bool Tokenizer::SkipToOneOfChars(const wxChar* chars, bool supportNesting)
{
// skip everything until we find any one of chars
while (1)
{
while (NotEOF() && !CharInString(CurrentChar(), chars))
{
if (CurrentChar() == '"' || CurrentChar() == '\'')
{
// this is the case that match is inside a string!
wxChar ch = CurrentChar();
MoveToNextChar();
SkipToChar(ch);
}
MoveToNextChar();
// make sure we skip comments
if (CurrentChar() == '/')
SkipComment(); // this will decide if it is a comment
// use 'while' here to cater for consecutive blocks to skip (e.g. sometemplate<foo>(bar)
// must skip <foo> and immediately after (bar))
// because if we don't, the next block won't be skipped ((bar) in the example) leading to weird
// parsing results
bool done = false;
while (supportNesting && !done)
{
switch (CurrentChar())
{
case '{': SkipBlock('{'); break;
case '(': SkipBlock('('); break;
case '[': SkipBlock('['); break;
case '<': // don't skip if << operator
if (NextChar() == '<')
MoveToNextChar(2); // skip it and also the next '<' or the next '<' leads to a SkipBlock('<');
else
SkipBlock('<');
break;
default: done = true; break;
}
}
}
if (PreviousChar() != '\\')
break;
else
{
// check for "\\"
if (m_TokenIndex - 2 >= 0 && m_Buffer.GetChar(m_TokenIndex - 2) == '\\')
break;
}
MoveToNextChar();
}
if (IsEOF())
return false;
return true;
}
void TextCutter::GetChunk(wxString& chunk, bool& isWord)
{
isWord = false;
chunk = wxEmptyString;
if (IsEOF())
return;
unsigned int start = m_CurIdx;
if (isalpha(CurrentChar()) || CurrentChar() == '_')
{
while (!IsEOF() &&
(isalnum(CurrentChar()) || CurrentChar() == '_'))
MoveToNextChar();
chunk = m_Text.Mid(start, m_CurIdx - start);
isWord = true;
}
else
{
SkipWhiteSpace();
SkipUnwanted();
if (start != m_CurIdx)
{
chunk = m_Text.Mid(start, m_CurIdx - start);
return;
}
if (isdigit(CurrentChar()))
{
// numbers
while (!IsEOF() && CharInString(CurrentChar(), "0123456789.abcdefABCDEFXxLl"))
MoveToNextChar();
}
else if (CurrentChar() == '"' ||
CurrentChar() == '\'')
{
// string, char, etc.
wxChar match = CurrentChar();
MoveToNextChar(); // skip starting ' or "
SkipToChar(match);
MoveToNextChar(); // skip ending ' or "
}
else
{
MoveToNextChar();
}
chunk = m_Text.Mid(start, m_CurIdx - start);
}
return;
}
// if we really move forward, return true, which means we have the new m_TokenIndex
// if we stay here, return false
bool Tokenizer::SkipComment()
{
if (IsEOF())
return false;
bool cstyle; // C or C++ style comments
//check the comment prompt
if (CurrentChar() == '/')
{
if (NextChar() == '*')
cstyle = true;
else if (NextChar() == '/')
cstyle = false;
else
return false; // Not a comment, return false;
}
else
return false; // Not a comment, return false;
TRACE(_T("SkipComment() : Start from line = %d"), m_LineNumber);
MoveToNextChar(2); // Skip the comment prompt
// Here, we are in the comment body
while (true)
{
if (cstyle) // C style comment
{
SkipToChar('/');
if (PreviousChar() == '*') // end of a C style comment
{
MoveToNextChar();
break;
}
if (!MoveToNextChar())
break;
}
else // C++ style comment
{
TRACE(_T("SkipComment() : Need to call SkipToEOL() here at line = %d"), m_LineNumber);
SkipToInlineCommentEnd();
break;
}
}
return true;
}
bool Tokenizer::SkipBlock(const wxChar& ch)
{
// skip blocks () [] {} <>
wxChar match;
switch (ch)
{
case '(': match = ')'; break;
case '[': match = ']'; break;
case '{': match = '}'; break;
case '<': match = '>'; break;
default : return false;
}
MoveToNextChar();
int count = 1; // counter for nested blocks (xxx())
while (NotEOF())
{
bool noMove = false;
if (CurrentChar() == '/')
SkipComment(); // this will decide if it is a comment
if (CurrentChar() == '"' || CurrentChar() == '\'')
{
// this is the case that match is inside a string!
