some != NULL checks

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

diff --git a/src/common/tokenzr.cpp b/src/common/tokenzr.cpp
index 95eb0d586072f8c5a10cf4741cd7abb0271e0304..b6272da149a720dcff508821acfa7b6aa53053ac 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: Gregory Pietsch
+// Modified by: Vadim Zeitlin

 // 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


@@ -22,86 +30,169 @@
 
 #include "wx/tokenzr.h"
 
 
 #include "wx/tokenzr.h"
 

-wxStringTokenizer::wxStringTokenizer(const wxString& to_tokenize,
+// Required for wxIs... functions
+#include <ctype.h>
+
+// ============================================================================
+// implementation
+// ============================================================================
+
+// ----------------------------------------------------------------------------
+// wxStringTokenizer construction
+// ----------------------------------------------------------------------------
+
+wxStringTokenizer::wxStringTokenizer(const wxString& str,

                                      const wxString& delims,
                                      const wxString& delims,

-                                     bool ret_delims)
+                                     wxStringTokenizerMode mode)

 {
 {

-    m_string = to_tokenize;
-    m_delims = delims;
-    m_retdelims = ret_delims;
-    m_pos = 0;
+    SetString(str, delims, mode);

 }
 
 }
 

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

 {
 {

-}
+    if ( mode == wxTOKEN_DEFAULT )
+    {
+        // 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 ( !wxIsspace(*p) )
+                break;
+        }

 
 

-int wxStringTokenizer::CountTokens() const
-{
-    size_t pos = 0;
-    int count = 0;
-    bool at_delim;
-
-    while (pos < m_string.length()) {
-        // while we're still counting ...
-        at_delim = (m_delims.find(m_string.at(pos)) < m_delims.length());
-        // are we at a delimiter? if so, move to the next nondelimiter;
-        // if not, move to the next delimiter.  If the find_first_of
-        // and find_first_not_of methods fail, pos will be assigned
-        // npos (0xFFFFFFFF) which will terminate the loop on the next
-        // go-round unless we have a really long string, which is unlikely
-        pos = at_delim ? m_string.find_first_not_of(m_delims, pos)
-                       : m_string.find_first_of(m_delims, pos);
-        if (m_retdelims)
+        if ( *p )

         {
         {

-            // if we're retaining delimiters, increment count
-            count++;
+            // not whitespace char in delims
+            mode = wxTOKEN_RET_EMPTY;

         }
         else
         {
         }
         else
         {

-            // if we're not retaining delimiters and at a token, inc count
-            count += (!at_delim);
+            // only whitespaces
+            mode = wxTOKEN_STRTOK;

         }
     }
         }
     }

-    return count;
+
+    m_delims = delims;
+    m_mode = mode;
+
+    Reinit(str);
+}
+
+void wxStringTokenizer::Reinit(const wxString& str)
+{
+    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();

 }
 
 }
 

-bool wxStringTokenizer::HasMoreTokens()
+// ----------------------------------------------------------------------------
+// access to the tokens
+// ----------------------------------------------------------------------------
+
+// do we have more of them?
+bool wxStringTokenizer::HasMoreTokens() const

 {
 {

-    return (m_retdelims
-            ? !m_string.IsEmpty()
-            : m_string.find_first_not_of(m_delims) < m_string.length());
+    wxCHECK_MSG( IsOk(), FALSE, _T("you should call SetString() first") );
+
+    if ( m_string.find_first_not_of(m_delims) == wxString::npos )
+    {
+        // no non empty tokens left, but in wxTOKEN_RET_EMPTY_ALL mode we
+        // still may return TRUE if GetNextToken() wasn't called yet for the
+        // last trailing empty token
+        return m_mode == wxTOKEN_RET_EMPTY_ALL ? 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

 {
 {

-    size_t pos;
-    wxString r_string;
-
-    if ( m_string.IsEmpty() )
-        return m_string;
-        pos = m_string.find_first_not_of(m_delims);
-    if ( m_retdelims ) {
-        // we're retaining delimiters (unusual behavior, IMHO)
-        if (pos == 0)
-            // first char is a non-delimiter
-            pos = m_string.find_first_of(m_delims);
-    } else {
-        // we're not retaining delimiters
-        m_string.erase(0, pos);
-        m_pos += pos;
-        if (m_string.IsEmpty())
-            return m_string;
-        pos = m_string.find_first_of(m_delims);
+    wxCHECK_MSG( IsOk(), 0, _T("you should call SetString() first") );
+
+    // 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;
+
+    size_t count = 0;
+    while ( self->HasMoreTokens() )
+    {
+        count++;
+
+        (void)self->GetNextToken();

     }
     }

-    if (pos <= m_string.length()) {
-        r_string = m_string.substr(0, pos);
-        m_string.erase(0, pos);
-        m_pos += pos;
-    } else {
-        r_string = m_string;
-        m_pos += m_string.length();
-        m_string.Empty();
+
+    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 ( !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 )
+        {
+            // 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
+        {
+            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;
+        }

     }
     }

-    return r_string;
+    while ( !allowEmpty && token.empty() );
+
+    return token;

 }
 }