X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/36acb8807245f6245a8d8b202b343115b7bdc16b..bbee1b4abc13379569607df8f6985ec68dc12231:/src/common/strconv.cpp diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp index 41e50679a5..c8207ca44e 100644 --- a/src/common/strconv.cpp +++ b/src/common/strconv.cpp @@ -28,8 +28,13 @@ #pragma hdrstop #endif +#ifndef WX_PRECOMP + #include "wx/intl.h" + #include "wx/log.h" +#endif // WX_PRECOMP + #ifdef __WXMSW__ - #include "wx/msw/private.h" + #include "wx/msw/private.h" #endif #include @@ -37,36 +42,42 @@ #include #include - -#include "wx/debug.h" #include "wx/strconv.h" -#include "wx/intl.h" -#include "wx/log.h" // ---------------------------------------------------------------------------- // globals // ---------------------------------------------------------------------------- -WXDLLEXPORT_DATA(wxMBConv *) wxConvCurrent = &wxConvLibc; +#if wxUSE_WCHAR_T + WXDLLEXPORT_DATA(wxMBConv) wxConvLibc; + WXDLLEXPORT_DATA(wxCSConv) wxConvLocal((const wxChar *)NULL); +#else + // stand-ins in absence of wchar_t + WXDLLEXPORT_DATA(wxMBConv) wxConvLibc, wxConvFile; +#endif // wxUSE_WCHAR_T +WXDLLEXPORT_DATA(wxMBConv *) wxConvCurrent = &wxConvLibc; -// ============================================================================ -// implementation -// ============================================================================ +// ---------------------------------------------------------------------------- +// headers +// ---------------------------------------------------------------------------- #if wxUSE_WCHAR_T #ifdef __SALFORDC__ - #include + #include #endif -#ifdef HAVE_ICONV_H - #include +#ifdef HAVE_ICONV + #include #endif -#ifdef __WXMSW__ - #include -#endif +#include "wx/encconv.h" +#include "wx/fontmap.h" + +// ---------------------------------------------------------------------------- +// macros +// ---------------------------------------------------------------------------- #define BSWAP_UCS4(str, len) { unsigned _c; for (_c=0; _c=0x110000) @@ -117,8 +135,8 @@ static size_t encode_utf16(wxUint32 input, wchar_t *output) { if (output) { - *output++ = (input >> 10)+0xd7c0; - *output++ = (input&0x3ff)+0xdc00; + *output++ = (wchar_t) ((input >> 10)+0xd7c0); + *output++ = (wchar_t) ((input&0x3ff)+0xdc00); } return 2; } @@ -149,8 +167,6 @@ static size_t decode_utf16(const wchar_t* input, wxUint32& output) // wxMBConv // ---------------------------------------------------------------------------- -WXDLLEXPORT_DATA(wxMBConv) wxConvLibc; - size_t wxMBConv::MB2WC(wchar_t *buf, const char *psz, size_t n) const { return wxMB2WC(buf, psz, n); @@ -341,11 +357,11 @@ size_t wxMBConvUTF8::MB2WC(wchar_t *buf, const char *psz, size_t n) const if (buf) buf += pa; len += pa; -#else +#else // !WC_UTF16 if (buf) *buf++ = res; len++; -#endif +#endif // WC_UTF16/!WC_UTF16 } } } @@ -373,7 +389,7 @@ size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const { // plain ASCII char if (buf) - *buf++ = cc; + *buf++ = (char) cc; len++; } @@ -382,9 +398,9 @@ size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const len += cnt + 1; if (buf) { - *buf++ = (-128 >> cnt) | ((cc >> (cnt * 6)) & (0x3f >> cnt)); + *buf++ = (char) ((-128 >> cnt) | ((cc >> (cnt * 6)) & (0x3f >> cnt))); while (cnt--) - *buf++ = 0x80 | ((cc >> (cnt * 6)) & 0x3f); + *buf++ = (char) (0x80 | ((cc >> (cnt * 6)) & 0x3f)); } } } @@ -393,110 +409,6 @@ size_t wxMBConvUTF8::WC2MB(char *buf, const wchar_t *psz, size_t n) const return len; } -// ---------------------------------------------------------------------------- -// specified character set -// ---------------------------------------------------------------------------- - -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 - -#if defined(__WIN32__) && !defined(__WXMICROWIN__) - -#if wxUSE_GUI - -// VZ: the new version of wxCharsetToCodepage() is more politically correct -// and should work on other Windows versions as well but the old version is -// still needed for !wxUSE_FONTMAP || !wxUSE_GUI case - -extern long wxEncodingToCodepage(wxFontEncoding encoding) -{ - // translate encoding into the Windows CHARSET - wxNativeEncodingInfo natveEncInfo; - if ( !wxGetNativeFontEncoding(encoding, &natveEncInfo) ) - return -1; - - // translate CHARSET to code page - CHARSETINFO csetInfo; - if ( !::TranslateCharsetInfo((DWORD *)(DWORD)natveEncInfo.charset, - &csetInfo, - TCI_SRCCHARSET) ) - { - wxLogLastError(_T("TranslateCharsetInfo(TCI_SRCCHARSET)")); - - return -1; - } - - return csetInfo.ciACP; -} - -#if wxUSE_FONTMAP - -extern long wxCharsetToCodepage(const wxChar *name) -{ - // first get the font encoding for this charset - if ( !name ) - return -1; - - wxFontEncoding enc = wxTheFontMapper->CharsetToEncoding(name, FALSE); - if ( enc == wxFONTENCODING_SYSTEM ) - return -1; - - // the use the helper function - return wxEncodingToCodepage(enc); -} - -#endif // wxUSE_FONTMAP - -#endif // wxUSE_GUI - -// include old wxCharsetToCodepage() by OK if needed -#if !wxUSE_GUI || !wxUSE_FONTMAP - -#include "wx/msw/registry.h" - -// this should work if Internet Exploiter is installed -extern long wxCharsetToCodepage(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); - - if (!key.Exists()) break; - - // 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.HasValue(wxT("InternetEncoding")) && - key.QueryValue(wxT("InternetEncoding"), &CP)) break; - - // no encoding, see if it's an alias - if (!key.HasValue(wxT("AliasForCharset")) || - !key.QueryValue(wxT("AliasForCharset"), cn)) break; - } while (1); - - return CP; -} - -#endif // !wxUSE_GUI || !wxUSE_FONTMAP - -#endif // Win32 - // ============================================================================ // wxCharacterSet and derived classes // ============================================================================ @@ -508,16 +420,11 @@ extern long wxCharsetToCodepage(const wxChar *name) class wxCharacterSet { public: - 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; } + wxCharacterSet(const wxChar*name) : cname(name) {} + virtual ~wxCharacterSet() {} + virtual size_t MB2WC(wchar_t *buf, const char *psz, size_t n) = 0; + virtual size_t WC2MB(char *buf, const wchar_t *psz, size_t n) = 0; + virtual bool usable() const = 0; public: const wxChar*cname; }; @@ -538,7 +445,7 @@ public: size_t WC2MB(char *buf, const wchar_t *psz, size_t n) { return work ? work->WC2MB(buf,psz,n) : (size_t)-1; } - bool usable() + bool usable() const { return work!=NULL; } public: wxMBConv*work; @@ -549,7 +456,7 @@ public: // The classes doing conversion using the iconv_xxx() functions // ============================================================================ -#ifdef HAVE_ICONV_H +#ifdef HAVE_ICONV // VS: glibc 2.1.3 is broken in that iconv() conversion to/from UCS4 fails with E2BIG // if output buffer is _exactly_ as big as needed. Such case is (unless there's @@ -565,11 +472,7 @@ public: #define ICONV_FAILED(cres, bufLeft) (cres == (size_t)-1) #endif -#ifdef WX_ICONV_TAKES_CHAR - #define ICONV_CHAR_CAST(x) (char **)x -#else - #define ICONV_CHAR_CAST(x) (const char **)x -#endif +#define ICONV_CHAR_CAST(x) ((ICONV_CONST char **)(x)) // ---------------------------------------------------------------------------- // IC_CharSet: encapsulates an iconv character set @@ -600,10 +503,11 @@ private: // true if the wide char encoding we use (i.e. ms_wcCharsetName) has // different endian-ness than the native one - static bool ms_wcNeedsSwap = FALSE; + static bool ms_wcNeedsSwap; }; const char *IC_CharSet::ms_wcCharsetName = NULL; +bool IC_CharSet::ms_wcNeedsSwap = FALSE; IC_CharSet::IC_CharSet(const wxChar *name) : wxCharacterSet(name) @@ -663,7 +567,11 @@ IC_CharSet::IC_CharSet(const wxChar *name) else { ms_wcCharsetName = NULL; - wxLogError(_("Impossible to convert to/from charset '%s'."), name); + + // VS: we must not output an error here, since wxWindows will safely + // fall back to using wxEncodingConverter. + wxLogTrace(wxT("strconv"), wxT("Impossible to convert to/from charset '%s' with iconv, falling back to wxEncodingConverter."), name); + //wxLogError( } } wxLogTrace(wxT("strconv"), wxT("wchar_t charset is '%s', needs swap: %i"), ms_wcCharsetName, ms_wcNeedsSwap); @@ -678,6 +586,10 @@ IC_CharSet::IC_CharSet(const wxChar *name) { w2m = iconv_open(wxConvLibc.cWX2MB(name), ms_wcCharsetName); } + else + { + w2m = (iconv_t)-1; + } } IC_CharSet::~IC_CharSet() @@ -800,13 +712,16 @@ size_t IC_CharSet::WC2MB(char *buf, const wchar_t *psz, size_t