X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/bbf8fc5391b2458d630d29a794df99c8e681e6eb..3d27b5746f360ea930bd1046fd61b1b27440ba7b:/src/common/tokenzr.cpp diff --git a/src/common/tokenzr.cpp b/src/common/tokenzr.cpp index 0601d5d0e5..937987241c 100644 --- a/src/common/tokenzr.cpp +++ b/src/common/tokenzr.cpp @@ -2,7 +2,7 @@ // 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 @@ -17,7 +17,7 @@ // headers // ---------------------------------------------------------------------------- -#ifdef __GNUG__ +#if defined(__GNUG__) && !defined(NO_GCC_PRAGMA) #pragma implementation "tokenzr.h" #endif @@ -29,6 +29,10 @@ #endif #include "wx/tokenzr.h" +#include "wx/arrstr.h" + +// Required for wxIs... functions +#include // ============================================================================ // implementation @@ -38,57 +42,111 @@ // wxStringTokenizer construction // ---------------------------------------------------------------------------- -wxStringTokenizer::wxStringTokenizer(const wxString& to_tokenize, +wxStringTokenizer::wxStringTokenizer(const wxString& str, const wxString& delims, - bool ret_delims) + wxStringTokenizerMode mode) { - SetString(to_tokenize, delims, ret_delims); + SetString(str, delims, mode); } -void wxStringTokenizer::SetString(const wxString& to_tokenize, +void wxStringTokenizer::SetString(const wxString& str, const wxString& delims, - bool ret_delim) + 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_delim; - m_pos = 0; + m_mode = mode; - // empty string doesn't have any tokens - m_hasMore = !m_string.empty(); + Reinit(str); } -wxStringTokenizer::~wxStringTokenizer() +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(); } // ---------------------------------------------------------------------------- -// count the number of tokens in the string +// access to the tokens // ---------------------------------------------------------------------------- -size_t wxStringTokenizer::CountTokens() const +// do we have more of them? +bool wxStringTokenizer::HasMoreTokens() const { - size_t pos = 0; - size_t count = 0; - for ( ;; ) + wxCHECK_MSG( IsOk(), FALSE, _T("you should call SetString() first") ); + + if ( m_string.find_first_not_of(m_delims) == wxString::npos ) { - pos = m_string.find_first_of(m_delims, pos); - if ( pos == wxString::npos ) - break; + // 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; + } +} - count++; // one more token found +// count the number of tokens in the string +size_t wxStringTokenizer::CountTokens() const +{ + wxCHECK_MSG( IsOk(), 0, _T("you should call SetString() first") ); - pos++; // skip delimiter - } + // 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; - // normally, we didn't count the last token in the loop above - so add it - // unless the string was empty from the very beginning, in which case it - // still has 0 (and not 1) tokens - if ( !m_string.empty() ) + size_t count = 0; + while ( self->HasMoreTokens() ) { count++; + + (void)self->GetNextToken(); } + self->Reinit(stringInitial); + return count; } @@ -98,33 +156,68 @@ size_t wxStringTokenizer::CountTokens() const wxString wxStringTokenizer::GetNextToken() { + // strtok() doesn't return empty tokens, all other modes do + bool allowEmpty = m_mode != wxTOKEN_STRTOK; + wxString token; - if ( HasMoreTokens() ) + do { - size_t pos = m_string.find_first_of(m_delims); // end of token - size_t pos2; // start of the next one - if ( pos != wxString::npos ) + if ( !HasMoreTokens() ) { - // return the delimiter too - pos2 = pos + 1; + break; } - else + // 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 ) { - pos2 = m_string.length(); + // no more delimiters, the token is everything till the end of + // string + token = m_string; - // no more tokens in this 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; - token = wxString(m_string, m_retdelims ? pos2 : pos); + // 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); + // 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; + // keep track of the position in the original string too + m_pos += pos2; + } } - //else: no more tokens, return empty token + 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; +}