X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/69c928ef472f89987f95e2fa798e2ba2d23eaf31..05b4b8ee882d86d1095efe72e5c866d80bb4cc95:/include/wx/strconv.h diff --git a/include/wx/strconv.h b/include/wx/strconv.h index 4e4afff9dc..1aa5f97550 100644 --- a/include/wx/strconv.h +++ b/include/wx/strconv.h @@ -14,7 +14,7 @@ #define _WX_STRCONV_H_ #include "wx/defs.h" -#include "wx/wxchar.h" +#include "wx/chartype.h" #include "wx/buffer.h" #ifdef __DIGITALMARS__ @@ -25,10 +25,14 @@ # undef __BSEXCPT__ #endif +#ifndef __WXPALMOS5__ #include +#endif // ! __WXPALMOS5__ #if wxUSE_WCHAR_T +class WXDLLIMPEXP_FWD_BASE wxString; + // the error value returned by wxMBConv methods #define wxCONV_FAILED ((size_t)-1) @@ -135,6 +139,12 @@ public: // encoding static size_t GetMaxMBNulLen() { return 4 /* for UTF-32 */; } +#if wxUSE_UNICODE_UTF8 + // return true if the converter's charset is UTF-8, i.e. char* strings + // decoded using this object can be directly copied to wxString's internal + // storage without converting to WC and than back to UTF-8 MB string + virtual bool IsUTF8() const { return false; } +#endif // The old conversion functions. The existing classes currently mostly // implement these ones but we're in transition to using To/FromWChar() @@ -175,6 +185,10 @@ public: virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const; virtual wxMBConv *Clone() const { return new wxMBConvLibc; } + +#if wxUSE_UNICODE_UTF8 + virtual bool IsUTF8() const { return wxLocaleIsUtf8; } +#endif }; #ifdef __UNIX__ @@ -189,7 +203,7 @@ public: class WXDLLIMPEXP_BASE wxConvBrokenFileNames : public wxMBConv { public: - wxConvBrokenFileNames(const wxChar *charset); + wxConvBrokenFileNames(const wxString& charset); wxConvBrokenFileNames(const wxConvBrokenFileNames& conv) : wxMBConv(), m_conv(conv.m_conv ? conv.m_conv->Clone() : NULL) @@ -213,6 +227,10 @@ public: return m_conv->GetMBNulLen(); } +#if wxUSE_UNICODE_UTF8 + virtual bool IsUTF8() const { return m_conv->IsUTF8(); } +#endif + virtual wxMBConv *Clone() const { return new wxConvBrokenFileNames(*this); } private: @@ -231,17 +249,113 @@ private: class WXDLLIMPEXP_BASE wxMBConvUTF7 : public wxMBConv { public: - virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const; - virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const; + wxMBConvUTF7() { } + + // compiler-generated copy ctor, assignment operator and dtor are ok + // (assuming it's ok to copy the shift state -- not really sure about it) + + virtual size_t ToWChar(wchar_t *dst, size_t dstLen, + const char *src, size_t srcLen = wxNO_LEN) const; + virtual size_t FromWChar(char *dst, size_t dstLen, + const wchar_t *src, size_t srcLen = wxNO_LEN) const; virtual wxMBConv *Clone() const { return new wxMBConvUTF7; } + +private: + // UTF-7 decoder/encoder may be in direct mode or in shifted mode after a + // '+' (and until the '-' or any other non-base64 character) + struct StateMode + { + enum Mode + { + Direct, // pass through state + Shifted // after a '+' (and before '-') + }; + }; + + // the current decoder state: this is only used by ToWChar() if srcLen + // parameter is not wxNO_LEN, when working on the entire NUL-terminated + // strings we neither update nor use the state + class DecoderState : StateMode + { + private: + // current state: this one is private as we want to enforce the use of + // ToDirect/ToShifted() methods below + Mode mode; + + public: + // the initial state is direct + DecoderState() { mode = Direct; } + + // switch to/from shifted mode + void ToDirect() { mode = Direct; } + void ToShifted() { mode = Shifted; accum = bit = 0; isLSB = false; } + + bool IsDirect() const { return mode == Direct; } + bool IsShifted() const { return mode == Shifted; } + + + // these variables are only used in shifted mode + + unsigned int accum; // accumulator of the bit we've already got + unsigned int bit; // the number of bits consumed mod 8 + unsigned char msb; // the high byte of UTF-16 word + bool isLSB; // whether we're decoding LSB or MSB of UTF-16 word + }; + + DecoderState m_stateDecoder; + + + // encoder state is simpler as we always receive entire Unicode characters + // on input + class EncoderState : StateMode + { + private: + Mode mode; + + public: + EncoderState() { mode = Direct; } + + void ToDirect() { mode = Direct; } + void ToShifted() { mode = Shifted; accum = bit = 0; } + + bool IsDirect() const { return mode == Direct; } + bool IsShifted() const { return mode == Shifted; } + + unsigned int accum; + unsigned int bit; + }; + + EncoderState m_stateEncoder; }; // ---------------------------------------------------------------------------- // wxMBConvUTF8 (for conversion using UTF8 encoding) // ---------------------------------------------------------------------------- -class WXDLLIMPEXP_BASE wxMBConvUTF8 : public wxMBConv +// this is the real UTF-8 conversion class, it has to be called "strict UTF-8" +// for compatibility reasons: the wxMBConvUTF8 class below also supports lossy +// conversions if it is created with non default options +class WXDLLIMPEXP_BASE wxMBConvStrictUTF8 : public wxMBConv +{ +public: + // compiler-generated default ctor and other methods are ok + + virtual size_t