n) return res; } -#endif // HAVE_ICONV_H +#endif // HAVE_ICONV // ============================================================================ // Win32 conversion classes // ============================================================================ #if defined(__WIN32__) && !defined(__WXMICROWIN__) + +extern long wxCharsetToCodepage(const wxChar *charset); // from utils.cpp + class CP_CharSet : public wxCharacterSet { public: @@ -834,7 +749,7 @@ public: return len ? (buf ? len : len-1) : (size_t)-1; } - bool usable() + bool usable() const { return m_CodePage != -1; } public: @@ -858,11 +773,12 @@ public: { if (name) enc = wxTheFontMapper->CharsetToEncoding(name, FALSE); - m2w.Init(enc, wxFONTENCODING_UNICODE); - w2m.Init(wxFONTENCODING_UNICODE, enc); + + m_ok = m2w.Init(enc, wxFONTENCODING_UNICODE) && + w2m.Init(wxFONTENCODING_UNICODE, enc); } - size_t MB2WC(wchar_t *buf, const char *psz, size_t n) + size_t MB2WC(wchar_t *buf, const char *psz, size_t WXUNUSED(n)) { size_t inbuf = strlen(psz); if (buf) @@ -870,7 +786,7 @@ public: return inbuf; } - size_t WC2MB(char *buf, const wchar_t *psz, size_t n) + size_t WC2MB(char *buf, const wchar_t *psz, size_t WXUNUSED(n)) { #if ( defined(__BORLANDC__) && (__BORLANDC__ > 0x530) ) \ || ( defined(__MWERKS__) && defined(__WXMSW__) ) @@ -884,12 +800,14 @@ public: return inbuf; } - bool usable() - { return (enc!=wxFONTENCODING_SYSTEM) && (enc!=wxFONTENCODING_DEFAULT); } + bool usable() const { return m_ok; } public: wxFontEncoding enc; wxEncodingConverter m2w, w2m; + + // were we initialized successfully? + bool m_ok; }; #endif // wxUSE_FONTMAP @@ -901,46 +819,67 @@ public: static wxCharacterSet *wxGetCharacterSet(const wxChar *name) { - wxCharacterSet *cset = NULL; - if (name) + // check for the special case of ASCII charset +#if wxUSE_FONTMAP + if ( wxTheFontMapper->CharsetToEncoding(name) == wxFONTENCODING_DEFAULT ) +#else // wxUSE_FONTMAP + if ( !name ) +#endif // wxUSE_FONTMAP/!wxUSE_FONTMAP { - if (wxStricmp(name, wxT("UTF8")) == 0 || wxStricmp(name, wxT("UTF-8")) == 0) - { - cset = new ID_CharSet(name, &wxConvUTF8); - } - else - { -#ifdef HAVE_ICONV_H - cset = new IC_CharSet(name); // may not take NULL -#endif - } + // don't convert at all + return NULL; } - if (cset && cset->usable()) - return cset; + // the test above must have taken care of this case + wxCHECK_MSG( name, NULL, _T("NULL name must be wxFONTENCODING_DEFAULT") ); + + wxCharacterSet *cset; + + if ( wxStricmp(name, wxT("UTF8")) == 0 || wxStricmp(name, wxT("UTF-8")) == 0) + { + cset = new ID_CharSet(name, &wxConvUTF8); + } + else + { +#ifdef HAVE_ICONV + cset = new IC_CharSet(name); +#else // !HAVE_ICONV + cset = NULL; +#endif // HAVE_ICONV/!HAVE_ICONV + } - if (cset) + // it can only be NULL in this case +#ifndef HAVE_ICONV + if ( cset ) +#endif // !HAVE_ICONV { + if ( cset->usable() ) + return cset; + delete cset; cset = NULL; } #if defined(__WIN32__) && !defined(__WXMICROWIN__) - cset = new CP_CharSet(name); // may take NULL - if (cset->usable()) + cset = new CP_CharSet(name); + if ( cset->usable() ) return cset; delete cset; + cset = NULL; #endif // __WIN32__ #if wxUSE_FONTMAP cset = new EC_CharSet(name); - if (cset->usable()) + if ( cset->usable() ) return cset; -#endif // wxUSE_FONTMAP delete cset; - wxLogError(_("Unknown encoding '%s'!"), name); + cset = NULL; +#endif // wxUSE_FONTMAP + + wxLogError(_("Cannot convert from encoding '%s'!"), name); + return NULL; } @@ -1031,69 +970,6 @@ size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const return len; } -// VZ: are the classes below used at all?? -#if 0 - -#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; - size_t res = iconv( cnv, ICONV_CHAR_CAST(&psz), &inbuf, &buf, &outbuf ); - if (res == (size_t)-1) - return (size_t)-1; - return (n - outbuf); - } - -public: - iconv_t cnv; -}; - -#endif // HAVE_ICONV_H - -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; -}; - -#endif // 0 - -#else // !wxUSE_WCHAR_T - -// ---------------------------------------------------------------------------- -// stand-ins in absence of wchar_t -// ---------------------------------------------------------------------------- - -WXDLLEXPORT_DATA(wxMBConv) wxConvLibc, wxConvFile; - #endif // wxUSE_WCHAR_T