X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/c6255a6eab37e9ac70b3b2edda8c271d4099fec6..1c0338c5615c351371f704b326634f58d0bf7da9:/src/common/string.cpp diff --git a/src/common/string.cpp b/src/common/string.cpp index 34453a87c3..d36f85bd5a 100644 --- a/src/common/string.cpp +++ b/src/common/string.cpp @@ -340,7 +340,7 @@ wxString::SubstrBufFromMB wxString::ConvertStr(const char *psz, size_t nLength, return SubstrBufFromMB("", 0); // and then to UTF-8: - SubstrBufFromMB buf(ConvertStr(wcBuf, wcLen, wxMBConvUTF8())); + SubstrBufFromMB buf(ConvertStr(wcBuf, wcLen, wxMBConvStrictUTF8())); // widechar -> UTF-8 conversion isn't supposed to ever fail: wxASSERT_MSG( buf.data, _T("conversion to UTF-8 failed") ); @@ -382,9 +382,12 @@ const wxCharBuffer wxString::mb_str(const wxMBConv& conv) const const wxWCharBuffer wxString::wc_str() const { - return wxMBConvUTF8().cMB2WC(m_impl.c_str(), - m_impl.length() + 1 /* size, not length */, - NULL); + return wxMBConvStrictUTF8().cMB2WC + ( + m_impl.c_str(), + m_impl.length() + 1, // size, not length + NULL + ); } const wxCharBuffer wxString::mb_str(const wxMBConv& conv) const @@ -395,10 +398,12 @@ const wxCharBuffer wxString::mb_str(const wxMBConv& conv) const // FIXME-UTF8: use wc_str() here once we have buffers with length size_t wcLen; - wxWCharBuffer wcBuf( - wxMBConvUTF8().cMB2WC(m_impl.c_str(), - m_impl.length() + 1 /* size, not length */, - &wcLen)); + wxWCharBuffer wcBuf(wxMBConvStrictUTF8().cMB2WC + ( + m_impl.c_str(), + m_impl.length() + 1, // size + &wcLen + )); if ( !wcLen ) return wxCharBuffer(""); @@ -565,6 +570,12 @@ wxString operator+(const wchar_t *pwz, const wxString& str) // string comparison // --------------------------------------------------------------------------- +bool wxString::IsSameAs(wxUniChar c, bool compareWithCase) const +{ + return (length() == 1) && (compareWithCase ? GetChar(0u) == c + : wxToupper(GetChar(0u)) == wxToupper(c)); +} + #ifdef HAVE_STD_STRING_COMPARE // NB: Comparison code (both if HAVE_STD_STRING_COMPARE and if not) works with @@ -967,35 +978,35 @@ int wxString::CmpNoCase(const wxString& s) const #endif #endif -wxString wxString::FromAscii(const char *ascii) +wxString wxString::FromAscii(const char *ascii, size_t len) { - if (!ascii) + if (!ascii || len == 0) return wxEmptyString; - size_t len = strlen(ascii); wxString res; - if ( len ) { wxImplStringBuffer buf(res, len); wxStringCharType *dest = buf; - for ( ;; ) + for ( ; len > 0; --len ) { unsigned char c = (unsigned char)*ascii++; wxASSERT_MSG( c < 0x80, _T("Non-ASCII value passed to FromAscii().") ); *dest++ = (wchar_t)c; - - if ( c == '\0' ) - break; } } return res; } +wxString wxString::FromAscii(const char *ascii) +{ + return FromAscii(ascii, wxStrlen(ascii)); +} + wxString wxString::FromAscii(const char ascii) { // What do we do with '\0' ? @@ -1376,14 +1387,15 @@ int wxString::Find(wxUniChar ch, bool bFromEnd) const // conversion to numbers // ---------------------------------------------------------------------------- -// the implementation of all the functions below is exactly the same so factor -// it out +// The implementation of all the functions below is exactly the same so factor +// it out. Note that number extraction works correctly on UTF-8 strings, so +// we can use wxStringCharType and wx_str() for maximum efficiency. -template -bool wxStringToIntType(const wxChar *start, +template +bool wxStringToIntType(const wxStringCharType *start, T *val, int base, - F func) + T (*func)(const wxStringCharType*, wxStringCharType**, int)) { wxCHECK_MSG( val, false, _T("NULL output pointer") ); wxASSERT_MSG( !base || (base > 1 && base <= 36), _T("invalid base") ); @@ -1392,7 +1404,7 @@ bool wxStringToIntType(const wxChar *start, errno = 0; #endif - wxChar *end; + wxStringCharType *end; *val = (*func)(start, &end, base); // return true only if scan was stopped by the terminating NUL and if the @@ -1406,22 +1418,22 @@ bool wxStringToIntType(const wxChar *start, bool wxString::ToLong(long *val, int base) const { - return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtol); + return wxStringToIntType(wx_str(), val, base, wxStrtol); } bool wxString::ToULong(unsigned long *val, int base) const { - return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtoul); + return wxStringToIntType(wx_str(), val, base, wxStrtoul); } bool wxString::ToLongLong(wxLongLong_t *val, int base) const { - return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtoll); + return wxStringToIntType(wx_str(), val, base, wxStrtoll); } bool wxString::ToULongLong(wxULongLong_t *val, int base) const { - return wxStringToIntType((const wxChar*)c_str(), val, base, wxStrtoull); + return wxStringToIntType(wx_str(), val, base, wxStrtoull); } bool wxString::ToDouble(double *val) const @@ -1559,6 +1571,11 @@ static int DoStringPrintfV(wxString& str, if ( !buf ) { // out of memory + + // in UTF-8 build, leaving uninitialized junk in the buffer + // could result in invalid non-empty UTF-8 string, so just + // reset the string to empty on failure: + buf[0] = '\0'; return -1; } @@ -1579,14 +1596,20 @@ static int DoStringPrintfV(wxString& str, // buffer were large enough (newer standards such as Unix98) if ( len < 0 ) { + // NB: wxVsnprintf() may call either wxCRT_VsnprintfW or + // wxCRT_VsnprintfA in UTF-8 build; wxUSE_WXVSNPRINTF + // is true if *both* of them use our own implementation, + // otherwise we can't be sure #if wxUSE_WXVSNPRINTF // we know that our own implementation of wxVsnprintf() returns -1 // only for a format error - thus there's something wrong with // the user's format string + buf[0] = '\0'; return -1; -#else // assume that system version only returns error if not enough space - // still not enough, as we don't know how much we need, double the - // current size of the buffer +#else // possibly using system version + // assume it only returns error if there is not enough space, but + // as we don't know how much we need, double the current size of + // the buffer size *= 2; #endif // wxUSE_WXVSNPRINTF/!wxUSE_WXVSNPRINTF }