X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/d36c9347ea16171bc2f855076d8b9b11801ec622..82c6387837096ddd8c0d8f3a2ef7b9d889ec8f75:/include/wx/strconv.h diff --git a/include/wx/strconv.h b/include/wx/strconv.h index 73620ad125..96dc8d8a70 100644 --- a/include/wx/strconv.h +++ b/include/wx/strconv.h @@ -1,11 +1,12 @@ /////////////////////////////////////////////////////////////////////////////// // Name: strconv.h // Purpose: conversion routines for char sets any Unicode -// Author: Robert Roebling, Ove Kaaven +// Author: Ove Kaaven, Robert Roebling, Vadim Zeitlin // Modified by: // Created: 29/01/98 // RCS-ID: $Id$ -// Copyright: (c) 1998 Ove Kaaven, Robert Roebling, Vadim Zeitlin +// Copyright: (c) 1998 Ove Kaaven, Robert Roebling +// (c) 1998-2006 Vadim Zeitlin // Licence: wxWindows licence /////////////////////////////////////////////////////////////////////////////// @@ -13,7 +14,7 @@ #define _WX_STRCONV_H_ #include "wx/defs.h" -#include "wx/wxchar.h" +#include "wx/chartype.h" #include "wx/buffer.h" #ifdef __DIGITALMARS__ @@ -28,9 +29,15 @@ #if wxUSE_WCHAR_T +class WXDLLIMPEXP_FWD_BASE wxString; + // the error value returned by wxMBConv methods #define wxCONV_FAILED ((size_t)-1) +// the default value for some length parameters meaning that the string is +// NUL-terminated +#define wxNO_LEN ((size_t)-1) + // ---------------------------------------------------------------------------- // wxMBConv (abstract base class for conversions) // ---------------------------------------------------------------------------- @@ -54,12 +61,13 @@ public: // characters, not bytes) of the converted string including any trailing // L'\0' or (possibly multiple) '\0'(s). If the conversion fails or if // there is not enough space for everything, including the trailing NUL - // character(s), in the output buffer, (size_t)-1 is returned. + // character(s), in the output buffer, wxCONV_FAILED is returned. // // In the special case when dstLen is 0 (outputBuf may be NULL then) the // return value is the length of the needed buffer but nothing happens - // otherwise. If srcLen is -1, the entire string, up to and including the - // trailing NUL(s), is converted, otherwise exactly srcLen bytes are. + // otherwise. If srcLen is wxNO_LEN, the entire string, up to and + // including the trailing NUL(s), is converted, otherwise exactly srcLen + // bytes are. // // Typical usage: // @@ -70,10 +78,10 @@ public: // conv.ToWChar(wbuf, dstLen, src); // virtual size_t ToWChar(wchar_t *dst, size_t dstLen, - const char *src, size_t srcLen = -1) const; + 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 = -1) const; + const wchar_t *src, size_t srcLen = wxNO_LEN) const; // Convenience functions for translating NUL-terminated strings: returns @@ -129,6 +137,12 @@ 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() @@ -169,6 +183,10 @@ 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__ @@ -183,9 +201,10 @@ public: class WXDLLIMPEXP_BASE wxConvBrokenFileNames : public wxMBConv { public: - wxConvBrokenFileNames(const wxChar *charset); + wxConvBrokenFileNames(const wxString& charset); wxConvBrokenFileNames(const wxConvBrokenFileNames& conv) - : m_conv(conv.m_conv ? conv.m_conv->Clone() : NULL) + : wxMBConv(), + m_conv(conv.m_conv ? conv.m_conv->Clone() : NULL) { } virtual ~wxConvBrokenFileNames() { delete m_conv; } @@ -206,6 +225,10 @@ 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: @@ -234,7 +257,29 @@ public: // 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 @@ -245,11 +290,20 @@ 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; }; @@ -261,7 +315,16 @@ private: class WXDLLIMPEXP_BASE wxMBConvUTF16Base : public wxMBConv { public: - virtual size_t GetMBNulLen() const { return 2; } + enum { BYTES_PER_CHAR = 2 }; + + virtual size_t GetMBNulLen() const { return BYTES_PER_CHAR; } + +protected: + // return the length of the buffer using srcLen if it's not wxNO_LEN and + // computing the length ourselves if it is; also checks that the length is + // even if specified as we need an entire number of UTF-16 characters and + // returns wxNO_LEN which indicates error if it is odd + static size_t GetLength(const char *src, size_t srcLen); }; // ---------------------------------------------------------------------------- @@ -271,8 +334,10 @@ public: class WXDLLIMPEXP_BASE wxMBConvUTF16LE : public wxMBConvUTF16Base { 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; + 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 wxMBConvUTF16LE; } }; @@ -283,8 +348,10 @@ public: class WXDLLIMPEXP_BASE wxMBConvUTF16BE : public wxMBConvUTF16Base { 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; + 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 wxMBConvUTF16BE; } }; @@ -295,7 +362,15 @@ public: class WXDLLIMPEXP_BASE wxMBConvUTF32Base : public wxMBConv { public: - virtual size_t GetMBNulLen() const { return 4; } + enum { BYTES_PER_CHAR = 4 }; + + virtual size_t GetMBNulLen() const { return BYTES_PER_CHAR; } + +protected: + // this is similar to wxMBConvUTF16Base method with the same name except + // that, of course, it verifies that length is divisible by 4 if given and + // not by 2 + static size_t GetLength(const char *src, size_t srcLen); }; // ---------------------------------------------------------------------------- @@ -305,8 +380,10 @@ public: class WXDLLIMPEXP_BASE wxMBConvUTF32LE : public wxMBConvUTF32Base { 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; + 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 wxMBConvUTF32LE; } }; @@ -317,8 +394,10 @@ public: class WXDLLIMPEXP_BASE wxMBConvUTF32BE : public wxMBConvUTF32Base { 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; + 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 wxMBConvUTF32BE; } }; @@ -333,7 +412,7 @@ 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); @@ -341,13 +420,25 @@ public: wxCSConv& operator=(const wxCSConv& conv); + 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 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 GetMBNulLen() const; + +#if wxUSE_UNICODE_UTF8 + virtual bool IsUTF8() const; +#endif + virtual wxMBConv *Clone() const { return new wxCSConv(*this); } void Clear(); + // return true if the conversion could be initilized successfully + bool IsOk() const; + private: // common part of all ctors void Init(); @@ -360,12 +451,12 @@ 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! @@ -378,14 +469,35 @@ 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) @@ -403,9 +515,17 @@ extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvFileName; // default in a couple of places inside wx (initially same as wxConvLibc) extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent; -// ??? -extern WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvLocal; +// the conversion corresponding to the current locale +WX_DECLARE_GLOBAL_CONV(wxCSConv, wxConvLocal) +#define wxConvLocal wxGet_wxConvLocal() +// the conversion corresponding to the encoding of the standard UI elements +// +// by default this is the same as wxConvLocal but may be changed if the program +// needs to use a fixed encoding +extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvUI; + +#undef WX_DECLARE_GLOBAL_CONV // ---------------------------------------------------------------------------- // endianness-dependent conversions @@ -423,13 +543,7 @@ extern WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvLocal; // filename conversion macros // ---------------------------------------------------------------------------- -// filenames are multibyte on Unix and probably widechar on Windows? -#if defined(__UNIX__) || defined(__BORLANDC__) || defined(__WXMAC__ ) - #define wxMBFILES 1 -#else - #define wxMBFILES 0 -#endif - +// filenames are multibyte on Unix and widechar on Windows #if wxMBFILES && wxUSE_UNICODE #define wxFNCONV(name) wxConvFileName->cWX2MB(name) #define wxFNSTRINGCAST wxMBSTRINGCAST @@ -476,10 +590,25 @@ extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent; #if wxUSE_UNICODE #define wxConvertWX2MB(s) wxConvCurrent->cWX2MB(s) #define wxConvertMB2WX(s) wxConvCurrent->cMB2WX(s) + + // these functions should be used when the conversions really, really have + // to succeed (usually because we pass their results to a standard C + // function which would crash if we passed NULL to it), so these functions + // always return a valid pointer if their argument is non-NULL + + // this function safety is achieved by trying wxConvLibc first, wxConvUTF8 + // next if it fails and, finally, wxConvISO8859_1 which always succeeds + extern WXDLLIMPEXP_BASE wxWCharBuffer wxSafeConvertMB2WX(const char *s); + + // this function uses wxConvLibc and wxConvUTF8(MAP_INVALID_UTF8_TO_OCTAL) + // if it fails + extern WXDLLIMPEXP_BASE wxCharBuffer wxSafeConvertWX2MB(const wchar_t *ws); #else // ANSI // no conversions to do #define wxConvertWX2MB(s) (s) #define wxConvertMB2WX(s) (s) + #define wxSafeConvertMB2WX(s) (s) + #define wxSafeConvertWX2MB(s) (s) #endif // Unicode/ANSI #endif // _WX_STRCONV_H_