X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/907173e51df6d2a91d65463ddf6ce810d3fd921a..3698ae71952311f7f0dc4594e0f666e07cc0a103:/src/common/strconv.cpp diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp index 667874bdef..db0147d6bc 100644 --- a/src/common/strconv.cpp +++ b/src/common/strconv.cpp @@ -1,12 +1,14 @@ ///////////////////////////////////////////////////////////////////////////// // Name: strconv.cpp // Purpose: Unicode conversion classes -// Author: Ove Kaaven, Robert Roebling, Vadim Zeitlin, Vaclav Slavik +// Author: Ove Kaaven, Robert Roebling, Vadim Zeitlin, Vaclav Slavik, +// Ryan Norton, Fredrik Roubert (UTF7) // Modified by: // Created: 29/01/98 // RCS-ID: $Id$ // Copyright: (c) 1999 Ove Kaaven, Robert Roebling, Vaclav Slavik // (c) 2000-2003 Vadim Zeitlin +// (c) 2004 Ryan Norton, Fredrik Roubert // Licence: wxWindows licence ///////////////////////////////////////////////////////////////////////////// @@ -53,6 +55,9 @@ #include #include #include +#ifdef HAVE_LANGINFO_H + #include +#endif #if defined(__WIN32__) && !defined(__WXMICROWIN__) #define wxHAVE_WIN32_MB2WC @@ -68,6 +73,7 @@ #ifdef HAVE_ICONV #include + #include "wx/thread.h" #endif #include "wx/encconv.h" @@ -168,7 +174,7 @@ static size_t decode_utf16(const wxUint16* input, wxUint32& output) wxMBConv::~wxMBConv() { - // nothing to do here + // nothing to do here (necessary for Darwin linking probably) } const wxWCharBuffer wxMBConv::cMB2WC(const char *psz) const @@ -215,6 +221,131 @@ const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *pwz) const return buf; } +const wxWCharBuffer wxMBConv::cMB2WC(const char *szString, size_t nStringLen, size_t* pOutSize) const +{ + wxASSERT(pOutSize != NULL); + + const char* szEnd = szString + nStringLen + 1; + const char* szPos = szString; + const char* szStart = szPos; + + size_t nActualLength = 0; + size_t nCurrentSize = nStringLen; //try normal size first (should never resize?) + + wxWCharBuffer theBuffer(nCurrentSize); + + //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 = MB2WC(NULL, szPos, 0); + + //Invalid conversion? + if( nLen == (size_t)-1 ) + { + *pOutSize = 0; + theBuffer.data()[0u] = wxT('\0'); + return theBuffer; + } + + + //Increase the actual length (+1 for current null character) + nActualLength += nLen + 1; + + //if buffer too big, realloc the buffer + if (nActualLength > (nCurrentSize+1)) + { + wxWCharBuffer theNewBuffer(nCurrentSize << 1); + memcpy(theNewBuffer.data(), theBuffer.data(), nCurrentSize * sizeof(wchar_t)); + theBuffer = theNewBuffer; + nCurrentSize <<= 1; + } + + //Convert the current (sub)string + if ( MB2WC(&theBuffer.data()[szPos - szStart], szPos, nLen + 1) == (size_t)-1 ) + { + *pOutSize = 0; + theBuffer.data()[0u] = wxT('\0'); + return theBuffer; + } + + //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; + } + + //success - return actual length and the buffer + *pOutSize = nActualLength; + return theBuffer; +} + +const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *szString, size_t nStringLen, size_t* pOutSize) const +{ + wxASSERT(pOutSize != NULL); + + const wchar_t* szEnd = szString + nStringLen + 1; + const wchar_t* szPos = szString; + const wchar_t* szStart = szPos; + + size_t nActualLength = 0; + size_t nCurrentSize = nStringLen << 2; //try * 4 first + + wxCharBuffer theBuffer(nCurrentSize); + + //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 = WC2MB(NULL, szPos, 0); + + //Invalid conversion? + if( nLen == (size_t)-1 ) + { + *pOutSize = 0; + theBuffer.data()[0u] = wxT('\0'); + return theBuffer; + } + + //Increase the actual length (+1 for current null character) + nActualLength += nLen + 1; + + //if buffer too big, realloc the buffer + if (nActualLength > (nCurrentSize+1)) + { + wxCharBuffer theNewBuffer(nCurrentSize << 1); + memcpy(theNewBuffer.data(), theBuffer.data(), nCurrentSize); + theBuffer = theNewBuffer; + nCurrentSize <<= 1; + } + + //Convert the current (sub)string + if(WC2MB(&theBuffer.data()[szPos - szStart], szPos, nLen + 1) == (size_t)-1 ) + { + *pOutSize = 0; + theBuffer.data()[0u] = wxT('\0'); + return theBuffer; + } + + //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; + } + + //success - return actual length and the buffer + *pOutSize = nActualLength; + return theBuffer; +} + // ---------------------------------------------------------------------------- // wxMBConvLibc // ---------------------------------------------------------------------------- @@ -230,32 +361,251 @@ size_t wxMBConvLibc::WC2MB(char *buf, const wchar_t *psz, size_t n) const } // ---------------------------------------------------------------------------- -// UTF-7 +// wxConvBrokenFileNames is made for GTK2 in Unicode mode when +// files are accidentally written in an encoding which is not +// the system encoding. Typically, the system encoding will be +// UTF8 but there might be files stored in ISO8859-1 on disk. // ---------------------------------------------------------------------------- -#if 0 -static char utf7_setD[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ" - "abcdefghijklmnopqrstuvwxyz" - "0123456789'(),-./:?"; -static char utf7_setO[]="!\"#$%&*;<=>@[]^_`{|}"; -static char utf7_setB[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ" - "abcdefghijklmnopqrstuvwxyz" - "0123456789+/"; +class wxConvBrokenFileNames: public wxMBConvLibc +{ +public: + wxConvBrokenFileNames() : m_utf8conv(wxMBConvUTF8::MAP_INVALID_UTF8_TO_OCTAL) { } + virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const; + virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const; + inline bool UseUTF8() const; +private: + wxMBConvUTF8 m_utf8conv; +}; + +bool wxConvBrokenFileNames::UseUTF8() const +{ +#if defined HAVE_LANGINFO_H && defined CODESET + char *codeset = nl_langinfo(CODESET); + return strcmp(codeset, "UTF-8") == 0; +#else + return false; #endif +} -// TODO: write actual implementations of UTF-7 here -size_t wxMBConvUTF7::MB2WC(wchar_t * WXUNUSED(buf), - const char * WXUNUSED(psz), - size_t WXUNUSED(n)) const +size_t wxConvBrokenFileNames::MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const { - return 0; + if (UseUTF8()) + return m_utf8conv.MB2WC( outputBuf, psz, outputSize ); + else + return wxMBConvLibc::MB2WC( outputBuf, psz, outputSize ); } -size_t wxMBConvUTF7::WC2MB(char * WXUNUSED(buf), - const wchar_t * WXUNUSED(psz), - size_t WXUNUSED(n)) const +size_t wxConvBrokenFileNames::WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const { - return 0; + if (UseUTF8()) + return m_utf8conv.WC2MB( outputBuf, psz, outputSize ); + else + return wxMBConvLibc::WC2MB( outputBuf, psz, outputSize ); +} + +// ---------------------------------------------------------------------------- +// UTF-7 +// ---------------------------------------------------------------------------- + +// Implementation (C) 2004 Fredrik Roubert + +// +// BASE64 decoding table +// +static const unsigned char utf7unb64[] = +{ + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0x3e, 0xff, 0xff, 0xff, 0x3f, + 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, 0x3a, 0x3b, + 0x3c, 0x3d, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, + 0x07, 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, + 0x0f, 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, + 0x17, 0x18, 0x19, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f, 0x20, + 0x21, 0x22, 0x23, 0x24, 0x25, 0x26, 0x27, 0x28, + 0x29, 0x2a, 0x2b, 0x2c, 0x2d, 0x2e, 0x2f, 0x30, + 0x31, 0x32, 0x33, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, + 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff, 0xff +}; + +size_t wxMBConvUTF7::MB2WC(wchar_t *buf, const char *psz, size_t n) const +{ + size_t len = 0; + + while (*psz && ((!buf) || (len < n))) + { + unsigned char cc = *psz++; + if (cc != '+') + { + // plain ASCII char + if (buf) + *buf++ = cc; + len++; + } + else if (*psz == '-') + { + // encoded plus sign + if (buf) + *buf++ = cc; + len++; + psz++; + } + else + { + // BASE64 encoded string + bool lsb; + unsigned char c; + unsigned int d, l; + for (lsb = false, d = 0, l = 0; + (cc = utf7unb64[(unsigned char)*psz]) != 0xff; psz++) + { + d <<= 6; + d += cc; + for (l += 6; l >= 8; lsb = !lsb) + { + c = (unsigned char)((d >> (l -= 8)) % 256); + if (lsb) + { + if (buf) + *buf++ |= c; + len ++; + } + else + if (buf) + *buf = (wchar_t)(c << 8); + } + } + if (*psz == '-') + psz++; + } + } + if (buf && (len < n)) + *buf = 0; + return len; +} + +// +// BASE64 encoding table +// +static const unsigned char utf7enb64[] = +{ + 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', + 'I', 'J', 'K', 'L', 'M', 'N', 'O', 'P', + 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', + 'Y', 'Z', 'a', 'b', 'c', 'd', 'e', 'f', + 'g', 'h', 'i', 'j', 'k', 'l', 'm', 'n', + 'o', 'p', 'q', 'r', 's', 't', 'u', 'v', + 'w', 'x', 'y', 'z', '0', '1', '2', '3', + '4', '5', '6', '7', '8', '9', '+', '/' +}; + +// +// UTF-7 encoding table +// +// 0 - Set D (directly encoded characters) +// 1 - Set O (optional direct characters) +// 2 - whitespace characters (optional) +// 3 - special characters +// +static const unsigned char utf7encode[128] = +{ + 3, 3, 3, 3, 3, 3, 3, 3, 3, 2, 2, 3, 3, 2, 3, 3, + 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, + 2, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 3, 0, 0, 0, 3, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, + 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 3, 1, 1, 1, + 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 3, 3 +}; + +size_t wxMBConvUTF7::WC2MB(char *buf, const wchar_t *psz, size_t n) const +{ + + + size_t len = 0; + + while (*psz && ((!buf) || (len < n))) + { + wchar_t cc = *psz++; + if (cc < 0x80 && utf7encode[cc] < 1) + { + // plain ASCII char + if (buf) + *buf++ = (char)cc; + len++; + } +#ifndef WC_UTF16 + else if (((wxUint32)cc) > 0xffff) + { + // no surrogate pair generation (yet?) + return (size_t)-1; + } +#endif + else + { + if (buf) + *buf++ = '+'; + len++; + if (cc != '+') + { + // BASE64 encode string + unsigned int lsb, d, l; + for (d = 0, l = 0;; psz++) + { + for (lsb = 0; lsb < 2; lsb ++) + { + d <<= 8; + d += lsb ? cc & 0xff : (cc & 0xff00) >> 8; + + for (l += 8; l >= 6; ) + { + l -= 6; + if (buf) + *buf++ = utf7enb64[(d >> l) % 64]; + len++; + } + } + cc = *psz; + if (!(cc) || (cc < 0x80 && utf7encode[cc] < 1)) + break; + } + if (l != 0) + { + if (buf) + *buf++ = utf7enb64[((d % 16) << (6 - l)) % 64]; + len++; + } + } + if (buf) + *buf++ = '-'; + len++; + } + } + if (buf && (len < n)) + *buf = 0; + return len; } // ---------------------------------------------------------------------------- @@ -265,12 +615,19 @@ size_t wxMBConvUTF7::WC2MB(char * WXUNUSED(buf), static wxUint32 utf8_max[]= { 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff, 0xffffffff }; +// boundaries of the private use area we use to (temporarily) remap invalid +// characters invalid in a UTF-8 encoded string +const wxUint32 wxUnicodePUA = 0x100000; +const wxUint32 wxUnicodePUAEnd = wxUnicodePUA + 256; + size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const { size_t len = 0; while (*psz && ((!buf) || (len < n))) { + const char *opsz = psz; + bool invalid = false; unsigned char cc = *psz++, fc = cc; unsigned cnt; for (cnt = 0; fc & 0x80; cnt++) @@ -288,7 +645,7 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const if (!cnt) { // invalid UTF-8 sequence - return (size_t)-1; + invalid = true; } else { @@ -296,32 +653,93 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const wxUint32 res = cc & (0x3f >> cnt); while (cnt--) { - cc = *psz++; + cc = *psz; if ((cc & 0xC0) != 0x80) { // invalid UTF-8 sequence - return (size_t)-1; + invalid = true; + break; } + psz++; res = (res << 6) | (cc & 0x3f); } - if (res <= utf8_max[ocnt]) + if (invalid || res <= utf8_max[ocnt]) { // illegal UTF-8 encoding - return (size_t)-1; + invalid = true; + } + else if ((m_options & MAP_INVALID_UTF8_TO_PUA) && + res >= wxUnicodePUA && res < wxUnicodePUAEnd) + { + // if one