///////////////////////////////////////////////////////////////////////////////
// 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
///////////////////////////////////////////////////////////////////////////////
-#ifndef _WX_WXSTRCONVH__
-#define _WX_WXSTRCONVH__
-
-#if defined(__GNUG__) && !defined(NO_GCC_PRAGMA)
- #pragma interface "strconv.h"
-#endif
+#ifndef _WX_STRCONV_H_
+#define _WX_STRCONV_H_
#include "wx/defs.h"
#include "wx/wxchar.h"
#if wxUSE_WCHAR_T
+// 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)
// ----------------------------------------------------------------------------
+// When deriving a new class from wxMBConv you must reimplement ToWChar() and
+// FromWChar() methods which are not pure virtual only for historical reasons,
+// don't let the fact that the existing classes implement MB2WC/WC2MB() instead
+// confuse you.
+//
+// You also have to implement Clone() to allow copying the conversions
+// polymorphically.
+//
+// And you might need to override GetMBNulLen() as well.
class WXDLLIMPEXP_BASE wxMBConv
{
public:
- // the actual conversion takes place here
+ // The functions doing actual conversion from/to narrow to/from wide
+ // character strings.
+ //
+ // On success, the return value is the length (i.e. the number of
+ // 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, wxCONV_FAILED is returned.
//
- // note that n is the size of the output buffer, not the length of input
- // (the latter is always supposed to be NUL-terminated)
- virtual size_t MB2WC(wchar_t *outputBuf, const char *psz, size_t outputSize) const = 0;
- virtual size_t WC2MB(char *outputBuf, const wchar_t *psz, size_t outputSize) const = 0;
+ // 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 wxNO_LEN, the entire string, up to and
+ // including the trailing NUL(s), is converted, otherwise exactly srcLen
+ // bytes are.
+ //
+ // Typical usage:
+ //
+ // size_t dstLen = conv.ToWChar(NULL, 0, src);
+ // if ( dstLen != wxCONV_FAILED )
+ // ... handle error ...
+ // wchar_t *wbuf = new wchar_t[dstLen];
+ // conv.ToWChar(wbuf, dstLen, src);
+ //
+ 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;
+
- // MB <-> WC
- const wxWCharBuffer cMB2WC(const char *psz) const;
- const wxCharBuffer cWC2MB(const wchar_t *psz) const;
+ // Convenience functions for translating NUL-terminated strings: returns
+ // the buffer containing the converted string or NULL pointer if the
+ // conversion failed.
+ const wxWCharBuffer cMB2WC(const char *in) const;
+ const wxCharBuffer cWC2MB(const wchar_t *in) const;
+
+ // Convenience functions for converting strings which may contain embedded
+ // NULs and don't have to be NUL-terminated.
+ //
+ // inLen is the length of the buffer including trailing NUL if any: if the
+ // last 4 bytes of the buffer are all NULs, these functions are more
+ // efficient as they avoid copying the string, but otherwise a copy is made
+ // internally which could be quite bad for (very) long strings.
+ //
+ // outLen receives, if not NULL, the length of the converted string or 0 if
+ // the conversion failed (returning 0 and not -1 in this case makes it
+ // difficult to distinguish between failed conversion and empty input but
+ // this is done for backwards compatibility)
+ const wxWCharBuffer
+ cMB2WC(const char *in, size_t inLen, size_t *outLen) const;
+ const wxCharBuffer
+ cWC2MB(const wchar_t *in, size_t inLen, size_t *outLen) const;
// convenience functions for converting MB or WC to/from wxWin default
#if wxUSE_UNICODE
const wxWCharBuffer cWX2WC(const char *psz) const { return cMB2WC(psz); }
#endif // Unicode/ANSI
+ // this function is used in the implementation of cMB2WC() to distinguish
+ // between the following cases:
+ //
+ // a) var width encoding with strings terminated by a single NUL
+ // (usual multibyte encodings): return 1 in this case
+ // b) fixed width encoding with 2 bytes/char and so terminated by
+ // 2 NULs (UTF-16/UCS-2 and variants): return 2 in this case
+ // c) fixed width encoding with 4 bytes/char and so terminated by
+ // 4 NULs (UTF-32/UCS-4 and variants): return 4 in this case
+ //
+ // anything else is not supported currently and -1 should be returned
+ virtual size_t GetMBNulLen() const { return 1; }
+
+ // return the maximal value currently returned by GetMBNulLen() for any
+ // encoding
+ static size_t GetMaxMBNulLen() { return 4 /* for UTF-32 */; }
+
+
+ // The old conversion functions. The existing classes currently mostly
+ // implement these ones but we're in transition to using To/FromWChar()
+ // instead and any new classes should implement just the new functions.
