more compilation fixes for VC6 (#9492)

[wxWidgets.git] / include / wx / strconv.h

diff --git a/include/wx/strconv.h b/include/wx/strconv.h
index 2cd26bd30d1dbe6848d7eb841afa326812fd5819..1aa5f97550b5971d46cb004aaaf4a5bd8cd95437 100644 (file)
--- a/include/wx/strconv.h
+++ b/include/wx/strconv.h
@@ -25,7 +25,9 @@
 #  undef __BSEXCPT__
 #endif
 
+#ifndef __WXPALMOS5__
 #include <stdlib.h>
+#endif // ! __WXPALMOS5__
 
 #if wxUSE_WCHAR_T
 
@@ -247,21 +249,115 @@ 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:
-    // FIXME-UTF8: split this class into multiple classes, one strict and
-    //             other lossy (PUA, OCTAL mappings)
     enum
     {
         MAP_INVALID_UTF8_NOT = 0,
@@ -270,8 +366,11 @@ 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); }
 
@@ -470,7 +569,7 @@ WX_DECLARE_GLOBAL_CONV(wxMBConv, wxConvLibc)
 WX_DECLARE_GLOBAL_CONV(wxCSConv, wxConvISO8859_1)
 #define wxConvISO8859_1 wxGet_wxConvISO8859_1()
 
-WX_DECLARE_GLOBAL_CONV(wxMBConvUTF8, wxConvUTF8)
+WX_DECLARE_GLOBAL_CONV(wxMBConvStrictUTF8, wxConvUTF8)
 #define wxConvUTF8 wxGet_wxConvUTF8()
 
 WX_DECLARE_GLOBAL_CONV(wxMBConvUTF7, wxConvUTF7)