bool CFParser :: applyRule(size_t ruleId, TokenInfo& token, _ScriptReader& reader)
{
ident_t name = retrieveKey(_names.start(), ruleId, DEFAULT_STR);
size_t readerRollback = reader.Position();
size_t coderRollback = token.LogPosition();
RuleMap::Iterator it = _rules.getIt(ruleId);
while(!it.Eof()) {
CFParser::TokenInfo tokenCopy(token);
if ((*it).apply(*it, tokenCopy, reader)) {
tokenCopy.save(token);
return true;
}
// roll back
reader.seek(readerRollback);
token.trimLog(coderRollback);
it = _rules.getNextIt(ruleId, it);
}
return false;
}
void TreeScriptParser :: parseScope(_ScriptReader& reader, ScriptBookmark& bm, SyntaxWriter& writer)
{
bm = reader.read();
while (!reader.Eof() && !reader.compare(")")) {
parseStatement(reader, bm, writer);
bm = reader.read();
}
}
bool VMTapeParser :: parse(TapeWriter& writer, _ScriptReader& reader)
{
bool idle = true;
ScriptBookmark bm = reader.read();
while (!reader.Eof()) {
idle = false;
parseStatement(reader, bm, writer);
bm = reader.read();
}
return !idle;
}
bool CFParser :: parseGrammarMode(_ScriptReader& reader)
{
ScriptBookmark bm = reader.read();
while (!reader.compare(";")) {
if (reader.compare("symbolic")) {
_symbolMode = true;
}
// else throw EParseError(reader.info.column, reader.info.row);
bm = reader.read();
}
return true;
}
int CFParser :: buildDerivationTree(_ScriptReader& reader, size_t startRuleId, MemoryWriter& writer)
{
DerivationQueue predictions;
predictions.push(DerivationItem(startRuleId, 0, -1));
ScriptBookmark bm;
while (predictions.Count() > 0) {
predictions.push(DerivationItem(0));
bm = reader.read();
int terminalOffset = writer.Position();
writer.write(&bm, sizeof(ScriptBookmark));
DerivationItem current = predictions.pop();
while (current.ruleId != 0) {
if (current.ruleId == -1) {
return current.trace;
}
predict(predictions, current, reader, bm, terminalOffset, writer);
current = predictions.pop();
}
}
throw EParseError(bm.column, bm.row);
}
void CFParser :: parse(_ScriptReader& reader, TapeWriter& writer)
{
// if (_symbolMode)
// reader.switchDFA(dfaSymbolic);
ScriptLog log;
TokenInfo token(this, &log);
token.read(reader);
size_t readerRollback = reader.Position();
//size_t outputRollback = compiler->Position();
size_t startId = mapRuleId("start");
if (token.state != dfaEOF) {
if (!applyRule(startId, token, reader))
throw EParseError(token.column, token.row);
if (token.state != dfaEOF)
throw EParseError(token.column, token.row);
}
IdentifierTextReader logReader((ident_t)log.getBody());
CachedScriptReader scriptReader(&logReader);
_baseParser->parse(scriptReader, writer);
}
void saveReference(_ScriptReader& scriptReader, CFParser* parser, ref_t ptr, ScriptLog& log)
{
ScriptBookmark bm;
parser->readScriptBookmark(ptr, bm);
log.write(scriptReader.lookup(bm));
}
bool CFParser :: compareToken(_ScriptReader& reader, ScriptBookmark& bm, int rule)
{
ident_t terminal = reader.lookup(bm);
ident_t ruleTerminal = getBodyText(rule);
return terminal.compare(ruleTerminal);
}
bool CFParser :: parseGrammarRule(_ScriptReader& reader)
{
ScriptBookmark bm = reader.read();
if (bm.state != dfaIdentifier)
throw EParseError(bm.column, bm.row);
size_t ruleId = mapRuleId(reader.lookup(bm));
reader.read();
if (!reader.compare("::="))
throw EParseError(bm.column, bm.row);
