From 7c968cee840eda3b7340cdb4ba822065721cbdf6 Mon Sep 17 00:00:00 2001 From: Vadim Zeitlin Date: Fri, 4 Feb 2000 18:31:26 +0000 Subject: [PATCH 1/1] another new version of wxStringTokenizer (with tests and docs) git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@5839 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775 --- docs/latex/wx/tokenizr.tex | 68 +++++++++++------ include/wx/tokenzr.h | 92 +++++++++++++++++----- samples/console/console.cpp | 136 +++++++++++++++++++++++++++------ src/common/tokenzr.cpp | 147 ++++++++++++++++++++++++++---------- 4 files changed, 334 insertions(+), 109 deletions(-) diff --git a/docs/latex/wx/tokenizr.tex b/docs/latex/wx/tokenizr.tex index b066262a52..0b8490ce1d 100644 --- a/docs/latex/wx/tokenizr.tex +++ b/docs/latex/wx/tokenizr.tex @@ -1,6 +1,8 @@ \section{\class{wxStringTokenizer}}\label{wxstringtokenizer} -wxStringTokenizer helps you to break a string up into a number of tokens. +wxStringTokenizer helps you to break a string up into a number of tokens. It +replaces the standard C function {\tt strtok()} and also extends it in a +number of ways. To use this class, you should create a wxStringTokenizer object, give it the string to tokenize and also the delimiters which separate tokens in the string @@ -14,7 +16,7 @@ For example: \begin{verbatim} -wxStringTokenizer tkz("first:second:third::fivth", ":"); +wxStringTokenizer tkz("first:second:third:fourth", ":"); while ( tkz.HasMoreTokens() ) { wxString token = tkz.GetNextToken(); @@ -23,14 +25,36 @@ while ( tkz.HasMoreTokens() ) } \end{verbatim} -Another feature of this class is that it may return the delimiter which -was found after the token with it. In a simple case like above, you are not -interested in this because the delimiter is always {\tt ':'}, but if the -delimiters string has several characters, you might need to know which of them -follows the current token. In this case, pass {\tt TRUE} to wxStringTokenizer -constructor or \helpref{SetString}{wxstringtokenizersetstring} method and -the delimiter will be appended to each returned token (except for the last -one). +By default, wxStringTokenizer will behave in the same way as {\tt strtok()} if +the delimiters string only contains white space characters but, unlike the +standard function, it will return empty tokens if this is not the case. This +is helpful for parsing strictly formatted data where the number of fields is +fixed but some of them may be empty (i.e. {\tt TAB} or comma delimited text +files). + +The behaviour is governed by the last +\helpref{constructor}{wxstringtokenizerwxstringtokenizer}/\helpref{SetString}{wxstringtokenizersetstring} +parameter {\tt mode} which may be one of the following: + +\twocolwidtha{5cm}% +\begin{twocollist}\itemsep=0pt +\twocolitem{{\tt wxTOKEN\_DEFAULT}}{Default behaviour (as described above): +same as {\tt wxTOKEN\_STRTOK} if the delimiter string contains only +whitespaces, same as {\tt wxTOKEN\_RET\_EMPTY} otherwise} +\twocolitem{{\tt wxTOKEN\_RET\_EMPTY}}{In this mode, the empty tokens in the +middle of the string will be returned, i.e. {\tt "a::b:"} will be tokenized in +three tokens `a', `' and `b'.} +\twocolitem{{\tt wxTOKEN\_RET\_EMPTY\_ALL}}{In this mode, empty trailing token +(after the last delimiter character) will be returned as well. The string as +above will contain four tokens: the already mentioned ones and another empty +one as the last one.