X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/f6bcfd974ef26faf6f91a62cac09827e09463fd1..079c842cae5a3ad08ad930ebc29a33c59acf634d:/src/common/strconv.cpp diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp index 997899e769..95fc0d9c07 100644 --- a/src/common/strconv.cpp +++ b/src/common/strconv.cpp @@ -28,6 +28,11 @@ #pragma hdrstop #endif +#ifdef __WXMSW__ + #include "wx/msw/private.h" +#endif + +#include #include #include #include @@ -36,9 +41,40 @@ #include #endif +#ifdef HAVE_ICONV_H + #include +#endif +#ifdef HAVE_LANGINFO_H + #include +#endif + +#ifdef __WXMSW__ + #include +#endif + #include "wx/debug.h" #include "wx/strconv.h" +#if defined(WORDS_BIGENDIAN) || defined(__STDC_ISO_10646__) +#define BSWAP_UCS4(str, len) +#define BSWAP_UCS2(str, len) +#else +#define BSWAP_UCS4(str, len) { unsigned _c; for (_c=0; _c=0x110000) { + return (size_t)-1; + } else { + if (output) { + *output++ = (input >> 10)+0xd7c0; + *output++ = (input&0x3ff)+0xdc00; + } + return 2; + } +} + +static size_t decode_utf16(wxUint16*input,wxUint32&output) +{ + if ((*input<0xd800) || (*input>0xdfff)) { + output = *input; + return 1; + } else + if ((input[1]<0xdc00) || (input[1]>=0xdfff)) { + output = *input; + return (size_t)-1; + } else { + output = ((input[0] - 0xd7c0) << 10) + (input[1] - 0xdc00); + return 2; + } +} + +#endif // WC_UTF16 + // ---------------------------------------------------------------------------- // wxMBConv // ---------------------------------------------------------------------------- @@ -187,7 +259,7 @@ size_t wxMBConvUTF7::WC2MB(char * WXUNUSED(buf), 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 { @@ -208,7 +280,7 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const return (size_t)-1; } else { unsigned ocnt=cnt-1; - unsigned long res=cc&(0x3f>>cnt); + wxUint32 res=cc&(0x3f>>cnt); while (cnt--) { cc = *psz++; if ((cc&0xC0)!=0x80) { @@ -221,8 +293,16 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const // illegal UTF-8 encoding return (size_t)-1; } +#ifdef WC_UTF16 + size_t pa = encode_utf16(res, buf); + if (pa == (size_t)-1) + return (size_t)-1; + if (buf) buf+=pa; + len+=pa; +#else if (buf) *buf++=res; len++; +#endif } } } @@ -235,7 +315,13 @@ size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const size_t len = 0; while (*psz && ((!buf) || (lenutf8_max[cnt]; cnt++); if (!cnt) { @@ -264,104 +350,451 @@ WXDLLEXPORT_DATA(wxCSConv) wxConvLocal((const wxChar *)NULL); #include "wx/encconv.h" #include "wx/fontmap.h" +// TODO: add some tables here +// - perhaps common encodings to common codepages (for Win32) +// - perhaps common encodings to objects ("UTF8" -> wxConvUTF8) +// - move wxEncodingConverter meat in here + +#ifdef __WIN32__ +#include "wx/msw/registry.h" +// this should work if M$ Internet Exploiter is installed +static long CharsetToCodepage(const wxChar *name) +{ + if (!name) + return GetACP(); + + long CP=-1; + + wxString cn(name); + do { + wxString path( wxT("MIME\\Database\\Charset\\") ); + path += cn; + wxRegKey key( wxRegKey::HKCR, path ); + + /* two cases: either there's an AliasForCharset string, + * or there are Codepage and InternetEncoding dwords. + * The InternetEncoding gives us the actual encoding, + * the Codepage just says which Windows character set to + * use when displaying the data. + */ + if (key.QueryValue( wxT("InternetEncoding"), &CP )) break; + + // no encoding, see if it's an alias + if (!key.QueryValue( wxT("AliasForCharset"), cn )) break; + } while (1); + + return CP; +} +#endif + class wxCharacterSet { public: - // temporarily just use wxEncodingConverter stuff, - // so that it works while a better implementation is built - wxFontEncoding enc; - wxEncodingConverter m2w, w2m; - wxCharacterSet(wxFontEncoding e) : enc(e) - { - m2w.Init(enc, wxFONTENCODING_UNICODE); - w2m.Init(wxFONTENCODING_UNICODE, enc); - } + wxCharacterSet(const wxChar*name) + : cname(name) {} + virtual ~wxCharacterSet() + {} + virtual size_t MB2WC(wchar_t*buf, const char*psz, size_t n) + { return (size_t)-1; } + virtual size_t WC2MB(char*buf, const wchar_t*psz, size_t n) + { return (size_t)-1; } + virtual bool usable() + { return FALSE; } +public: + const wxChar*cname; +}; + +class ID_CharSet : public wxCharacterSet +{ +public: + ID_CharSet(const wxChar*name,wxMBConv*cnv) + : wxCharacterSet(name), work(cnv) {} + + size_t MB2WC(wchar_t*buf, const char*psz, size_t n) + { return work ? work->MB2WC(buf,psz,n) : (size_t)-1; } + + size_t WC2MB(char*buf, const wchar_t*psz, size_t n) + { return work ? work->WC2MB(buf,psz,n) : (size_t)-1; } + + bool usable() + { return work!=NULL; } +public: + wxMBConv*work; +}; + +#ifdef HAVE_ICONV_H +class IC_CharSet : public wxCharacterSet +{ +public: + IC_CharSet(const wxChar*name) + : wxCharacterSet(name), m2w((iconv_t)-1), w2m((iconv_t)-1) {} + ~IC_CharSet() + { + if (m2w!=(iconv_t)-1) iconv_close(m2w); + if (w2m!=(iconv_t)-1) iconv_close(w2m); + } + + void LoadM2W() + { + if (m2w==(iconv_t)-1) + m2w=iconv_open(WC_NAME,wxConvLibc.cWX2MB(cname)); + } + + void LoadW2M() + { + if (w2m==(iconv_t)-1) + w2m=iconv_open(wxConvLibc.cWX2MB(cname),WC_NAME); + } + + size_t MB2WC(wchar_t*buf, const char*psz, size_t n) + { + LoadM2W(); + size_t inbuf = strlen(psz); + size_t outbuf = n*SIZEOF_WCHAR_T; + size_t res, cres; + fprintf(stderr,"IC Convert to WC using %s\n",(const char*)wxConvLibc.cWX2MB(cname)); + if (buf) + { + // have destination buffer, convert there +#ifdef WX_ICONV_TAKES_CHAR + cres = iconv( m2w, (char**)&psz, &inbuf, (char**)&buf, &outbuf ); +#else + cres = iconv( m2w, &psz, &inbuf, (char**)&buf, &outbuf ); +#endif + res = n-(outbuf/SIZEOF_WCHAR_T); + // convert to native endianness + WC_BSWAP(buf, res) + } + else + { + // no destination buffer... convert using temp buffer + // to calculate destination buffer requirement + wchar_t tbuf[8]; + res = 0; + do { + buf = tbuf; outbuf = 8*SIZEOF_WCHAR_T; +#ifdef WX_ICONV_TAKES_CHAR + cres = iconv( m2w, (char**)&psz, &inbuf, (char**)&buf, &outbuf ); +#else + cres = iconv( m2w, &psz, &inbuf, (char**)&buf, &outbuf ); +#endif + res += 8-(outbuf/SIZEOF_WCHAR_T); + } while ((cres==(size_t)-1) && (errno==E2BIG)); + } + + if (cres==(size_t)-1) + return (size_t)-1; + + return res; + } + + size_t WC2MB(char*buf, const wchar_t*psz, size_t n) + { + LoadW2M(); +#if defined(__BORLANDC__) && (__BORLANDC__ > 0x530) + size_t inbuf = std::wcslen(psz); +#else + size_t inbuf = ::wcslen(psz); +#endif + size_t outbuf = n; + size_t res, cres; + fprintf(stderr,"IC Convert from WC using %s\n",(const char*)wxConvLibc.cWX2MB(cname)); +#ifdef WC_NEED_BSWAP + // need to copy to temp buffer to switch endianness + // this absolutely doesn't rock! + // (no, doing WC_BSWAP twice on the original buffer won't help, as it + // could be in read-only memory, or be accessed in some other thread) + wchar_t*tmpbuf=(wchar_t*)malloc((inbuf+1)*SIZEOF_WCHAR_T); + memcpy(tmpbuf,psz,(inbuf+1)*SIZEOF_WCHAR_T); + WC_BSWAP(tmpbuf, inbuf) + psz=tmpbuf; +#endif + if (buf) + { + // have destination buffer, convert there +#ifdef WX_ICONV_TAKES_CHAR + cres = iconv( w2m, (char**)&psz, &inbuf, &buf, &outbuf ); +#else + cres = iconv( w2m, (const char**)&psz, &inbuf, &buf, &outbuf ); +#endif + res = n-outbuf; + } + else + { + // no destination buffer... convert using temp buffer + // to calculate destination buffer requirement + char tbuf[16]; + res = 0; + do { + buf = tbuf; outbuf = 16; +#ifdef WX_ICONV_TAKES_CHAR + cres = iconv( w2m, (char**)&psz, &inbuf, &buf, &outbuf ); +#else + cres = iconv( w2m, (const char**)&psz, &inbuf, &buf, &outbuf ); +#endif + res += 16 - outbuf; + } while ((cres==(size_t)-1) && (errno==E2BIG)); + } +#ifdef WC_NEED_BSWAP + free(tmpbuf); +#endif + if (cres==(size_t)-1) + return (size_t)-1; + + return res; + } + + bool usable() + { return TRUE; } + +public: + iconv_t m2w, w2m; +}; +#endif + +#ifdef __WIN32__ +class CP_CharSet : public wxCharacterSet +{ +public: + CP_CharSet(const wxChar*name) + : wxCharacterSet(name), CodePage(CharsetToCodepage(name)) {} + + size_t MB2WC(wchar_t*buf, const char*psz, size_t n) + { + size_t len = MultiByteToWideChar(CodePage,0,psz,-1,buf,buf?n:0); + return len ? len : (size_t)-1; + } + + size_t WC2MB(char*buf, const wchar_t*psz, size_t n) + { + size_t len = WideCharToMultiByte(CodePage,0,psz,-1,buf,buf?n:0,NULL,NULL); + return len ? len : (size_t)-1; + } + + bool usable() + { return CodePage!=-1; } + +public: + long CodePage; +}; +#endif + +class EC_CharSet : public wxCharacterSet +{ +public: + // temporarily just use wxEncodingConverter stuff, + // so that it works while a better implementation is built + EC_CharSet(const wxChar*name) : wxCharacterSet(name), enc(wxFONTENCODING_SYSTEM) + { + if (name) + enc = wxTheFontMapper->CharsetToEncoding(name, FALSE); + m2w.Init(enc, wxFONTENCODING_UNICODE); + w2m.Init(wxFONTENCODING_UNICODE, enc); + } + + size_t MB2WC(wchar_t*buf, const char*psz, size_t n) + { + size_t inbuf = strlen(psz); + fprintf(stderr,"EC Convert to WC using %d\n",enc); + if (buf) m2w.Convert(psz,buf); + return inbuf; + } + + size_t WC2MB(char*buf, const wchar_t*psz, size_t n) + { +#if defined(__BORLANDC__) && (__BORLANDC__ > 0x530) + size_t inbuf = std::wcslen(psz); +#else + size_t inbuf = ::wcslen(psz); +#endif + fprintf(stderr,"EC Convert from WC using %d\n",enc); + if (buf) + w2m.Convert(psz,buf); + + return inbuf; + } + + bool usable() + { return (enc!=wxFONTENCODING_SYSTEM) && (enc!=wxFONTENCODING_DEFAULT); } + +public: + wxFontEncoding enc; + wxEncodingConverter m2w, w2m; }; static wxCharacterSet *wxGetCharacterSet(const wxChar *name) { - wxFontEncoding enc = name ? wxTheFontMapper->CharsetToEncoding(name, FALSE) - : wxFONTENCODING_SYSTEM; - wxCharacterSet *cset = (enc != wxFONTENCODING_SYSTEM) ? new wxCharacterSet(enc) - : (wxCharacterSet *)NULL; - return cset; + wxCharacterSet *cset = NULL; + if (name) + { + if (!wxStricmp(name, wxT("UTF8")) || !wxStricmp(name, wxT("UTF-8"))) + { + cset = new ID_CharSet(name, &wxConvUTF8); + } + else + { +#ifdef HAVE_ICONV_H + cset = new IC_CharSet(name); // may not take NULL +#endif + } + } + + if (cset && cset->usable()) return cset; + if (cset) delete cset; +#ifdef __WIN32__ + cset = new CP_CharSet(name); // may take NULL + if (cset->usable()) return cset; +#endif + if (cset) delete cset; + cset = new EC_CharSet(name); + if (cset->usable()) return cset; + delete cset; + return NULL; } wxCSConv::wxCSConv(const wxChar *charset) { - m_name = (wxChar *) NULL; - m_cset = (wxCharacterSet *) NULL; - m_deferred = TRUE; - SetName(charset); + m_name = (wxChar *) NULL; + m_cset = (wxCharacterSet *) NULL; + m_deferred = TRUE; + SetName(charset); } wxCSConv::~wxCSConv() { - if (m_name) free(m_name); - if (m_cset) delete m_cset; + if (m_name) free(m_name); + if (m_cset) delete m_cset; } void wxCSConv::SetName(const wxChar *charset) { - if (charset) { - m_name = wxStrdup(charset); - m_deferred = TRUE; - } + if (charset) + { + m_name = wxStrdup(charset); + m_deferred = TRUE; + } } void wxCSConv::LoadNow() { // wxPrintf(wxT("Conversion request\n")); - if (m_deferred) { - if (!m_name) { + if (m_deferred) + { + if (!m_name) + { #ifdef __UNIX__ - wxChar *lang = wxGetenv(wxT("LC_ALL")); - if (!lang) lang = wxGetenv(wxT("LANG")); - wxChar *dot = lang ? wxStrchr(lang, wxT('.')) : (wxChar *)NULL; - if (dot) SetName(dot+1); +#if defined(HAVE_LANGINFO_H) && defined(CODESET) + // GNU libc provides current character set this way + char*alang = nl_langinfo(CODESET); + if (alang) + { + SetName(wxConvLibc.cMB2WX(alang)); + } + else +#endif + { + // if we can't get at the character set directly, + // try to see if it's in the environment variables + // (in most cases this won't work, but I was out of ideas) + wxChar *lang = wxGetenv(wxT("LC_ALL")); + if (!lang) lang = wxGetenv(wxT("LC_CTYPE")); + if (!lang) lang = wxGetenv(wxT("LANG")); + wxChar *dot = lang ? wxStrchr(lang, wxT('.')) : (wxChar *)NULL; + if (dot) SetName(dot+1); + } #endif + } + m_cset = wxGetCharacterSet(m_name); + m_deferred = FALSE; } - m_cset = wxGetCharacterSet(m_name); - m_deferred = FALSE; - } } size_t wxCSConv::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 (m_cset) + return m_cset->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 n; - } - return strlen(psz); + + return len; } size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const { - ((wxCSConv *)this)->LoadNow(); // 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; - } + ((wxCSConv *)this)->LoadNow(); // discard constness + + if (m_cset) + return m_cset->WC2MB(buf, psz, n); + + // latin-1 (direct) #if defined(__BORLANDC__) && (__BORLANDC__ > 0x530) - return std::wcslen(psz); + size_t len=std::wcslen(psz); #else - return ::wcslen(psz); + size_t len=::wcslen(psz); #endif + if (buf) + { + for (size_t c=0; c<=len; c++) + buf[c] = (psz[c]>0xff) ? '?' : psz[c]; + } + + return len; } +#ifdef HAVE_ICONV_H +class IC_CharSetConverter +{ +public: + IC_CharSetConverter(IC_CharSet*from,IC_CharSet*to) + { cnv = iconv_open(wxConvLibc.cWX2MB(to->cname),wxConvLibc.cWX2MB(from->cname)); } + + ~IC_CharSetConverter() + { if (cnv!=(iconv_t)-1) iconv_close(cnv); } + + size_t Convert(char*buf, const char*psz, size_t n) + { + size_t inbuf = strlen(psz); + size_t outbuf = n; +#ifdef WX_ICONV_TAKES_CHAR + size_t res = iconv( cnv, (char**)&psz, &inbuf, &buf, &outbuf ); +#else + size_t res = iconv( cnv, &psz, &inbuf, &buf, &outbuf ); +#endif + if (res==(size_t)-1) return (size_t)-1; + return n-outbuf; + } + +public: + iconv_t cnv; +}; +#endif + +class EC_CharSetConverter +{ +public: + EC_CharSetConverter(EC_CharSet*from,EC_CharSet*to) + { cnv.Init(from->enc,to->enc); } + + size_t Convert(char*buf, const char*psz, size_t n) + { + size_t inbuf = strlen(psz); + if (buf) cnv.Convert(psz,buf); + return inbuf; + } + +public: + wxEncodingConverter cnv; +}; + #else // !wxUSE_WCHAR_T // ----------------------------------------------------------------------------