X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/e95354ec548e2ef40f14734970911c0c2a14eb82..4eb7c4b180d2ddf44f9b82e5d56c81fcdc530f64:/src/common/strconv.cpp diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp index 4c9d38cf19..6363e4001c 100644 --- a/src/common/strconv.cpp +++ b/src/common/strconv.cpp @@ -34,6 +34,10 @@ #include "wx/log.h" #endif // WX_PRECOMP +#include "wx/strconv.h" + +#if wxUSE_WCHAR_T + #ifdef __WXMSW__ #include "wx/msw/private.h" #endif @@ -46,57 +50,14 @@ #include #include -#include "wx/module.h" -#include "wx/strconv.h" - #if defined(__WIN32__) && !defined(__WXMICROWIN__) #define wxHAVE_WIN32_MB2WC #endif // __WIN32__ but !__WXMICROWIN__ -// ---------------------------------------------------------------------------- -// globals -// ---------------------------------------------------------------------------- - -#if wxUSE_WCHAR_T - WXDLLIMPEXP_DATA_BASE(wxMBConv) wxConvLibc; - WXDLLIMPEXP_DATA_BASE(wxCSConv) wxConvLocal((const wxChar *)NULL); - WXDLLIMPEXP_DATA_BASE(wxCSConv) wxConvISO8859_1(_T("iso-8859-1")); -#else - // stand-ins in absence of wchar_t - WXDLLIMPEXP_DATA_BASE(wxMBConv) wxConvLibc, - wxConvFile, - wxConvISO8859_1, - wxConvLocal, - wxConvUTF8; -#endif // wxUSE_WCHAR_T - -WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent = &wxConvLibc; - -class wxStrConvModule: public wxModule -{ -public: - wxStrConvModule() : wxModule() { } - virtual bool OnInit() { return true; } - virtual void OnExit() - { -#if wxUSE_WCHAR_T - wxConvLocal.Clear(); - wxConvISO8859_1.Clear(); -#endif - } - - DECLARE_DYNAMIC_CLASS(wxStrConvModule) -}; - -IMPLEMENT_DYNAMIC_CLASS(wxStrConvModule, wxModule) - - // ---------------------------------------------------------------------------- // headers // ---------------------------------------------------------------------------- -#if wxUSE_WCHAR_T - #ifdef __SALFORDC__ #include #endif @@ -108,6 +69,13 @@ IMPLEMENT_DYNAMIC_CLASS(wxStrConvModule, wxModule) #include "wx/encconv.h" #include "wx/fontmap.h" +#ifdef __WXMAC__ +#include "ATSUnicode.h" +#include "TextCommon.h" +#include "TextEncodingConverter.h" + +#include "wx/mac/private.h" // includes mac headers +#endif // ---------------------------------------------------------------------------- // macros // ---------------------------------------------------------------------------- @@ -115,12 +83,6 @@ IMPLEMENT_DYNAMIC_CLASS(wxStrConvModule, wxModule) #define BSWAP_UCS4(str, len) { unsigned _c; for (_c=0; _c=0x110000) @@ -168,7 +131,7 @@ static size_t encode_utf16(wxUint32 input, wxUint16 *output) if (output) { *output++ = (wxUint16) ((input >> 10)+0xd7c0); - *output++ = (wxUint16) ((input&0x3ff)+0xdc00); + *output = (wxUint16) ((input&0x3ff)+0xdc00); } return 2; } @@ -198,49 +161,11 @@ static size_t decode_utf16(const wxUint16* input, wxUint32& output) // wxMBConv // ---------------------------------------------------------------------------- -#define IGNORE_LIBC 0 - wxMBConv::~wxMBConv() { // nothing to do here } -size_t wxMBConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const -{ -#if IGNORE_LIBC - if (buf) - { - for (size_t i = 0; i < strlen( psz )+1; i++) - buf[i] = (wchar_t) psz[i]; - return strlen( psz ); - } - else - { - return strlen( psz ); - } -#else - return wxMB2WC(buf, psz, n); -#endif -} - -size_t wxMBConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const -{ -#if IGNORE_LIBC - if (buf) - { - for (size_t i = 0; i < wxStrlen( psz )+1; i++) - buf[i] = (char) psz[i]; - return wxStrlen( psz ); - } - else - { - return wxStrlen( psz ); - } -#else - return