X-Git-Url: https://git.saurik.com/wxWidgets.git/blobdiff_plain/7c968cee840eda3b7340cdb4ba822065721cbdf6..aa4919edd16047bb9590658f7a9fbe50fc659bc1:/src/common/tokenzr.cpp diff --git a/src/common/tokenzr.cpp b/src/common/tokenzr.cpp index 64f49e5822..2822025bb7 100644 --- a/src/common/tokenzr.cpp +++ b/src/common/tokenzr.cpp @@ -1,8 +1,8 @@ ///////////////////////////////////////////////////////////////////////////// -// Name: tokenzr.cpp +// Name: src/common/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 @@ -17,10 +17,6 @@ // headers // ---------------------------------------------------------------------------- -#ifdef __GNUG__ - #pragma implementation "tokenzr.h" -#endif - // For compilers that support precompilation, includes "wx.h". #include "wx/wxprec.h" @@ -30,10 +26,54 @@ #include "wx/tokenzr.h" +#ifndef WX_PRECOMP + #include "wx/arrstr.h" + #include "wx/crt.h" +#endif + +// Required for wxIs... functions +#include + // ============================================================================ // implementation // ============================================================================ +// ---------------------------------------------------------------------------- +// helpers +// ---------------------------------------------------------------------------- + +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, wxT("invalid index") ); + + for ( wxString::const_iterator i = from; i != end; ++i ) + { + if ( wxTmemchr(delims, *i, len) ) + return i; + } + + return end; +} + +static wxString::const_iterator +find_first_not_of(const wxChar *delims, size_t len, + const wxString::const_iterator& from, + const wxString::const_iterator& end) +{ + wxASSERT_MSG( from <= end, wxT("invalid index") ); + + for ( wxString::const_iterator i = from; i != end; ++i ) + { + if ( !wxTmemchr(delims, *i, len) ) + return i; + } + + return end; +} + // ---------------------------------------------------------------------------- // wxStringTokenizer construction // ---------------------------------------------------------------------------- @@ -55,14 +95,14 @@ void wxStringTokenizer::SetString(const wxString& str, // 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++ ) + wxString::const_iterator p; + for ( p = delims.begin(); p != delims.end(); ++p ) { if ( !wxIsspace(*p) ) break; } - if ( *p ) + if ( p != delims.end() ) { // not whitespace char in delims mode = wxTOKEN_RET_EMPTY; @@ -74,7 +114,13 @@ void wxStringTokenizer::SetString(const wxString& str, } } - m_delims = delims; +#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); @@ -82,13 +128,13 @@ void wxStringTokenizer::SetString(const wxString& str, void wxStringTokenizer::Reinit(const wxString& str) { - wxASSERT_MSG( IsOk(), _T("you should call SetString() first") ); + wxASSERT_MSG( IsOk(), wxT("you should call SetString() first") ); m_string = str; - m_pos = 0; - - // empty string doesn't have any tokens - m_hasMore = !m_string.empty(); + m_stringEnd = m_string.end(); + m_pos = m_string.begin(); + m_lastDelim = wxT('\0'); + m_hasMoreTokens = MoreTokens_Unknown; } // ---------------------------------------------------------------------------- @@ -98,45 +144,82 @@ void wxStringTokenizer::Reinit(const wxString& str) // 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 ) + // 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 ) { - // 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; + bool r = DoHasMoreTokens(); + wxConstCast(this, wxStringTokenizer)->m_hasMoreTokens = + r ? MoreTokens_Yes : MoreTokens_No; + return r; } else + return m_hasMoreTokens == MoreTokens_Yes; +} + +bool wxStringTokenizer::DoHasMoreTokens() const +{ + wxCHECK_MSG( IsOk(), false, wxT("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, hence we do have more - // tokens - return TRUE; + // there are non delimiter characters left, so we do have more tokens + return true; } + + 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(); + + 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 != wxT('\0'); + + case wxTOKEN_INVALID: + case wxTOKEN_DEFAULT: + wxFAIL_MSG( wxT("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") ); + wxCHECK_MSG( IsOk(), 0, wxT("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(wxString(m_pos, m_stringEnd), 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; } @@ -146,9 +229,6 @@ 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; do { @@ -156,40 +236,61 @@ wxString wxStringTokenizer::GetNextToken() { break; } + + m_hasMoreTokens = MoreTokens_Unknown; + // find the end of this token - size_t pos = m_string.find_first_of(m_delims); + wxString::const_iterator pos = + find_first_of(m_delims, m_delimsLen, m_pos, m_stringEnd); // and the start of the next one - if ( pos == wxString::npos ) + if ( pos == m_stringEnd ) { // no more delimiters, the token is everything till the end of // string - token = m_string; + token.assign(m_pos, m_stringEnd); - m_pos += m_string.length(); - m_string.clear(); + // skip the token + m_pos = m_stringEnd; - // 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 = wxT('\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 + wxString::const_iterator tokenEnd(pos); + if ( m_mode == wxTOKEN_RET_DELIMS ) + ++tokenEnd; + + token.assign(m_pos, tokenEnd); - // remove token with the following it delimiter from string - m_string.erase(0, pos2); + // skip the token and the trailing delimiter + m_pos = pos + 1; - // keep track of the position in the original string too - m_pos += pos2; + m_lastDelim = (pos == m_stringEnd) ? wxT('\0') : (wxChar)*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; +}