Fix for Unicode conversion.

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

diff --git a/src/common/tokenzr.cpp b/src/common/tokenzr.cpp
index c4f4bfb899ba68b87ed7f5afea0a2ebd082c1e0b..32c220a69afd31af0b304f612c10223e947d30c2 100644 (file)
--- a/src/common/tokenzr.cpp
+++ b/src/common/tokenzr.cpp
@@ -2,13 +2,21 @@
 // Name:        tokenzr.cpp
 // Purpose:     String tokenizer
 // Author:      Guilhem Lavaux
 // Name:        tokenzr.cpp
 // Purpose:     String tokenizer
 // Author:      Guilhem Lavaux

-// Modified by:
+// Modified by: Vadim Zeitlin (almost full rewrite)

 // Created:     04/22/98
 // RCS-ID:      $Id$
 // Copyright:   (c) Guilhem Lavaux
 // Licence:     wxWindows licence
 /////////////////////////////////////////////////////////////////////////////
 
 // Created:     04/22/98
 // RCS-ID:      $Id$
 // Copyright:   (c) Guilhem Lavaux
 // Licence:     wxWindows licence
 /////////////////////////////////////////////////////////////////////////////
 

+// ============================================================================
+// declarations
+// ============================================================================
+
+// ----------------------------------------------------------------------------
+// headers
+// ----------------------------------------------------------------------------
+

 #ifdef __GNUG__
     #pragma implementation "tokenzr.h"
 #endif
 #ifdef __GNUG__
     #pragma implementation "tokenzr.h"
 #endif


@@ -22,117 +30,193 @@
 
 #include "wx/tokenzr.h"
 
 
 #include "wx/tokenzr.h"
 

-wxStringTokenizer::wxStringTokenizer(const wxString& to_tokenize,
-                                     const wxString& delims,
-                                     bool ret_delims)
-{
-    m_string = to_tokenize;
-    m_delims = delims;
-    m_retdelims = ret_delims;
-}
+// Required for wxIs... functions
+#include <ctype.h>
+
+// ============================================================================
+// implementation
+// ============================================================================

 
 

-wxStringTokenizer::~wxStringTokenizer()
+// ----------------------------------------------------------------------------
+// wxStringTokenizer construction
+// ----------------------------------------------------------------------------
+
+wxStringTokenizer::wxStringTokenizer(const wxString& str,
+                                     const wxString& delims,
+                                     wxStringTokenizerMode mode)

 {
 {

+    SetString(str, delims, mode);

 }
 
 }
 

-off_t wxStringTokenizer::FindDelims(const wxString& str, const wxString& delims) const
+void wxStringTokenizer::SetString(const wxString& str,
+                                  const wxString& delims,
+                                  wxStringTokenizerMode mode)

 {
 {

-    for ( int i = 0; i < str.Length(); i++ )
+    if ( mode == wxTOKEN_DEFAULT )

     {
     {

-        char c = str[i];
-
-        for ( int j = 0; j < delims.Length() ; j++ )
+        // by default, we behave like strtok() if the delimiters are only
+        // whitespace characters and as wxTOKEN_RET_EMPTY otherwise (for
+        // whitespace delimiters, strtok() behaviour is better because we want
+        // to count consecutive spaces as one delimiter)
+        const wxChar *p;
+        for ( p = delims.c_str(); *p; p++ )

         {
         {

-            if ( delims[j] == c )
-                return i;
+            if ( !wxIsspace(*p) )
+                break;

         }
         }

-    }
-
-    return -1;
-}

 
 

-int wxStringTokenizer::CountTokens() const
-{
-    wxString p_string = m_string;
-    bool found = TRUE;
-    int pos, count = 1;
-
-    if (p_string.Length() == 0)
-        return 0;
-
-    while (found)
-    {
-        pos = FindDelims(p_string, m_delims);
-        if (pos != -1)
+        if ( *p )

         {
         {

-            count++;
-            p_string = p_string(0, pos);
+            // not whitespace char in delims
+            mode = wxTOKEN_RET_EMPTY;

         }
         else
         {
         }
         else
         {

-            found = FALSE;
+            // only whitespaces
+            mode = wxTOKEN_STRTOK;

         }
     }
 
         }
     }
 

-    return count;
+    m_delims = delims;
+    m_mode = mode;
+
+    Reinit(str);

 }
 
 }
 

-bool wxStringTokenizer::HasMoreTokens()
+void wxStringTokenizer::Reinit(const wxString& str)

 {
 {

-    return !m_string.IsEmpty();
+    wxASSERT_MSG( IsOk(), _T("you should call SetString() first") );
+
+    m_string = str;
+    m_pos = 0;
+
+    // empty string doesn't have any tokens
+    m_hasMore = !m_string.empty();