} +\twocolitem{{\tt wxTOKEN\_RET\_DELIMS}}{In this mode, the delimiter character +after the end of the current token (there may be none if this is the last +token) is returned appended to the token. Otherwise, it is the same mode as +{\tt wxTOKEN\_RET\_EMPTY}.} +\twocolitem{{\tt wxTOKEN\_STRTOK}}{In this mode the class behaves exactly like +the standard {\tt strtok()} function. The empty tokens are never returned.} +\end{twocollist} \wxheading{Derived from} @@ -46,18 +70,14 @@ one). \func{}{wxStringTokenizer}{\void} -Default constructor. +Default constructor. You must call +\helpref{SetString}{wxstringtokenizersetstring} before calling any other +methods. -\func{}{wxStringTokenizer}{\param{const wxString\& }{to\_tokenize}, \param{const wxString\& }{delims = " $\backslash$t$\backslash$r$\backslash$n"}, \param{bool }{ret\_delim = FALSE}} +\func{}{wxStringTokenizer}{\param{const wxString\& }{str}, \param{const wxString\& }{delims = " $\backslash$t$\backslash$r$\backslash$n"}, \param{wxStringTokenizerMode }{mode = wxTOKEN\_DEFAULT}} -Constructor. Pass the string to tokenize, a string containing delimiters, -a flag specifying whether to return delimiters with tokens. - -\membersection{wxStringTokenizer::\destruct{wxStringTokenizer}}\label{wxstringtokenizerdtor} - -\func{}{\destruct{wxStringTokenizer}}{\void} - -Destructor. +Constructor. Pass the string to tokenize, a string containing delimiters +and the mode specifying how the string should be tokenized. \membersection{wxStringTokenizer::CountTokens}\label{wxstringtokenizercounttokens} @@ -69,11 +89,11 @@ Returns the number of tokens in the input string. \constfunc{bool}{HasMoreTokens}{\void} -Returns TRUE if the tokenizer has further tokens. +Returns TRUE if the tokenizer has further tokens, FALSE if none are left. \membersection{wxStringTokenizer::GetNextToken}\label{wxstringtokenizergetnexttoken} -\constfunc{wxString}{GetNextToken}{\void} +\func{wxString}{GetNextToken}{\void} Returns the next token or empty string if the end of string was reached. @@ -93,10 +113,10 @@ Returns the part of the starting string without all token already extracted. \membersection{wxStringTokenizer::SetString}\label{wxstringtokenizersetstring} -\func{void}{SetString}{\param{const wxString\& }{to\_tokenize}, \param{const wxString\& }{delims = " $\backslash$t$\backslash$r$\backslash$n"}, \param{bool }{ret\_delim = FALSE}} +\func{void}{SetString}{\param{const wxString\& }{to\_tokenize}, \param{const wxString\& }{delims = " $\backslash$t$\backslash$r$\backslash$n"}, \param{wxStringTokenizerMode }{mode = wxTOKEN\_DEFAULT}} Initializes the tokenizer. Pass the string to tokenize, a string containing delimiters, -a flag specifying whether to return delimiters with tokens. +and the mode specifying how the string should be tokenized. diff --git a/include/wx/tokenzr.h b/include/wx/tokenzr.h index 9ff109d46d..9a08ddb460 100644 --- a/include/wx/tokenzr.h +++ b/include/wx/tokenzr.h @@ -1,6 +1,6 @@ ///////////////////////////////////////////////////////////////////////////// -// Name: tokenzr.h -// Purpose: String tokenizer +// Name: wx/tokenzr.h +// Purpose: String tokenizer - a C++ replacement for strtok(3) // Author: Guilhem Lavaux // Modified by: Vadim Zeitlin // Created: 04/22/98 @@ -19,48 +19,102 @@ #include "wx/object.h" #include "wx/string.h" +// ---------------------------------------------------------------------------- +// constants +// ---------------------------------------------------------------------------- + // default: delimiters are usual white space characters #define wxDEFAULT_DELIMITERS (_T(" \t\r\n")) +// wxStringTokenizer mode flags which determine its behaviour +enum wxStringTokenizerMode +{ + wxTOKEN_INVALID = -1, // set by def ctor until SetString() is called + wxTOKEN_DEFAULT, // strtok() for whitespace delims, RET_EMPTY else + wxTOKEN_RET_EMPTY, // return empty token in the middle of the string + wxTOKEN_RET_EMPTY_ALL, // return trailing empty tokens too + wxTOKEN_RET_DELIMS, // return the delim with token (implies RET_EMPTY) + wxTOKEN_STRTOK // behave exactly like strtok(3) +}; + +// ---------------------------------------------------------------------------- +// wxStringTokenizer: replaces infamous strtok() and has some other features +// ---------------------------------------------------------------------------- + class WXDLLEXPORT wxStringTokenizer : public wxObject { public: - // ctors and such - wxStringTokenizer() { m_retdelims = FALSE; m_pos = 0; } - wxStringTokenizer(const wxString& to_tokenize, + // ctors and initializers + // default ctor, call SetString() later + wxStringTokenizer() { m_mode = wxTOKEN_INVALID; } + // ctor which gives us the string + wxStringTokenizer(const wxString& str, const wxString& delims = wxDEFAULT_DELIMITERS, - bool ret_delim = FALSE); - void SetString(const wxString& to_tokenize, + wxStringTokenizerMode mode = wxTOKEN_DEFAULT); + + // args are same as for the non default ctor above + void SetString(const wxString& str, const wxString& delims = wxDEFAULT_DELIMITERS, - bool ret_delim = FALSE); - virtual ~wxStringTokenizer(); + wxStringTokenizerMode mode = wxTOKEN_DEFAULT); + + // reinitialize the tokenizer with the same delimiters/mode + void Reinit(const wxString& str); - // count tokens/get next token + // tokens access + // count them size_t CountTokens() const; - bool HasMoreTokens() { return m_hasMore; } + // did we reach the end of the string? + bool HasMoreTokens() const; + // get the next token, will return empty string if !HasMoreTokens() wxString GetNextToken(); - // One note about GetString -- it returns the string - // remaining after the previous tokens have been removed, - // not the original string + // get current tokenizer state + // returns the part of the string which remains to tokenize (*not* the + // initial string) wxString GetString() const { return m_string; } - // returns the current position (i.e. one index after the last returned - // token or 0 if GetNextToken() has never been called) in the original - // string + // returns the current position (i.e. one index after the last + // returned token or 0 if GetNextToken() has never been called) in the + // original string size_t GetPosition() const { return m_pos; } + // misc + // get the current mode - can be different from the one passed to the + // ctor if it was wxTOKEN_DEFAULT + wxStringTokenizerMode GetMode() const { return m_mode; } + + // backwards compatibility section from now on + // ------------------------------------------- + // for compatibility only, use GetNextToken() instead wxString NextToken() { return GetNextToken(); } + // compatibility only, don't use + void SetString(const wxString& to_tokenize, + const wxString& delims, + bool