bool Compiler::compile(Error& error)
{
error.clear();
error.step = Error::COMPILATION;
error.state = Error::ERROR;
m_tokenList = MUON_NEW(std::vector<parser::Token>);
m_nodeRoot = MUON_NEW(parser::ASTNode);
// avoid a macro, and avoid duplicating code
auto clearVariable = [&]()
{
m_loadBuffer.clear();
MUON_DELETE(m_tokenList);
MUON_DELETE(m_nodeRoot);
};
if (!lexical(error))
{
clearVariable();
return false;
}
if (!syntaxic(error))
{
clearVariable();
return false;
}
if (!semantic(error))
{
clearVariable();
return false;
}
clearVariable();
error.state = Error::SUCCESS;
return true;
}
int main(int argc, char* argv[])
{
#ifdef _MSC_VER
_CrtSetDbgFlag( _CRTDBG_ALLOC_MEM_DF | _CRTDBG_LEAK_CHECK_DF );
#endif
if (argc < 3) {
fprintf (stdout, "usage: dLexCompiler [inputRulesFileName] [outputFileName]\n");
fprintf (stdout, "[inputRulesFileName] name of the file containing a Lex or Flex rules\n");
fprintf (stdout, "[outputFileName] name of the file cpp output file\n");
exit (0);
}
const char* const inputRulesFileName = argv[1];
const char* const outputFileName = argv[2];
FILE* const rules = fopen (inputRulesFileName, "rb");
if (!rules) {
fprintf (stdout, "Rule file \"%s\" not found\n", inputRulesFileName);
exit (0);
}
fseek (rules, 0, SEEK_END);
int size = ftell (rules);
fseek (rules, 0, SEEK_SET);
string buffer;
buffer.resize (size + 1);
fread ((void*) buffer.c_str(), 1, size, rules);
fclose (rules);
buffer.erase(strlen (buffer.c_str()));
dLexCompiler lexical (buffer.c_str(), outputFileName, inputRulesFileName);
return 0;
}
void dParserCompiler::ScanGrammarFile(
const dString& inputRules,
dProductionRule& ruleList,
dTree<dTokenInfo, dCRCTYPE>& symbolList,
dOperatorsPrecedence& operatorPrecedence,
dString& userCodeBlock,
dString& userVariableClass,
dString& endUserCode,
int& lastTokenEnum)
{
dString startSymbol ("");
int tokenEnumeration = 256;
int operatorPrecedencePriority = 0;
dParserLexical lexical (inputRules.GetStr());
LoadTemplateFile("dParserUserVariableTemplate_cpp.txt", userVariableClass);
// scan the definition segment
for (dToken token = dToken(lexical.NextToken()); token != GRAMMAR_SEGMENT; ) {
switch (int (token))
{
case START:
{
token = dToken(lexical.NextToken());
startSymbol = lexical.GetTokenString();
token = dToken(lexical.NextToken());
break;
}
case TOKEN:
{
for (token = dToken(lexical.NextToken()); token == LITERAL; token = dToken(lexical.NextToken())) {
const char* const name = lexical.GetTokenString();
symbolList.Insert(dTokenInfo (tokenEnumeration, TERMINAL, name), dCRC64 (name));
tokenEnumeration ++;
}
break;
}
case LEFT:
case RIGHT:
{
dOperatorsAssociation& association = operatorPrecedence.Append()->GetInfo();
association.m_prioprity = operatorPrecedencePriority;
operatorPrecedencePriority ++;
switch (int (token))
{
case LEFT:
association.m_associativity = dOperatorsAssociation::m_left;
break;
case RIGHT:
association.m_associativity = dOperatorsAssociation::m_right;
break;
}
for (token = dToken(lexical.NextToken()); (token == LITERAL) || ((token < 256) && !isalnum (token)); token = dToken(lexical.NextToken())) {
association.Append(dCRC64 (lexical.GetTokenString()));
}
break;
}
case UNION:
{
token = dToken(lexical.NextToken());
dAssert (token == SEMANTIC_ACTION);
userVariableClass = lexical.GetTokenString() + 1;
userVariableClass.Replace(userVariableClass.Size() - 1, 1, "");
token = dToken(lexical.NextToken());
break;
}
case CODE_BLOCK:
{
userCodeBlock += lexical.GetTokenString();
token = dToken(lexical.NextToken());
break;
}
case EXPECT:
{
token = dToken(lexical.NextToken());
dAssert (token == INTEGER);
m_shiftReduceExpectedWarnings = atoi (lexical.GetTokenString());
token = dToken(lexical.NextToken());
break;
}
default:;
{
dAssert (0);
token = dToken(lexical.NextToken());
}
}
}
int ruleNumber = 1;
lastTokenEnum = tokenEnumeration;
// scan the production rules segment
dToken token1 = dToken(lexical.NextToken());
for (; (token1 != GRAMMAR_SEGMENT) && (token1 != -1); token1 = dToken(lexical.NextToken())) {
//dTrace (("%s\n", lexical.GetTokenString()));
switch (int (token1))
{
case LITERAL:
{
// add the first Rule;
dRuleInfo& rule = ruleList.Append()->GetInfo();
rule.m_token = token1;
rule.m_type = NONTERMINAL;
rule.m_name = lexical.GetTokenString();
rule.m_nameCRC = dCRC64 (lexical.GetTokenString());
dTree<dTokenInfo, dCRCTYPE>::dTreeNode* nonTerminalIdNode = symbolList.Find(rule.m_nameCRC);
if (!nonTerminalIdNode) {
nonTerminalIdNode = symbolList.Insert(dTokenInfo (tokenEnumeration, NONTERMINAL, rule.m_name), rule.m_nameCRC);
tokenEnumeration ++;
}
rule.m_ruleId = nonTerminalIdNode->GetInfo().m_tokenId;
rule.m_ruleNumber = ruleNumber;
ruleNumber ++;
token1 = ScanGrammarRule(lexical, ruleList, symbolList, ruleNumber, tokenEnumeration, operatorPrecedence);
break;
}
default:
dAssert (0);
}
}
dProductionRule::dListNode* firtRuleNode = ruleList.GetFirst();
if (startSymbol != "") {
firtRuleNode = ruleList.Find (dCRC64 (startSymbol.GetStr()));
}
//Expand the Grammar Rule by adding an empty start Rule;
const dRuleInfo& firstRule = firtRuleNode->GetInfo();
dRuleInfo& rule = ruleList.Addtop()->GetInfo();
rule.m_ruleNumber = 0;
rule.m_ruleId = tokenEnumeration;
rule.m_token = firstRule.m_token;
rule.m_type = NONTERMINAL;
rule.m_name = firstRule.m_name + dString("__");
rule.m_nameCRC = dCRC64 (rule.m_name.GetStr());
symbolList.Insert(dTokenInfo (tokenEnumeration, rule.m_type, rule.m_name), rule.m_nameCRC);
tokenEnumeration ++;
dSymbol& symbol = rule.Append()->GetInfo();
symbol.m_token = firstRule.m_token;
symbol.m_type = firstRule.m_type;
symbol.m_name = firstRule.m_name;
symbol.m_nameCRC = firstRule.m_nameCRC;
// scan literal use code
if (token1 == GRAMMAR_SEGMENT) {
endUserCode = lexical.GetNextBuffer();
//endUserCode += "\n";
}
}
