X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/f6bcfd974ef26faf6f91a62cac09827e09463fd1..eccd1992237c9b8bbe02a9eda66b709fd7ecea73:/src/common/strconv.cpp diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp index 997899e769..bcb1a76a3a 100644 --- a/src/common/strconv.cpp +++ b/src/common/strconv.cpp @@ -1,12 +1,13 @@ ///////////////////////////////////////////////////////////////////////////// // Name: strconv.cpp // Purpose: Unicode conversion classes -// Author: Ove Kaaven, Robert Roebling, Vadim Zeitlin +// Author: Ove Kaaven, Robert Roebling, Vadim Zeitlin, Vaclav Slavik // Modified by: // Created: 29/01/98 // RCS-ID: $Id$ -// Copyright: (c) 1999 Ove Kaaven, Robert Roebling, Vadim Zeitlin -// Licence: wxWindows license +// Copyright: (c) 1999 Ove Kaaven, Robert Roebling, Vaclav Slavik +// (c) 2000-2003 Vadim Zeitlin +// Licence: wxWindows licence ///////////////////////////////////////////////////////////////////////////// // ============================================================================ @@ -17,7 +18,7 @@ // headers // ---------------------------------------------------------------------------- -#ifdef __GNUG__ +#if defined(__GNUG__) && !defined(NO_GCC_PRAGMA) #pragma implementation "strconv.h" #endif @@ -28,134 +29,193 @@ #pragma hdrstop #endif +#ifndef WX_PRECOMP + #include "wx/intl.h" + #include "wx/log.h" +#endif // WX_PRECOMP + +#include "wx/strconv.h" + +#if wxUSE_WCHAR_T + +#ifdef __WXMSW__ + #include "wx/msw/private.h" +#endif + +#ifndef __WXWINCE__ +#include +#endif + #include #include #include +#if defined(__WIN32__) && !defined(__WXMICROWIN__) + #define wxHAVE_WIN32_MB2WC +#endif // __WIN32__ but !__WXMICROWIN__ + +// ---------------------------------------------------------------------------- +// headers +// ---------------------------------------------------------------------------- + #ifdef __SALFORDC__ - #include + #include #endif -#include "wx/debug.h" -#include "wx/strconv.h" +#ifdef HAVE_ICONV + #include +#endif + +#include "wx/encconv.h" +#include "wx/fontmap.h" // ---------------------------------------------------------------------------- -// globals +// macros // ---------------------------------------------------------------------------- -WXDLLEXPORT_DATA(wxMBConv *) wxConvCurrent = &wxConvLibc; +#define BSWAP_UCS4(str, len) { unsigned _c; for (_c=0; _c=0x110000) { - size_t nLen = MB2WC((wchar_t *) NULL, psz, 0); - if (nLen == (size_t)-1) - return wxWCharBuffer((wchar_t *) NULL); - wxWCharBuffer buf(nLen); - MB2WC((wchar_t *)(const wchar_t *) buf, psz, nLen); - return buf; + return (size_t)-1; } else - return wxWCharBuffer((wchar_t *) NULL); + { + if (output) + { + *output++ = (wxUint16) ((input >> 10)+0xd7c0); + *output++ = (wxUint16) ((input&0x3ff)+0xdc00); + } + return 2; + } } -const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *psz) const +static size_t decode_utf16(const wxUint16* input, wxUint32& output) { - if (psz) + if ((*input<0xd800) || (*input>0xdfff)) { - size_t nLen = WC2MB((char *) NULL, psz, 0); - if (nLen == (size_t)-1) - return