/////////////////////////////////////////////////////////////////////////////
-// Name: tokenzr.cpp
+// Name: src/common/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
/////////////////////////////////////////////////////////////////////////////
-#ifdef __GNUG__
- #pragma implementation "tokenzr.h"
-#endif
+// ============================================================================
+// declarations
+// ============================================================================
+
+// ----------------------------------------------------------------------------
+// headers
+// ----------------------------------------------------------------------------
// For compilers that support precompilation, includes "wx.h".
#include "wx/wxprec.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;
-}
+#ifndef WX_PRECOMP
+ #include "wx/arrstr.h"
+ #include "wx/crt.h"
+#endif
+
+// Required for wxIs... functions
+#include <ctype.h>
+
+// ============================================================================
+// implementation
+// ============================================================================
+
+// ----------------------------------------------------------------------------
+// helpers
+// ----------------------------------------------------------------------------
-wxStringTokenizer::~wxStringTokenizer()
+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, _T("invalid index") );
+
+ for ( wxString::const_iterator i = from; i != end; ++i )
+ {
+ if ( wxTmemchr(delims, *i, len) )
+ return i;
+ }
+
+ return end;
}
-off_t wxStringTokenizer::FindDelims(const wxString& str, const wxString& delims) const
+static wxString::const_iterator
+find_first_not_of(const wxChar *delims, size_t len,
+ const wxString::const_iterator& from,
+ const wxString::const_iterator& end)
{
- for ( size_t i = 0; i < str.Length(); i++ )
- {
- wxChar c = str[i];
+ wxASSERT_MSG( from <= end, _T("invalid index") );
- for ( size_t j = 0; j < delims.Length() ; j++ )
- {
- if ( delims[j] == c )
- return i;
- }
+ for ( wxString::const_iterator i = from; i != end; ++i )
+ {
+ if ( !wxTmemchr(delims, *i, len) )
+ return i;
}
- return -1;
+ return end;
}
-int wxStringTokenizer::CountTokens() const
-{
- wxString p_string = m_string;
- bool found = TRUE;
- int pos, count = 1;
+// ----------------------------------------------------------------------------
+// wxStringTokenizer construction
+// ----------------------------------------------------------------------------
- if (p_string.Length() == 0)
- return 0;
+wxStringTokenizer::wxStringTokenizer(const wxString& str,
+ const wxString& delims,
+ wxStringTokenizerMode mode)
+{
+ SetString(str, delims, mode);
+}
- while (found)
+void wxStringTokenizer::SetString(const wxString& str,
+ const wxString& delims,
+ wxStringTokenizerMode mode)
+{
+ if ( mode == wxTOKEN_DEFAULT )
{
- pos = FindDelims(p_string, m_delims);
- if (pos != -1)
+ // 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)
+ wxString::const_iterator p;
+ for ( p = delims.begin(); p != delims.end(); ++p )
{
- count++;
- p_string = p_string(pos+1, p_string.Length());
+ if ( !wxIsspace(*p) )
+ break;
+ }
+
+ if ( p != delims.end() )
+ {
+ // not whitespace char in delims
+ mode = wxTOKEN_RET_EMPTY;
}
else
{
- found = FALSE;
+ // only whitespaces
+ mode = wxTOKEN_STRTOK;
}
}
- return count;
+#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);
}
-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_stringEnd = m_string.end();
+ m_pos = m_string.begin();
+ m_lastDelim = _T('\0');
+ m_hasMoreTokens = MoreTokens_Unknown;
}
-// needed to fix leading whitespace / mult. delims bugs
-void wxStringTokenizer::EatLeadingDelims()
-{
- int pos;
+// ----------------------------------------------------------------------------
+// access to the tokens
+// ----------------------------------------------------------------------------
- // while leading delims trim 'em from the left
- while ( ( pos = FindDelims(m_string, m_delims)) == 0 )
+// do we have more of them?
+bool wxStringTokenizer::HasMoreTokens() const
+{
+ // 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 )
{
- m_string = m_string.Mid((size_t)1);
+ bool r = DoHasMoreTokens();
+ wxConstCast(this, wxStringTokenizer)->m_hasMoreTokens =
+ r ? MoreTokens_Yes : MoreTokens_No;
+ return r;
}
+ else
+ return m_hasMoreTokens == MoreTokens_Yes;
}
-wxString wxStringTokenizer::NextToken()
+bool wxStringTokenizer::DoHasMoreTokens() const
{
- off_t pos, pos2;
- wxString r_string;
+ wxCHECK_MSG( IsOk(), false, _T("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;
+ }
- if ( m_string.IsEmpty() )
- return m_string;
+ switch ( m_mode )
+ {
+ 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_string.empty() && m_pos == m_string.begin();
- if ( !m_retdelims )
- EatLeadingDelims();
+ 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_stringEnd || m_lastDelim != _T('\0');
- pos = FindDelims(m_string, m_delims);
- if (pos == -1)
+ 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 (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 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(wxString(m_pos, m_stringEnd), m_delims, m_mode);
+
+ size_t count = 0;
+ while ( tkz.HasMoreTokens() )
{
- r_string = m_string;
- m_string = wxEmptyString;
+ count++;
- return r_string;
+ (void)tkz.GetNextToken();
}
- if (m_retdelims)
+ return count;
+}
+
+// ----------------------------------------------------------------------------
+// token extraction
+// ----------------------------------------------------------------------------
+
+wxString wxStringTokenizer::GetNextToken()
+{
+ wxString token;
+ do
{
- if (!pos)
+ if ( !HasMoreTokens() )
{
- pos++;
- pos2 = 1;
+ break;
}
- else
+
+ m_hasMoreTokens = MoreTokens_Unknown;
+
+ // find the end of this token
+ wxString::const_iterator pos =
+ find_first_of(m_delims, m_delimsLen, m_pos, m_stringEnd);
+
+ // and the start of the next one
+ if ( pos == m_stringEnd )
{
- pos2 = pos;
+ // no more delimiters, the token is everything till the end of
+ // string
+ token.assign(m_pos, m_stringEnd);
+
+ // skip the token
+ m_pos = m_stringEnd;
+
+ // it wasn't terminated
+ m_lastDelim = _T('\0');
+ }
+ else // we found a delimiter at pos
+ {
+ // in wxTOKEN_RET_DELIMS mode we return the delimiter character
+ // with token, otherwise leave it out
+ if ( m_mode == wxTOKEN_RET_DELIMS )
+ ++pos;
+
+ token.assign(m_pos, pos);
+
+ // skip the token and the trailing delimiter
+ m_pos = pos + 1;
+
+ m_lastDelim = (pos == m_stringEnd) ? _T('\0') : (wxChar)*pos;
}
}
- 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;
}
projects / wxWidgets.git / blobdiff