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 nestLevel = 1; // counter for nested blocks (xxx())
while (NotEOF())
{
while (SkipWhiteSpace() || SkipString() || SkipComment())
;
if (CurrentChar() == ch)
++nestLevel;
else if (CurrentChar() == match)
--nestLevel;
MoveToNextChar();
if (nestLevel == 0)
break;
}
return NotEOF();
}
bool Tokenizer::SkipToEOL(bool nestBraces, bool skippingComment)
{
// skip everything until we find EOL
while (1)
{
while (NotEOF() && CurrentChar() != '\n')
{
if (CurrentChar() == '/' && NextChar() == '*')
{
SkipComment(false); // don't skip whitespace after the comment
if (skippingComment && CurrentChar() == '\n')
{
continue; // early exit from the loop
}
}
if (nestBraces && CurrentChar() == _T('{'))
++m_NestLevel;
else if (nestBraces && CurrentChar() == _T('}'))
--m_NestLevel;
MoveToNextChar();
}
wxChar last = PreviousChar();
// if DOS line endings, we 've hit \r and we skip to \n...
if (last == '\r')
last = m_Buffer.GetChar(m_TokenIndex - 2);
if (IsEOF() || last != '\\')
break;
else
MoveToNextChar();
}
if (IsEOF())
return false;
return true;
}
void Tokenizer::SkipToEndConditionPreprocessor()
{
do
{
wxChar ch = CurrentChar();
if (ch == _T('\'') || ch == _T('"') || ch == _T('/') || ch <= _T(' '))
{
while (SkipWhiteSpace() || SkipString() || SkipComment())
;
ch = CurrentChar();
}
if (ch == _T('#'))
{
MoveToNextChar();
while (SkipWhiteSpace() || SkipComment())
;
const wxChar current = CurrentChar();
const wxChar next = NextChar();
// #if
if (current == _T('i') && next == _T('f'))
SkipToEndConditionPreprocessor();
// #endif
else if (current == _T('e') && next == _T('n'))
{
SkipToEOL(false);
break;
}
}
}
while (MoveToNextChar());
}
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;
}
// For example: X"ABCDEFG\"HIJKLMN"Y
// We are now at A, and would skip to Y
// The double quote before H is a "C-escaped-character", We shouldn't quite from that
bool Tokenizer::SkipToStringEnd(const wxChar& ch)
{
while (true)
{
while (CurrentChar() != ch && MoveToNextChar()) // don't check EOF when MoveToNextChar already does
;
if (IsEOF())
return false;
if (IsEscapedChar()) break;
MoveToNextChar();
}
return true;
}
// 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;
}
// expect we are not in a C-string
bool Tokenizer::SkipToChar(const wxChar& ch)
{
// skip everything until we find ch
while (CurrentChar() != ch && MoveToNextChar()) // don't check EOF when MoveToNextChar already does
;
return NotEOF();
}
// return true if we really skip a string, that means m_TokenIndex has changed.
bool Tokenizer::SkipString()
{
if (IsEOF())
return false;
const wxChar ch = CurrentChar();
if (ch == _T('"') || ch == _T('\''))
{
// this is the case that match is inside a string!
MoveToNextChar();
SkipToStringEnd(ch);
MoveToNextChar();
return true;
}
return false;
}
bool TextCutter::SkipWhiteSpace()
{
if (IsEOF())
return false;
while (!IsEOF() && isspace(CurrentChar()))
MoveToNextChar();
return true;
}
// expect we are not in a C-string.
bool Tokenizer::SkipToOneOfChars(const wxChar* chars, bool supportNesting, bool skipPreprocessor, bool skipAngleBrace)
{
while (NotEOF() && !CharInString(CurrentChar(), chars))
{
MoveToNextChar();
while (SkipString() || SkipComment())
;
// 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 '#':
if (skipPreprocessor)
SkipToEOL(true);
else
done = true;
break;
case '{': SkipBlock('{'); break;
case '(': SkipBlock('('); break;
case '[': SkipBlock('['); break;
case '<': // don't skip if << operator
if (skipAngleBrace)
{
if (NextChar() == '<')
MoveToNextChar(2); // skip it and also the next '<' or the next '<' leads to a SkipBlock('<');
else
SkipBlock('<');
break;
}
default: done = true; break;
}
}
}
return NotEOF();
}
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::SkipWhiteSpace()
{
// skip spaces, tabs, etc.
