X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/a4a44612a654667fbd40e01d39c7482820fb851b..df08bf5ab3a0eb24d53b2c002f473799c5622bbb:/src/common/string.cpp diff --git a/src/common/string.cpp b/src/common/string.cpp index ab47f8c300..717ce0b52c 100644 --- a/src/common/string.cpp +++ b/src/common/string.cpp @@ -10,13 +10,6 @@ // Licence: wxWindows licence ///////////////////////////////////////////////////////////////////////////// -/* - * About ref counting: - * 1) all empty strings use g_strEmpty, nRefs = -1 (set in Init()) - * 2) AllocBuffer() sets nRefs to 1, Lock() increments it by one - * 3) Unlock() decrements nRefs and frees memory if it goes to 0 - */ - // =========================================================================== // headers, declarations, constants // =========================================================================== @@ -30,6 +23,7 @@ #ifndef WX_PRECOMP #include "wx/string.h" + #include "wx/wxcrtvararg.h" #endif #include @@ -80,9 +74,9 @@ wxSTD ostream& operator<<(wxSTD ostream& os, const wxCStrData& str) { // FIXME-UTF8: always, not only if wxUSE_UNICODE #if wxUSE_UNICODE && !defined(__BORLANDC__) - return os << str.AsWChar(); + return os << (const wchar_t*)str.AsWCharBuf(); #else - return os << str.AsChar(); + return os << (const char*)str.AsCharBuf(); #endif } @@ -111,394 +105,6 @@ wxSTD ostream& operator<<(wxSTD ostream& os, const wxWCharBuffer& str) #if wxUSE_UNICODE_UTF8 -// --------------------------------------------------------------------------- -// UTF-8 operations -// --------------------------------------------------------------------------- - -// -// Table 3.1B from Unicode spec: Legal UTF-8 Byte Sequences -// -// Code Points | 1st Byte | 2nd Byte | 3rd Byte | 4th Byte | -// -------------------+----------+----------+----------+----------+ -// U+0000..U+007F | 00..7F | | | | -// U+0080..U+07FF | C2..DF | 80..BF | | | -// U+0800..U+0FFF | E0 | A0..BF | 80..BF | | -// U+1000..U+FFFF | E1..EF | 80..BF | 80..BF | | -// U+10000..U+3FFFF | F0 | 90..BF | 80..BF | 80..BF | -// U+40000..U+FFFFF | F1..F3 | 80..BF | 80..BF | 80..BF | -// U+100000..U+10FFFF | F4 | 80..8F | 80..BF | 80..BF | -// -------------------+----------+----------+----------+----------+ - -bool wxString::IsValidUtf8String(const char *str) -{ - if ( !str ) - return true; // empty string is UTF8 string - - const unsigned char *c = (const unsigned char*)str; - - for ( ; *c; ++c ) - { - unsigned char b = *c; - - if ( b <= 0x7F ) // 00..7F - continue; - - else if ( b < 0xC2 ) // invalid lead bytes: 80..C1 - return false; - - // two-byte sequences: - else if ( b <= 0xDF ) // C2..DF - { - b = *(++c); - if ( !(b >= 0x80 && b <= 0xBF ) ) - return false; - } - - // three-byte sequences: - else if ( b == 0xE0 ) - { - b = *(++c); - if ( !(b >= 0xA0 && b <= 0xBF ) ) - return false; - b = *(++c); - if ( !(b >= 0x80 && b <= 0xBF ) ) - return false; - } - else if ( b <= 0xEF ) // E1..EF - { - for ( int i = 0; i < 2; ++i ) - { - b = *(++c); - if ( !(b >= 0x80 && b <= 0xBF ) ) - return false; - } - } - - // four-byte sequences: - else if ( b == 0xF0 ) - { - b = *(++c); - if ( !(b >= 0x90 && b <= 0xBF ) ) - return false; - for ( int i = 0; i < 2; ++i ) - { - b = *(++c); - if ( !(b >= 0x80 && b <= 0xBF ) ) - return false; - } - } - else if ( b <= 0xF3 ) // F1..F3 - { - for ( int i = 0; i < 3; ++i ) - { - b = *(++c); - if ( !