of our PUA characters turns up externally + // it must also be treated as an illegal sequence + // (a bit like you have to escape an escape character) + invalid = true; } + else + { #ifdef WC_UTF16 - // cast is ok because wchar_t == wxUuint16 if WC_UTF16 - size_t pa = encode_utf16(res, (wxUint16 *)buf); - if (pa == (size_t)-1) - return (size_t)-1; - if (buf) - buf += pa; - len += pa; + // cast is ok because wchar_t == wxUuint16 if WC_UTF16 + size_t pa = encode_utf16(res, (wxUint16 *)buf); + if (pa == (size_t)-1) + { + invalid = true; + } + else + { + if (buf) + buf += pa; + len += pa; + } #else // !WC_UTF16 - if (buf) - *buf++ = res; - len++; + if (buf) + *buf++ = res; + len++; #endif // WC_UTF16/!WC_UTF16 + } + } + if (invalid) + { + if (m_options & MAP_INVALID_UTF8_TO_PUA) + { + while (opsz < psz && (!buf || len < n)) + { +#ifdef WC_UTF16 + // cast is ok because wchar_t == wxUuint16 if WC_UTF16 + size_t pa = encode_utf16((unsigned char)*opsz + wxUnicodePUA, (wxUint16 *)buf); + wxASSERT(pa != (size_t)-1); + if (buf) + buf += pa; + opsz++; + len += pa; +#else + if (buf) + *buf++ = wxUnicodePUA + (unsigned char)*opsz; + opsz++; + len++; +#endif + } + } + else if (m_options & MAP_INVALID_UTF8_TO_OCTAL) + { + while (opsz < psz && (!buf || len < n)) + { + if ( buf && len + 3 < n ) + { + unsigned char n = *opsz; + *buf++ = L'\\'; + *buf++ = L'0' + n / 0100; + *buf++ = L'0' + (n % 0100) / 010; + *buf++ = L'0' + n % 010; + } + opsz++; + len += 4; + } + } + else // MAP_INVALID_UTF8_NOT + { + return (size_t)-1; + } } } } @@ -330,6 +748,11 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const return len; } +static inline bool isoctal(wchar_t wch) +{ + return L'0' <= wch && wch <= L'7'; +} + size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const { size_t len = 0; @@ -344,36 +767,59 @@ size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const #else cc=(*psz++) & 0x7fffffff; #endif - unsigned cnt; - for (cnt = 0; cc > utf8_max[cnt]; cnt++) {} - if (!cnt) + + if ( (m_options & MAP_INVALID_UTF8_TO_PUA) + && cc >= wxUnicodePUA && cc < wxUnicodePUAEnd ) { - // plain ASCII char if (buf) - *buf++ = (char) cc; + *buf++ = (char)(cc - wxUnicodePUA); len++; } + else if ( (m_options & MAP_INVALID_UTF8_TO_OCTAL) && + cc == L'\\' && + isoctal(psz[0]) && isoctal(psz[1]) && isoctal(psz[2]) ) + { + if (buf) + { + *buf++ = (char) (psz[0] - L'0')*0100 + + (psz[1] - L'0')*010 + + (psz[2] - L'0'); + } + psz += 3; + len++; + } else { - len += cnt + 1; - if (buf) + unsigned cnt; + for (cnt = 0; cc > utf8_max[cnt]; cnt++) {} + if (!cnt) { - *buf++ = (char) ((-128 >> cnt) | ((cc >> (cnt * 6)) & (0x3f >> cnt))); - while (cnt--) - *buf++ = (char) (0x80 | ((cc >> (cnt * 6)) & 0x3f)); + // plain ASCII char + if (buf) + *buf++ = (char) cc; + len++; + } + + else + { + len += cnt + 1; + if (buf) + { + *buf++ = (char) ((-128 >> cnt) | ((cc >> (cnt * 6)) & (0x3f >> cnt))); + while (cnt--) + *buf++ = (char) (0x80 | ((cc >> (cnt * 6)) & 0x3f)); + } } } } - if (buf && (lenWC conversion would fail "randomly". + wxMutexLocker lock(wxConstCast(this, wxMBConv_iconv)->m_iconvMutex); +#endif + size_t inbuf = strlen(psz); size_t outbuf = n * SIZEOF_WCHAR_T; size_t res, cres; @@ -1050,6 +1511,11 @@ size_t wxMBConv_iconv::MB2WC(wchar_t *buf, const char *psz, size_t n) const size_t wxMBConv_iconv::WC2MB(char *buf, const wchar_t *psz, size_t n) const { +#if wxUSE_THREADS + // NB: explained in MB2WC + wxMutexLocker lock(wxConstCast(this, wxMBConv_iconv)->m_iconvMutex); +#endif + size_t inbuf = wxWcslen(psz) * SIZEOF_WCHAR_T; size_t outbuf = n; size_t res, cres; @@ -1153,10 +1619,20 @@ public: // and break the library itself, e.g. wxTextInputStream::NextChar() // wouldn't work if reading an incomplete MB char didn't result in an // error + // + // note however that using MB_ERR_INVALID_CHARS with CP_UTF7 results in + // an error (tested under Windows Server 2003) and apparently it is + // done on purpose, i.e. the function accepts any input in this case + // and although I'd prefer to return error on ill-formed output, our + // own wxMBConvUTF7 doesn't detect errors (e.g. lone "+" which is + // explicitly ill-formed according to RFC 2152) neither so we don't + // even have any fallback here... + int flags = m_CodePage == CP_UTF7 ? 