// Name: 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"
+// ============================================================================
+// declarations
+// ============================================================================
+
+// ----------------------------------------------------------------------------
+// headers
+// ----------------------------------------------------------------------------
+
+#if defined(__GNUG__) && !defined(NO_GCC_PRAGMA)
+ #pragma implementation "tokenzr.h"
#endif
// For compilers that support precompilation, includes "wx.h".
#include "wx/wxprec.h"
#ifdef __BORLANDC__
-#pragma hdrstop
+ #pragma hdrstop
#endif
-#ifndef WX_PRECOMP
-#endif
-
-#include "wx/object.h"
-#include "wx/string.h"
#include "wx/tokenzr.h"
+#include "wx/arrstr.h"
+
+// Required for wxIs... functions
+#include <ctype.h>
-wxStringTokenizer::wxStringTokenizer(const wxString& to_tokenize,
+// ============================================================================
+// implementation
+// ============================================================================
+
+// ----------------------------------------------------------------------------
+// wxStringTokenizer construction
+// ----------------------------------------------------------------------------
+
+wxStringTokenizer::wxStringTokenizer(const wxString& str,
const wxString& delims,
- bool ret_delims)
- : wxObject()
+ wxStringTokenizerMode mode)
{
- m_string = to_tokenize;
- m_delims = delims;
- m_retdelims = ret_delims;
+ 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;
+ }
+
+ if ( *p )
+ {
+ // not whitespace char in delims
+ mode = wxTOKEN_RET_EMPTY;
+ }
+ else
+ {
+ // only whitespaces
+ mode = wxTOKEN_STRTOK;
+ }
+ }
+
+ m_delims = delims;
+ m_mode = mode;
+
+ Reinit(str);
}
-off_t wxStringTokenizer::FindDelims(const wxString& str, const wxString& delims)
+void wxStringTokenizer::Reinit(const wxString& str)
{
- int i, j;
- register int s_len = str.Length(),
- len = delims.Length();
-
- for (i=0;i<s_len;i++) {
- register char c = str[i];
-
- for (j=0;j<len;j++)
- if (delims[j] == c)
- return i;
- }
- return -1;
+ 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();
}
-int wxStringTokenizer::CountTokens()
+// ----------------------------------------------------------------------------
+// access to the tokens
+// ----------------------------------------------------------------------------
+
+// do we have more of them?
+bool wxStringTokenizer::HasMoreTokens() const
{
- wxString p_string = m_string;
- bool found = TRUE;
- int pos, count = 1;
-
- if (p_string.Length() == 0)
- return 0;
-
- while (found) {
- pos = FindDelims(p_string, m_delims);
- if (pos != -1) {
- count++;
- p_string = p_string(0, pos);
- } else
- found = FALSE;
- }
- return count;
+ 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;
+ }
}
-bool wxStringTokenizer::HasMoreToken()
+// count the number of tokens in the string
+size_t wxStringTokenizer::CountTokens() const
{
- return (m_string.Length() != 0);
+ 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;
}
-// AVS - added to fix leading whitespace / mult. delims bugs
-void wxStringTokenizer::EatLeadingDelims()
-{
- int pos;
+// ----------------------------------------------------------------------------
+// token extraction
+// ----------------------------------------------------------------------------
- while ((pos=FindDelims(m_string, m_delims))==0) { // while leading delims
- m_string = m_string.Mid((size_t)1); // trim 'em from the left
- }
+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;
+ }
+ }
+ while ( !allowEmpty && token.empty() );
+
+ return token;
}
-wxString wxStringTokenizer::NextToken()
+// ----------------------------------------------------------------------------
+// public functions
+// ----------------------------------------------------------------------------
+
+wxArrayString wxStringTokenize(const wxString& str,
+ const wxString& delims,
+ wxStringTokenizerMode mode)
{
- register off_t pos, pos2;
- wxString r_string;
-
- if (m_string.IsNull())
- return m_string;
-
- if (!m_retdelims)
- EatLeadingDelims(); // AVS - added to fix leading whitespace /
- // mult. delims bugs
-
- pos = FindDelims(m_string, m_delims);
- if (pos == -1) {
- r_string = m_string;
- m_string = wxEmptyString;
-
- return r_string;
- }
-
- if (m_retdelims) {
- if (!pos) {
- pos++;
- pos2 = 1;
- } else
- pos2 = pos;
- } else
- pos2 = pos + 1;
-
- r_string = m_string.Left((size_t)pos);
- m_string = m_string.Mid((size_t)pos2);
-
- return r_string;
+ wxArrayString tokens;
+ wxStringTokenizer tk(str, delims, mode);
+ while ( tk.HasMoreTokens() )
+ {
+ tokens.Add(tk.GetNextToken());
+ }
+
+ return tokens;
}
projects / wxWidgets.git / blobdiff