while (CurrentChar() <= _T(' ') && MoveToNextChar()) // don't check EOF when MoveToNextChar already does, also replace isspace() which calls msvcrt.dll
; // with a dirty hack: CurrentChar() <= ' ' is "good enough" here
if (IsEOF())
return false;
return true;
}
void TextCutter::SkipToChar(const wxChar& ch)
{
// skip everything until we find ch
while (1)
{
while (!IsEOF() && CurrentChar() != ch && CurrentChar() != '\n')
MoveToNextChar();
break;
}
}
PreprocessorType Tokenizer::GetPreprocessorType()
{
const unsigned int undoIndex = m_TokenIndex;
const unsigned int undoLine = m_LineNumber;
MoveToNextChar();
while (SkipWhiteSpace() || SkipComment())
;
const wxString token = DoGetToken();
switch (token.Len())
{
case 2:
if (token == TokenizerConsts::kw_if)
return ptIf;
break;
case 4:
if (token == TokenizerConsts::kw_else)
return ptElse;
else if (token == TokenizerConsts::kw_elif)
return ptElif;
break;
case 5:
if (token == TokenizerConsts::kw_ifdef)
return ptIfdef;
else if (token == TokenizerConsts::kw_endif)
return ptEndif;
break;
case 6:
if (token == TokenizerConsts::kw_ifndef)
return ptIfndef;
break;
case 7:
if (token == TokenizerConsts::kw_elifdef)
return ptElifdef;
break;
case 8:
if (token == TokenizerConsts::kw_elifndef)
return ptElifndef;
break;
}
m_TokenIndex = undoIndex;
m_LineNumber = undoLine;
return ptOthers;
}
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
}
uLONG VLexerFileInput::FindCommandStart()
{
uLONG nSkipped = 0;
UniChar nChar;
while (HasMoreChars()) {
nSkipped++;
nChar = MoveToNextChar();
if (nChar == CHAR_SEMICOLON) break;
}
fCommandStart = fCurrentChar;
return nSkipped;
}
void Tokenizer::SkipToNextConditionPreprocessor()
{
do
{
wxChar ch = CurrentChar();
if (ch == _T('\'') || ch == _T('"') || ch == _T('/') || ch <= _T(' '))
{
while (SkipWhiteSpace() || SkipString() || SkipComment())
;
ch = CurrentChar();
}
if (ch == _T('#'))
{
const unsigned int undoIndex = m_TokenIndex;
const unsigned int undoLine = m_LineNumber;
MoveToNextChar();
while (SkipWhiteSpace() || SkipComment())
;
const wxChar current = CurrentChar();
const wxChar next = NextChar();
// #if
if (current == _T('i') && next == _T('f'))
SkipToEndConditionPreprocessor();
// #else #elif #elifdef #elifndef #endif
else if (current == _T('e') && (next == _T('l') || next == _T('n')))
{
m_TokenIndex = undoIndex;
m_LineNumber = undoLine;
break;
}
}
}
while (MoveToNextChar());
}
bool Tokenizer::SkipToChar(const wxChar& ch)
{
// skip everything until we find ch
while(true)
{
while (CurrentChar() != ch && MoveToNextChar()) // don't check EOF when MoveToNextChar already does
;
if (IsEOF())
return false;
if (PreviousChar() != '\\')
break;
else
{
// check for "\\"
if (m_TokenIndex - 2 >= 0 && m_Buffer.GetChar(m_TokenIndex - 2) == '\\')
break;
}
MoveToNextChar();
}
return true;
}
bool Tokenizer::SkipToEOL(bool nestBraces)
{
TRACE(_T("%s : line=%d, CurrentChar='%c', PreviousChar='%c', NextChar='%c', nestBrace(%d)"),
wxString(__PRETTY_FUNCTION__, wxConvUTF8).wc_str(), m_LineNumber, CurrentChar(),
PreviousChar(), NextChar(), nestBraces ? 1 : 0);
// skip everything until we find EOL
for (;;)
{
while (NotEOF() && CurrentChar() != '\n')
{
if (CurrentChar() == '/' && NextChar() == '*')
{
SkipComment();
if (CurrentChar() == _T('\n'))
break;
}
if (nestBraces && CurrentChar() == _T('{'))
++m_NestLevel;
else if (nestBraces && CurrentChar() == _T('}'))
--m_NestLevel;
MoveToNextChar();
}
if (!IsBackslashBeforeEOL() || IsEOF())
break;
else
MoveToNextChar();
}