wxCharBuffer((char *) NULL); - wxCharBuffer buf(nLen); - WC2MB((char *)(const char *) buf, psz, nLen); - return buf; + output = *input; + return 1; + } + else if ((input[1]<0xdc00) || (input[1]>=0xdfff)) + { + output = *input; + return (size_t)-1; } else - return wxCharBuffer((char *) NULL); + { + output = ((input[0] - 0xd7c0) << 10) + (input[1] - 0xdc00); + return 2; + } } + // ---------------------------------------------------------------------------- -// standard file conversion +// wxMBConv // ---------------------------------------------------------------------------- -WXDLLEXPORT_DATA(wxMBConvFile) wxConvFile; - -// just use the libc conversion for now -size_t wxMBConvFile::MB2WC(wchar_t *buf, const char *psz, size_t n) const +wxMBConv::~wxMBConv() { - return wxMB2WC(buf, psz, n); + // nothing to do here } -size_t wxMBConvFile::WC2MB(char *buf, const wchar_t *psz, size_t n) const +const wxWCharBuffer wxMBConv::cMB2WC(const char *psz) const { - return wxWC2MB(buf, psz, n); + if ( psz ) + { + // calculate the length of the buffer needed first + size_t nLen = MB2WC(NULL, psz, 0); + if ( nLen != (size_t)-1 ) + { + // now do the actual conversion + wxWCharBuffer buf(nLen); + MB2WC(buf.data(), psz, nLen + 1); // with the trailing NUL + + return buf; + } + } + + wxWCharBuffer buf((wchar_t *)NULL); + + return buf; } -// ---------------------------------------------------------------------------- -// standard gdk conversion -// ---------------------------------------------------------------------------- +const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *pwz) const +{ + if ( pwz ) + { + size_t nLen = WC2MB(NULL, pwz, 0); + if ( nLen != (size_t)-1 ) + { + wxCharBuffer buf(nLen+3); // space for a wxUint32 trailing zero + WC2MB(buf.data(), pwz, nLen + 4); + + return buf; + } + } -#ifdef __WXGTK12__ + wxCharBuffer buf((char *)NULL); -WXDLLEXPORT_DATA(wxMBConvGdk) wxConvGdk; + return buf; +} -#include +// ---------------------------------------------------------------------------- +// wxMBConvLibc +// ---------------------------------------------------------------------------- -size_t wxMBConvGdk::MB2WC(wchar_t *buf, const char *psz, size_t n) const +size_t wxMBConvLibc::MB2WC(wchar_t *buf, const char *psz, size_t n) const { - if (buf) { - return gdk_mbstowcs((GdkWChar *)buf, psz, n); - } else { - GdkWChar *nbuf = new GdkWChar[n=strlen(psz)]; - size_t len = gdk_mbstowcs(nbuf, psz, n); - delete [] nbuf; - return len; - } + return wxMB2WC(buf, psz, n); } -size_t wxMBConvGdk::WC2MB(char *buf, const wchar_t *psz, size_t n) const +size_t wxMBConvLibc::WC2MB(char *buf, const wchar_t *psz, size_t n) const { - char *mbstr = gdk_wcstombs((GdkWChar *)psz); - size_t len = mbstr ? strlen(mbstr) : 0; - if (buf) { - if (len > n) len = n; - memcpy(buf, psz, len); - if (len < n) buf[len] = 0; - } - return len; + return wxWC2MB(buf, psz, n); } -#endif // GTK > 1.0 - // ---------------------------------------------------------------------------- // UTF-7 // ---------------------------------------------------------------------------- -WXDLLEXPORT_DATA(wxMBConvUTF7) wxConvUTF7; - #if 