(b >= 0x80 && b <= 0xBF ) ) - return false; - } - } - else if ( b == 0xF4 ) - { - b = *(++c); - if ( !(b >= 0x80 && b <= 0x8F ) ) - return false; - for ( int i = 0; i < 2; ++i ) - { - b = *(++c); - if ( !(b >= 0x80 && b <= 0xBF ) ) - return false; - } - } - else // otherwise, it's invalid lead byte - return false; - } - - return true; -} - -#ifdef __WXDEBUG__ -/* static */ -bool wxString::IsValidUtf8LeadByte(unsigned char c) -{ - return (c <= 0x7F) || (c >= 0xC2 && c <= 0xF4); -} -#endif - -unsigned char wxString::ms_utf8IterTable[256] = { - // single-byte sequences (ASCII): - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 00..0F - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 10..1F - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 20..2F - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 30..3F - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 40..4F - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 50..5F - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 60..6F - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 70..7F - - // these are invalid, we use step 1 to skip - // over them (should never happen): - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 80..8F - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // 90..9F - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // A0..AF - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, // B0..BF - 1, 1, // C0,C1 - - // two-byte sequences: - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // C2..CF - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, // D0..DF - - // three-byte sequences: - 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, // E0..EF - - // four-byte sequences: - 4, 4, 4, 4, 4, // F0..F4 - - // these are invalid again (5- or 6-byte - // sequences and sequences for code points - // above U+10FFFF, as restricted by RFC 3629): - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1 // F5..FF -}; - -/* static */ -void wxString::DecIter(wxStringImpl::const_iterator& i) -{ - wxASSERT( IsValidUtf8LeadByte(*i) ); - - // Non-lead bytes are all in the 0x80..0xBF range (i.e. 10xxxxxx in - // binary), so we just have to go back until we hit a byte that is either - // < 0x80 (i.e. 0xxxxxxx in binary) or 0xC0..0xFF (11xxxxxx in binary; this - // includes some invalid values, but we can ignore it here, because we - // assume valid UTF-8 input for the purpose of efficient implementation). - --i; - while ( ((*i) & 0xC0) == 0x80 /* 2 highest bits are '10' */ ) - --i; -} - -/* static */ -void wxString::DecIter(wxStringImpl::iterator& i) -{ - // FIXME-UTF8: use template instead - wxASSERT( IsValidUtf8LeadByte(*i) ); - --i; - while ( ((*i) & 0xC0) == 0x80 /* 2 highest bits are '10' */ ) - --i; -} - -/* static */ -wxStringImpl::const_iterator -wxString::AddToIter(wxStringImpl::const_iterator i, int n) -{ - wxStringImpl::const_iterator out(i); - - if ( n > 0 ) - { - for ( int j = 0; j < n; ++j ) - IncIter(out); - } - else if ( n < 0 ) - { - for ( int j = 0; j > n; --j ) - DecIter(out); - } - - return out; -} - -wxStringImpl::iterator -wxString::AddToIter(wxStringImpl::iterator i, int n) -{ - // FIXME-UTF8: use template instead - wxStringImpl::iterator out(i); - - if ( n > 0 ) - { - for ( int j = 0; j < n; ++j ) - IncIter(out); - } - else if ( n < 0 ) - { - for ( int j = 0; j > n; --j ) - DecIter(out); - } - - return out; -} - - -/* static */ -int wxString::DiffIters(wxStringImpl::const_iterator i1, - wxStringImpl::const_iterator i2) -{ - int dist = 0; - - if ( i1 < i2 ) - { - while ( i1 != i2 ) - { - IncIter(i1); - dist--; - } - } - else if ( i2 < i1 ) - { - while ( i2 != i1 ) - { - IncIter(i2); - dist++; - } - } - - return dist; -} - -int wxString::DiffIters(wxStringImpl::iterator i1, wxStringImpl::iterator i2) -{ - // FIXME-UTF8: use template instead - int dist = 0; - - if ( i1 < i2 ) - { - while ( i1 != i2 ) - { - IncIter(i1); - dist--; - } - } - else if ( i2 < i1 ) - { - while ( i2 != i1 ) - { - IncIter(i2); - dist++; - } - } - - return dist; -} - -/* static */ -wxString::Utf8CharBuffer wxString::EncodeChar(wxUniChar ch) -{ - Utf8CharBuffer buf; - char *out = buf.data; - - wxUniChar::value_type code = ch.GetValue(); - - // Char. number range | UTF-8 octet sequence - // (hexadecimal) | (binary) - // ----------------------+--------------------------------------------- - // 0000 0000 - 0000 007F | 0xxxxxxx - // 0000 0080 - 0000 07FF | 110xxxxx 10xxxxxx - // 0000 0800 - 0000 FFFF | 1110xxxx 10xxxxxx 10xxxxxx - // 0001 0000 - 0010 FFFF | 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx - // - // Code point value is stored in bits marked with 'x', lowest-order bit - // of the value on the right side in the diagram above. - // (from RFC 3629) - - if ( code <= 0x7F ) - { - out[1] = 0; - out[0] = (char)code; - } - else if ( code <= 0x07FF ) - { - out[2] = 0; - // NB: this line takes 6 least significant bits, encodes them as - // 10xxxxxx and discards them so that the next byte can be encoded: - out[1] = 0x80 | (code & 0x3F); code >>= 6; - out[0] = 0xC0 | code; - } - else if ( code < 0xFFFF ) - { - out[3] = 0; - out[2] = 0x80 | (code & 0x3F); code >>= 6; - out[1] = 0x80 | (code & 0x3F); code >>= 6; - out[0] = 0xE0 | code; - } - else if ( code <= 0x10FFFF ) - { - out[4] = 0; - out[3] = 0x80 | (code & 0x3F); code >>= 6; - out[2] = 0x80 | (code & 0x3F); code >>= 6; - out[1] = 0x80 | (code & 0x3F); code >>= 6; - out[0] = 0xF0 | code; - } - else - { - wxFAIL_MSG( _T("trying to encode undefined Unicode character") ); - out[0] = 0; - } - - return buf; -} - -/* static */ -wxUniChar wxUniCharRef::DecodeChar(wxStringImpl::const_iterator i) -{ - wxASSERT( wxString::IsValidUtf8LeadByte(*i) ); // FIXME-UTF8: no "wxString::" - - wxUniChar::value_type code = 0; - size_t len = wxString::GetUtf8CharLength(*i); - wxASSERT_MSG( len <= 4, _T("invalid UTF-8 sequence length") ); - - // Char. number range | UTF-8 octet sequence - // (hexadecimal) | (binary) - // ----------------------+--------------------------------------------- - // 0000 0000 - 0000 007F | 0xxxxxxx - // 0000 0080 - 0000 07FF | 110xxxxx 10xxxxxx - // 0000 0800 - 0000 FFFF | 1110xxxx 10xxxxxx 10xxxxxx - // 0001 0000 - 0010 FFFF | 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx - // - // Code point value is stored in bits marked with 'x', lowest-order bit - // of the value on the right side in the diagram above. - // (from RFC 3629) - - // mask to extract lead byte's value ('x' bits above), by sequence's length: - static const unsigned char s_leadValueMask[4] = { 0x7F, 0x1F, 