wxChar ch = CurrentChar();
MoveToNextChar();
SkipToChar(ch);
MoveToNextChar();
// don't move to next char below if concatenating strings (e.g. printf("" ""))
if (CurrentChar() == '"' || CurrentChar() == '\'')
noMove = true;
}
if (CurrentChar() == ch)
++count;
else if (CurrentChar() == match)
--count;
if (!noMove)
MoveToNextChar();
if (count == 0)
break;
}
if (IsEOF())
return false;
return true;
}
bool Tokenizer::SkipToInlineCommentEnd()
{
TRACE(_T("%s : line=%d, CurrentChar='%c', PreviousChar='%c', NextChar='%c'"),
wxString(__PRETTY_FUNCTION__, wxConvUTF8).wc_str(), m_LineNumber, CurrentChar(),
PreviousChar(), NextChar());
// skip everything until we find EOL
while (true)
{
SkipToChar(_T('\n'));
if (!IsBackslashBeforeEOL() || IsEOF())
break;
else
MoveToNextChar();
}
TRACE(_T("SkipToInlineCommentEnd(): (END) We are now at line %d, CurrentChar='%c', PreviousChar='%c',")
_T(" NextChar='%c'"), m_LineNumber, CurrentChar(), PreviousChar(), NextChar());
return NotEOF();
}
bool Tokenizer::SkipComment(bool skipWhiteAtEnd) // = true
{
// C/C++ style comments
bool is_comment = CurrentChar() == '/' && (NextChar() == '/' || NextChar() == '*');
if (!is_comment)
return true;
bool cstyle = NextChar() == '*';
MoveToNextChar(2);
while (1)
{
if (!cstyle)
{
if (!SkipToEOL(false, true))
return false;
MoveToNextChar();
break;
}
else
{
if (SkipToChar('/'))
{
if (PreviousChar() == '*')
{
MoveToNextChar();
break;
}
MoveToNextChar();
}
else
return false;
}
}
if (IsEOF())
return false;
if (skipWhiteAtEnd && !SkipWhiteSpace())
return false;
return CurrentChar() == '/' ? SkipComment() : true; // handle chained comments
}
//vfc add bGetValue
wxString Tokenizer::DoGetToken(bool bGetValue, bool bTemplate)
{
if (IsEOF())
return wxEmptyString;
if (!SkipWhiteSpace())
return wxEmptyString;
if (m_SkipUnwantedTokens && !SkipUnwanted(bGetValue))
return wxEmptyString;
// if m_SkipUnwantedTokens is false, we need to handle comments here too
if (!m_SkipUnwantedTokens)
SkipComment();
int start = m_TokenIndex;
wxString m_Str;
wxChar c = CurrentChar();
if (c == '_' || wxIsalpha(c))
{
// keywords, identifiers, etc.
// operator== is cheaper than wxIsalnum, also MoveToNextChar already includes IsEOF
while ( ( CurrentChar() == '_' ||
wxIsalnum(CurrentChar()) ) && MoveToNextChar() )
;
if (IsEOF())
return wxEmptyString;
m_Str = m_Buffer.Mid(start, m_TokenIndex - start);
m_IsOperator = m_Str.IsSameAs(TokenizerConsts::operator_str);
}
#ifdef __WXMSW__ // This is a Windows only bug!
else if (c == 178 || c == 179 || c == 185) // fetch ?and ?
{
m_Str = c;
MoveToNextChar();
}
#endif
else if (wxIsdigit(CurrentChar()))
{
// numbers
while (NotEOF() && CharInString(CurrentChar(), _T("0123456789.abcdefABCDEFXxLl")))
MoveToNextChar();
if (IsEOF())
return wxEmptyString;
m_Str = m_Buffer.Mid(start, m_TokenIndex - start);
m_IsOperator = false;
}
else if (CurrentChar() == '"' ||
CurrentChar() == '\'')
{
// string, char, etc.
wxChar match = CurrentChar();
MoveToNextChar(); // skip starting ' or "
if (!SkipToChar(match))
return wxEmptyString;
MoveToNextChar(); // skip ending ' or "
m_Str = m_Buffer.Mid(start, m_TokenIndex - start);
}
else if (CurrentChar() == ':')
{
if (NextChar() == ':')
{
MoveToNextChar();
MoveToNextChar();
m_Str.assign(TokenizerConsts::colon_colon); // this only copies a pointer, but operator= allocates memory and does a memcpy!
