bool Parser::parseExpressionList(std::vector<Expression>& list, int min, int max)
{
bool valid = true;
list.clear();
list.reserve(max >= 0 ? max : 32);
const Token& start = peekToken();
Expression exp = parseExpression();
list.push_back(exp);
if (exp.isLoaded() == false)
{
printError(start,L"Parameter failure");
getTokenizer()->skipLookahead();
valid = false;
}
while (peekToken().type == TokenType::Comma)
{
eatToken();
exp = parseExpression();
list.push_back(exp);
if (exp.isLoaded() == false)
{
printError(start,L"Parameter failure");
getTokenizer()->skipLookahead();
valid = false;
}
}
if (list.size() < (size_t) min)
{
printError(start,L"Not enough parameters (min %d)",min);
return false;
}
if (max != -1 && (size_t) max < list.size())
{
printError(start,L"Too many parameters (max %d)",max);
return false;
}
return valid;
}
CAssemblerCommand* Parser::parseMacroCall()
{
const Token& start = peekToken();
if (start.type != TokenType::Identifier)
return nullptr;
auto it = macros.find(start.getStringValue());
if (it == macros.end())
return nullptr;
ParserMacro& macro = it->second;
eatToken();
// create a token stream for the macro content,
// registering replacements for parameter values
TokenStreamTokenizer macroTokenizer;
std::set<std::wstring> identifierParameters;
for (size_t i = 0; i < macro.parameters.size(); i++)
{
if (i != 0)
{
if (nextToken().type != TokenType::Comma)
return nullptr;
}
if (i == macro.parameters.size())
{
size_t count = macro.parameters.size();
while (peekToken().type == TokenType::Comma)
{
eatToken();
parseExpression();
}
printError(start,L"Not enough macro arguments (%d vs %d)",count,macro.parameters.size());
return nullptr;
}
TokenizerPosition startPos = getTokenizer()->getPosition();
Expression exp = parseExpression();
if (exp.isLoaded() == false)
return nullptr;
TokenizerPosition endPos = getTokenizer()->getPosition();
std::vector<Token> tokens = getTokenizer()->getTokens(startPos,endPos);
// remember any single identifier parameters for the label replacement
if (tokens.size() == 1 && tokens[0].type == TokenType::Identifier)
identifierParameters.insert(tokens[0].getStringValue());
// give them as a replacement to new tokenizer
macroTokenizer.registerReplacement(macro.parameters[i],tokens);
}
if (peekToken().type == TokenType::Comma)
{
size_t count = macro.parameters.size();
while (peekToken().type == TokenType::Comma)
{
eatToken();
parseExpression();
count++;
}
printError(start,L"Too many macro arguments (%d vs %d)",count,macro.parameters.size());
return nullptr;
}
// a macro is fully parsed once when it's loaded
// to gather all labels. it's not necessary to
// instantiate other macros at that time
if (initializingMacro)
return new DummyCommand();
// the first time a macro is instantiated, it needs to be analyzed
// for labels
if (macro.counter == 0)
{
initializingMacro = true;
// parse the short lived next command
macroTokenizer.init(macro.content);
CAssemblerCommand* command = parse(¯oTokenizer,true);
delete command;
macro.labels = macroLabels;
macroLabels.clear();
initializingMacro = false;
}
// register labels and replacements
for (const std::wstring& label: macro.labels)
{
// check if the label is using the name of a parameter
// in that case, don't register a unique replacement
if (identifierParameters.find(label) != identifierParameters.end())
continue;
// otherwise make sure the name is unique
std::wstring fullName;
if (Global.symbolTable.isLocalSymbol(label))
fullName = formatString(L"@@%s_%s_%08X",macro.name,label.substr(2),macro.counter);
else if (Global.symbolTable.isStaticSymbol(label))
fullName = formatString(L"@%s_%s_%08X",macro.name,label.substr(1),macro.counter);
else
fullName = formatString(L"%s_%s_%08X",macro.name,label,macro.counter);
macroTokenizer.registerReplacement(label,fullName);
}
macroTokenizer.init(macro.content);
macro.counter++;
return parse(¯oTokenizer,true);
}
bool Parser::checkMacroDefinition()
{
const Token& first = peekToken();
if (first.type != TokenType::Identifier)
return false;
if (!first.stringValueStartsWith(L'.') || first.getStringValue() != L".macro")
return false;
eatToken();
// nested macro definitions are not allowed
if (initializingMacro)
{
printError(first,L"Nested macro definitions not allowed");
while (!atEnd())
{
const Token& token = nextToken();
if (token.type == TokenType::Identifier && token.getStringValue() == L".endmacro")
break;
}
return true;
}
std::vector<Expression> parameters;
if (parseExpressionList(parameters,1,-1) == false)
return false;
// load name
std::wstring macroName;
if (parameters[0].evaluateIdentifier(macroName) == false)
return false;
// duplicate check the macro
ParserMacro ¯o = macros[macroName];
if (macro.name.length() != 0)
{
