Add some version checks to help compiling on OSX.

[wxWidgets.git] / src / common / tokenzr.cpp

diff --git a/src/common/tokenzr.cpp b/src/common/tokenzr.cpp
index ea615aaf5781311e978135ccd306aa927cf428de..2822025bb7099b878aca5e2b0885ce876d0d96e5 100644 (file)
--- a/src/common/tokenzr.cpp
+++ b/src/common/tokenzr.cpp
@@ -38,6 +38,42 @@
 // implementation
 // ============================================================================
 
 // implementation
 // ============================================================================
 

+// ----------------------------------------------------------------------------
+// helpers
+// ----------------------------------------------------------------------------
+
+static wxString::const_iterator
+find_first_of(const wxChar *delims, size_t len,
+              const wxString::const_iterator& from,
+              const wxString::const_iterator& end)
+{
+    wxASSERT_MSG( from <= end,  wxT("invalid index") );
+
+    for ( wxString::const_iterator i = from; i != end; ++i )
+    {
+        if ( wxTmemchr(delims, *i, len) )
+            return i;
+    }
+
+    return end;
+}
+
+static wxString::const_iterator
+find_first_not_of(const wxChar *delims, size_t len,
+                  const wxString::const_iterator& from,
+                  const wxString::const_iterator& end)
+{
+    wxASSERT_MSG( from <= end,  wxT("invalid index") );
+
+    for ( wxString::const_iterator i = from; i != end; ++i )
+    {
+        if ( !wxTmemchr(delims, *i, len) )
+            return i;
+    }
+
+    return end;
+}
+

 // ----------------------------------------------------------------------------
 // wxStringTokenizer construction
 // ----------------------------------------------------------------------------
 // ----------------------------------------------------------------------------
 // wxStringTokenizer construction
 // ----------------------------------------------------------------------------


@@ -78,7 +114,13 @@ void wxStringTokenizer::SetString(const wxString& str,
         }
     }
 
         }
     }
 

-    m_delims = delims;
+#if wxUSE_UNICODE // FIXME-UTF8: only wc_str()
+    m_delims = delims.wc_str();
+#else
+    m_delims = delims.mb_str();
+#endif
+    m_delimsLen = delims.length();
+

     m_mode = mode;
 
     Reinit(str);
     m_mode = mode;
 
     Reinit(str);


@@ -86,11 +128,13 @@ void wxStringTokenizer::SetString(const wxString& str,
 
 void wxStringTokenizer::Reinit(const wxString& str)
 {
 
 void wxStringTokenizer::Reinit(const wxString& str)
 {

-    wxASSERT_MSG( IsOk(), _T("you should call SetString() first") );
+    wxASSERT_MSG( IsOk(), wxT("you should call SetString() first") );

 
     m_string = str;
 
     m_string = str;

-    m_pos = 0;
-    m_lastDelim = _T('\0');
+    m_stringEnd = m_string.end();
+    m_pos = m_string.begin();
+    m_lastDelim = wxT('\0');
+    m_hasMoreTokens = MoreTokens_Unknown;

 }
 
 // ----------------------------------------------------------------------------
 }
 
 // ----------------------------------------------------------------------------


@@ -100,9 +144,28 @@ void wxStringTokenizer::Reinit(const wxString& str)
 // do we have more of them?
 bool wxStringTokenizer::HasMoreTokens() const
 {
 // do we have more of them?
 bool wxStringTokenizer::HasMoreTokens() const
 {

-    wxCHECK_MSG( IsOk(), false, _T("you should call SetString() first") );
+    // GetNextToken() calls HasMoreTokens() and so HasMoreTokens() is called
+    // twice in every interation in the following common usage patten:
+    //     while ( HasMoreTokens() )
+    //        GetNextToken();
+    // We optimize this case by caching HasMoreTokens() return value here:
+    if ( m_hasMoreTokens == MoreTokens_Unknown )
+    {
+        bool r = DoHasMoreTokens();
+        wxConstCast(this, wxStringTokenizer)->m_hasMoreTokens =
+            r ? MoreTokens_Yes : MoreTokens_No;
+        return r;
+    }
+    else
+        return m_hasMoreTokens == MoreTokens_Yes;
+}

 
 

-    if ( m_string.find_first_not_of(m_delims, m_pos) != wxString::npos )
+bool wxStringTokenizer::DoHasMoreTokens() const
+{
+    wxCHECK_MSG( IsOk(), false, wxT("you should call SetString() first") );
+
+    if ( find_first_not_of(m_delims, m_delimsLen, m_pos, m_stringEnd)
+         != m_stringEnd )

     {
         // there are non delimiter characters left, so we do have more tokens
         return true;
     {
         // there are non delimiter characters left, so we do have more tokens
         return true;


@@ -114,7 +177,7 @@ bool wxStringTokenizer::HasMoreTokens() const
         case wxTOKEN_RET_DELIMS:
             // special hack for wxTOKEN_RET_EMPTY: we should return the initial
             // empty token even if there are only delimiters after it
         case wxTOKEN_RET_DELIMS:
             // special hack for wxTOKEN_RET_EMPTY: we should return the initial
             // empty token even if there are only delimiters after it