ret_delim) + { + SetString(to_tokenize, delims, wxTOKEN_RET_DELIMS); + } + + wxStringTokenizer(const wxString& to_tokenize, + const wxString& delims, + bool ret_delim) + { + SetString(to_tokenize, delims, ret_delim); + } + protected: + bool IsOk() const { return m_mode != wxTOKEN_INVALID; } + wxString m_string, // the (rest of) string to tokenize m_delims; // all delimiters size_t m_pos; // the position in the original string - bool m_retdelims; // if TRUE, return delims with tokens - bool m_hasMore; // do we have more tokens? + wxStringTokenizerMode m_mode; // see wxTOKEN_XXX values + + bool m_hasMore; // do we have more (possible empty) tokens? }; #endif // _WX_TOKENZRH diff --git a/samples/console/console.cpp b/samples/console/console.cpp index bd8b550139..d073cb5511 100644 --- a/samples/console/console.cpp +++ b/samples/console/console.cpp @@ -425,7 +425,7 @@ static void TestAddition() #if wxUSE_LONGLONG_NATIVE wxASSERT_MSG( c == wxLongLongNative(a.GetHi(), a.GetLo()) + wxLongLongNative(b.GetHi(), b.GetLo()), - "addition failure" ); + "addition failure" ); #else // !wxUSE_LONGLONG_NATIVE wxASSERT_MSG( c - b == a, "addition failure" ); #endif // wxUSE_LONGLONG_NATIVE @@ -921,7 +921,7 @@ def GetMonthWeek(dt): if weekNumMonth < 0: weekNumMonth = weekNumMonth + 53 return weekNumMonth - + def GetLastSundayBefore(dt): if dt.iso_week[2] == 7: return dt @@ -1322,7 +1322,7 @@ static void TestTimeArithmetics() { wxDateSpan span; const char *name; - } testArithmData[] = + } testArithmData[] = { { wxDateSpan::Day(), "day" }, { wxDateSpan::Week(), "week" }, @@ -1330,7 +1330,7 @@ static void TestTimeArithmetics() { wxDateSpan::Year(), "year" }, { wxDateSpan(1, 2, 3, 4), "year, 2 months, 3 weeks, 4 days" }, }; - + wxDateTime dt(29, wxDateTime::Dec, 1999), dt1, dt2; for ( size_t n = 0; n < WXSIZEOF(testArithmData); n++ ) @@ -1701,6 +1701,37 @@ void PrintArray(const char* name, const wxArrayString& array) #include "wx/timer.h" #include "wx/tokenzr.h" +static void TestStringConstruction() +{ + puts("*** Testing wxString constructores ***"); + + #define TEST_CTOR(args, res) \ + { \ + wxString s args ; \ + printf("wxString%s = %s ", #args, s.c_str()); \ + if ( s == res ) \ + { \ + puts("(ok)"); \ + } \ + else \ + { \ + printf("(ERROR: should be %s)\n", res); \ + } \ + } + + TEST_CTOR((_T('Z'), 4), _T("ZZZZ")); + TEST_CTOR((_T("Hello"), 4), _T("Hell")); + TEST_CTOR((_T("Hello"), 5), _T("Hello")); + // TEST_CTOR((_T("Hello"), 6), _T("Hello")); -- should give assert failure + + static const wxChar *s = _T("?really!"); + const wxChar *start = wxStrchr(s, _T('r')); + const wxChar *end = wxStrchr(s, _T('!')); + TEST_CTOR((start, end), _T("really")); + + puts(""); +} + static void TestString() { wxStopWatch sw; @@ -1842,33 +1873,49 @@ static void TestStringTokenizer() { puts("*** Testing wxStringTokenizer ***"); + static const wxChar *modeNames[] = + { + _T("default"), + _T("return empty"), + _T("return all empty"), + _T("with delims"), + _T("like strtok"), + }; + static const struct StringTokenizerTest { - const wxChar *str; // string to tokenize - const wxChar *delims; // delimiters to use - size_t count; // count of token - bool with; // return tokens with delimiters? - } tokenizerTestData[] = - { - { _T(""), _T(" "), 0, FALSE }, - { _T("Hello, world"), _T(" "), 2, FALSE }, - { _T("Hello, world"), _T(","), 2, FALSE }, - { _T("Hello, world!"), _T(",!"), 3, TRUE }, - { _T("username:password:uid:gid:gecos:home:shell"), _T(":"), 7, FALSE }, - { _T("1 \t3\t4 6 "), wxDEFAULT_DELIMITERS, 9, TRUE }, - { _T("01/02/99"), _T("/-"), 3, FALSE }, + const wxChar *str; // string to tokenize + const wxChar *delims; // delimiters to use + size_t count; // count of token + wxStringTokenizerMode mode; // how should we tokenize it + } tokenizerTestData[] = + { + { _T(""), _T(" "), 0 }, + { _T("Hello, world"), _T(" "), 2 }, + { _T("Hello, world "), _T(" "), 2 }, + { _T("Hello, world"), _T(","), 2 }, + { _T("Hello, world!"), _T(",!"), 2 }, + { _T("Hello,, world!"), _T(",!"), 3 }, + { _T("Hello, world!"), _T(",!"), 3, wxTOKEN_RET_EMPTY_ALL }, + { _T("username:password:uid:gid:gecos:home:shell"), _T(":"), 7 }, + { _T("1 \t3\t4 6 "), wxDEFAULT_DELIMITERS, 4 }, + { _T("1 \t3\t4 6 "), wxDEFAULT_DELIMITERS, 6, wxTOKEN_RET_EMPTY }, + { _T("1 \t3\t4 6 "), wxDEFAULT_DELIMITERS, 9, wxTOKEN_RET_EMPTY_ALL }, + { _T("01/02/99"), _T("/-"), 3 }, + { _T("01-02/99"), _T("/-"), 3, wxTOKEN_RET_DELIMS }, }; for ( size_t n = 0; n < WXSIZEOF(tokenizerTestData); n++ ) { const StringTokenizerTest& tt = tokenizerTestData[n]; - wxStringTokenizer tkz(tt.str, tt.delims, tt.with); + wxStringTokenizer tkz(tt.str, tt.delims, tt.mode); size_t count = tkz.CountTokens(); - printf(_T("String '%s' has %u tokens delimited by '%s' "), - tt.str, + printf(_T("String '%s' has %u tokens delimited by '%s' (mode = %s) "), + MakePrintable(tt.str).c_str(), count, - MakePrintable(tt.delims).c_str()); + MakePrintable(tt.delims).c_str(), + modeNames[tkz.GetMode()]); if ( count == tt.count ) { puts(_T("(ok)")); @@ -1880,19 +1927,57 @@ static void TestStringTokenizer() continue; } + // if we emulate strtok(), check that we do it correctly + wxChar *buf, *s, *last; + + if ( tkz.GetMode() == wxTOKEN_STRTOK ) + { + buf = new wxChar[wxStrlen(tt.str) + 1]; + wxStrcpy(buf, tt.str); + + s = wxStrtok(buf, tt.delims, &last); + } + else + { + buf = NULL; + } + // now show the tokens themselves size_t count2 = 0; while ( tkz.HasMoreTokens() ) { - printf(_T("\ttoken %u: '%s'\n"), + wxString token = tkz.GetNextToken(); + + printf(_T("\ttoken %u: '%s'"), ++count2, - MakePrintable(tkz.GetNextToken()).c_str()); + MakePrintable(token).c_str()); + + if ( buf ) + { + if ( token == s ) + { + puts(" (ok)"); + } + else + { + printf(" (ERROR: should be %s)\n", s); + } + + s = wxStrtok(NULL, tt.delims, &last); + } + else + { + // nothing to compare with + puts(""); + } } if ( count2 != count ) { - puts(_T("ERROR: token count mismatch")); + puts(_T("\tERROR: token count mismatch")); } + + delete [] buf; } puts(""); @@ -1959,11 +2044,12 @@ int main(int argc, char **argv) } if ( 0 ) { + TestStringConstruction(); TestStringSub(); TestStringFormat(); TestStringFind(); } - TestStringTokenizer(); + TestStringTokenizer(); #endif // TEST_STRINGS #ifdef TEST_ARRAYS diff --git a/src/common/tokenzr.cpp b/src/common/tokenzr.cpp index 0601d5d0e5..64f49e5822 100644 --- a/src/common/tokenzr.cpp +++ b/src/common/tokenzr.cpp @@ -38,57 +38,105 @@ // 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 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; + } + 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 +146,50 @@ 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; + + m_pos += m_string.length(); + m_string.clear(); - // no more tokens in this string + // 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; } -- 2.45.2