+ // For now, however, we provide default implementation of To/FromWChar() in
+ // this base class in terms of MB2WC/WC2MB() to avoid having to rewrite all
+ // the conversions at once.
+ //
+ // On success, the return value is the length (i.e. the number of
+ // characters, not bytes) not counting the trailing NUL(s) of the converted
+ // string. On failure, (size_t)-1 is returned. In the special case when
+ // outputBuf is NULL the return value is the same one but nothing is
+ // written to the buffer.
+ //
+ // Note that outLen is the length of the output buffer, not the length of
+ // the input (which is always supposed to be terminated by one or more
+ // NULs, as appropriate for the encoding)!
+ virtual size_t MB2WC(wchar_t *out, const char *in, size_t outLen) const;
+ virtual size_t WC2MB(char *out, const wchar_t *in, size_t outLen) const;
+
+
+ // make a heap-allocated copy of this object
+ virtual wxMBConv *Clone() const = 0;
+
// virtual dtor for any base class
virtual ~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;
+
+ virtual wxMBConv *Clone() const { return new wxMBConvLibc; }
+};
+
+#ifdef __UNIX__
+
+// ----------------------------------------------------------------------------
+// wxConvBrokenFileNames is made for Unix in Unicode mode when
+// files are accidentally written in an encoding which is not
+// the system encoding. Typically, the system encoding will be
+// UTF8 but there might be files stored in ISO8859-1 on disk.
+// ----------------------------------------------------------------------------
+
+class WXDLLIMPEXP_BASE wxConvBrokenFileNames : public wxMBConv
+{
+public:
+ wxConvBrokenFileNames(const wxChar *charset);
+ wxConvBrokenFileNames(const wxConvBrokenFileNames& conv)
+ : wxMBConv(),
+ m_conv(conv.m_conv ? conv.m_conv->Clone() : NULL)
+ {
+ }
+ virtual ~wxConvBrokenFileNames() { delete m_conv; }
+
+ virtual size_t MB2WC(wchar_t *out, const char *in, size_t outLen) const
+ {
+ return m_conv->MB2WC(out, in, outLen);
+ }
+
+ virtual size_t WC2MB(char *out, const wchar_t *in, size_t outLen) const
+ {
+ return m_conv->WC2MB(out, in, outLen);
+ }
+
+ virtual size_t GetMBNulLen() const
+ {
+ // cast needed to call a private function
+ return m_conv->GetMBNulLen();
+ }
+
+ virtual wxMBConv *Clone() const { return new wxConvBrokenFileNames(*this); }
+
+private:
+ // the conversion object we forward to
+ wxMBConv *m_conv;
+
+ DECLARE_NO_ASSIGN_CLASS(wxConvBrokenFileNames)
};
-// not very accurately named because it is not necessarily of type wxMBConvLibc
-// (but the name can't eb changed because of backwards compatibility) default
-// conversion
-WXDLLIMPEXP_DATA_BASE(extern wxMBConv&) wxConvLibc;
+#endif // __UNIX__
// ----------------------------------------------------------------------------
// wxMBConvUTF7 (for conversion using UTF7 encoding)
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;
-};
-WXDLLIMPEXP_DATA_BASE(extern wxMBConvUTF7&) wxConvUTF7;
+ virtual wxMBConv *Clone() const { return new wxMBConvUTF7; }
+};
// ----------------------------------------------------------------------------
// wxMBConvUTF8 (for conversion using UTF8 encoding)
class WXDLLIMPEXP_BASE wxMBConvUTF8 : public wxMBConv
{
public:
+ enum
+ {
+ MAP_INVALID_UTF8_NOT = 0,
+ MAP_INVALID_UTF8_TO_PUA = 1,
+ MAP_INVALID_UTF8_TO_OCTAL = 2
+ };
+
+ 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 wxMBConv *Clone() const { return new wxMBConvUTF8(m_options); }
+
+private:
+ int m_options;
};
-WXDLLIMPEXP_DATA_BASE(extern wxMBConvUTF8&) wxConvUTF8;
+// ----------------------------------------------------------------------------
+// wxMBConvUTF16Base: for both LE and BE variants
+// ----------------------------------------------------------------------------
+
+class WXDLLIMPEXP_BASE wxMBConvUTF16Base : public wxMBConv
+{
+public:
+ 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);
+};
// ----------------------------------------------------------------------------
// wxMBConvUTF16LE (for conversion using UTF16 Little Endian encoding)
// ----------------------------------------------------------------------------
-class WXDLLIMPEXP_BASE wxMBConvUTF16LE : public wxMBConv
+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; }
};
// ----------------------------------------------------------------------------