ToWChar(wchar_t *dst, size_t dstLen, + const char *src, size_t srcLen = wxNO_LEN) const; + virtual size_t FromWChar(char *dst, size_t dstLen, + const wchar_t *src, size_t srcLen = wxNO_LEN) const; + + virtual wxMBConv *Clone() const { return new wxMBConvStrictUTF8(); } + +#if wxUSE_UNICODE_UTF8 + // NB: other mapping modes are not, strictly speaking, UTF-8, so we can't + // take the shortcut in that case + virtual bool IsUTF8() const { return true; } +#endif +}; + +class WXDLLIMPEXP_BASE wxMBConvUTF8 : public wxMBConvStrictUTF8 { public: enum @@ -252,11 +366,20 @@ public: }; wxMBConvUTF8(int options = MAP_INVALID_UTF8_NOT) : m_options(options) { } - virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const; - virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const; + + virtual size_t ToWChar(wchar_t *dst, size_t dstLen, + const char *src, size_t srcLen = wxNO_LEN) const; + virtual size_t FromWChar(char *dst, size_t dstLen, + const wchar_t *src, size_t srcLen = wxNO_LEN) const; virtual wxMBConv *Clone() const { return new wxMBConvUTF8(m_options); } +#if wxUSE_UNICODE_UTF8 + // NB: other mapping modes are not, strictly speaking, UTF-8, so we can't + // take the shortcut in that case + virtual bool IsUTF8() const { return m_options == MAP_INVALID_UTF8_NOT; } +#endif + private: int m_options; }; @@ -365,7 +488,7 @@ class WXDLLIMPEXP_BASE wxCSConv : public wxMBConv public: // we can be created either from charset name or from an encoding constant // but we can't have both at once - wxCSConv(const wxChar *charset); + wxCSConv(const wxString& charset); wxCSConv(wxFontEncoding encoding); wxCSConv(const wxCSConv& conv); @@ -381,6 +504,10 @@ public: virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const; virtual size_t GetMBNulLen() const; +#if wxUSE_UNICODE_UTF8 + virtual bool IsUTF8() const; +#endif + virtual wxMBConv *Clone() const { return new wxCSConv(*this); } void Clear(); @@ -400,12 +527,12 @@ private: // set the name (may be only called when m_name == NULL), makes copy of // charset string - void SetName(const wxChar *charset); + void SetName(const char *charset); // note that we can't use wxString here because of compilation // dependencies: we're included from wx/string.h - wxChar *m_name; + char *m_name; wxFontEncoding m_encoding; // use CreateConvIfNeeded() before accessing m_convReal! @@ -418,14 +545,35 @@ private: // declare predefined conversion objects // ---------------------------------------------------------------------------- +// Note: this macro is an implementation detail (see the comment in +// strconv.cpp). The wxGet_XXX() and wxGet_XXXPtr() functions shouldn't be +// used by user code and neither should XXXPtr, use the wxConvXXX macro +// instead. +#define WX_DECLARE_GLOBAL_CONV(klass, name) \ + extern WXDLLIMPEXP_DATA_BASE(klass*) name##Ptr; \ + extern WXDLLIMPEXP_BASE klass* wxGet_##name##Ptr(); \ + inline klass& wxGet_##name() \ + { \ + if ( !name##Ptr ) \ + name##Ptr = wxGet_##name##Ptr(); \ + return *name##Ptr; \ + } + + // conversion to be used with all standard functions affected by locale, e.g. // strtol(), strftime(), ... -extern WXDLLIMPEXP_DATA_BASE(wxMBConv&) wxConvLibc; +WX_DECLARE_GLOBAL_CONV(wxMBConv, wxConvLibc) +#define wxConvLibc wxGet_wxConvLibc() // conversion ISO-8859-1/UTF-7/UTF-8 <-> wchar_t -extern WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvISO8859_1; -extern WXDLLIMPEXP_DATA_BASE(wxMBConvUTF7&) wxConvUTF7; -extern WXDLLIMPEXP_DATA_BASE(wxMBConvUTF8&) wxConvUTF8; +WX_DECLARE_GLOBAL_CONV(wxCSConv, wxConvISO8859_1) +#define wxConvISO8859_1 wxGet_wxConvISO8859_1() + +WX_DECLARE_GLOBAL_CONV(wxMBConvStrictUTF8, wxConvUTF8) +#define wxConvUTF8 wxGet_wxConvUTF8() + +WX_DECLARE_GLOBAL_CONV(wxMBConvUTF7, wxConvUTF7) +#define wxConvUTF7 wxGet_wxConvUTF7() // conversion used for the file names on the systems where they're not Unicode // (basically anything except Windows) @@ -444,7 +592,8 @@ extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvFileName; extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent; // the conversion corresponding to the current locale -extern WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvLocal; +WX_DECLARE_GLOBAL_CONV(wxCSConv, wxConvLocal) +#define wxConvLocal wxGet_wxConvLocal() // the conversion corresponding to the encoding of the standard UI elements // @@ -452,6 +601,8 @@ extern WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvLocal; // needs to use a fixed encoding extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvUI; +#undef WX_DECLARE_GLOBAL_CONV + // ---------------------------------------------------------------------------- // endianness-dependent conversions // ---------------------------------------------------------------------------- @@ -469,12 +620,6 @@ extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvUI; // ---------------------------------------------------------------------------- // filenames are multibyte on Unix and widechar on Windows -#if defined(__UNIX__) || defined(__WXMAC__) - #define wxMBFILES 1 -#else - #define wxMBFILES 0 -#endif - #if wxMBFILES && wxUSE_UNICODE #define wxFNCONV(name) wxConvFileName->cWX2MB(name) #define wxFNSTRINGCAST wxMBSTRINGCAST