// Name: tokenzr.cpp
// Purpose: String tokenizer
// Author: Guilhem Lavaux
-// Modified by: Vadim Zeitlin
+// Modified by: Vadim Zeitlin (almost full rewrite)
// Created: 04/22/98
// RCS-ID: $Id$
// Copyright: (c) Guilhem Lavaux
// headers
// ----------------------------------------------------------------------------
-#ifdef __GNUG__
- #pragma implementation "tokenzr.h"
-#endif
-
// For compilers that support precompilation, includes "wx.h".
#include "wx/wxprec.h"
#endif
#include "wx/tokenzr.h"
+#include "wx/arrstr.h"
// Required for wxIs... functions
#include <ctype.h>
m_string = str;
m_pos = 0;
-
- // empty string doesn't have any tokens
- m_hasMore = !m_string.empty();
+ m_lastDelim = _T('\0');
}
// ----------------------------------------------------------------------------
// do we have more of them?
bool wxStringTokenizer::HasMoreTokens() const
{
- wxCHECK_MSG( IsOk(), FALSE, _T("you should call SetString() first") );
+ wxCHECK_MSG( IsOk(), false, _T("you should call SetString() first") );
- if ( m_string.find_first_not_of(m_delims) == wxString::npos )
+ if ( m_string.find_first_not_of(m_delims, m_pos) != 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;
+ // there are non delimiter characters left, so we do have more tokens
+ return true;
}
- else
+
+ switch ( m_mode )
{
- // there are non delimiter characters left, hence we do have more
- // tokens
- return TRUE;
+ case wxTOKEN_RET_EMPTY:
+ 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();
+
+ case wxTOKEN_RET_EMPTY_ALL:
+ // special hack for wxTOKEN_RET_EMPTY_ALL: we can know if we had
+ // already returned the trailing empty token after the last
+ // delimiter by examining m_lastDelim: it is set to NUL if we run
+ // 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');
+
+ case wxTOKEN_INVALID:
+ case wxTOKEN_DEFAULT:
+ wxFAIL_MSG( _T("unexpected tokenizer mode") );
+ // fall through
+
+ case wxTOKEN_STRTOK:
+ // never return empty delimiters
+ break;
}
+
+ return false;
}
-// count the number of tokens in the string
+// count the number of (remaining) tokens in the string
size_t wxStringTokenizer::CountTokens() const
{
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
+ // 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 *self = (wxStringTokenizer *)this; // const_cast
- wxString stringInitial = m_string;
+ wxStringTokenizer tkz(m_string.c_str() + m_pos, m_delims, m_mode);
size_t count = 0;
- while ( self->HasMoreTokens() )
+ while ( tkz.HasMoreTokens() )
{
count++;
- (void)self->GetNextToken();
+ (void)tkz.GetNextToken();
}
- self->Reinit(stringInitial);
-
return count;
}
wxString wxStringTokenizer::GetNextToken()
{
- // strtok() doesn't return empty tokens, all other modes do
- bool allowEmpty = m_mode != wxTOKEN_STRTOK;
-
wxString token;
do
{
{
break;
}
+
// find the end of this token
- size_t pos = m_string.find_first_of(m_delims);
+ size_t pos = m_string.find_first_of(m_delims, m_pos);
// 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;
+ token.assign(m_string, m_pos, wxString::npos);
- m_pos += m_string.length();
- m_string.clear();
+ // skip the token
+ m_pos = m_string.length();
- // 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;
+ // it wasn't terminated
+ m_lastDelim = _T('\0');
}
- else
+ else // we found a delimiter at 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);
+ // with token, otherwise leave it out
+ size_t len = pos - m_pos;
+ if ( m_mode == wxTOKEN_RET_DELIMS )
+ len++;
- // remove token with the following it delimiter from string
- m_string.erase(0, pos2);
+ token.assign(m_string, m_pos, len);
- // keep track of the position in the original string too
- m_pos += pos2;
+ // skip the token and the trailing delimiter
+ m_pos = pos + 1;
+
+ m_lastDelim = m_string[pos];
}
}
- while ( !allowEmpty && token.empty() );
+ 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() )
+ {
+ tokens.Add(tk.GetNextToken());
+ }
+
+ return tokens;
+}
projects / wxWidgets.git / blobdiff