bm = reader.read();
Rule rule;
defineGrammarRule(reader, bm, rule, ruleId);
addRule(ruleId, rule);
return true;
}
void TreeScriptParser :: parse(_ScriptReader& reader, MemoryDump* output)
{
MemoryWriter writer(output);
writer.writeDWord(0); // HOTFIX : end of tape
int sizePos = writer.Position();
writer.writeDWord(0);
SyntaxTree tree;
SyntaxWriter treeWriter(tree);
ScriptBookmark bm = reader.read();
while (!reader.Eof()) {
parseStatement(reader, bm, treeWriter);
bm = reader.read();
}
if (tree.save(output)) {
writer.seek(sizePos);
writer.writeDWord(output->Length() - sizePos);
}
else output->clear();
}
void CFParser :: parse(_ScriptReader& reader, MemoryDump* output)
{
if (_symbolMode)
reader.switchDFA(dfaSymbolic);
ScriptLog log;
size_t startId = mapRuleId("start");
while (!reader.Eof()) {
MemoryWriter writer(&_body);
int presaved = writer.Position();
int trace = buildDerivationTree(reader, startId, writer);
generateOutput(trace, reader, log);
IdentifierTextReader logReader((const char*)log.getBody());
ScriptReader scriptReader(&logReader);
_baseParser->parse(scriptReader, output);
_body.trim(presaved);
log.clear();
}
}
void TreeScriptParser :: parseStatement(_ScriptReader& reader, ScriptBookmark& bm, SyntaxWriter& writer)
{
int type = _tokens.get(reader.lookup(bm));
if (type != 0) {
writer.newNode((LexicalType)type);
bm = reader.read();
if (reader.compare("(")) {
parseScope(reader, bm, writer);
}
else if (reader.compare("=")) {
bm = reader.read();
writer.appendNode(lxTerminal, reader.lookup(bm));
}
else throw EParseError(bm.column, bm.row);
writer.closeNode();
}
else throw EParseError(bm.column, bm.row);
}
void VMTapeParser:: parseStatement(_ScriptReader& reader, ScriptBookmark& bm, TapeWriter& writer)
{
if (reader.compare("*")) {
bm = reader.read();
writeArgument(writer, bm.state, reader.lookup(bm));
}
else if (reader.compare("=")) {
bm = reader.read();
writeArray(writer, bm.state, reader.lookup(bm));
}
else if (reader.compare("^")) {
bm = reader.read();
writeMessage(writer, reader.lookup(bm), SEND_TAPE_MESSAGE_ID);
}
else if (reader.compare("%")) {
bm = reader.read();
ident_t message = reader.lookup(bm);
if (message.find('.') != NOTFOUND_POS) {
writeExtension(writer, reader.lookup(bm), PUSHE_TAPE_MESSAGE_ID);
}
else writeMessage(writer, reader.lookup(bm), PUSHM_TAPE_MESSAGE_ID);
}
else writeObject(writer, bm.state, reader.lookup(bm));
}
bool normalLetterApplyRule(CFParser::Rule& rule, ScriptBookmark& bm, _ScriptReader& reader, CFParser*)
{
ident_t value = reader.lookup(bm);
return (value[0] >= 'A' && value[0] <= 'Z') || (value[0] >= 'a' && value[0] <= 'z');
}
bool normalEOLApplyRule(CFParser::Rule& rule, ScriptBookmark& bm, _ScriptReader& reader, CFParser*)
{
ident_t value = reader.lookup(bm);
return (value[0] == '\n');
}
void CFParser :: defineGrammarRule(_ScriptReader& reader, ScriptBookmark& bm, Rule& rule, ref_t ruleId)
{
// read: terminal [nonterminal] ;
// read: nonterminal [nonterminal2] ;
rule.type = rtNormal;
int applyMode = 0;
while (!reader.compare(";") || bm.state == dfaQuote) {
if (bm.state == dfaQuote) {
if (rule.terminal) {
rule.nonterminal = defineGrammarRule(reader, bm, ruleId, rule.nonterminal);
break;
}
else if (rule.nonterminal) {
rule.type = rtChomski;
rule.terminal = defineGrammarRule(reader, bm, ruleId);
