wxCharBuffer buffer(length()); @@ -1276,7 +1334,7 @@ wxString wxString::Right(size_t nCount) const return dest; } -// get all characters after the last occurence of ch +// get all characters after the last occurrence of ch // (returns the whole string if ch not found) wxString wxString::AfterLast(wxUniChar ch) const { @@ -1303,7 +1361,7 @@ wxString wxString::Left(size_t nCount) const return dest; } -// get all characters before the first occurence of ch +// get all characters before the first occurrence of ch // (returns the whole string if ch not found) wxString wxString::BeforeFirst(wxUniChar ch) const { @@ -1313,7 +1371,7 @@ wxString wxString::BeforeFirst(wxUniChar ch) const return wxString(*this, 0, iPos); } -/// get all characters before the last occurence of ch +/// get all characters before the last occurrence of ch /// (returns empty string if ch not found) wxString wxString::BeforeLast(wxUniChar ch) const { @@ -1325,7 +1383,7 @@ wxString wxString::BeforeLast(wxUniChar ch) const return str; } -/// get all characters after the first occurence of ch +/// get all characters after the first occurrence of ch /// (returns empty string if ch not found) wxString wxString::AfterFirst(wxUniChar ch) const { @@ -1337,7 +1395,7 @@ wxString wxString::AfterFirst(wxUniChar ch) const return str; } -// replace first (or all) occurences of some substring with another one +// replace first (or all) occurrences of some substring with another one size_t wxString::Replace(const wxString& strOld, const wxString& strNew, bool bReplaceAll) { @@ -1375,30 +1433,62 @@ size_t wxString::Replace(const wxString& strOld, break; } } - else // general case + else if ( !bReplaceAll) + { + size_t pos = m_impl.find(strOld, 0); + if ( pos != npos ) + { + m_impl.replace(pos, strOld.m_impl.length(), strNew.m_impl); + uiCount = 1; + } + } + else // replace all occurrences { const size_t uiOldLen = strOld.m_impl.length(); const size_t uiNewLen = strNew.m_impl.length(); - for ( size_t pos = 0; ; ) + // first scan the string to find all positions at which the replacement + // should be made + wxVector replacePositions; + + size_t pos; + for ( pos = m_impl.find(strOld.m_impl, 0); + pos != npos; + pos = m_impl.find(strOld.m_impl, pos + uiOldLen)) { - pos = m_impl.find(strOld.m_impl, pos); - if ( pos == npos ) - break; + replacePositions.push_back(pos); + ++uiCount; + } - // replace this occurrence of the old string with the new one - m_impl.replace(pos, uiOldLen, strNew.m_impl); + if ( !uiCount ) + return 0; - // move up pos past the string that was replaced - pos += uiNewLen; + // allocate enough memory for the whole new string + wxString tmp; + tmp.m_impl.reserve(m_impl.length() + uiCount*(uiNewLen - uiOldLen)); - // increase replace count - uiCount++; + // copy this string to tmp doing replacements on the fly + size_t replNum = 0; + for ( pos = 0; replNum < uiCount; replNum++ ) + { + const size_t nextReplPos = replacePositions[replNum]; - // stop after the first one? - if ( !bReplaceAll ) - break; + if ( pos != nextReplPos ) + { + tmp.m_impl.append(m_impl, pos, nextReplPos - pos); + } + + tmp.m_impl.append(strNew.m_impl); + pos = nextReplPos + uiOldLen; } + + if ( pos != m_impl.length() ) + { + // append the rest of the string unchanged + tmp.m_impl.append(m_impl, pos, m_impl.length() - pos); + } + + swap(tmp); } return uiCount; @@ -1589,64 +1679,105 @@ int wxString::Find(wxUniChar ch, bool bFromEnd) const #define DO_IF_NOT_WINCE(x) #endif -#define WX_STRING_TO_INT_TYPE(out, base, func, T) \ - wxCHECK_MSG( out, false, _T("NULL output pointer") ); \ - wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") ); \ - \ +#define WX_STRING_TO_X_TYPE_START \ + wxCHECK_MSG( pVal, false, _T("NULL output pointer") ); \ DO_IF_NOT_WINCE( errno = 0; ) \ - \ const wxStringCharType *start = wx_str(); \ - wxStringCharType *end; \ - T val = func(start, &end, base); \ - \ + wxStringCharType *end; + +#define WX_STRING_TO_X_TYPE_END \ /* return true only if scan was stopped by the terminating NUL and */ \ /* if the string was not empty to start with and no under/overflow */ \ /* occurred: */ \ if ( *end || end == start DO_IF_NOT_WINCE(|| errno == ERANGE) ) \ return false; \ - *out = val; \ - return true + *pVal = val; \ + return true; bool wxString::ToLong(long *pVal, int base) const { - WX_STRING_TO_INT_TYPE(pVal, base, wxStrtol, long); + wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") ); + + WX_STRING_TO_X_TYPE_START + long val = wxStrtol(start, &end, base); + WX_STRING_TO_X_TYPE_END } bool wxString::ToULong(unsigned long *pVal, int base) const { - WX_STRING_TO_INT_TYPE(pVal, base, wxStrtoul, unsigned long); + wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") ); + + WX_STRING_TO_X_TYPE_START + unsigned long val = wxStrtoul(start, &end, base); + WX_STRING_TO_X_TYPE_END } bool wxString::ToLongLong(wxLongLong_t *pVal, int base) const { - WX_STRING_TO_INT_TYPE(pVal, base, wxStrtoll, wxLongLong_t); + wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") ); + + WX_STRING_TO_X_TYPE_START + wxLongLong_t val = wxStrtoll(start, &end, base); + WX_STRING_TO_X_TYPE_END } bool wxString::ToULongLong(wxULongLong_t *pVal, int base) const { - WX_STRING_TO_INT_TYPE(pVal, base, wxStrtoull, wxULongLong_t); + wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") ); + + WX_STRING_TO_X_TYPE_START + wxULongLong_t val = wxStrtoull(start, &end, base); + WX_STRING_TO_X_TYPE_END } bool wxString::ToDouble(double *pVal) const { - wxCHECK_MSG( pVal, false, _T("NULL output pointer") ); + WX_STRING_TO_X_TYPE_START + double val = wxStrtod(start, &end); + WX_STRING_TO_X_TYPE_END +} - DO_IF_NOT_WINCE( errno = 0; ) +#if wxUSE_XLOCALE - const wxChar *start = c_str(); - wxChar *end; - double val = wxStrtod(start, &end); +bool wxString::ToCLong(long *pVal, int base) const +{ + wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") ); - // return true only if scan was stopped by the terminating NUL and if the - // string was not empty to start with and no under/overflow occurred - if ( *end || end == start DO_IF_NOT_WINCE(|| errno == ERANGE) ) - return false; + WX_STRING_TO_X_TYPE_START +#if wxUSE_UNICODE_UTF8 || !wxUSE_UNICODE + long val = wxStrtol_lA(start, &end, base, wxCLocale); +#else + long val = wxStrtol_l(start, &end, base, wxCLocale); +#endif + WX_STRING_TO_X_TYPE_END +} - *pVal = val; +bool wxString::ToCULong(unsigned long *pVal, int base) const +{ + wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") ); - return true; + WX_STRING_TO_X_TYPE_START +#if wxUSE_UNICODE_UTF8 || !wxUSE_UNICODE + unsigned long val = wxStrtoul_lA(start, &end, base, wxCLocale); +#else + unsigned long val = wxStrtoul_l(start, &end, base, wxCLocale); +#endif + WX_STRING_TO_X_TYPE_END } +bool wxString::ToCDouble(double *pVal) const +{ + WX_STRING_TO_X_TYPE_START +#if wxUSE_UNICODE_UTF8 || !wxUSE_UNICODE + double val = wxStrtod_lA(start, &end, wxCLocale); +#else + double val = wxStrtod_l(start, &end, wxCLocale); +#endif + WX_STRING_TO_X_TYPE_END +} + +#endif // wxUSE_XLOCALE + // --------------------------------------------------------------------------- // formatted output // --------------------------------------------------------------------------- @@ -1988,8 +2119,8 @@ bool wxString::Matches(const wxString& mask) const // FIXME-UTF8: implement using iterators, remove #if #if wxUSE_UNICODE_UTF8 - wxWCharBuffer maskBuf = mask.wc_str(); - wxWCharBuffer txtBuf = wc_str(); + const wxScopedWCharBuffer maskBuf = mask.wc_str(); + const wxScopedWCharBuffer txtBuf = wc_str(); const wxChar *pszMask = maskBuf.data(); const wxChar *pszTxt = txtBuf.data(); #else @@ -2091,47 +2222,3 @@ int wxString::Freq(wxUniChar ch) const return count; } -// ---------------------------------------------------------------------------- -// wxUTF8StringBuffer -// ---------------------------------------------------------------------------- - -#if wxUSE_UNICODE_WCHAR -wxUTF8StringBuffer::~wxUTF8StringBuffer() -{ - wxMBConvStrictUTF8 conv; - size_t wlen = conv.ToWChar(NULL, 0, m_buf); - wxCHECK_RET( wlen != wxCONV_FAILED, "invalid UTF-8 data in string buffer?" ); - - wxStringInternalBuffer wbuf(m_str, wlen); - conv.ToWChar(wbuf, wlen, m_buf); -} - -wxUTF8StringBufferLength::~wxUTF8StringBufferLength() -{ - wxCHECK_RET(m_lenSet, "length not set"); - - wxMBConvStrictUTF8 conv; - size_t wlen = conv.ToWChar(NULL, 0, m_buf, m_len); - wxCHECK_RET( wlen != wxCONV_FAILED, "invalid UTF-8 data in string buffer?" ); - - wxStringInternalBufferLength wbuf(m_str, wlen); - conv.ToWChar(wbuf, wlen, m_buf, m_len); - wbuf.SetLength(wlen); -} -#endif // wxUSE_UNICODE_WCHAR - -// ---------------------------------------------------------------------------- -// wxCharBufferType -// ---------------------------------------------------------------------------- - -#ifndef __VMS_BROKEN_TEMPLATES -template<> -#endif -wxCharTypeBuffer::Data -wxCharTypeBuffer::NullData(NULL); - -#ifndef __VMS_BROKEN_TEMPLATES -template<> -#endif -wxCharTypeBuffer::Data -wxCharTypeBuffer::NullData(NULL);