0x0F, 0x07 }; -#ifdef __WXDEBUG__ - // mask and value of lead byte's most significant bits, by length: - static const unsigned char s_leadMarkerMask[4] = { 0x80, 0xE0, 0xF0, 0xF8 }; - static const unsigned char s_leadMarkerVal[4] = { 0x00, 0xC0, 0xE0, 0xF0 }; -#endif - - // extract the lead byte's value bits: - wxASSERT_MSG( ((unsigned char)*i & s_leadMarkerMask[len-1]) == - s_leadMarkerVal[len-1], - _T("invalid UTF-8 lead byte") ); - code = (unsigned char)*i & s_leadValueMask[len-1]; - - // all remaining bytes, if any, are handled in the same way regardless of - // sequence's length: - for ( ++i ; len > 1; --len, ++i ) - { - wxASSERT_MSG( ((unsigned char)*i & 0xC0) == 0x80, - _T("invalid UTF-8 byte") ); - - code <<= 6; - code |= (unsigned char)*i & 0x3F; - } - - return wxUniChar(code); -} - -/* static */ -wxCharBuffer wxString::EncodeNChars(size_t n, wxUniChar ch) -{ - Utf8CharBuffer once(EncodeChar(ch)); - // the IncIter() table can be used to determine the length of ch's encoding: - size_t len = ms_utf8IterTable[(unsigned char)once.data[0]]; - - wxCharBuffer buf(n * len); - char *ptr = buf.data(); - for ( size_t i = 0; i < n; i++, ptr += len ) - { - memcpy(ptr, once.data, len); - } - - return buf; -} - - void wxString::PosLenToImpl(size_t pos, size_t len, size_t *implPos, size_t *implLen) const { @@ -507,7 +113,7 @@ void wxString::PosLenToImpl(size_t pos, size_t len, else { const_iterator i = begin() + pos; - *implPos = wxStringImpl::const_iterator(i) - m_impl.begin(); + *implPos = wxStringImpl::const_iterator(i.impl()) - m_impl.begin(); if ( len == npos ) *implLen = npos; else @@ -518,8 +124,7 @@ void wxString::PosLenToImpl(size_t pos, size_t len, if ( pos + len > length() ) len = length() - pos; - *implLen = wxStringImpl::const_iterator(i + len) - - wxStringImpl::const_iterator(i); + *implLen = (i + len).impl() - i.impl(); } } } @@ -804,7 +409,7 @@ bool wxString::Shrink() // deprecated compatibility code: #if WXWIN_COMPATIBILITY_2_8 && !wxUSE_STL_BASED_WXSTRING && !wxUSE_UNICODE_UTF8 -wxChar *wxString::GetWriteBuf(size_t nLen) +wxStringCharType *wxString::GetWriteBuf(size_t nLen) { return DoGetWriteBuf(nLen); } @@ -1350,19 +955,24 @@ wxString wxString::FromAscii(const char *ascii) if (!ascii) return wxEmptyString; - size_t len = strlen( ascii ); + size_t len = strlen(ascii); wxString res; if ( len ) { - wxStringBuffer buf(res, len); - - wchar_t *dest = buf; + wxImplStringBuffer buf(res, len); + wxStringCharType *dest = buf; for ( ;; ) { - if ( (*dest++ = (wchar_t)(unsigned char)*ascii++) == L'\0' ) - break; + 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; } } @@ -1373,35 +983,36 @@ wxString wxString::FromAscii(const char ascii) { // What do we do with '\0' ? - wxString res; - res += (wchar_t)(unsigned char) ascii; + unsigned char c = (unsigned char)ascii; - return res; + wxASSERT_MSG( c < 0x80, _T("Non-ASCII value passed to FromAscii().") ); + + // NB: the cast to wchar_t causes interpretation of 'ascii' as Latin1 value + return wxString(wxUniChar((wchar_t)c)); } const wxCharBuffer wxString::ToAscii() const { // this will allocate enough space for the terminating NUL too wxCharBuffer buffer(length()); - - char *dest = buffer.data(); - const wchar_t *pwc = c_str(); - for ( ;; ) + for ( const_iterator i = begin(); i != end(); ++i ) { - *dest++ = (char)(*pwc > SCHAR_MAX ? wxT('_') : *pwc); + wxUniChar c(*i); + // FIXME-UTF8: unify substituted char ('_') with wxUniChar ('?') + *dest++ = c.IsAscii() ? (char)c : '_'; // the output string can't have embedded NULs anyhow, so we can safely // stop at first of them even if we do have any - if ( !