wxWC2MB(buf, psz, n); -#endif -} - const wxWCharBuffer wxMBConv::cMB2WC(const char *psz) const { if ( psz ) @@ -282,10 +207,22 @@ const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *pwz) const } // ---------------------------------------------------------------------------- -// UTF-7 +// wxMBConvLibc // ---------------------------------------------------------------------------- -WXDLLIMPEXP_DATA_BASE(wxMBConvUTF7) wxConvUTF7; +size_t wxMBConvLibc::MB2WC(wchar_t *buf, const char *psz, size_t n) const +{ + return wxMB2WC(buf, psz, n); +} + +size_t wxMBConvLibc::WC2MB(char *buf, const wchar_t *psz, size_t n) const +{ + return wxWC2MB(buf, psz, n); +} + +// ---------------------------------------------------------------------------- +// UTF-7 +// ---------------------------------------------------------------------------- #if 0 static char utf7_setD[]="ABCDEFGHIJKLMNOPQRSTUVWXYZ" @@ -316,8 +253,6 @@ size_t wxMBConvUTF7::WC2MB(char * WXUNUSED(buf), // UTF-8 // ---------------------------------------------------------------------------- -WXDLLIMPEXP_DATA_BASE(wxMBConvUTF8) wxConvUTF8; - static wxUint32 utf8_max[]= { 0x7f, 0x7ff, 0xffff, 0x1fffff, 0x3ffffff, 0x7fffffff, 0xffffffff }; @@ -366,7 +301,8 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const return (size_t)-1; } #ifdef WC_UTF16 - size_t pa = encode_utf16(res, buf); + // 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) @@ -393,7 +329,8 @@ size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const { wxUint32 cc; #ifdef WC_UTF16 - size_t pa = decode_utf16(psz, cc); + // cast is ok for WC_UTF16 + size_t pa = decode_utf16((const wxUint16 *)psz, cc); psz += (pa == (size_t)-1) ? 1 : pa; #else cc=(*psz++) & 0x7fffffff; @@ -433,24 +370,16 @@ size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const // ---------------------------------------------------------------------------- #ifdef WORDS_BIGENDIAN -#define wxMBConvUTF16straight wxMBConvUTF16BE -#define wxMBConvUTF16swap wxMBConvUTF16LE + #define wxMBConvUTF16straight wxMBConvUTF16BE + #define wxMBConvUTF16swap wxMBConvUTF16LE #else -#define wxMBConvUTF16swap wxMBConvUTF16BE -#define wxMBConvUTF16straight wxMBConvUTF16LE + #define wxMBConvUTF16swap wxMBConvUTF16BE + #define wxMBConvUTF16straight wxMBConvUTF16LE #endif -WXDLLIMPEXP_DATA_BASE(wxMBConvUTF16LE) wxConvUTF16LE; -WXDLLIMPEXP_DATA_BASE(wxMBConvUTF16BE) wxConvUTF16BE; - - - - - #ifdef WC_UTF16 - // copy 16bit MB to 16bit String size_t wxMBConvUTF16straight::MB2WC(wchar_t *buf, const char *psz, size_t n) const { @@ -576,7 +505,7 @@ size_t wxMBConvUTF16straight::WC2MB(char *buf, const wchar_t *psz, size_t n) con if (buf) { *(wxUint16*)buf = cc[0]; - buf += sizeof(wxUint16); + buf += sizeof(wxUint16); if (pa > 1) { *(wxUint16*)buf = cc[1]; @@ -712,7 +641,8 @@ size_t wxMBConvUTF32straight::WC2MB(char *buf, const wchar_t *psz, size_t n) con { wxUint32 cc; - size_t pa=decode_utf16(psz, cc); + // cast is ok for WC_UTF16 + size_t pa = decode_utf16((const wxUint16 *)psz, cc); if (pa == (size_t)-1) return pa; @@ -724,7 +654,9 @@ size_t wxMBConvUTF32straight::WC2MB(char *buf, const wchar_t *psz, size_t n) con len += sizeof(wxUint32); psz += pa; } - if (buf && len<=n-sizeof(wxUint32)) *(wxUint32*)buf=0; + + if (buf && len<=n-sizeof(wxUint32)) + *(wxUint32*)buf=0; return len; } @@ -758,7 +690,9 @@ size_t wxMBConvUTF32swap::MB2WC(wchar_t *buf, const char *psz, size_t n) const len += pa; psz += sizeof(wxUint32); } - if (buf && lenCharsetToEncoding(name, FALSE) ) ) ; + } + + wxMBConv_mac(wxFontEncoding encoding) + { + Init( wxMacGetSystemEncFromFontEnc(encoding) ); + } + + ~wxMBConv_mac() + { + OSStatus