0 static char utf7_setD[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ" "abcdefghijklmnopqrstuvwxyz" @@ -185,191 +245,1298 @@ size_t wxMBConvUTF7::WC2MB(char * WXUNUSED(buf), // UTF-8 // ---------------------------------------------------------------------------- -WXDLLEXPORT_DATA(wxMBConvUTF8) wxConvUTF8; - -static unsigned long utf8_max[]={0x7f,0x7ff,0xffff,0x1fffff,0x3ffffff,0x7fffffff,0xffffffff}; +static wxUint32 utf8_max[]= + { 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff, 0xffffffff }; size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const { - size_t len = 0; - - while (*psz && ((!buf) || (len>cnt); - while (cnt--) { - cc = *psz++; - if ((cc&0xC0)!=0x80) { - // invalid UTF-8 sequence - return (size_t)-1; - } - res=(res<<6)|(cc&0x3f); + size_t len = 0; + + while (*psz && ((!buf) || (len < n))) + { + unsigned char cc = *psz++, fc = cc; + unsigned cnt; + for (cnt = 0; fc & 0x80; cnt++) + fc <<= 1; + if (!cnt) + { + // plain ASCII char + if (buf) + *buf++ = cc; + len++; } - if (res<=utf8_max[ocnt]) { - // illegal UTF-8 encoding - return (size_t)-1; + else + { + cnt--; + if (!cnt) + { + // invalid UTF-8 sequence + return (size_t)-1; + } + else + { + unsigned ocnt = cnt - 1; + wxUint32 res = cc & (0x3f >> cnt); + while (cnt--) + { + cc = *psz++; + if ((cc & 0xC0) != 0x80) + { + // invalid UTF-8 sequence + return (size_t)-1; + } + res = (res << 6) | (cc & 0x3f); + } + if (res <= utf8_max[ocnt]) + { + // illegal UTF-8 encoding + return (size_t)-1; + } +#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; +#else // !WC_UTF16 + if (buf) + *buf++ = res; + len++; +#endif // WC_UTF16/!WC_UTF16 + } } - if (buf) *buf++=res; - len++; - } } - } - if (buf && (lenutf8_max[cnt]; cnt++); - if (!cnt) { - // plain ASCII char - if (buf) *buf++=cc; - len++; - } else { - len+=cnt+1; - if (buf) { - *buf++=(-128>>cnt)|((cc>>(cnt*6))&(0x3f>>cnt)); - while (cnt--) - *buf++=0x80|((cc>>(cnt*6))&0x3f); - } - } - } - if (buf && (len utf8_max[cnt]; cnt++) {} + if (!cnt) + { + // 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 && (lenCharsetToEncoding(name, FALSE) - : wxFONTENCODING_SYSTEM; - wxCharacterSet *cset = (enc != wxFONTENCODING_SYSTEM) ? new wxCharacterSet(enc) - : (wxCharacterSet *)NULL; - return cset; + size_t len=0; + + while (*psz && (!buf || len < n)) + { + if (buf) + { + *(wxUint16*)buf = *psz; + buf += sizeof(wxUint16); + } + len += sizeof(wxUint16); + psz++; + } + if (buf && len<=n-sizeof(wxUint16)) *(wxUint16*)buf=0; + + return len; } -wxCSConv::wxCSConv(const wxChar *charset) + +// swap 16bit MB to 16bit String +size_t wxMBConvUTF16swap::MB2WC(wchar_t *buf, const char *psz, size_t n) const { - m_name = (wxChar *) NULL; - m_cset = (wxCharacterSet *) NULL; - m_deferred = TRUE; - SetName(charset); + size_t len=0; + + while (*(wxUint16*)psz && (!buf || len < n)) + { + if (buf) + { + ((char *)buf)[0] = psz[1]; + ((char *)buf)[1] = psz[0]; + buf++; + } + len++; + psz += sizeof(wxUint16); + } + if (buf && len 1) + { + *(wxUint16*)buf = cc[1]; + buf += sizeof(wxUint16); + } + } + + len += pa*sizeof(wxUint16); + psz++; } - m_cset = wxGetCharacterSet(m_name); - m_deferred = FALSE; - } + if (buf && len<=n-sizeof(wxUint16)) *(wxUint16*)buf=0; + + return len; } -size_t wxCSConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const + +// swap 16bit MB to 32bit String +size_t wxMBConvUTF16swap::MB2WC(wchar_t *buf, const char *psz, size_t n) const { - ((wxCSConv *)this)->LoadNow(); // discard constness - if (buf) { - if (m_cset) { - m_cset->m2w.Convert(psz, buf); - } else { - // latin-1 (direct) - for (size_t c=0; cLoadNow(); // discard constness - if (buf) { - if (m_cset) { - m_cset->w2m.Convert(psz, buf); - } else { - // latin-1 (direct) - for (size_t c=0; c0xff) ? '?' : psz[c]; - } - return n; - } -#if defined(__BORLANDC__) && (__BORLANDC__ > 0x530) - return std::wcslen(psz); -#else - return ::wcslen(psz); -#endif + size_t len=0; + + while (*psz && (!buf || len < n)) + { + wxUint16 cc[2]; + size_t pa=encode_utf16(*psz, cc); + + if (pa == (size_t)-1) + return pa; + + if (buf) + { + *buf++ = ((char*)cc)[1]; + *buf++ = ((char*)cc)[0]; + if (pa > 1) + { + *buf++ = ((char*)cc)[3]; + *buf++ = ((char*)cc)[2]; + } + } + + len += pa*sizeof(wxUint16); + psz++; + } + if (buf && len<=n-sizeof(wxUint16)) *(wxUint16*)buf=0; + + return len; } -#else // !wxUSE_WCHAR_T +#endif // WC_UTF16 + // ---------------------------------------------------------------------------- -// stand-ins in absence of wchar_t +// UTF-32 // ---------------------------------------------------------------------------- -WXDLLEXPORT_DATA(wxMBConv) wxConvLibc, wxConvFile; +#ifdef WORDS_BIGENDIAN +#define wxMBConvUTF32straight wxMBConvUTF32BE +#define wxMBConvUTF32swap wxMBConvUTF32LE +#else +#define wxMBConvUTF32swap wxMBConvUTF32BE +#define wxMBConvUTF32straight wxMBConvUTF32LE +#endif + + +WXDLLIMPEXP_DATA_BASE(wxMBConvUTF32LE) wxConvUTF32LE; +WXDLLIMPEXP_DATA_BASE(wxMBConvUTF32BE) wxConvUTF32BE; + + +#ifdef WC_UTF16 + +// copy 32bit MB to 16bit String +size_t wxMBConvUTF32straight::MB2WC(wchar_t *buf, const char *psz, size_t n) const +{ + size_t len=0; + + while (*(wxUint32*)psz && (!buf || len < n)) + { + wxUint16 cc[2]; + + size_t pa=encode_utf16(*(wxUint32*)psz, cc); + if (pa == (size_t)-1) + return pa; + + if (buf) + { + *buf++ = cc[0]; + if (pa > 1) + *buf++ = cc[1]; + } + len += pa; + psz += sizeof(wxUint32); + } + if (buf && len 1) + *buf++ = cc[1]; + } + len += pa; + psz += sizeof(wxUint32); + } + + if (buf && lenCharsetToEncoding(name, false); + else + m_enc = wxFONTENCODING_SYSTEM; + + Init(); + } + + wxMBConv_wxwin(wxFontEncoding enc) + { + m_enc = enc; + + Init(); + } + + size_t MB2WC(wchar_t *buf, const char *psz, size_t WXUNUSED(n)) const + { + size_t inbuf = strlen(psz); + if (buf) + m2w.Convert(psz,buf); + return inbuf; + } + + size_t WC2MB(char *buf, const wchar_t *psz, size_t WXUNUSED(n)) const + { + const size_t inbuf = wxWcslen(psz); + if (buf) + w2m.Convert(psz,buf); + + return inbuf; + } + + bool IsOk() const { return m_ok; } + +public: + wxFontEncoding m_enc; + wxEncodingConverter