}
else
{
MoveToNextChar();
m_Str.assign(TokenizerConsts::colon);
}
}
else if (CurrentChar() == '<' && bTemplate)
{
wxChar match = _T('>');
MoveToNextChar();
if (!SkipToOneOfChars(_T(">\r\n")),false)
return wxEmptyString;
MoveToNextChar();
wxString tmp = m_Buffer.Mid(start+1,m_TokenIndex-start-2);
tmp.Trim();
m_Str = _T("<");
m_Str += tmp;
m_Str += _T(">");//m_Buffer.Mid(start, m_TokenIndex - start);
}
else if (CurrentChar() == '(')
{
m_IsOperator = false;
// skip blocks () []
if (!SkipBlock(CurrentChar()))
return wxEmptyString;
wxString tmp = m_Buffer.Mid(start, m_TokenIndex - start);
// tmp.Replace(_T("\t"), _T(" ")); // replace tabs with spaces
// tmp.Replace(_T("\n"), _T(" ")); // replace LF with spaces
// tmp.Replace(_T("\r"), _T(" ")); // replace CR with spaces
{ // this is much faster:
size_t i;
while((i = tmp.find_first_of(TokenizerConsts::tabcrlf)) != wxString::npos)
//tmp[i] = _T(' ');
tmp.SetAt(i,_T(' '));
}
// fix-up arguments (remove excessive spaces/tabs/newlines)
for (unsigned int i = 0; i < tmp.Length() - 1; ++i)
{
//skip spaces before '=' and ','
if (tmp.GetChar(i) == ' ' && (tmp.GetChar(i + 1) == ',' || tmp.GetChar(i + 1) == '='))
continue;
if (tmp.GetChar(i) == '/' && tmp.GetChar(i + 1) == '*')
{
// skip C comments
i += 2;
while (i < tmp.Length() - 1)
{
if (tmp.GetChar(i) == '*' && tmp.GetChar(i + 1) == '/')
break;
++i;
}
if (i >= tmp.Length() - 1 || tmp.GetChar(i + 1) != '/')
continue; // we failed...
i += 2;
}
else if (tmp.GetChar(i) == '=')
{
// skip default assignments
++i;
int level = 0; // nesting parenthesis
while (i < tmp.Length())
{
if (tmp.GetChar(i) == '(')
++level;
else if (tmp.GetChar(i) == ')')
--level;
if ((tmp.GetChar(i) == ',' && level == 0) ||
(tmp.GetChar(i) == ')' && level < 0))
break;
++i;
}
if (i < tmp.Length() && tmp.GetChar(i) == ',')
--i;
continue; // we are done here
}
if (i < tmp.Length() - 1)
{
if ((tmp.GetChar(i) == ' ') && (tmp.GetChar(i + 1) == ' '))
continue; // skip excessive spaces
// in case of c-style comments "i" might already be tmp.Length()
// thus do only add the current char otherwise.
// otherwise the following statement:
// m_Str << _T(')');
// below would add another closing bracket.
m_Str << tmp.GetChar(i);
}
}
m_Str << _T(')'); // add closing parenthesis (see "i < tmp.Length() - 1" in previous "for")
// m_Str.Replace(_T(" "), _T(" ")); // replace two-spaces with single-space (introduced if it skipped comments or assignments)
// m_Str.Replace(_T("( "), _T("("));
// m_Str.Replace(_T(" )"), _T(")"));
//Str.Replace is massive overkill here since it has to allocate one new block per replacement
CompactSpaces(m_Str);
}
else
{
if (CurrentChar() == '{')
++m_NestLevel;
else if (CurrentChar() == '}')
--m_NestLevel;
m_Str = CurrentChar();
MoveToNextChar();
}
if (m_LastWasPreprocessor && !m_Str.IsSameAs(_T("#")) && !m_LastPreprocessor.IsSameAs(_T("#")))
{
if (!m_LastPreprocessor.IsSameAs(TokenizerConsts::include_str))
{
// except for #include and #if[[n]def], all other preprocessor directives need only
// one word exactly after the directive, e.g. #define THIS_WORD
SkipToEOL();
}
m_LastPreprocessor.Clear();
}
if (m_LastWasPreprocessor)
m_LastPreprocessor << m_Str;
m_LastWasPreprocessor = false;
return m_Str;
}