TRACE(_T("SkipToEOL(): (END) We are now at line %d, CurrentChar='%c', PreviousChar='%c', NextChar='%c'"),
m_LineNumber, CurrentChar(), PreviousChar(), NextChar());
return NotEOF();
}
void Tokenizer::SpliteArguments(wxArrayString& results)
{
while (SkipWhiteSpace() || SkipComment())
;
if (CurrentChar() != _T('('))
return;
MoveToNextChar(); // Skip the '('
int level = 1; // include '('
wxString piece;
while (NotEOF())
{
wxString token = DoGetToken();
if (token.IsEmpty())
break;
if (token == _T("("))
++level;
else if (token == _T(")"))
--level;
if (token == _T(","))
{
results.Add(piece);
piece.Clear();
}
else if (level != 0)
{
if (!piece.IsEmpty() && piece.Last() > _T(' '))
piece << _T(" ");
piece << token;
}
if (level == 0)
{
if (!piece.IsEmpty())
results.Add(piece);
break;
}
while (SkipWhiteSpace() || SkipComment())
;
}
}
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();
}
void TextCutter::SkipToEOL()
{
while (!IsEOF() && CurrentChar() != '\n')
MoveToNextChar();
}
//vfc add bGetValue
bool Tokenizer::SkipUnwanted(bool bGetValue)
{
while (CurrentChar() == '#' ||
(!m_IsOperator && CurrentChar() == '=') ||
(!m_IsOperator && CurrentChar() == '[') ||
CurrentChar() == '?' ||
(CurrentChar() == '/' && (NextChar() == '/' || NextChar() == '*') ))
{
bool skipPreprocessor = false; // used for #include
while (m_Buffer.Mid(m_TokenIndex, 2) == _T("//") ||
m_Buffer.Mid(m_TokenIndex, 2) == _T("/*"))
{
// C/C++ style comments
SkipComment();
if (IsEOF())
return false;
if (!SkipWhiteSpace())
return false;
}
while (CurrentChar() == '#')
{
// preprocessor directives
// we only care for #include and #define, for now
unsigned int backupIdx = m_TokenIndex;
MoveToNextChar();
SkipWhiteSpace();
if ((CurrentChar() == 'i' && NextChar() == 'n') || // in(clude)
(CurrentChar() == 'i' && NextChar() == 'f') || // if(|def|ndef)
(CurrentChar() == 'e' && NextChar() == 'l') || // el(se|if)
(CurrentChar() == 'e' && NextChar() == 'n') || // en(dif)
(m_Options.wantPreprocessor && CurrentChar() == 'd' && NextChar() == 'e')) // de(fine)
{
// ok, we have something like #in(clude)
m_LastWasPreprocessor = true;
m_LastPreprocessor.Clear();
m_TokenIndex = backupIdx; // keep #
skipPreprocessor = true;
break;
}
else
{
// skip the rest for now...
SkipToEOL(false);
if (!SkipWhiteSpace())
return false;
}
if (skipPreprocessor)
break;
}
while (CurrentChar() == '[')
{
// array subscripts
// skip them for now...
SkipBlock('[');
if (!SkipWhiteSpace())
return false;
}
while (CurrentChar() == '=')
{
// skip assignments
// TODO: what happens with operators?
if (bGetValue == true)
{
MoveToNextChar();
SkipWhiteSpace();
return true;
}
else if (!SkipToOneOfChars(_T(",;}"), true))
return false;
}
while (CurrentChar() == '?')
{
// skip "condition ? true : false"
// TODO: what happens with operators?
if (!SkipToOneOfChars(_T(";}")))
return false;
}
if (skipPreprocessor)
break;
}
return true;
}
//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;
}
void Tokenizer::ReadParentheses(wxString& str)
{
static const size_t maxBufferLen = 4093;
wxChar buffer[maxBufferLen + 3];
buffer[0] = _T('$'); // avoid segfault error
wxChar* realBuffer = buffer + 1;
wxChar* p = realBuffer;
int level = 0;
while (NotEOF())
{
while (SkipComment())
;
wxChar ch = CurrentChar();
while (ch == _T('#')) // do not use if
{
const PreprocessorType type = GetPreprocessorType();
if (type == ptOthers)
break;
HandleConditionPreprocessor(type);
ch = CurrentChar();
}
const unsigned int startIndex = m_TokenIndex;
switch(ch)
{
case _T('('):
{
++level;
*p = ch;
