// Name: tokenzr.cpp
// Purpose: String tokenizer
// Author: Guilhem Lavaux
-// Modified by: Gregory Pietsch
+// Modified by: Vadim Zeitlin (almost full rewrite)
// Created: 04/22/98
// RCS-ID: $Id$
// Copyright: (c) Guilhem Lavaux
// Licence: wxWindows licence
/////////////////////////////////////////////////////////////////////////////
-#ifdef __GNUG__
+// ============================================================================
+// declarations
+// ============================================================================
+
+// ----------------------------------------------------------------------------
+// headers
+// ----------------------------------------------------------------------------
+
+#if defined(__GNUG__) && !defined(NO_GCC_PRAGMA)
#pragma implementation "tokenzr.h"
#endif
#endif
#include "wx/tokenzr.h"
+#include "wx/arrstr.h"
+
+// Required for wxIs... functions
+#include <ctype.h>
+
+// ============================================================================
+// implementation
+// ============================================================================
-wxStringTokenizer::wxStringTokenizer(const wxString& to_tokenize,
+// ----------------------------------------------------------------------------
+// wxStringTokenizer construction
+// ----------------------------------------------------------------------------
+
+wxStringTokenizer::wxStringTokenizer(const wxString& str,
const wxString& delims,
- bool ret_delims)
+ wxStringTokenizerMode mode)
+{
+ SetString(str, delims, mode);
+}
+
+void wxStringTokenizer::SetString(const wxString& str,
+ const wxString& delims,
+ wxStringTokenizerMode mode)
{
- m_string = to_tokenize;
+ 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;
+ }
+
+ if ( *p )
+ {
+ // not whitespace char in delims
+ mode = wxTOKEN_RET_EMPTY;
+ }
+ else
+ {
+ // only whitespaces
+ mode = wxTOKEN_STRTOK;
+ }
+ }
+
m_delims = delims;
- m_retdelims = ret_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();
+}
+
+// ----------------------------------------------------------------------------
+// access to the tokens
+// ----------------------------------------------------------------------------
+
+// do we have more of them?
+bool wxStringTokenizer::HasMoreTokens() const
+{
+ 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 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;
+ }
}
-wxStringTokenizer::~wxStringTokenizer()
+// count the number of 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
+ // in all modes
+ wxStringTokenizer *self = (wxStringTokenizer *)this; // const_cast
+ wxString stringInitial = m_string;
+
+ size_t count = 0;
+ while ( self->HasMoreTokens() )
+ {
+ count++;
+
+ (void)self->GetNextToken();
+ }
+
+ self->Reinit(stringInitial);
+
+ return count;
}
-int wxStringTokenizer::CountTokens() const
+// ----------------------------------------------------------------------------
+// token extraction
+// ----------------------------------------------------------------------------
+
+wxString wxStringTokenizer::GetNextToken()
{
- 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)
+ // strtok() doesn't return empty tokens, all other modes do
+ bool allowEmpty = m_mode != wxTOKEN_STRTOK;
+
+ wxString token;
+ do
+ {
+ if ( !HasMoreTokens() )
{
- // if we're retaining delimiters, increment count
- count++;
+ 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
{
- // if we're not retaining delimiters and at a token, inc count
- count += (!at_delim);
+ 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 count;
-}
+ while ( !allowEmpty && token.empty() );
-bool wxStringTokenizer::HasMoreTokens()
-{
- return (m_retdelims
- ? !m_string.IsEmpty()
- : m_string.find_first_not_of(m_delims) < m_string.length());
+ return token;
}
-wxString wxStringTokenizer::NextToken()
+// ----------------------------------------------------------------------------
+// public functions
+// ----------------------------------------------------------------------------
+
+wxArrayString wxStringTokenize(const wxString& str,
+ const wxString& delims,
+ wxStringTokenizerMode mode)
{
- 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);
- }
- 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();
+ wxArrayString tokens;
+ wxStringTokenizer tk(str, delims, mode);
+ while ( tk.HasMoreTokens() )
+ {
+ tokens.Add(tk.GetNextToken());
}
- return r_string;
+
+ return tokens;
}
projects / wxWidgets.git / blobdiff