X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/4f61e22c77b1a7cb148f6567996d816f72ac854f..65dc921d3f1080b8b5c6ca49d2cce228d6924480:/src/common/strconv.cpp diff --git a/src/common/strconv.cpp b/src/common/strconv.cpp index 17a15ca5a8..bbd8ce8369 100644 --- a/src/common/strconv.cpp +++ b/src/common/strconv.cpp @@ -6,7 +6,7 @@ // Created: 29/01/98 // RCS-ID: $Id$ // Copyright: (c) 1999 Ove Kaaven, Robert Roebling, Vadim Zeitlin, Vaclav Slavik -// Licence: wxWindows license +// Licence: wxWindows licence ///////////////////////////////////////////////////////////////////////////// // ============================================================================ @@ -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 +42,67 @@ #include #include - -#include "wx/debug.h" +#include "wx/module.h" #include "wx/strconv.h" -#include "wx/intl.h" -#include "wx/log.h" // ---------------------------------------------------------------------------- // globals // ---------------------------------------------------------------------------- +#if wxUSE_WCHAR_T + WXDLLEXPORT_DATA(wxMBConv) wxConvLibc; + WXDLLEXPORT_DATA(wxCSConv) wxConvLocal((const wxChar *)NULL); + WXDLLEXPORT_DATA(wxCSConv) wxConvISO8859_1(_T("iso-8859-1")); +#else + // stand-ins in absence of wchar_t + WXDLLEXPORT_DATA(wxMBConv) wxConvLibc, + wxConvFile, + wxConvISO8859_1, + wxConvLocal, + wxConvUTF8; +#endif // wxUSE_WCHAR_T + WXDLLEXPORT_DATA(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) -// ============================================================================ -// implementation -// ============================================================================ + +// ---------------------------------------------------------------------------- +// headers +// ---------------------------------------------------------------------------- #if wxUSE_WCHAR_T #ifdef __SALFORDC__ - #include + #include #endif #ifdef HAVE_ICONV - #include + #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 +160,8 @@ 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; } @@ -149,63 +192,87 @@ static size_t decode_utf16(const wchar_t* input, wxUint32& output) // wxMBConv // ---------------------------------------------------------------------------- -WXDLLEXPORT_DATA(wxMBConv) wxConvLibc; +#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) + if ( psz ) { - 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; + // 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; + } } - else - return wxWCharBuffer((wchar_t *) NULL); + + wxWCharBuffer buf((wchar_t *)NULL); + + return buf; } -const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *psz) const +const wxCharBuffer wxMBConv::cWC2MB(const wchar_t *pwz) const { - if (psz) + if ( pwz ) { - 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; - } - else - return wxCharBuffer((char *) NULL); -} - -// ---------------------------------------------------------------------------- -// standard file conversion -// ---------------------------------------------------------------------------- + size_t nLen = WC2MB(NULL, pwz, 0); + if ( nLen != (size_t)-1 ) + { + wxCharBuffer buf(nLen); + WC2MB(buf.data(), pwz, nLen + 1); -WXDLLEXPORT_DATA(wxMBConvFile) wxConvFile; + return buf; + } + } -// just use the libc conversion for now -size_t wxMBConvFile::MB2WC(wchar_t *buf, const char *psz, size_t n) const -{ - return wxMB2WC(buf, psz, n); -} + wxCharBuffer buf((char *)NULL); -size_t wxMBConvFile::WC2MB(char *buf, const wchar_t *psz, size_t n) const -{ - return wxWC2MB(buf, psz, n); + return buf; } // ---------------------------------------------------------------------------- @@ -341,11 +408,11 @@ 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 +440,7 @@ 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++; } @@ -382,121 +449,18 @@ 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)); } } } if (buf && (len 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; + return len; } -#endif // !wxUSE_GUI || !wxUSE_FONTMAP - -#endif // Win32 - // ============================================================================ // wxCharacterSet and derived classes // ============================================================================ @@ -600,6 +564,11 @@ bool IC_CharSet::ms_wcNeedsSwap = FALSE; IC_CharSet::IC_CharSet(const wxChar *name) : wxCharacterSet(name) { + // Do it the hard way + char cname[100]; + for (size_t i = 0; i < wxStrlen(name)+1; i++) + cname[i] = (char) name[i]; + // check for charset that represents wchar_t: if (ms_wcCharsetName == NULL) { @@ -607,20 +576,20 @@ IC_CharSet::IC_CharSet(const wxChar *name) // try charset with explicit bytesex info (e.g. "UCS-4LE"): ms_wcCharsetName = WC_NAME_BEST; - m2w = iconv_open(ms_wcCharsetName, wxConvLibc.cWX2MB(name)); + m2w = iconv_open(ms_wcCharsetName, cname); if (m2w == (iconv_t)-1) { // try charset w/o bytesex info (e.g. "UCS4") // and check for bytesex ourselves: ms_wcCharsetName = WC_NAME; - m2w = iconv_open(ms_wcCharsetName, wxConvLibc.cWX2MB(name)); + m2w = iconv_open(ms_wcCharsetName, cname); // last bet, try if it knows WCHAR_T pseudo-charset if (m2w == (iconv_t)-1) { ms_wcCharsetName = "WCHAR_T"; - m2w = iconv_open(ms_wcCharsetName, wxConvLibc.cWX2MB(name)); + m2w = iconv_open(ms_wcCharsetName, cname); } if (m2w != (iconv_t)-1) @@ -645,7 +614,7 @@ IC_CharSet::IC_CharSet(const wxChar *name) { ms_wcCharsetName = NULL; wxLogLastError(wxT("iconv")); - wxLogError(_("Convertion to charset '%s' doesn't work."), name); + wxLogError(_("Conversion to charset '%s' doesn't work."), name); } else { @@ -655,7 +624,7 @@ IC_CharSet::IC_CharSet(const wxChar *name) else { ms_wcCharsetName = NULL; - + // 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); @@ -666,13 +635,13 @@ IC_CharSet::IC_CharSet(const wxChar *name) } else // we already have ms_wcCharsetName { - m2w = iconv_open(ms_wcCharsetName, wxConvLibc.cWX2MB(name)); + m2w = iconv_open(ms_wcCharsetName, cname); } // NB: don't ever pass NULL to iconv_open(), it may crash! if ( ms_wcCharsetName ) { - w2m = iconv_open(wxConvLibc.cWX2MB(name), ms_wcCharsetName); + w2m = iconv_open( cname, ms_wcCharsetName); } else { @@ -710,6 +679,12 @@ size_t IC_CharSet::MB2WC(wchar_t *buf, const char *psz, size_t n) // convert to native endianness WC_BSWAP(buf /* _not_ bufPtr */, res) } + + // NB: iconv was given only strlen(psz) characters on input, and so + // it couldn't convert the trailing zero. Let's do it ourselves + // if there's some room left for it in the output buffer. + if (res < n) + buf[res] = 0; } else { @@ -741,11 +716,7 @@ size_t IC_CharSet::MB2WC(wchar_t *buf, const char *psz, size_t n) size_t IC_CharSet::WC2MB(char *buf, const wchar_t *psz, size_t n) { -#if defined(__BORLANDC__) && (__BORLANDC__ > 0x530) - size_t inbuf = std::wcslen(psz) * SIZEOF_WCHAR_T; -#else - size_t inbuf = ::wcslen(psz) * SIZEOF_WCHAR_T; -#endif + size_t inbuf = wxWcslen(psz) * SIZEOF_WCHAR_T; size_t outbuf = n; size_t res, cres; @@ -769,6 +740,12 @@ size_t IC_CharSet::WC2MB(char *buf, const wchar_t *psz, size_t n) cres = iconv( w2m, ICONV_CHAR_CAST(&psz), &inbuf, &buf, &outbuf ); res = n-outbuf; + + // NB: iconv was given only wcslen(psz) characters on input, and so + // it couldn't convert the trailing zero. Let's do it ourselves + // if there's some room left for it in the output buffer. + if (res < n) + buf[0] = 0; } else { @@ -806,7 +783,10 @@ size_t IC_CharSet::WC2MB(char *buf, const wchar_t *psz, size_t n) // Win32 conversion classes // ============================================================================ -#if defined(__WIN32__) && !defined(__WXMICROWIN__) +#if defined(__WIN32__) && !defined(__WXMICROWIN__) && !defined(__WXUNIVERSAL__) + +extern long wxCharsetToCodepage(const wxChar *charset); // from utils.cpp + class CP_CharSet : public wxCharacterSet { public: @@ -818,20 +798,37 @@ public: size_t MB2WC(wchar_t *buf, const char *psz, size_t n) { - size_t len = - MultiByteToWideChar(m_CodePage, 0, psz, -1, buf, buf ? n : 0); - //VS: returns # of written chars for buf!=NULL and *size* - // needed buffer for buf==NULL - return len ? (buf ? len : len-1) : (size_t)-1; + const size_t len = ::MultiByteToWideChar + ( + m_CodePage, // code page + 0, // flags (none) + psz, // input string + -1, // its length (NUL-terminated) + buf, // output string + buf ? n : 0 // size of output buffer + ); + + // note that it returns # of written chars for buf != NULL and *size* + // of the needed buffer for buf == NULL + return len ? (buf ? len : len - 1) : (size_t)-1; } size_t WC2MB(char *buf, const wchar_t *psz, size_t n) { - size_t len = WideCharToMultiByte(m_CodePage, 0, psz, -1, buf, - buf ? n : 0, NULL, NULL); - //VS: returns # of written chars for buf!