0 : MB_ERR_INVALID_CHARS; + const size_t len = ::MultiByteToWideChar ( m_CodePage, // code page - MB_ERR_INVALID_CHARS, // flags: fall on error + flags, // flags: fall on error psz, // input string -1, // its length (NUL-terminated) buf, // output string @@ -1286,6 +1762,418 @@ private: #endif // wxHAVE_WIN32_MB2WC +// ============================================================================ +// Cocoa conversion classes +// ============================================================================ + +#if defined(__WXCOCOA__) + +// RN: There is no UTF-32 support in either Core Foundation or +// Cocoa. Strangely enough, internally Core Foundation uses +// UTF 32 internally quite a bit - its just not public (yet). + +#include +#include + +CFStringEncoding wxCFStringEncFromFontEnc(wxFontEncoding encoding) +{ + CFStringEncoding enc = kCFStringEncodingInvalidId ; + if ( encoding == wxFONTENCODING_DEFAULT ) + { + enc = CFStringGetSystemEncoding(); + } + else switch( encoding) + { + case wxFONTENCODING_ISO8859_1 : + enc = kCFStringEncodingISOLatin1 ; + break ; + case wxFONTENCODING_ISO8859_2 : + enc = kCFStringEncodingISOLatin2; + break ; + case wxFONTENCODING_ISO8859_3 : + enc = kCFStringEncodingISOLatin3 ; + break ; + case wxFONTENCODING_ISO8859_4 : + enc = kCFStringEncodingISOLatin4; + break ; + case wxFONTENCODING_ISO8859_5 : + enc = kCFStringEncodingISOLatinCyrillic; + break ; + case wxFONTENCODING_ISO8859_6 : + enc = kCFStringEncodingISOLatinArabic; + break ; + case wxFONTENCODING_ISO8859_7 : + enc = kCFStringEncodingISOLatinGreek; + break ; + case wxFONTENCODING_ISO8859_8 : + enc = kCFStringEncodingISOLatinHebrew; + break ; + case wxFONTENCODING_ISO8859_9 : + enc = kCFStringEncodingISOLatin5; + break ; + case wxFONTENCODING_ISO8859_10 : + enc = kCFStringEncodingISOLatin6; + break ; + case wxFONTENCODING_ISO8859_11 : + enc = kCFStringEncodingISOLatinThai; + break ; + case wxFONTENCODING_ISO8859_13 : + enc = kCFStringEncodingISOLatin7; + break ; + case wxFONTENCODING_ISO8859_14 : + enc = kCFStringEncodingISOLatin8; + break ; + case wxFONTENCODING_ISO8859_15 : + enc = kCFStringEncodingISOLatin9; + break ; + + case wxFONTENCODING_KOI8 : + enc = kCFStringEncodingKOI8_R; + break ; + case wxFONTENCODING_ALTERNATIVE : // MS-DOS CP866 + enc = kCFStringEncodingDOSRussian; + break ; + +// case wxFONTENCODING_BULGARIAN : +// enc = ; +// break ; + + case wxFONTENCODING_CP437 : + enc =kCFStringEncodingDOSLatinUS ; + break ; + case wxFONTENCODING_CP850 : + enc = kCFStringEncodingDOSLatin1; + break ; + case wxFONTENCODING_CP852 : + enc = kCFStringEncodingDOSLatin2; + break ; + case wxFONTENCODING_CP855 : + enc = kCFStringEncodingDOSCyrillic; + break ; + case wxFONTENCODING_CP866 : + enc =kCFStringEncodingDOSRussian ; + break ; + case wxFONTENCODING_CP874 : + enc = kCFStringEncodingDOSThai; + break ; + case wxFONTENCODING_CP932 : + enc = kCFStringEncodingDOSJapanese; + break ; + case wxFONTENCODING_CP936 : + enc =kCFStringEncodingDOSChineseSimplif ; + break ; + case wxFONTENCODING_CP949 : + enc = kCFStringEncodingDOSKorean; + break ; + case wxFONTENCODING_CP950 : + enc = kCFStringEncodingDOSChineseTrad; + break ; + case wxFONTENCODING_CP1250 : + enc = kCFStringEncodingWindowsLatin2; + break ; + case wxFONTENCODING_CP1251 : + enc =kCFStringEncodingWindowsCyrillic ; + break ; + case wxFONTENCODING_CP1252 : + enc =kCFStringEncodingWindowsLatin1 ; + break ; + case wxFONTENCODING_CP1253 : + enc = kCFStringEncodingWindowsGreek; + break ; + case wxFONTENCODING_CP1254 : + enc = kCFStringEncodingWindowsLatin5; + break ; + case wxFONTENCODING_CP1255 : + enc =kCFStringEncodingWindowsHebrew ; + break ; + case wxFONTENCODING_CP1256 : + enc =kCFStringEncodingWindowsArabic ; + break ; + case wxFONTENCODING_CP1257 : + enc = kCFStringEncodingWindowsBalticRim; + break ; +// This only really encodes to UTF7 (if that) evidently +// case wxFONTENCODING_UTF7 : +// enc = kCFStringEncodingNonLossyASCII ; +// break ; + case wxFONTENCODING_UTF8 : + enc = kCFStringEncodingUTF8 ; + break ; + case wxFONTENCODING_EUC_JP : + enc = kCFStringEncodingEUC_JP; + break ; + case wxFONTENCODING_UTF16 : + enc = kCFStringEncodingUnicode ; + break ; + case wxFONTENCODING_MACROMAN : + enc = kCFStringEncodingMacRoman ; + break ; + case wxFONTENCODING_MACJAPANESE : + enc = kCFStringEncodingMacJapanese ; + break ; + case wxFONTENCODING_MACCHINESETRAD : + enc = kCFStringEncodingMacChineseTrad ; + break ; + case wxFONTENCODING_MACKOREAN : + enc = kCFStringEncodingMacKorean ; + break ; + case wxFONTENCODING_MACARABIC : + enc = kCFStringEncodingMacArabic ; + break ; + case wxFONTENCODING_MACHEBREW : + enc = kCFStringEncodingMacHebrew ; + break ; + case wxFONTENCODING_MACGREEK : + enc = kCFStringEncodingMacGreek ; + break ; + case wxFONTENCODING_MACCYRILLIC : + enc = kCFStringEncodingMacCyrillic ; + break ; + case wxFONTENCODING_MACDEVANAGARI : + enc = kCFStringEncodingMacDevanagari ; + break ; + case wxFONTENCODING_MACGURMUKHI : + enc = kCFStringEncodingMacGurmukhi ; + break ; + case wxFONTENCODING_MACGUJARATI : + enc = kCFStringEncodingMacGujarati ; + break ; + case wxFONTENCODING_MACORIYA : + enc = kCFStringEncodingMacOriya ; + break ; + case wxFONTENCODING_MACBENGALI : + enc = kCFStringEncodingMacBengali ; + break ; + case wxFONTENCODING_MACTAMIL : + enc = kCFStringEncodingMacTamil ; + break ; + case wxFONTENCODING_MACTELUGU : + enc = kCFStringEncodingMacTelugu ; + break ; + case wxFONTENCODING_MACKANNADA : + enc = kCFStringEncodingMacKannada ; + break ; + case wxFONTENCODING_MACMALAJALAM : + enc = kCFStringEncodingMacMalayalam ; + break ; + case wxFONTENCODING_MACSINHALESE : + enc = kCFStringEncodingMacSinhalese ; + break ; + case wxFONTENCODING_MACBURMESE : + enc = kCFStringEncodingMacBurmese ; + break ; + case wxFONTENCODING_MACKHMER : + enc = kCFStringEncodingMacKhmer ; + break ; + case wxFONTENCODING_MACTHAI : + enc = kCFStringEncodingMacThai ; + break ; + case wxFONTENCODING_MACLAOTIAN : + enc = kCFStringEncodingMacLaotian ; + break ; + case wxFONTENCODING_MACGEORGIAN : + enc = kCFStringEncodingMacGeorgian ; + break ; + case wxFONTENCODING_MACARMENIAN : + enc = kCFStringEncodingMacArmenian ; + break ; + case wxFONTENCODING_MACCHINESESIMP : + enc = kCFStringEncodingMacChineseSimp ; + break ; + case wxFONTENCODING_MACTIBETAN : + enc = kCFStringEncodingMacTibetan ; + break ; + case wxFONTENCODING_MACMONGOLIAN : + enc = kCFStringEncodingMacMongolian ; + break ; + case wxFONTENCODING_MACETHIOPIC : + enc = kCFStringEncodingMacEthiopic ; + break ; + case wxFONTENCODING_MACCENTRALEUR : + enc = kCFStringEncodingMacCentralEurRoman ; + break ; + case wxFONTENCODING_MACVIATNAMESE : + enc = kCFStringEncodingMacVietnamese ; + break ; + case wxFONTENCODING_MACARABICEXT : + enc = kCFStringEncodingMacExtArabic ; + break ; + case wxFONTENCODING_MACSYMBOL : + enc = kCFStringEncodingMacSymbol ; + break ; + case wxFONTENCODING_MACDINGBATS : + enc = kCFStringEncodingMacDingbats ; + break ; + case wxFONTENCODING_MACTURKISH : + enc = kCFStringEncodingMacTurkish ; + break ; + case wxFONTENCODING_MACCROATIAN : + enc = kCFStringEncodingMacCroatian ; + break ; + case wxFONTENCODING_MACICELANDIC : + enc = kCFStringEncodingMacIcelandic ; + break ; + case wxFONTENCODING_MACROMANIAN : + enc = kCFStringEncodingMacRomanian ; + break ; + case wxFONTENCODING_MACCELTIC : + enc = kCFStringEncodingMacCeltic ; + break ; + case wxFONTENCODING_MACGAELIC : + enc = kCFStringEncodingMacGaelic ; + break ; +// case wxFONTENCODING_MACKEYBOARD : +// enc = kCFStringEncodingMacKeyboardGlyphs ; +// break ; + default : + // because gcc is picky + break ; + } ; + return enc ; +} + +class wxMBConv_cocoa : public wxMBConv +{ +public: + wxMBConv_cocoa() + { + Init(CFStringGetSystemEncoding()) ; + } + +#if