m2w, w2m; + + // were we initialized successfully? + bool m_ok; + + DECLARE_NO_COPY_CLASS(wxMBConv_wxwin) +}; + +#endif // wxUSE_FONTMAP + +// ============================================================================ +// wxCSConv implementation +// ============================================================================ + +void wxCSConv::Init() +{ + m_name = NULL; + m_convReal = NULL; + m_deferred = true; +} + +// find a valid value for the encoding +void wxCSConv::SetEncoding() +{ +#if wxUSE_INTL + m_encoding = wxLocale::GetSystemEncoding(); +#else + m_encoding = wxFONTENCODING_SYSTEM; +#endif +} + +wxCSConv::wxCSConv(const wxChar *charset) +{ + Init(); + + if ( charset ) + { + // not used + m_encoding = wxFONTENCODING_SYSTEM; + + SetName(charset); + } + else // no charset specified + { + SetEncoding(); + } +} + +wxCSConv::wxCSConv(wxFontEncoding encoding) +{ + if ( encoding == wxFONTENCODING_MAX || + encoding == wxFONTENCODING_DEFAULT ) + { + wxFAIL_MSG( _T("invalid encoding value in wxCSConv ctor") ); + + encoding = wxFONTENCODING_SYSTEM; + } + + Init(); + + if ( encoding == wxFONTENCODING_SYSTEM ) + { + SetEncoding(); + } + else // have valid encoding, use it + { + m_encoding = encoding; + } +} + +wxCSConv::~wxCSConv() +{ + Clear(); +} + +wxCSConv::wxCSConv(const wxCSConv& conv) + : wxMBConv() +{ + Init(); + + SetName(conv.m_name); + m_encoding = conv.m_encoding; +} + +wxCSConv& wxCSConv::operator=(const wxCSConv& conv) +{ + Clear(); + + SetName(conv.m_name); + m_encoding = conv.m_encoding; + + return *this; +} + +void wxCSConv::Clear() +{ + free(m_name); + delete m_convReal; + + m_name = NULL; + m_convReal = NULL; +} + +void wxCSConv::SetName(const wxChar *charset) +{ + if (charset) + { + m_name = wxStrdup(charset); + m_deferred = true; + } +} + +static inline bool DoesntNeedConv(wxFontEncoding enc) +{ + return enc == wxFONTENCODING_DEFAULT || + enc == wxFONTENCODING_SYSTEM || + enc == wxFONTENCODING_ISO8859_1; +} + +wxMBConv *wxCSConv::DoCreate() const +{ +#if wxUSE_FONTMAP + wxFontMapper * const fontMapper = wxFontMapper::Get(); + + wxFontEncoding encFromName = m_name ? fontMapper->CharsetToEncoding(m_name) + : wxFONTENCODING_SYSTEM; +#endif // wxUSE_FONTMAP + + // check for the special case of ASCII charset + if ( (!m_name && DoesntNeedConv(m_encoding)) +#if wxUSE_FONTMAP + || (m_name && DoesntNeedConv(encFromName)) +#endif // wxUSE_FONTMAP + ) + { + // don't convert at all + return NULL; + } + + // we trust OS to do conversion better than we can so try external + // conversion methods first + // + // the full order is: + // 1. OS conversion (iconv() under Unix or Win32 API) + // 2. hard coded conversions for UTF + // 3. wxEncodingConverter as fall back + + // step (1) +#ifdef HAVE_ICONV + if ( m_name ) + { + wxMBConv_iconv *conv = new wxMBConv_iconv(m_name); + if ( conv->IsOk() ) + return conv; + + delete conv; + } +#endif // HAVE_ICONV + +#ifdef wxHAVE_WIN32_MB2WC + { + wxMBConv_win32 *conv = m_name ? new wxMBConv_win32(m_name) + : new wxMBConv_win32(m_encoding); + if ( conv->IsOk() ) + return