=NULL and *size* - // needed buffer for buf==NULL - return len ? (buf ? len : len-1) : (size_t)-1; + const size_t len = ::WideCharToMultiByte + ( + m_CodePage, // code page + 0, // flags (none) + psz, // input string + -1, // it is (wide) NUL-terminated + buf, // output buffer + buf ? n : 0, // and its size + NULL, // default "replacement" char + NULL // [out] was it used? + ); + + // see the comment above! + return len ? (buf ? len : len - 1) : (size_t)-1; } bool usable() const @@ -840,7 +837,7 @@ public: public: long m_CodePage; }; -#endif // __WIN32__ +#endif // defined(__WIN32__) && !defined(__WXMICROWIN__) && !defined(__WXUNIVERSAL__) // ============================================================================ // wxEncodingConverter based conversion classes @@ -857,12 +854,13 @@ public: enc(wxFONTENCODING_SYSTEM) { if (name) - enc = wxTheFontMapper->CharsetToEncoding(name, FALSE); - m2w.Init(enc, wxFONTENCODING_UNICODE); - w2m.Init(wxFONTENCODING_UNICODE, enc); + enc = wxFontMapper::Get()->CharsetToEncoding(name, FALSE); + + 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) @@ -870,26 +868,23 @@ 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__) ) - size_t inbuf = std::wcslen(psz); -#else - size_t inbuf = ::wcslen(psz); -#endif + const size_t inbuf = wxWcslen(psz); if (buf) w2m.Convert(psz,buf); return inbuf; } - bool usable() const - { 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 +896,67 @@ public: static wxCharacterSet *wxGetCharacterSet(const wxChar *name) { - wxCharacterSet *cset = NULL; - if (name) + // check for the special case of ASCII charset +#if wxUSE_FONTMAP + if ( wxFontMapper::Get()->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 - 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") ); - if (cset) + 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 + } + + // 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()) +#if defined(__WIN32__) && !defined(__WXMICROWIN__) && !defined(__WXUNIVERSAL__) + cset = new CP_CharSet(name); + if ( cset->usable() ) return cset; delete cset; -#endif // __WIN32__ + cset = NULL; +#endif // defined(__WIN32__) && !defined(__WXMICROWIN__) && !defined(__WXUNIVERSAL__) #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; } @@ -959,8 +975,31 @@ wxCSConv::wxCSConv(const wxChar *charset) wxCSConv::~wxCSConv() { - free(m_name); - delete m_cset; + Clear(); +} + +wxCSConv::wxCSConv(const wxCSConv& conv) + : wxMBConv() +{ + Clear(); + SetName(conv.m_name); +} + +wxCSConv& wxCSConv::operator=(const wxCSConv& conv) +{ + Clear(); + SetName(conv.m_name); + return *this; +} + +void wxCSConv::Clear() +{ + if (m_name) + free(m_name); + if (m_cset) + delete m_cset; + m_name = NULL; + m_cset = NULL; } void wxCSConv::SetName(const wxChar *charset) @@ -974,14 +1013,20 @@ void wxCSConv::SetName(const wxChar *charset) void wxCSConv::LoadNow() { - if (m_deferred) + if ( m_deferred ) { + // it would probably be better to make GetSystemEncodingName() always + // available (i.e. even when wxUSE_INTL == 0)? +#if wxUSE_INTL if ( !m_name ) { wxString name = wxLocale::GetSystemEncodingName(); if ( !name.empty() ) + { SetName(name); + } } +#endif // wxUSE_INTL // wxGetCharacterSet() complains about NULL name m_cset = m_name ? wxGetCharacterSet(m_name) : NULL; @@ -1016,12 +1061,7 @@ size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const return m_cset->WC2MB(buf, psz, n); // latin-1 (direct) -#if ( defined(__BORLANDC__) && (__BORLANDC__ > 0x530) ) \ - || ( defined(__MWERKS__) && defined(__WXMSW__) ) - size_t len=std::wcslen(psz); -#else - size_t len=::wcslen(psz); -#endif + const size_t len = wxWcslen(psz); if (buf) { for (size_t c = 0; c <= len; c++) @@ -1031,14 +1071,6 @@ size_t wxCSConv::WC2MB(char *buf, const wchar_t *psz, size_t n) const return len; } -#else // !wxUSE_WCHAR_T - -// ---------------------------------------------------------------------------- -// stand-ins in absence of wchar_t -// ---------------------------------------------------------------------------- - -WXDLLEXPORT_DATA(wxMBConv) wxConvLibc, wxConvFile; - #endif // wxUSE_WCHAR_T