src/common/tokenzr.cpp

   1 /////////////////////////////////////////////////////////////////////////////
   2 // Name:        tokenzr.cpp
   3 // Purpose:     String tokenizer
   4 // Author:      Guilhem Lavaux
   5 // Modified by: Vadim Zeitlin
   6 // Created:     04/22/98
   7 // RCS-ID:      $Id$
   8 // Copyright:   (c) Guilhem Lavaux
   9 // Licence:     wxWindows licence
  10 /////////////////////////////////////////////////////////////////////////////
  11
  12 // ============================================================================
  13 // declarations
  14 // ============================================================================
  15
  16 // ----------------------------------------------------------------------------
  17 // headers
  18 // ----------------------------------------------------------------------------
  19
  20 #ifdef __GNUG__
  21     #pragma implementation "tokenzr.h"
  22 #endif
  23
  24 // For compilers that support precompilation, includes "wx.h".
  25 #include "wx/wxprec.h"
  26
  27 #ifdef __BORLANDC__
  28     #pragma hdrstop
  29 #endif
  30
  31 #include "wx/tokenzr.h"
  32
  33 // ============================================================================
  34 // implementation
  35 // ============================================================================
  36
  37 // ----------------------------------------------------------------------------
  38 // wxStringTokenizer construction
  39 // ----------------------------------------------------------------------------
  40
  41 wxStringTokenizer::wxStringTokenizer(const wxString& str,
  42                                      const wxString& delims,
  43                                      wxStringTokenizerMode mode)
  44 {
  45     SetString(str, delims, mode);
  46 }
  47
  48 void wxStringTokenizer::SetString(const wxString& str,
  49                                   const wxString& delims,
  50                                   wxStringTokenizerMode mode)
  51 {
  52     if ( mode == wxTOKEN_DEFAULT )
  53     {
  54         // by default, we behave like strtok() if the delimiters are only
  55         // whitespace characters and as wxTOKEN_RET_EMPTY otherwise (for
  56         // whitespace delimiters, strtok() behaviour is better because we want
  57         // to count consecutive spaces as one delimiter)
  58         const wxChar *p;
  59         for ( p = delims.c_str(); *p; p++ )
  60         {
  61             if ( !wxIsspace(*p) )
  62                 break;
  63         }
  64
  65         if ( *p )
  66         {
  67             // not whitespace char in delims
  68             mode = wxTOKEN_RET_EMPTY;
  69         }
  70         else
  71         {
  72             // only whitespaces
  73             mode = wxTOKEN_STRTOK;
  74         }
  75     }
  76
  77     m_delims = delims;
  78     m_mode = mode;
  79
  80     Reinit(str);
  81 }
  82
  83 void wxStringTokenizer::Reinit(const wxString& str)
  84 {
  85     wxASSERT_MSG( IsOk(), _T("you should call SetString() first") );
  86
  87     m_string = str;
  88     m_pos = 0;
  89
  90     // empty string doesn't have any tokens
  91     m_hasMore = !m_string.empty();
  92 }
  93
  94 // ----------------------------------------------------------------------------
  95 // access to the tokens
  96 // ----------------------------------------------------------------------------
  97
  98 // do we have more of them?
  99 bool wxStringTokenizer::HasMoreTokens() const
 100 {
 101     wxCHECK_MSG( IsOk(), FALSE, _T("you should call SetString() first") );
 102
 103     if ( m_string.find_first_not_of(m_delims) == wxString::npos )
 104     {
 105         // no non empty tokens left, but in wxTOKEN_RET_EMPTY_ALL mode we
 106         // still may return TRUE if GetNextToken() wasn't called yet for the
 107         // last trailing empty token
 108         return m_mode == wxTOKEN_RET_EMPTY_ALL ? m_hasMore : FALSE;
 109     }
 110     else
 111     {
 112         // there are non delimiter characters left, hence we do have more
 113         // tokens
 114         return TRUE;
 115     }
 116 }
 117
 118 // count the number of tokens in the string
 119 size_t wxStringTokenizer::CountTokens() const
 120 {
 121     wxCHECK_MSG( IsOk(), 0, _T("you should call SetString() first") );
 122
 123     // VZ: this function is IMHO not very useful, so it's probably not very
 124     //     important if it's implementation here is not as efficient as it
 125     //     could be - but OTOH like this we're sure to get the correct answer
 126     //     in all modes
 127     wxStringTokenizer *self = (wxStringTokenizer *)this;    // const_cast
 128     wxString stringInitial = m_string;
 129
 130     size_t count = 0;
 131     while ( self->HasMoreTokens() )
 132     {
 133         count++;
 134
 135         (void)self->GetNextToken();
 136     }
 137
 138     self->Reinit(stringInitial);
 139
 140     return count;
 141 }
 142
 143 // ----------------------------------------------------------------------------
 144 // token extraction
 145 // ----------------------------------------------------------------------------
 146
 147 wxString wxStringTokenizer::GetNextToken()
 148 {
 149     // strtok() doesn't return empty tokens, all other modes do
 150     bool allowEmpty = m_mode != wxTOKEN_STRTOK;
 151
 152     wxString token;
 153     do
 154     {
 155         if ( !HasMoreTokens() )
 156         {
 157             break;
 158         }
 159         // find the end of this token
 160         size_t pos = m_string.find_first_of(m_delims);
 161
 162         // and the start of the next one
 163         if ( pos == wxString::npos )
 164         {
 165             // no more delimiters, the token is everything till the end of
 166             // string
 167             token = m_string;
 168
 169             m_pos += m_string.length();
 170             m_string.clear();
 171
 172             // no more tokens in this string, even in wxTOKEN_RET_EMPTY_ALL
 173             // mode (we will return the trailing one right now in this case)
 174             m_hasMore = FALSE;
 175         }
 176         else
 177         {
 178             size_t pos2 = pos + 1;
 179
 180             // in wxTOKEN_RET_DELIMS mode we return the delimiter character
 181             // with token
 182             token = wxString(m_string, m_mode == wxTOKEN_RET_DELIMS ? pos2
 183                                                                     : pos);
 184
 185             // remove token with the following it delimiter from string
 186             m_string.erase(0, pos2);
 187
 188             // keep track of the position in the original string too
 189             m_pos += pos2;
 190         }
 191     }
 192     while ( !allowEmpty && token.empty() );
 193
 194     return token;
 195 }