++p;
}
break;
case _T(')'):
{
if (*(p - 1) <= _T(' '))
--p;
--level;
*p = ch;
++p;
}
break;
case _T('\''):
case _T('"'):
{
MoveToNextChar();
SkipToStringEnd(ch);
MoveToNextChar();
const size_t writeLen = m_TokenIndex - startIndex;
const size_t usedLen = p - realBuffer;
if (usedLen + writeLen > maxBufferLen)
{
if (writeLen > maxBufferLen)
{
TRACE(_T("ReadParentheses, Catch Exception 1: %d"), writeLen);
return;
}
if (p != realBuffer)
{
str.Append(realBuffer, usedLen);
p = realBuffer;
}
str.Append((const wxChar*)m_Buffer + startIndex, writeLen);
}
else
{
memcpy(p, (const wxChar*)m_Buffer + startIndex, writeLen * sizeof(wxChar));
p += writeLen;
}
continue;
}
break;
case _T(','):
{
if (*(p - 1) <= _T(' '))
--p;
*p = _T(',');
*++p = _T(' ');
++p;
}
break;
case _T('*'):
{
if (*(p - 1) <= _T(' '))
--p;
*p = _T('*');
*++p = _T(' ');
++p;
}
break;
case _T('&'):
{
if (*(p - 1) <= _T(' '))
--p;
*p = _T('&');
*++p = _T(' ');
++p;
}
break;
case _T('='):
{
if (*(p - 1) <= _T(' '))
{
*p = _T('=');
*++p = _T(' ');
++p;
}
else
{
switch (*(p - 1))
{
case _T('='):
case _T('!'):
case _T('>'):
case _T('<'):
{
*p = _T('=');
*++p = _T(' ');
++p;
}
default:
{
*p = _T(' ');
*++p = _T('=');
*++p = _T(' ');
++p;
}
}
}
}
break;
case _T(' '):
{
if (*(p - 1) != _T(' ') && *(p - 1) != _T('('))
{
*p = _T(' ');
++p;
}
}
break;
case _T('\r'):
case _T('\t'):
break;
case _T('\n'): // we need keep the \n for records paras correct position
if (*(p - 1) == _T(' '))
--p;
if (*(p - 1) != _T('('))
{
*p = ch;
++p;
}
break;
default:
{
*p = ch;
++p;
}
break;
}
if (p >= realBuffer + maxBufferLen)
{
str.Append(realBuffer, p - realBuffer);
p = realBuffer;
}
MoveToNextChar();
if (level == 0)
break;
}
if (p > realBuffer)
str.Append(realBuffer, p - realBuffer);
TRACE(_T("ReadParentheses(): %s, line=%d"), str.wx_str(), m_LineNumber);
if (str.Len() > 512)
TRACE(_T("ReadParentheses: Catch Exception 2?: %d"), str.Len());
}
wxString Tokenizer::ReadToEOL(bool nestBraces, bool stripUnneeded)
{
if (stripUnneeded)
{
TRACE(_T("%s : line=%d, CurrentChar='%c', PreviousChar='%c', NextChar='%c', nestBrace(%d)"),
wxString(__PRETTY_FUNCTION__, wxConvUTF8).wc_str(), m_LineNumber, CurrentChar(),
PreviousChar(), NextChar(), nestBraces ? 1 : 0);
static const size_t maxBufferLen = 4094;
wxChar buffer[maxBufferLen + 2];
wxChar* p = buffer;
wxString str;
for (;;)
{
while (NotEOF() && CurrentChar() != _T('\n'))
{
while (SkipComment())
;
const wxChar ch = CurrentChar();
if (ch == _T('\n'))
break;
if (ch <= _T(' ') && (p == buffer || *(p - 1) == ch))
{
MoveToNextChar();
continue;
}
*p = ch;
++p;
if (p >= buffer + maxBufferLen)
{
str.Append(buffer, p - buffer);
p = buffer;
}
if (nestBraces)
{
if (ch == _T('{'))
++m_NestLevel;
else if (ch == _T('}'))
--m_NestLevel;
}
MoveToNextChar();
}
if (!IsBackslashBeforeEOL() || IsEOF())
break;
else
{
while (p > buffer && *(--p) <= _T(' '))
;
MoveToNextChar();
}
}
while (p > buffer && *(p - 1) <= _T(' '))
--p;
if (p > buffer)
str.Append(buffer, p - buffer);
TRACE(_T("ReadToEOL(): (END) We are now at line %d, CurrentChar='%c', PreviousChar='%c', NextChar='%c'"),
m_LineNumber, CurrentChar(), PreviousChar(), NextChar());
TRACE(_T("ReadToEOL(): %s"), str.wx_str());
return str;
}
else
{
const unsigned int idx = m_TokenIndex;
SkipToEOL(nestBraces);
return m_Buffer.Mid(idx, m_TokenIndex - idx);
}
}
wxString Tokenizer::DoGetToken()
{
int start = m_TokenIndex;
bool needReplace = false;
wxString str;
wxChar c = CurrentChar();
if (c == '_' || wxIsalpha(c))
{
// keywords, identifiers, etc.