break;
}
else rule.terminal = writeBodyText(reader.lookup(bm));
}
else if (reader.compare("<=")) {
saveScript(reader, rule, applyMode);
}
else if (bm.state == dfaPrivate) {
if (rule.terminal) {
rule.nonterminal = defineGrammarRule(reader, bm, ruleId, rule.nonterminal);
break;
}
else if (rule.nonterminal) {
rule.type = rtChomski;
rule.terminal = defineGrammarRule(reader, bm, ruleId);
break;
}
else {
rule.terminal = -1;
// rule.prefixPtr = defineDSARule(token, reader);
if (reader.compare(LITERAL_KEYWORD)) {
applyMode = LITERAL_MODE;
}
else if (reader.compare(NUMERIC_KEYWORD)) {
applyMode = NUMERIC_MODE;
}
else if (reader.compare(EPS_KEYWORD)) {
if (rule.nonterminal)
throw EParseError(bm.column, bm.row);
rule.nonterminal = -1;
rule.terminal = -1;
applyMode = IDLE_MODE;
}
else if (reader.compare(EOF_KEYWORD)) {
applyMode = EOF_MODE;
}
else if (reader.compare(REFERENCE_KEYWORD)) {
applyMode = REFERENCE_MODE;
}
else if (reader.compare(IDENTIFIER_KEYWORD)) {
applyMode = IDENTIFIER_MODE;
}
else if (reader.compare(EOL_KEYWORD)) {
applyMode = EOL_MODE;
}
else if (reader.compare(ANYCHR_KEYWORD)) {
applyMode = LETTER_MODE;
}
}
}
else if (bm.state == dfaIdentifier) {
if (rule.nonterminal == 0) {
//rule.prefixPtr = defineDSARule(token, reader);
rule.nonterminal = mapRuleId(reader.lookup(bm));
}
else if (rule.terminal == 0) {
rule.type = rtChomski;
rule.terminal = mapRuleId(reader.lookup(bm));
}
else {
if (rule.type == rtChomski) {
rule.terminal = defineGrammarRule(reader, bm, ruleId, rule.terminal);
}
else rule.nonterminal = defineGrammarRule(reader, bm, ruleId, rule.nonterminal);
break;
}
}
bm = reader.read();
}
// rule.postfixPtr = defineDSARule(token, reader);
defineApplyRule(rule, applyMode);
}
void CFParser :: saveScript(_ScriptReader& reader, Rule& rule, int& mode)
{
bool prefixMode = false;
if (rule.terminal == 0 && rule.nonterminal == 0) {
prefixMode = true;
rule.prefixPtr = _body.Length();
}
else rule.postfixPtr = _body.Length();
MemoryWriter writer(&_body);
ScriptBookmark bm = reader.read();
while (!reader.compare("=>") || bm.state == dfaQuote) {
if (bm.state == dfaPrivate) {
if (rule.saveTo != NULL)
throw EParseError(bm.column, bm.row);
if (reader.compare(REFERENCE_KEYWORD)) {
rule.terminal = -1;
rule.saveTo = saveReference;
mode = REFERENCE_MODE;
}
else if (reader.compare(IDENTIFIER_KEYWORD)) {
rule.terminal = -1;
rule.saveTo = saveReference;
mode = IDENTIFIER_MODE;
}
else if (reader.compare(LITERAL_KEYWORD)) {
rule.terminal = -1;
rule.saveTo = saveLiteral;
mode = LITERAL_MODE;
}
else if (reader.compare(NUMERIC_KEYWORD)) {
rule.terminal = -1;
rule.saveTo = saveReference;
mode = NUMERIC_MODE;
}
writer.writeChar((char)0);
rule.postfixPtr = _body.Length();
}
else if (bm.state == dfaQuote && reader.compare(REFERENCE_KEYWORD)) {
rule.terminal = -1;
rule.saveTo = saveLiteral;
mode = REFERENCE_MODE;
}
else if (bm.state == dfaQuote && reader.compare(IDENTIFIER_KEYWORD)) {
rule.terminal = -1;
rule.saveTo = saveLiteral;
mode = IDENTIFIER_MODE;
}
else {
ident_t token = reader.lookup(bm);
if (bm.state == dfaQuote) {
writer.writeChar('\"');
writer.writeLiteral(token, getlength(token));
writer.writeChar('\"');
}
else writer.writeLiteral(token, getlength(token));
writer.writeChar(' ');
}
bm = reader.read();
}
writer.writeChar((char)0);
}