*pwc++ ) + if ( !c ) break; } return buffer; } -#endif // Unicode +#endif // wxUSE_UNICODE // extract string of length nCount starting at nFirst wxString wxString::Mid(size_t nFirst, size_t nCount) const @@ -1838,9 +1449,9 @@ bool wxString::ToDouble(double *val) const /* static */ #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN -wxString wxStringPrintfMixinBase::DoFormat(const wxChar *format, ...) +wxString wxStringPrintfMixinBase::DoFormat(const wxString& format, ...) #else -wxString wxString::DoFormat(const wxChar *format, ...) +wxString wxString::DoFormat(const wxString& format, ...) #endif { va_list argptr; @@ -1863,9 +1474,9 @@ wxString wxString::FormatV(const wxString& format, va_list argptr) } #ifdef wxNEEDS_WXSTRING_PRINTF_MIXIN -int wxStringPrintfMixinBase::DoPrintf(const wxChar *format, ...) +int wxStringPrintfMixinBase::DoPrintf(const wxString& format, ...) #else -int wxString::DoPrintf(const wxChar *format, ...) +int wxString::DoPrintf(const wxString& format, ...) #endif { va_list argptr; @@ -1887,14 +1498,27 @@ int wxString::DoPrintf(const wxChar *format, ...) return iLen; } -int wxString::PrintfV(const wxString& format, va_list argptr) +#if wxUSE_UNICODE_UTF8 +template +#else +// we only need one version in non-UTF8 builds and at least two Windows +// compilers have problems with this function template, so use just one +// normal function here +#endif +static int DoStringPrintfV(wxString& str, + const wxString& format, va_list argptr) { int size = 1024; for ( ;; ) { - wxStringBuffer tmp(*this, size + 1); +#if wxUSE_UNICODE_UTF8 + BufferType tmp(str, size + 1); + typename BufferType::CharType *buf = tmp; +#else + wxStringBuffer tmp(str, size + 1); wxChar *buf = tmp; +#endif if ( !buf ) { @@ -1906,7 +1530,7 @@ int wxString::PrintfV(const wxString& format, va_list argptr) // only a copy va_list argptrcopy; wxVaCopy(argptrcopy, argptr); - int len = wxVsnprintf(buf, size, (const wxChar*)/*FIXME-UTF8*/format, argptrcopy); + int len = wxVsnprintf(buf, size, format, argptrcopy); va_end(argptrcopy); // some implementations of vsnprintf() don't NUL terminate @@ -1950,9 +1574,37 @@ int wxString::PrintfV(const wxString& format, va_list argptr) } // we could have overshot - Shrink(); + str.Shrink(); + + return str.length(); +} + +int wxString::PrintfV(const wxString& format, va_list argptr) +{ + va_list argcopy; + wxVaCopy(argcopy, argptr); + +#if wxUSE_UNICODE_UTF8 + #if wxUSE_STL_BASED_WXSTRING + typedef wxStringTypeBuffer Utf8Buffer; + #else + typedef wxImplStringBuffer Utf8Buffer; + #endif +#endif - return length(); +#if wxUSE_UTF8_LOCALE_ONLY + return DoStringPrintfV(*this, format, argcopy); +#else + #if wxUSE_UNICODE_UTF8 + if ( wxLocaleIsUtf8 ) + return DoStringPrintfV(*this, format, argcopy); + else + // wxChar* version + return DoStringPrintfV(*this, format, argcopy); + #else + return DoStringPrintfV(*this, format, argcopy); + #endif // UTF8/WCHAR +#endif } // ----------------------------------------------------------------------------