// operator== is cheaper than wxIsalnum, also MoveToNextChar already includes IsEOF
while ( ( (c == '_') || (wxIsalnum(c)) )
&& MoveToNextChar() )
c = CurrentChar(); // repeat
if (IsEOF())
return wxEmptyString;
needReplace = true;
str = m_Buffer.Mid(start, m_TokenIndex - start);
}
#ifdef __WXMSW__ // This is a Windows only bug!
// fetch non-English characters, see more details in: http://forums.codeblocks.org/index.php/topic,11387.0.html
else if (c == 178 || c == 179 || c == 185)
{
str = c;
MoveToNextChar();
}
#endif
else if (wxIsdigit(c))
{
// numbers
while (NotEOF() && CharInString(CurrentChar(), _T("0123456789.abcdefABCDEFXxLl")))
MoveToNextChar();
if (IsEOF())
return wxEmptyString;
str = m_Buffer.Mid(start, m_TokenIndex - start);
}
else if ( (c == '"') || (c == '\'') )
{
SkipString();
//Now, we are after the end of the C-string, so return the whole string as a token.
str = m_Buffer.Mid(start, m_TokenIndex - start);
}
else if (c == ':')
{
if (NextChar() == ':')
{
MoveToNextChar();
MoveToNextChar();
// this only copies a pointer, but operator= allocates memory and does a memcpy!
str.assign(TokenizerConsts::colon_colon);
}
else
{
MoveToNextChar();
str.assign(TokenizerConsts::colon);
}
}
else if (c == '<')
{
if (m_State&tsSingleAngleBrace)
{
if ( !SkipToOneOfChars( _T(">"), true, true) )
return wxEmptyString;
MoveToNextChar();
str= m_Buffer.Mid(start, m_TokenIndex - start);
}
else
{
str = c;
MoveToNextChar();
}
}
else if (c == '(')
{
if (m_State & tsReadRawExpression)
{
str = c;
MoveToNextChar();
}
else
{
ReadParentheses(str);
}
}
else
{
if (c == '{')
++m_NestLevel;
else if (c == '}')
--m_NestLevel;
str = c;
MoveToNextChar();
}
if (m_FirstRemainingLength != 0 && m_BufferLen - m_FirstRemainingLength < m_TokenIndex)
{
m_FirstRemainingLength = 0;
m_IsReplaceParsing = false;
m_RepeatReplaceCount = 0;
}
if (needReplace && m_State ^ tsReadRawExpression)
MacroReplace(str);
return str;
}
void Tokenizer::MacroReplace(wxString& str)
{
if (m_IsReplaceParsing)
{
const int id = m_TokensTree->TokenExists(str, -1, tkPreprocessor);
if (id != -1)
{
Token* tk = m_TokensTree->at(id);
if (tk)
{
bool replaced = false;
if (!tk->m_Args.IsEmpty())
replaced = ReplaceMacroActualContext(tk, false);
else if (tk->m_Type != tk->m_Name)
replaced = ReplaceBufferForReparse(tk->m_Type, false);
if (replaced || tk->m_Type.IsEmpty())
{
SkipUnwanted();
str = DoGetToken();
}
}
}
}
wxStringHashMap::const_iterator it = s_Replacements.find(str);
if (it == s_Replacements.end())
return;
TRACE(_T("MacroReplace() : Replacing '%s' with '%s' (file='%s', line='%d')."), it->first.wx_str(),
it->second.wx_str(), m_Filename.wx_str(), m_LineNumber);
if (it->second.IsEmpty())
{
SkipUnwanted();
str = DoGetToken();
}
else if (it->second[0] == _T('+'))
{
while (SkipWhiteSpace() || SkipComment())
;
DoGetToken(); // eat (...)
wxString target = (const wxChar*)it->second + 1;
if (target.IsEmpty())
{
while (SkipWhiteSpace() || SkipComment())
;
str = DoGetToken();
}
else if (target != str && ReplaceBufferForReparse(target, false))
str = DoGetToken();
}
else if (it->second[0] == _T('-'))
{
wxString end((const wxChar*)it->second + 1);
if (end.IsEmpty())
return;
while (NotEOF())
{
while (SkipComment() && SkipWhiteSpace())
;
if (CurrentChar() == end[0])
{
if (DoGetToken() == end)
break;
}
else
MoveToNextChar();
}
// eat ()
SkipUnwanted();
str = DoGetToken();
if (str[0] == _T('('))
{
SkipUnwanted();
str = DoGetToken();
}
}
else
{
if (it->second != str && ReplaceBufferForReparse(it->second, false))
str = DoGetToken();
}
}