private int mapped_charset_size; // reduced charset size
public void simplify(CSpec m_spec)
{
computeClasses(m_spec); // initialize fields.
// now rewrite the NFA using our character class mapping.
IEnumerator e = m_spec.m_nfa_states.elements();
while (e.MoveNext())
{
CNfa nfa = (CNfa)e.Current;
if (nfa.m_edge == CNfa.EMPTY || nfa.m_edge == CNfa.EPSILON)
{
continue; // no change.
}
if (nfa.m_edge == CNfa.CCL)
{
CSet ncset = new CSet();
ncset.map(nfa.m_set, ccls); // map it.
nfa.m_set = ncset;
}
else
{ // single character
nfa.m_edge = ccls[nfa.m_edge]; // map it.
}
}
// now update m_spec with the mapping.
m_spec.m_ccls_map = ccls;
m_spec.m_dtrans_ncols = mapped_charset_size;
}
/********************************************************
* Function: mimic
* Description: Converts this NFA state into a copy of
* the input one.
*******************************************************/
public void mimic
(
CNfa nfa
)
{
m_edge = nfa.m_edge;
if (null != nfa.m_set)
{
if (null == m_set)
{
m_set = new CSet();
}
m_set.mimic(nfa.m_set);
}
else
{
m_set = null;
}
m_next = nfa.m_next;
m_next2 = nfa.m_next2;
m_accept = nfa.m_accept;
m_anchor = nfa.m_anchor;
if (null != nfa.m_states)
{
m_states = (SparseBitSet)nfa.m_states.Clone();
}
else
{
m_states = null;
}
}
/***************************************************************
* Function: CNfaPair
**************************************************************/
public CNfaPair
(
)
{
m_start = null;
m_end = null;
}
/***************************************************************
* Function: discardCNfa
* Description:
**************************************************************/
private void discardCNfa
(
CNfa nfa
)
{
m_spec.m_nfa_states.removeElement(nfa);
}
/********************************************************
* Function: CNfa
*******************************************************/
public CNfa
(
)
{
m_edge = EMPTY;
m_set = null;
m_next = null;
m_next2 = null;
m_accept = null;
m_anchor = CSpec.NONE;
m_label = NO_LABEL;
m_states = null;
}
/***************************************************************
* Function: factor
* Description: Recursive descent regular expression parser.
**************************************************************/
private void factor
(
CNfaPair pair
)
{
CNfa start = null;
CNfa end = null;
#if (DESCENT_DEBUG)
{
CUtility.enter("factor", m_spec.m_lexeme, m_spec.m_current_token);
}
#endif
term(pair);
if (CLexGen.CLOSURE == m_spec.m_current_token ||
CLexGen.PLUS_CLOSE == m_spec.m_current_token ||
CLexGen.OPTIONAL == m_spec.m_current_token)
{
start = CAlloc.newCNfa(m_spec);
end = CAlloc.newCNfa(m_spec);
start.m_next = pair.m_start;
pair.m_end.m_next = end;
if (CLexGen.CLOSURE == m_spec.m_current_token ||
CLexGen.OPTIONAL == m_spec.m_current_token)
{
start.m_next2 = end;
}
if (CLexGen.CLOSURE == m_spec.m_current_token ||
CLexGen.PLUS_CLOSE == m_spec.m_current_token)
{
pair.m_end.m_next2 = pair.m_start;
}
pair.m_start = start;
pair.m_end = end;
m_lexGen.advance();
}
#if (DESCENT_DEBUG)
{
CUtility.leave("factor", m_spec.m_lexeme, m_spec.m_current_token);
}
#endif
}
/***************************************************************
* Function: newCNfa
* Description:
**************************************************************/
public static CNfa newCNfa
(
CSpec spec
)
{
CNfa p;
/* UNDONE: Buffer this? */
p = new CNfa();
/*p.m_label = spec.m_nfa_states.size();*/
spec.m_nfa_states.addElement(p);
p.m_edge = CNfa.EPSILON;
return(p);
}
/***************************************************************
* Function: processStates
* Description:
**************************************************************/
private void processStates
(
SparseBitSet states,
CNfa current
)
{
int size;
int i;
size = m_spec.m_states.Count;
for (i = 0; i < size; ++i)
{
if (states.Get(i))
{
m_spec.m_state_rules[i].addElement(current);
}
}
}
/***************************************************************
* Function: CSpec
* Description: Constructor.