// wxMBConvUTF16BE (for conversion using UTF16 Big Endian encoding)
// ----------------------------------------------------------------------------
-class WXDLLIMPEXP_BASE wxMBConvUTF16BE : public wxMBConv
+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; }
};
// ----------------------------------------------------------------------------
-// wxMBConvUCS4LE (for conversion using UTF32 Little Endian encoding)
+// wxMBConvUTF32Base: base class for both LE and BE variants
// ----------------------------------------------------------------------------
-class WXDLLIMPEXP_BASE wxMBConvUTF32LE : public wxMBConv
+class WXDLLIMPEXP_BASE wxMBConvUTF32Base : 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;
+ 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);
};
// ----------------------------------------------------------------------------
-// wxMBConvUCS4BE (for conversion using UTF32 Big Endian encoding)
+// wxMBConvUTF32LE (for conversion using UTF32 Little Endian encoding)
// ----------------------------------------------------------------------------
-class WXDLLIMPEXP_BASE wxMBConvUTF32BE : public wxMBConv
+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; }
+};
+
+// ----------------------------------------------------------------------------
+// wxMBConvUTF32BE (for conversion using UTF32 Big Endian encoding)
+// ----------------------------------------------------------------------------
+
+class WXDLLIMPEXP_BASE wxMBConvUTF32BE : public wxMBConvUTF32Base
+{
+public:
+ 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; }
};
// ----------------------------------------------------------------------------
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;
+ virtual wxMBConv *Clone() const { return new wxCSConv(*this); }
- void Clear() ;
+ void Clear();
private:
// common part of all ctors
bool m_deferred;
};
-#define wxConvFile wxConvLocal
-WXDLLIMPEXP_DATA_BASE(extern wxCSConv&) wxConvLocal;
-WXDLLIMPEXP_DATA_BASE(extern wxCSConv&) wxConvISO8859_1;
-WXDLLIMPEXP_DATA_BASE(extern wxMBConv *) wxConvCurrent;
+
+// ----------------------------------------------------------------------------
+// declare predefined conversion objects
+// ----------------------------------------------------------------------------
+
+// conversion to be used with all standard functions affected by locale, e.g.
+// strtol(), strftime(), ...
+extern WXDLLIMPEXP_DATA_BASE(wxMBConv&) 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;
+
+// conversion used for the file names on the systems where they're not Unicode
+// (basically anything except Windows)
+//
+// this is used by all file functions, can be changed by the application
+//
+// by default UTF-8 under Mac OS X and wxConvLibc elsewhere (but it's not used
+// under Windows normally)
+extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvFileName;
+
+// backwards compatible define
+#define wxConvFile (*wxConvFileName)
+
+// the current conversion object, may be set to any conversion, is used by
+// default in a couple of places inside wx (initially same as wxConvLibc)
+extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent;
+
+// ???
+extern WXDLLIMPEXP_DATA_BASE(wxCSConv&) wxConvLocal;
+
// ----------------------------------------------------------------------------
// endianness-dependent conversions
#endif
#if wxMBFILES && wxUSE_UNICODE
- #define wxFNCONV(name) wxConvFile.cWX2MB(name)
+ #define wxFNCONV(name) wxConvFileName->cWX2MB(name)
#define wxFNSTRINGCAST wxMBSTRINGCAST
+#else
+#if defined( __WXOSX__ ) && wxMBFILES
+ #define wxFNCONV(name) wxConvFileName->cWC2MB( wxConvLocal.cWX2WC(name) )
#else
#define wxFNCONV(name) name
+#endif
#define wxFNSTRINGCAST WXSTRINGCAST
#endif
-#else
- // !wxUSE_WCHAR_T
+#else // !wxUSE_WCHAR_T
// ----------------------------------------------------------------------------
// stand-ins in absence of wchar_t
#define wxConvFile wxConvLocal
-WXDLLIMPEXP_DATA_BASE(extern wxMBConv) wxConvLibc,
+extern WXDLLIMPEXP_DATA_BASE(wxMBConv) wxConvLibc,
wxConvLocal,
wxConvISO8859_1,
wxConvUTF8;
-WXDLLIMPEXP_DATA_BASE(extern wxMBConv *) wxConvCurrent;
+extern WXDLLIMPEXP_DATA_BASE(wxMBConv *) wxConvCurrent;
#define wxFNCONV(name) name
#define wxFNSTRINGCAST WXSTRINGCAST
#define wxConvertMB2WX(s) (s)
#endif // Unicode/ANSI
-#endif
- // _WX_WXSTRCONVH__
+#endif // _WX_STRCONV_H_