-            return m_pos == 0 && !m_string.empty();
+            return !m_string.empty() && m_pos == m_string.begin();

 
         case wxTOKEN_RET_EMPTY_ALL:
             // special hack for wxTOKEN_RET_EMPTY_ALL: we can know if we had
 
         case wxTOKEN_RET_EMPTY_ALL:
             // special hack for wxTOKEN_RET_EMPTY_ALL: we can know if we had


@@ -123,11 +186,11 @@ bool wxStringTokenizer::HasMoreTokens() const
             // up to the end of the string in GetNextToken(), but if it is not
             // NUL yet we still have this last token to return even if m_pos is
             // already at m_string.length()
             // up to the end of the string in GetNextToken(), but if it is not
             // NUL yet we still have this last token to return even if m_pos is
             // already at m_string.length()

-            return m_pos < m_string.length() || m_lastDelim != _T('\0');
+            return m_pos < m_stringEnd || m_lastDelim != wxT('\0');

 
         case wxTOKEN_INVALID:
         case wxTOKEN_DEFAULT:
 
         case wxTOKEN_INVALID:
         case wxTOKEN_DEFAULT:

-            wxFAIL_MSG( _T("unexpected tokenizer mode") );
+            wxFAIL_MSG( wxT("unexpected tokenizer mode") );

             // fall through
 
         case wxTOKEN_STRTOK:
             // fall through
 
         case wxTOKEN_STRTOK:


@@ -141,13 +204,13 @@ bool wxStringTokenizer::HasMoreTokens() const
 // count the number of (remaining) tokens in the string
 size_t wxStringTokenizer::CountTokens() const
 {
 // count the number of (remaining) tokens in the string
 size_t wxStringTokenizer::CountTokens() const
 {

-    wxCHECK_MSG( IsOk(), 0, _T("you should call SetString() first") );
+    wxCHECK_MSG( IsOk(), 0, wxT("you should call SetString() first") );

 
     // VZ: this function is IMHO not very useful, so it's probably not very
     //     important if its implementation here is not as efficient as it
     //     could be -- but OTOH like this we're sure to get the correct answer
     //     in all modes
 
     // VZ: this function is IMHO not very useful, so it's probably not very
     //     important if its implementation here is not as efficient as it
     //     could be -- but OTOH like this we're sure to get the correct answer
     //     in all modes

-    wxStringTokenizer tkz(m_string.c_str() + m_pos, m_delims, m_mode);
+    wxStringTokenizer tkz(wxString(m_pos, m_stringEnd), m_delims, m_mode);

 
     size_t count = 0;
     while ( tkz.HasMoreTokens() )
 
     size_t count = 0;
     while ( tkz.HasMoreTokens() )


@@ -174,36 +237,39 @@ wxString wxStringTokenizer::GetNextToken()
             break;
         }
 
             break;
         }
 

+        m_hasMoreTokens = MoreTokens_Unknown;
+

         // find the end of this token
         // find the end of this token

-        size_t pos = m_string.find_first_of(m_delims, m_pos);
+        wxString::const_iterator pos =
+            find_first_of(m_delims, m_delimsLen, m_pos, m_stringEnd);

 
         // and the start of the next one
 
         // and the start of the next one

-        if ( pos == wxString::npos )
+        if ( pos == m_stringEnd )

         {
             // no more delimiters, the token is everything till the end of
             // string
         {
             // no more delimiters, the token is everything till the end of
             // string

-            token.assign(m_string, m_pos, wxString::npos);
+            token.assign(m_pos, m_stringEnd);

 
             // skip the token
 
             // skip the token

-            m_pos = m_string.length();
+            m_pos = m_stringEnd;

 
             // it wasn't terminated
 
             // it wasn't terminated

-            m_lastDelim = _T('\0');
+            m_lastDelim = wxT('\0');

         }
         else // we found a delimiter at pos
         {
             // in wxTOKEN_RET_DELIMS mode we return the delimiter character
             // with token, otherwise leave it out
         }
         else // we found a delimiter at pos
         {
             // in wxTOKEN_RET_DELIMS mode we return the delimiter character
             // with token, otherwise leave it out

-            size_t len = pos - m_pos;
+            wxString::const_iterator tokenEnd(pos);

             if ( m_mode == wxTOKEN_RET_DELIMS )
             if ( m_mode == wxTOKEN_RET_DELIMS )

-                len++;
+                ++tokenEnd;

 
 

-            token.assign(m_string, m_pos, len);
+            token.assign(m_pos, tokenEnd);

 
             // skip the token and the trailing delimiter
             m_pos = pos + 1;
 
 
             // skip the token and the trailing delimiter
             m_pos = pos + 1;
 

-            m_lastDelim = m_string[pos];
+            m_lastDelim = (pos == m_stringEnd) ? wxT('\0') : (wxChar)*pos;

         }
     }
     while ( !AllowEmpty() && token.empty() );
         }
     }
     while ( !AllowEmpty() && token.empty() );