printError(first,L"Macro \"%s\" already defined",macro.name);
return false;
}
// and register it
macro.name = macroName;
macro.counter = 0;
// load parameters
for (size_t i = 1; i < parameters.size(); i++)
{
std::wstring name;
if (parameters[i].evaluateIdentifier(name) == false)
return false;
macro.parameters.push_back(name);
}
// load macro content
TokenizerPosition start = getTokenizer()->getPosition();
bool valid = false;
while (atEnd() == false)
{
const Token& tok = nextToken();
if (tok.type == TokenType::Identifier && tok.getStringValue() == L".endmacro")
{
valid = true;
break;
}
}
// no .endmacro, not valid
if (valid == false)
return true;
// get content
TokenizerPosition end = getTokenizer()->getPosition().previous();
macro.content = getTokenizer()->getTokens(start,end);
return true;
}
Expression Parser::parseExpression()
{
return ::parseExpression(*getTokenizer());
}
bool IWORKParser::parse()
{
const shared_ptr<xmlTextReader> sharedReader(xmlReaderForIO(readFromStream, closeStream, m_input.get(), "", nullptr, 0), xmlFreeTextReader);
if (!sharedReader)
return false;
xmlTextReaderPtr reader = sharedReader.get();
assert(reader);
const IWORKTokenizer &tokenizer = getTokenizer();
stack<IWORKXMLContextPtr_t> contextStack;
int ret = xmlTextReaderRead(reader);
contextStack.push(createDocumentContext());
bool keynoteDocTypeChecked=false;
char const *defaultNS=nullptr;
while ((1 == ret))
{
switch (xmlTextReaderNodeType(reader))
{
case XML_READER_TYPE_ELEMENT:
{
if (!keynoteDocTypeChecked)
{
// check for keynote 1 file with doctype node and not a namespace in first node
keynoteDocTypeChecked=true;
if (xmlTextReaderNodeType(reader)==XML_READER_TYPE_ELEMENT &&
xmlTextReaderConstNamespaceUri(reader)==nullptr)
defaultNS="http://developer.apple.com/schemas/APXL";
}
const int id = tokenizer.getQualifiedId(char_cast(xmlTextReaderConstLocalName(reader)),
defaultNS ? defaultNS : char_cast(xmlTextReaderConstNamespaceUri(reader)));
IWORKXMLContextPtr_t newContext = contextStack.top()->element(id);
if (!newContext)
newContext = createDiscardContext();
const bool isEmpty = xmlTextReaderIsEmptyElement(reader);
newContext->startOfElement();
if (xmlTextReaderHasAttributes(reader))
{
ret = xmlTextReaderMoveToFirstAttribute(reader);
while (1 == ret)
{
processAttribute(reader, newContext, tokenizer);
ret = xmlTextReaderMoveToNextAttribute(reader);
}
}
if (isEmpty)
newContext->endOfElement();
else
contextStack.push(newContext);
break;
}
case XML_READER_TYPE_END_ELEMENT:
{
contextStack.top()->endOfElement();
contextStack.pop();
break;
}
case XML_READER_TYPE_CDATA :
{
const xmlChar *text = xmlTextReaderConstValue(reader);
if (text) contextStack.top()->CDATA(char_cast(text));
break;
}
case XML_READER_TYPE_TEXT :
{
xmlChar *const text = xmlTextReaderReadString(reader);
contextStack.top()->text(char_cast(text));
xmlFree(text);
break;
}
default:
break;
}
ret = xmlTextReaderRead(reader);
}
while (!contextStack.empty()) // finish parsing in case of broken XML
{
contextStack.top()->endOfElement();
contextStack.pop();
}
xmlTextReaderClose(reader);
return true;
}
bool IWORKParser::parse()
{
const shared_ptr<xmlTextReader> sharedReader(xmlReaderForIO(readFromStream, closeStream, m_input.get(), "", 0, 0), xmlFreeTextReader);
if (!sharedReader)
return false;
xmlTextReaderPtr reader = sharedReader.get();
assert(reader);
const IWORKTokenizer &tokenizer = getTokenizer();
stack<IWORKXMLContextPtr_t> contextStack;
int ret = xmlTextReaderRead(reader);
contextStack.push(createDocumentContext());
while ((1 == ret))
{
switch (xmlTextReaderNodeType(reader))
{
case XML_READER_TYPE_ELEMENT:
{
const int id = tokenizer.getQualifiedId(char_cast(xmlTextReaderConstLocalName(reader)), char_cast(xmlTextReaderConstNamespaceUri(reader)));
IWORKXMLContextPtr_t newContext = contextStack.top()->element(id);
if (!newContext)
newContext = createDiscardContext();
const bool isEmpty = xmlTextReaderIsEmptyElement(reader);
newContext->startOfElement();
if (xmlTextReaderHasAttributes(reader))
{
ret = xmlTextReaderMoveToFirstAttribute(reader);
while (1 == ret)
{
processAttribute(reader, newContext, tokenizer);
ret = xmlTextReaderMoveToNextAttribute(reader);
}
}
if (isEmpty)
newContext->endOfElement();
else
contextStack.push(newContext);
break;
}
case XML_READER_TYPE_END_ELEMENT:
{
contextStack.top()->endOfElement();
contextStack.pop();
break;
}
case XML_READER_TYPE_TEXT :
{
xmlChar *const text = xmlTextReaderReadString(reader);
contextStack.top()->text(char_cast(text));
xmlFree(text);
break;
}
default:
break;
}
ret = xmlTextReaderRead(reader);
}
while (!contextStack.empty()) // finish parsing in case of broken XML
{
contextStack.top()->endOfElement();
contextStack.pop();
}
xmlTextReaderClose(reader);
return true;
}