**************************************************************/
public CSpec
(
CLexGen lexGen
)
{
m_lexGen = lexGen;
/* Initialize regular expression token variables. */
m_current_token = CLexGen.EOS;
m_lexeme = '\0';
m_in_quote = false;
m_in_ccl = false;
/* Initialize hashtable for lexer states. */
m_states = new Hashtable();
m_states.Add("YYINITIAL", m_states.Count);
/* Initialize hashtable for lexical macros. */
m_macros = new Hashtable();
/* Initialize variables for lexer options. */
m_integer_type = false;
m_intwrap_type = false;
m_count_lines = false;
m_count_chars = true;
m_cup_compatible = false;
m_unix = true;
m_public = false;
m_yyeof = false;
m_ignorecase = false;
/* Initialize variables for JLex runtime options. */
m_verbose = false;
m_nfa_start = null;
m_nfa_states = new Vector();
m_dfa_states = new Vector();
m_dfa_sets = new Hashtable();
m_dtrans_vector = new Vector();
m_dtrans_ncols = CUtility.MAX_SEVEN_BIT + 1;
m_row_map = null;
m_col_map = null;
m_accept_vector = null;
m_anchor_array = null;
m_init_code = null;
m_init_read = 0;
m_init_throw_code = null;
m_init_throw_read = 0;
m_yylex_throw_code = null;
m_yylex_throw_read = 0;
m_class_code = null;
m_class_read = 0;
m_eof_code = null;
m_eof_read = 0;
m_eof_value_code = null;
m_eof_value_read = 0;
m_eof_throw_code = null;
m_eof_throw_read = 0;
m_state_dtrans = null;
m_state_rules = null;
}
/** Compute minimum Set of character classes needed to disambiguate
* edges. We optimistically assume that every character belongs to
* a single character class, and then incrementally split classes
* as we see edges that require discrimination between characters in
* the class. [CSA, 25-Jul-1999] */
private void computeClasses(CSpec m_spec)
{
this.original_charset_size = m_spec.m_dtrans_ncols;
this.ccls = new int[original_charset_size]; // initially all zero.
int nextcls = 1;
SparseBitSet clsA = new SparseBitSet(), clsB = new SparseBitSet();
Hashtable h = new Hashtable();
if (m_spec.m_verbose)
{
System.Console.Write("Working on character classes.");
}
IEnumerator e = m_spec.m_nfa_states.elements();
while (e.MoveNext())
{
CNfa nfa = (CNfa)e.Current;
if (nfa.m_edge == CNfa.EMPTY || nfa.m_edge == CNfa.EPSILON)
{
continue; // no discriminatory information.
}
clsA.clearAll(); clsB.clearAll();
for (int i = 0; i < ccls.Length; i++)
{
if (nfa.m_edge == i || // edge labeled with a character
nfa.m_edge == CNfa.CCL && nfa.m_set.contains(i)) // Set of characters
{
clsA.Set(ccls[i]);
}
else
{
clsB.Set(ccls[i]);
}
}
// now figure out which character classes we need to split.
clsA.and(clsB); // split the classes which show up on both sides of edge
if (m_spec.m_verbose)
{
System.Console.Write(clsA.size() == 0?".":":");
}
if (clsA.size() == 0)
{
continue; // nothing to do.
}
// and split them.
h.Clear(); // h will map old to new class name
for (int i = 0; i < ccls.Length; i++)
{
if (clsA.Get(ccls[i])) // a split class
{
if (nfa.m_edge == i ||
nfa.m_edge == CNfa.CCL && nfa.m_set.contains(i))
{ // on A side
int split = ccls[i];
if (!h.ContainsKey(split))
{
h.Add(split, (nextcls++)); // make new class
}
ccls[i] = (int)h[split];
}
}
}
}
if (m_spec.m_verbose)
{
System.Console.WriteLine();
System.Console.WriteLine("NFA has " + nextcls + " distinct character classes.");
}
this.mapped_charset_size = nextcls;
}
/***************************************************************
* Function: rule
* Description: Recursive descent regular expression parser.
**************************************************************/
private CNfa rule
(
)
{
CNfaPair pair;
//CNfa p;
CNfa start = null;
CNfa end = null;
int anchor = CSpec.NONE;
#if (DESCENT_DEBUG)
{
CUtility.enter("rule", m_spec.m_lexeme, m_spec.m_current_token);
}
#endif
pair = CAlloc.newCNfaPair();
if (CLexGen.AT_BOL == m_spec.m_current_token)
{
anchor = anchor | CSpec.START;
m_lexGen.advance();
expr(pair);
// CSA: fixed beginning-of-line operator. 8-aug-1999
start = CAlloc.newCNfa(m_spec);
start.m_edge = m_spec.BOL;
start.m_next = pair.m_start;
end = pair.m_end;
}
else
{
expr(pair);
start = pair.m_start;
end = pair.m_end;
}
if (CLexGen.AT_EOL == m_spec.m_current_token)
{
m_lexGen.advance();
// CSA: fixed end-of-line operator. 8-aug-1999
CNfaPair nlpair = CAlloc.newNLPair(m_spec);
end.m_next = CAlloc.newCNfa(m_spec);
end.m_next.m_next = nlpair.m_start;
end.m_next.m_next2 = CAlloc.newCNfa(m_spec);
end.m_next.m_next2.m_edge = m_spec.EOF;
end.m_next.m_next2.m_next = nlpair.m_end;
end = nlpair.m_end;
anchor = anchor | CSpec.END;
}
/* Check for null rules. Charles Fischer found this bug. [CSA] */
if (end == null)
{
CError.parse_error(CError.E_ZERO, m_input.m_line_number);
}
/* Handle end of regular expression. See page 103. */
end.m_accept = m_lexGen.packAccept();
end.m_anchor = anchor;
/* Begin: Removed for states. */
/*m_lexGen.advance();*/
/* End: Removed for states. */
#if (DESCENT_DEBUG)
{
CUtility.leave("rule", m_spec.m_lexeme, m_spec.m_current_token);
}
#endif
return(start);
}