conv; + + delete conv; + } +#endif // wxHAVE_WIN32_MB2WC + + // step (2) + wxFontEncoding enc = m_encoding; +#if wxUSE_FONTMAP + if ( enc == wxFONTENCODING_SYSTEM ) + enc = encFromName; +#endif // wxUSE_FONTMAP + + switch ( enc ) + { + case wxFONTENCODING_UTF7: + return new wxMBConvUTF7; + + case wxFONTENCODING_UTF8: + return new wxMBConvUTF8; + + case wxFONTENCODING_UTF16: + return new wxMBConvUTF16; + + case wxFONTENCODING_UTF16BE: + return new wxMBConvUTF16BE; + + case wxFONTENCODING_UTF16LE: + return new wxMBConvUTF16LE; + + case wxFONTENCODING_UTF32: + return new wxMBConvUTF32; + + case wxFONTENCODING_UTF32BE: + return new wxMBConvUTF32BE; + + case wxFONTENCODING_UTF32LE: + return new wxMBConvUTF32LE; + + default: + // nothing to do but put here to suppress gcc warnings + ; + } + + // step (3) +#if wxUSE_FONTMAP + { + wxMBConv_wxwin *conv = m_name ? new wxMBConv_wxwin(m_name) + : new wxMBConv_wxwin(m_encoding); + if ( conv->IsOk() ) + return conv; + + delete conv; + } +#endif // wxUSE_FONTMAP + + wxLogError(_("Cannot convert from the charset '%s'!"), + m_name ? m_name + : +#if wxUSE_FONTMAP + wxFontMapper::GetEncodingDescription(m_encoding).c_str() +#else // !wxUSE_FONTMAP + wxString::Format(_("encoding %s"), m_encoding).c_str() +#endif // wxUSE_FONTMAP/!wxUSE_FONTMAP + ); + + return NULL; +} + +void wxCSConv::CreateConvIfNeeded() const +{ + if ( m_deferred ) + { + wxCSConv *self = (wxCSConv *)this; // const_cast + self->m_convReal = DoCreate(); + self->m_deferred = false; + } +} + +size_t wxCSConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const +{ + CreateConvIfNeeded(); + + if (m_convReal) + return m_convReal->MB2WC(buf, psz, n); + + // latin-1 (direct) + size_t len = strlen(psz); + + if (buf) + { + for (size_t c = 0; c <= len; c++) + buf[c] = (unsigned char)(psz[c]); + } + + return len; +} + +size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const +{ + CreateConvIfNeeded(); + + if (m_convReal) + return m_convReal->WC2MB(buf, psz, n); + + // latin-1 (direct) + const size_t len = wxWcslen(psz); + if (buf) + { + for (size_t c = 0; c <= len; c++) + buf[c] = (psz[c] > 0xff) ? '?' : psz[c]; + } + + return len; +} + +// ---------------------------------------------------------------------------- +// globals +// ---------------------------------------------------------------------------- + +#ifdef __WINDOWS__ + static wxMBConv_win32 wxConvLibcObj; +#else + static wxMBConvLibc wxConvLibcObj; +#endif + +static wxCSConv wxConvLocalObj(wxFONTENCODING_SYSTEM); +static wxCSConv wxConvISO8859_1Obj(wxFONTENCODING_ISO8859_1); +static wxMBConvUTF7 wxConvUTF7Obj; +static wxMBConvUTF8 wxConvUTF8Obj; + + +WXDLLIMPEXP_DATA_BASE(wxMBConv&) wxConvLibc = wxConvLibcObj; +WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvLocal = wxConvLocalObj; +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; + +#else // !wxUSE_WCHAR_T + +// stand-ins in absence of wchar_t +WXDLLIMPEXP_DATA_BASE(wxMBConv) wxConvLibc, + wxConvISO8859_1, + wxConvLocal, + wxConvUTF8; -#endif // wxUSE_WCHAR_T +#endif // wxUSE_WCHAR_T/!wxUSE_WCHAR_T