static private int mapped_charset_size; // reduced charset size
static internal void simplify(Spec spec)
{
computeClasses(spec); // initialize fields.
/*
* now rewrite the NFA using our character class mapping.
*/
for (int i = 0; i < spec.nfa_states.Count; i++)
{
Nfa nfa = (Nfa)spec.nfa_states[i];
if (nfa.GetEdge() == Nfa.EMPTY || nfa.GetEdge() == Nfa.EPSILON)
{
continue; // no change.
}
if (nfa.GetEdge() == Nfa.CCL)
{
CharSet nset = new CharSet();
nset.map(nfa.GetCharSet(), ccls); // map it.
nfa.SetCharSet(nset);
}
else
{ // single character
nfa.SetEdge(ccls[nfa.GetEdge()]); // map it.
}
}
/*
* now update spec with the mapping.
*/
spec.ccls_map = ccls;
spec.dtrans_ncols = mapped_charset_size;
}
/*
* Function: rule
* Description: Recursive descent regular expression parser.
*/
private static Nfa rule()
{
NfaPair pair;
Nfa start = null;
Nfa end = null;
int anchor = Spec.NONE;
#if DESCENT_DEBUG
Utility.enter("rule", spec.lexeme, spec.current_token);
#endif
pair = Alloc.NewNfaPair();
if (Gen.AT_BOL == spec.current_token)
{
anchor = anchor | Spec.START;
gen.Advance();
expr(pair);
start = Alloc.NewNfa(spec);
start.SetEdge(spec.BOL);
start.SetNext(pair.start);
end = pair.end;
}
else
{
expr(pair);
start = pair.start;
end = pair.end;
}
if (Gen.AT_EOL == spec.current_token)
{
gen.Advance();
NfaPair nlpair = Alloc.NewNLPair(spec);
end.SetNext(Alloc.NewNfa(spec));
Nfa enext = end.GetNext();
enext.SetNext(nlpair.start);
enext.SetSib(Alloc.NewNfa(spec));
enext.GetSib().SetEdge(spec.EOF);
enext.GetSib().SetNext(nlpair.end);
end = nlpair.end;
anchor = anchor | Spec.END;
}
/* check for null rules */
if (end == null)
{
Error.parse_error(Error.E_ZERO, input.line_number);
}
/* Handle end of regular expression */
end.SetAccept(gen.packAccept());
end.SetAnchor(anchor);
#if DESCENT_DEBUG
Utility.leave("rule", spec.lexeme, spec.current_token);
#endif
return(start);
}
/*
* Function: machine
* Description: Recursive descent regular expression parser.
*/
private static Nfa machine()
{
Nfa start;
Nfa p;
BitSet states;
#if DESCENT_DEBUG
Utility.enter("machine", spec.lexeme, spec.current_token);
#endif
start = Alloc.NewNfa(spec);
p = start;
states = gen.GetStates();
/* Begin: Added for states. */
spec.current_token = Gen.EOS;
gen.Advance();
/* End: Added for states. */
if (Gen.END_OF_INPUT != spec.current_token)
{
p.SetNext(rule());
ProcessStates(states, p.GetNext());
}
while (Gen.END_OF_INPUT != spec.current_token)
{
/* Make state changes HERE. */
states = gen.GetStates();
/* Begin: Added for states. */
gen.Advance();
if (Gen.END_OF_INPUT == spec.current_token)
{
break;
}
/* End: Added for states. */
p.SetSib(Alloc.NewNfa(spec));
p = p.GetSib();
p.SetNext(rule());
ProcessStates(states, p.GetNext());
}
/*
* add pseudo-rules for BOL and EOF
*/
p.SetSib(Alloc.NewNfa(spec));
p = p.GetSib();
p.SetNext(Alloc.NewNfa(spec));
Nfa pnext = p.GetNext();
pnext.SetEdge(Nfa.CCL);
pnext.SetNext(Alloc.NewNfa(spec));
pnext.SetCharSet(new CharSet());
pnext.GetCharSet().add(spec.BOL);
pnext.GetCharSet().add(spec.EOF);
// do-nothing accept rule
pnext.GetNext().SetAccept(new Accept(null, input.line_number + 1));
/* add the pseudo rules */
for (int i = 0; i < spec.states.Count; i++)
{
ArrayList srule = spec.state_rules[i];
srule.Add(pnext);
}
#if DESCENT_DEBUG
Utility.leave("machine", spec.lexeme, spec.current_token);
#endif
return(start);
}