status = noErr ; + status = TECDisposeConverter(m_MB2WC_converter); + status = TECDisposeConverter(m_WC2MB_converter); + } + + + void Init( TextEncodingBase encoding) + { + OSStatus status = noErr ; + m_char_encoding = encoding ; +#if SIZEOF_WCHAR_T == 4 + m_unicode_encoding = CreateTextEncoding(kTextEncodingUnicodeDefault,0,kUnicode32BitFormat) ; +#else + m_unicode_encoding = CreateTextEncoding(kTextEncodingUnicodeDefault,0,kUnicode16BitFormat) ; +#endif + status = TECCreateConverter(&m_MB2WC_converter, + m_char_encoding, + m_unicode_encoding); + status = TECCreateConverter(&m_WC2MB_converter, + m_unicode_encoding, + m_char_encoding); + } + + size_t MB2WC(wchar_t *buf, const char *psz, size_t n) const + { + OSStatus status = noErr ; + ByteCount byteOutLen ; + ByteCount byteInLen = strlen(psz) ; + wchar_t *tbuf = NULL ; + + if (buf == NULL) + { + n = byteInLen ; + tbuf = (wchar_t*) malloc( n * SIZEOF_WCHAR_T) ; + } + + ByteCount byteBufferLen = n * SIZEOF_WCHAR_T ; + status = TECConvertText(m_MB2WC_converter, (ConstTextPtr) psz , byteInLen, &byteInLen, + (TextPtr) (buf ? buf : tbuf) , byteBufferLen, &byteOutLen); + + if ( buf == NULL ) + free(tbuf) ; + + size_t res = byteOutLen / SIZEOF_WCHAR_T ; + if ( buf && res < n) + buf[res] = 0; + + return res ; + } + + size_t WC2MB(char *buf, const wchar_t *psz, size_t n) const + { + OSStatus status = noErr ; + ByteCount byteOutLen ; + ByteCount byteInLen = wxWcslen(psz) * SIZEOF_WCHAR_T ; + + char *tbuf = NULL ; + + if (buf == NULL) + { + // worst case + n = byteInLen * 2 ; + tbuf = (char*) malloc( n ) ; + } + + ByteCount byteBufferLen = n ; + status = TECConvertText(m_WC2MB_converter, (ConstTextPtr) psz , byteInLen, &byteInLen, + (TextPtr) ( buf ? buf : tbuf ) , byteBufferLen, &byteOutLen); + + if ( buf == NULL ) + free(tbuf) ; + + size_t res = byteOutLen ; + if ( buf && res < n) + buf[res] = 0; + + return res ; + } + + bool IsOk() const + { return m_MB2WC_converter != NULL && m_WC2MB_converter != NULL ; } + +private: + TECObjectRef m_MB2WC_converter ; + TECObjectRef m_WC2MB_converter ; + + TextEncodingBase m_char_encoding ; + TextEncodingBase m_unicode_encoding ; +}; + +#endif // defined(__WXMAC__) && defined(TARGET_CARBON) // ============================================================================ // wxEncodingConverter based conversion classes @@ -1264,7 +1331,7 @@ public: Init(); } - size_t MB2WC(wchar_t *buf, const char *psz, size_t WXUNUSED(n)) + size_t MB2WC(wchar_t *buf, const char *psz, size_t WXUNUSED(n)) const { size_t inbuf = strlen(psz); if (buf) @@ -1272,7 +1339,7 @@ public: return inbuf; } - size_t WC2MB(char *buf, const wchar_t *psz, size_t WXUNUSED(n)) + size_t WC2MB(char *buf, const wchar_t *psz, size_t WXUNUSED(n)) const { const size_t inbuf = wxWcslen(psz); if (buf) @@ -1306,37 +1373,21 @@ void wxCSConv::Init() 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(); - } + + m_encoding = wxFONTENCODING_SYSTEM; } wxCSConv::wxCSConv(wxFontEncoding encoding) { - if ( encoding == wxFONTENCODING_MAX || - encoding == wxFONTENCODING_DEFAULT ) + if ( encoding == wxFONTENCODING_MAX || encoding == wxFONTENCODING_DEFAULT ) { wxFAIL_MSG( _T("invalid encoding value in wxCSConv ctor") ); @@ -1345,14 +1396,7 @@ wxCSConv::wxCSConv(wxFontEncoding encoding) Init(); - if ( encoding == wxFONTENCODING_SYSTEM ) - { - SetEncoding(); - } - else // have valid encoding, use it - { - m_encoding = encoding; - } + m_encoding = encoding; } wxCSConv::~wxCSConv() @@ -1397,28 +1441,12 @@ void wxCSConv::SetName(const wxChar *charset) } } -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 - ) + // check for the special case of ASCII or ISO8859-1 charset: as we have + // special knowledge of it anyhow, we don't need to create a special + // conversion object + if ( m_encoding == wxFONTENCODING_ISO8859_1 ) { // don't convert at all return NULL; @@ -1434,9 +1462,18 @@ wxMBConv *wxCSConv::DoCreate() const // step (1) #ifdef HAVE_ICONV +#if !wxUSE_FONTMAP if ( m_name ) +#endif // !wxUSE_FONTMAP { - wxMBConv_iconv *conv = new wxMBConv_iconv(m_name); + wxString name(m_name); + +#if wxUSE_FONTMAP + if ( name.empty() ) + name = wxFontMapper::Get()->GetEncodingName(m_encoding); +#endif // wxUSE_FONTMAP + + wxMBConv_iconv *conv = new wxMBConv_iconv(name); if ( conv->IsOk() ) return conv; @@ -1454,12 +1491,30 @@ wxMBConv *wxCSConv::DoCreate() const delete conv; } #endif // wxHAVE_WIN32_MB2WC +#if defined(__WXMAC__) + { + if ( m_name || ( m_encoding < wxFONTENCODING_UTF16BE ) ) + { + + wxMBConv_mac *conv = m_name ? new wxMBConv_mac(m_name) + : new wxMBConv_mac(m_encoding); + if ( conv->IsOk() ) + return conv; + delete conv; + } + } +#endif // step (2) wxFontEncoding enc = m_encoding; #if wxUSE_FONTMAP - if ( enc == wxFONTENCODING_SYSTEM ) - enc = encFromName; + if ( enc == wxFONTENCODING_SYSTEM && m_name ) + { + // 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); + } #endif // wxUSE_FONTMAP switch ( enc ) @@ -1470,18 +1525,12 @@ wxMBConv *wxCSConv::DoCreate() const 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; @@ -1505,8 +1554,19 @@ wxMBConv *wxCSConv::DoCreate() const } #endif // wxUSE_FONTMAP - wxLogError(_("Cannot convert from the charset '%s'!"), - m_name ? m_name + // NB: This is a hack to prevent deadlock. What could otherwise happen + // in Unicode build: wxConvLocal creation ends up being here + // because of some failure and logs the error. But wxLog will try to + // attach timestamp, for which it will need wxConvLocal (to convert + // time to char* and then wchar_t*), but that fails, tries to log + // error, but wxLog has a (already locked) critical section that + // guards static buffer. + static bool alreadyLoggingError = false; + if (!alreadyLoggingError) + { + alreadyLoggingError = true; + wxLogError(_("Cannot convert from the charset '%s'!"), + m_name ? m_name : #if wxUSE_FONTMAP wxFontMapper::GetEncodingDescription(m_encoding).c_str() @@ -1514,6 +1574,8 @@ wxMBConv *wxCSConv::DoCreate() const wxString::Format(_("encoding %s"), m_encoding).c_str() #endif // wxUSE_FONTMAP/!wxUSE_FONTMAP ); + alreadyLoggingError = false; + } return NULL; } @@ -1523,6 +1585,16 @@ void wxCSConv::CreateConvIfNeeded() const if ( m_deferred ) { wxCSConv *self = (wxCSConv *)this; // const_cast + +#if wxUSE_INTL + // if we don't have neither the name nor the encoding, use the default + // encoding for this system + if ( !m_name && m_encoding == wxFONTENCODING_SYSTEM ) + { + self->m_name = wxStrdup(wxLocale::GetSystemEncodingName()); + } +#endif // wxUSE_INTL + self->m_convReal = DoCreate(); self->m_deferred = false; } @@ -1559,12 +1631,55 @@ size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const if (buf) { for (size_t c = 0; c <= len; c++) - buf[c] = (psz[c] > 0xff) ? '?' : psz[c]; + { + if (psz[c] > 0xFF) + return (size_t)-1; + buf[c] = psz[c]; + } + } + else + { + for (size_t c = 0; c <= len; c++) + { + if (psz[c] > 0xFF) + return (size_t)-1; + } } return len; } -#endif // wxUSE_WCHAR_T +// ---------------------------------------------------------------------------- +// 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/!wxUSE_WCHAR_T