wxUSE_FONTMAP + wxMBConv_cocoa(const wxChar* name) + { + Init( wxCFStringEncFromFontEnc(wxFontMapperBase::Get()->CharsetToEncoding(name, false) ) ) ; + } +#endif + + wxMBConv_cocoa(wxFontEncoding encoding) + { + Init( wxCFStringEncFromFontEnc(encoding) ); + } + + ~wxMBConv_cocoa() + { + } + + void Init( CFStringEncoding encoding) + { + m_encoding = encoding ; + } + + size_t MB2WC(wchar_t * szOut, const char * szUnConv, size_t nOutSize) const + { + wxASSERT(szUnConv); + + CFStringRef theString = CFStringCreateWithBytes ( + NULL, //the allocator + (const UInt8*)szUnConv, + strlen(szUnConv), + m_encoding, + false //no BOM/external representation + ); + + wxASSERT(theString); + + size_t nOutLength = CFStringGetLength(theString); + + if (szOut == NULL) + { + CFRelease(theString); + return nOutLength; + } + + CFRange theRange = { 0, nOutSize }; + +#if SIZEOF_WCHAR_T == 4 + UniChar* szUniCharBuffer = new UniChar[nOutSize]; +#endif + + CFStringGetCharacters(theString, theRange, szUniCharBuffer); + + CFRelease(theString); + + szUniCharBuffer[nOutLength] = '\0' ; + +#if SIZEOF_WCHAR_T == 4 + wxMBConvUTF16 converter ; + converter.MB2WC(szOut, (const char*)szUniCharBuffer , nOutSize ) ; + delete[] szUniCharBuffer; +#endif + + return nOutLength; + } + + size_t WC2MB(char *szOut, const wchar_t *szUnConv, size_t nOutSize) const + { + wxASSERT(szUnConv); + + size_t nRealOutSize; + size_t nBufSize = wxWcslen(szUnConv); + UniChar* szUniBuffer = (UniChar*) szUnConv; + +#if SIZEOF_WCHAR_T == 4 + wxMBConvUTF16BE converter ; + nBufSize = converter.WC2MB( NULL , szUnConv , 0 ); + szUniBuffer = new UniChar[ (nBufSize / sizeof(UniChar)) + 1] ; + converter.WC2MB( (char*) szUniBuffer , szUnConv, nBufSize + sizeof(UniChar)) ; + nBufSize /= sizeof(UniChar); +#endif + + CFStringRef theString = CFStringCreateWithCharactersNoCopy( + NULL, //allocator + szUniBuffer, + nBufSize, + kCFAllocatorNull //deallocator - we want to deallocate it ourselves + ); + + wxASSERT(theString); + + //Note that CER puts a BOM when converting to unicode + //so we check and use getchars instead in that case + if (m_encoding == kCFStringEncodingUnicode) + { + if (szOut != NULL) + CFStringGetCharacters(theString, CFRangeMake(0, nOutSize - 1), (UniChar*) szOut); + + nRealOutSize = CFStringGetLength(theString) + 1; + } + else + { + CFStringGetBytes( + theString, + CFRangeMake(0, CFStringGetLength(theString)), + m_encoding, + 0, //what to put in characters that can't be converted - + //0 tells CFString to return NULL if it meets such a character + false, //not an external representation + (UInt8*) szOut, + nOutSize, + (CFIndex*) &nRealOutSize + ); + } + + CFRelease(theString); + +#if SIZEOF_WCHAR_T == 4 + delete[] szUniBuffer; +#endif + + return nRealOutSize - 1; + } + + bool IsOk() const + { + return m_encoding != kCFStringEncodingInvalidId && + CFStringIsEncodingAvailable(m_encoding); + } + +private: + CFStringEncoding m_encoding ; +}; + +#endif // defined(__WXCOCOA__) + // ============================================================================ // Mac conversion classes // ============================================================================ @@ -1300,10 +2188,12 @@ public: Init(CFStringGetSystemEncoding()) ; } +#if wxUSE_FONTMAP wxMBConv_mac(const wxChar* name) { - Init( wxMacGetSystemEncFromFontEnc(wxFontMapper::Get()->CharsetToEncoding(name, false) ) ) ; + Init( wxMacGetSystemEncFromFontEnc(wxFontMapperBase::Get()->CharsetToEncoding(name, false) ) ) ; } +#endif wxMBConv_mac(wxFontEncoding encoding) { @@ -1343,7 +2233,8 @@ public: if (buf == NULL) { - n = byteInLen ; + //apple specs say at least 32 + n = wxMax( 32 , byteInLen ) ; tbuf = (wchar_t*) malloc( n * SIZEOF_WCHAR_T) ; } ByteCount byteBufferLen = n * sizeof( UniChar ) ; @@ -1383,8 +2274,8 @@ public: if (buf == NULL) { - // worst case - n = byteInLen * 2 ; + //apple specs say at least 32 + n = wxMax( 32 , ((byteInLen / SIZEOF_WCHAR_T) * 8) + SIZEOF_WCHAR_T ); tbuf = (char*) malloc( n ) ; } @@ -1409,8 +2300,23 @@ public: size_t res = byteOutLen ; if ( buf && res < n) + { buf[res] = 0; + //we need to double-trip to verify it didn't insert any ? in place + //of bogus characters + wxWCharBuffer wcBuf(n); + size_t pszlen = wxWcslen(psz); + if ( MB2WC(wcBuf.data(), buf, n) == (size_t)-1 || + wxWcslen(wcBuf) != pszlen || + memcmp(wcBuf, psz, pszlen * sizeof(wchar_t)) != 0 ) + { + // we didn't obtain the same thing we started from, hence + // the conversion was lossy and we consider that it failed + return (size_t)-1; + } + } + return res ; } @@ -1448,7 +2354,7 @@ public: wxMBConv_wxwin(const wxChar* name) { if (name) - m_enc = wxFontMapper::Get()->CharsetToEncoding(name, false); + m_enc = wxFontMapperBase::Get()->CharsetToEncoding(name, false); else m_enc = wxFONTENCODING_SYSTEM; @@ -1466,7 +2372,10 @@ public: { size_t inbuf = strlen(psz); if (buf) - m2w.Convert(psz,buf); + { + if (!m2w.Convert(psz,buf)) + return (size_t)-1; + } return inbuf; } @@ -1474,7 +2383,10 @@ public: { const size_t inbuf = wxWcslen(psz); if (buf) - w2m.Convert(psz,buf); + { + if (!w2m.Convert(psz,buf)) + return (size_t)-1; + } return inbuf; } @@ -1601,7 +2513,7 @@ wxMBConv *wxCSConv::DoCreate() const #if wxUSE_FONTMAP if ( name.empty() ) - name = wxFontMapper::Get()->GetEncodingName(m_encoding); + name = wxFontMapperBase::Get()->GetEncodingName(m_encoding); #endif // wxUSE_FONTMAP wxMBConv_iconv *conv = new wxMBConv_iconv(name); @@ -1628,11 +2540,35 @@ wxMBConv *wxCSConv::DoCreate() const #endif // wxHAVE_WIN32_MB2WC #if defined(__WXMAC__) { - if ( m_name || ( m_encoding < wxFONTENCODING_UTF16BE ) ) + // leave UTF16 and UTF32 to the built-ins of wx + if ( m_name || ( m_encoding < wxFONTENCODING_UTF16BE || + ( m_encoding >= wxFONTENCODING_MACMIN && m_encoding <= wxFONTENCODING_MACMAX ) ) ) { +#if wxUSE_FONTMAP wxMBConv_mac *conv = m_name ? new wxMBConv_mac(m_name) : new wxMBConv_mac(m_encoding); +#else + wxMBConv_mac *conv = new wxMBConv_mac(m_encoding); +#endif + if ( conv->IsOk() ) + return conv; + + delete conv; + } + } +#endif +#if defined(__WXCOCOA__) + { + if ( m_name || ( m_encoding <= wxFONTENCODING_UTF16 ) ) + { + +#if wxUSE_FONTMAP + wxMBConv_cocoa *conv = m_name ? new wxMBConv_cocoa(m_name) + : new wxMBConv_cocoa(m_encoding); +#else + wxMBConv_cocoa *conv = new wxMBConv_cocoa(m_encoding); +#endif if ( conv->IsOk() ) return conv; @@ -1648,7 +2584,7 @@ wxMBConv *wxCSConv::DoCreate() const // use "false" to suppress interactive dialogs -- we can be called from // anywhere and popping up a dialog from here is the last thing we want to // do - enc = wxFontMapper::Get()->CharsetToEncoding(m_name, false); + enc = wxFontMapperBase::Get()->CharsetToEncoding(m_name, false); } #endif // wxUSE_FONTMAP @@ -1704,7 +2640,7 @@ wxMBConv *wxCSConv::DoCreate() const m_name ? m_name : #if wxUSE_FONTMAP - wxFontMapper::GetEncodingDescription(m_encoding).c_str() + wxFontMapperBase::GetEncodingDescription(m_encoding).c_str() #else // !wxUSE_FONTMAP wxString::Format(_("encoding %s"), m_encoding).c_str() #endif // wxUSE_FONTMAP/!wxUSE_FONTMAP @@ -1800,7 +2736,7 @@ static wxCSConv wxConvLocalObj(wxFONTENCODING_SYSTEM); static wxCSConv wxConvISO8859_1Obj(wxFONTENCODING_ISO8859_1); static wxMBConvUTF7 wxConvUTF7Obj; static wxMBConvUTF8 wxConvUTF8Obj; - +static wxConvBrokenFileNames wxConvBrokenFileNamesObj; WXDLLIMPEXP_DATA_BASE(wxMBConv&) wxConvLibc = wxConvLibcObj; WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvLocal = wxConvLocalObj; @@ -1808,6 +2744,15 @@ WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvISO8859_1 = wxConvISO8859_1Obj; WXDLLIMPEXP_DATA_BASE(wxMBConvUTF7&) wxConvUTF7 = wxConvUTF7Obj; WXDLLIMPEXP_DATA_BASE(wxMBConvUTF8&) wxConvUTF8 = wxConvUTF8Obj; WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent = &wxConvLibcObj; +WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvFileName = & +#ifdef __WXOSX__ + wxConvUTF8Obj; +#elif __WXGTK20__ + wxConvBrokenFileNamesObj; +#else + wxConvLibcObj; +#endif + #else // !wxUSE_WCHAR_T