1. changed all "wxMBConv& conv" parameters to "const wxMBConv&"

2. this allows to use wxConvAuto() instead of wxConvUTF8 as default value
   for this parameter in the classes which read text from the file: wxConvAuto
   automatically recognizes the BOM at the start of file and uses the correct
   conversion
3. don't use Windows for UTF-7 conversions as there is no way to make it
   fail on invalid UTF-7 strings; use our own wxMBConvUtf7 instead


git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@38570 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775

include/wx/convauto.h

@@ -0,0 +1,99 @@
+///////////////////////////////////////////////////////////////////////////////
+// Name:        wx/convauto.h
+// Purpose:     wxConvAuto class declaration
+// Author:      Vadim Zeitlin
+// Created:     2006-04-03
+// RCS-ID:      $Id$
+// Copyright:   (c) 2006 Vadim Zeitlin
+// Licence:     wxWindows licence
+///////////////////////////////////////////////////////////////////////////////
+
+#ifndef _WX_CONVAUTO_H_
+#define _WX_CONVAUTO_H_
+
+#include "wx/strconv.h"
+
+#if wxUSE_WCHAR_T
+
+// ----------------------------------------------------------------------------
+// wxConvAuto: uses BOM to automatically detect input encoding
+// ----------------------------------------------------------------------------
+
+class WXDLLIMPEXP_BASE wxConvAuto : public wxMBConv
+{
+public:
+    // default ctor, the real conversion will be created on demand
+    wxConvAuto() { m_conv = NULL; /* the rest will be initialized later */ }
+
+    // copy ctor doesn't initialize anything neither as conversion can only be
+    // deduced on first use
+    wxConvAuto(const wxConvAuto& WXUNUSED(other)) { m_conv = NULL; }
+
+    virtual ~wxConvAuto() { if ( m_conv && m_ownsConv ) delete m_conv; }
+
+    // override the base class virtual function(s) to use our m_conv
+    virtual size_t ToWChar(wchar_t *dst, size_t dstLen,
+                           const char *src, size_t srcLen = -1) const;
+
+    virtual size_t FromWChar(char *dst, size_t dstLen,
+                             const wchar_t *src, size_t srcLen = -1) const;
+
+    virtual size_t GetMBNulLen() const { return m_conv->GetMBNulLen(); }
+
+private:
+    // all currently recognized BOM values
+    enum BOMType
+    {
+        BOM_None,
+        BOM_UTF32BE,
+        BOM_UTF32LE,
+        BOM_UTF16BE,
+        BOM_UTF16LE,
+        BOM_UTF8
+    };
+
+    // return the BOM type of this buffer
+    static BOMType DetectBOM(const char *src, size_t srcLen);
+
+    // initialize m_conv with the conversion to use by default (UTF-8)
+    void InitWithDefault()
+    {
+        m_conv = &wxConvUTF8;
+        m_ownsConv = false;
+    }
+
+    // create the correct conversion object for the given BOM type
+    void InitFromBOM(BOMType bomType);
+
+    // create the correct conversion object for the BOM present in the
+    // beginning of the buffer; adjust the buffer to skip the BOM if found
+    void InitFromInput(const char **src, size_t *len);
+
+    // adjust src and len to skip over the BOM (identified by m_bomType) at the
+    // start of the buffer
+    void SkipBOM(const char **src, size_t *len) const;
+
+
+    // conversion object which we really use, NULL until the first call to
+    // either ToWChar() or FromWChar()
+    wxMBConv *m_conv;
+
+    // our BOM type
+    BOMType m_bomType;
+
+    // true if we allocated m_conv ourselves, false if we just use an existing
+    // global conversion
+    bool m_ownsConv;
+
+    // true if we already skipped BOM when converting (and not just calculating
+    // the size)
+    bool m_consumedBOM;
+
+
+    DECLARE_NO_ASSIGN_CLASS(wxConvAuto);
+};
+
+#endif // wxUSE_WCHAR_T
+
+#endif // _WX_CONVAUTO_H_
+