 }
 
 }
 

-// needed to fix leading whitespace / mult. delims bugs
-void wxStringTokenizer::EatLeadingDelims()
+// ----------------------------------------------------------------------------
+// access to the tokens
+// ----------------------------------------------------------------------------
+
+// do we have more of them?
+bool wxStringTokenizer::HasMoreTokens() const

 {
 {

-    int pos;
+    wxCHECK_MSG( IsOk(), FALSE, _T("you should call SetString() first") );

 
 

-    // while leading delims trim 'em from the left
-    while ( ( pos = FindDelims(m_string, m_delims)) == 0 )
+    if ( m_string.find_first_not_of(m_delims) == wxString::npos )

     {
     {

-        m_string = m_string.Mid((size_t)1);
+        // no non empty tokens left, but in 2 cases we still may return TRUE if
+        // GetNextToken() wasn't called yet for this empty token:
+        //
+        //   a) in wxTOKEN_RET_EMPTY_ALL mode we always do it
+        //   b) in wxTOKEN_RET_EMPTY mode we do it in the special case of a
+        //      string containing only the delimiter: then there is an empty
+        //      token just before it
+        return (m_mode == wxTOKEN_RET_EMPTY_ALL) ||
+               (m_mode == wxTOKEN_RET_EMPTY && m_pos == 0)
+                    ? m_hasMore : FALSE;
+    }
+    else
+    {
+        // there are non delimiter characters left, hence we do have more
+        // tokens
+        return TRUE;

     }
 }
 
     }
 }
 

-wxString wxStringTokenizer::NextToken()
+// count the number of tokens in the string
+size_t wxStringTokenizer::CountTokens() const

 {
 {

-    off_t pos, pos2;
-    wxString r_string;
-
-    if ( m_string.IsEmpty() )
-        return m_string;
+    wxCHECK_MSG( IsOk(), 0, _T("you should call SetString() first") );

 
 

-    if ( !m_retdelims )
-        EatLeadingDelims();
+    // VZ: this function is IMHO not very useful, so it's probably not very
+    //     important if it's 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 *self = (wxStringTokenizer *)this;    // const_cast
+    wxString stringInitial = m_string;

 
 

-    pos = FindDelims(m_string, m_delims);
-    if (pos == -1)
+    size_t count = 0;
+    while ( self->HasMoreTokens() )

     {
     {

-        r_string = m_string;
-        m_string = wxEmptyString;
+        count++;

 
 

-        return r_string;
+        (void)self->GetNextToken();

     }
 
     }
 

-    if (m_retdelims)
+    self->Reinit(stringInitial);
+
+    return count;
+}
+
+// ----------------------------------------------------------------------------
+// token extraction
+// ----------------------------------------------------------------------------
+
+wxString wxStringTokenizer::GetNextToken()
+{
+    // strtok() doesn't return empty tokens, all other modes do
+    bool allowEmpty = m_mode != wxTOKEN_STRTOK;
+
+    wxString token;
+    do

     {
     {

-        if (!pos)
+        if ( !HasMoreTokens() )
+        {
+            break;
+        }
+        // find the end of this token
+        size_t pos = m_string.find_first_of(m_delims);
+
+        // and the start of the next one
+        if ( pos == wxString::npos )

         {
         {

-            pos++;
-            pos2 = 1;
+            // no more delimiters, the token is everything till the end of
+            // string
+            token = m_string;
+
+            m_pos += m_string.length();
+            m_string.clear();
+
+            // no more tokens in this string, even in wxTOKEN_RET_EMPTY_ALL
+            // mode (we will return the trailing one right now in this case)
+            m_hasMore = FALSE;

         }
         else
         {
         }
         else
         {

-            pos2 = pos;
+            size_t pos2 = pos + 1;
+
+            // in wxTOKEN_RET_DELIMS mode we return the delimiter character
+            // with token
+            token = wxString(m_string, m_mode == wxTOKEN_RET_DELIMS ? pos2
+                                                                    : pos);
+
+            // remove token with the following it delimiter from string
+            m_string.erase(0, pos2);
+
+            // keep track of the position in the original string too
+            m_pos += pos2;

         }
     }
         }
     }

-    else
+    while ( !allowEmpty && token.empty() );
+
+    return token;
+}
+
+// ----------------------------------------------------------------------------
+// public functions
+// ----------------------------------------------------------------------------
+
+wxArrayString wxStringTokenize(const wxString& str,
+                               const wxString& delims,
+                               wxStringTokenizerMode mode)
+{
+    wxArrayString tokens;
+    wxStringTokenizer tk(str, delims, mode);
+    while ( tk.HasMoreTokens() )

     {
     {

-        pos2 = pos + 1;
+        tokens.Add(tk.GetNextToken());

     }
 
     }
 

-    r_string = m_string.Left((size_t)pos);
-    m_string = m_string.Mid((size_t)pos2);
-
-    return r_string;
+    return tokens;

 }
 }