public cs0tokens(YyLexer tks)
: base(tks)
{
}
void _Create()
{
m_outFile.WriteLine("using System;using Tools;");
m_tokens = new YyLexer(erh);
string buf = "";
string str = "";
string name = "";
string startsym;
Nfa nfa;
int p, q, max;
Console.WriteLine("Reading Input File");
while (!m_inFile.Eof())
{
buf = m_inFile.ReadLine();
startsym = "YYINITIAL";
max = buf.Length;
p = 0;
if (!White(buf, ref p, max))
{
continue;
}
if (buf[p] == '%')
{ // directive
// %lexer
if (buf.Length >= p + 6 && "%lexer".Equals(buf.Substring(p, 6)))
{
m_lexerseen = true;
p += 6;
if (!White(buf, ref p, max))
{
continue;
}
q = p;
NonWhite(buf, ref p, max);
if (q != p)
{
m_outname = buf.Substring(q, p - q);
}
continue;
}
// %encoding
if (buf.Length >= p + 9 && "%encoding".Equals(buf.Substring(p, 9)))
{
p += 9; White(buf, ref p, max);
q = p;
NonWhite(buf, ref p, max);
m_tokens.InputEncoding = buf.Substring(q, p - q);
continue;
}
// %namespace
if (buf.Length >= p + 10 && "%namespace".Equals(buf.Substring(p, 10)))
{
p += 10; White(buf, ref p, max);
q = p;
NonWhite(buf, ref p, max);
m_outFile.WriteLine("namespace " + buf.Substring(q, p - q) + " {");
m_namespace = true;
continue;
}
// %define
if (buf.Length >= p + 7 && "%define".Equals(buf.Substring(p, 7)))
{
p += 7; White(buf, ref p, max);
q = p;
if (!NonWhite(buf, ref p, max))
{
erh.Error(new CSToolsException(44, "Bad define"));
continue;
}
name = buf.Substring(q, p - q);
p++;
if (White(buf, ref p, max))
{
defines[name] = buf.Substring(p, max - p);
}
}
else
// % token/node
if (buf.Length >= p + 6 && "%token".Equals(buf.Substring(p, 6)))
{
EmitClassDefin(buf, ref p, max, m_inFile, "TOKEN", out str, out name, true);
}
else if (buf.Length >= p + 5 && "%node".Equals(buf.Substring(p, 5)))
{
EmitClassDefin(buf, ref p, max, m_inFile, "NODE", out str, out name, true);
}
else if (buf.Length >= p + 2 && "%{".Equals(buf.Substring(p, 2)))
{
CopyCode();
}
else if (buf.Length >= p + 9 && "%declare{".Equals(buf.Substring(p, 9)))
{
p += 8;
m_actvars = ToBraceIfFound(ref buf, ref p, ref max, m_inFile);
m_actvars = m_actvars.Substring(1, m_actvars.Length - 2);
}
else
{
m_tokens.erh.Error(new CSToolsException(8, "Unknown directive " + buf.Substring(p, max - p)));
}
continue;
}
else if (buf[p] == '<')
{ // startstate
q = p++;
while (p < max && buf[p] != '>')
{
p++;
}
if (p++ == max)
{
m_tokens.erh.Error(new CSToolsException(25, "Bad startsymbol"));
continue;
}
startsym = buf.Substring(q + 1, p - q - 2);
White(buf, ref p, max);
}
q = p; // can't simply look for nonwhite space here because embedded spaces
GetRegex(buf, ref p, max);
string trgx = buf.Substring(q, p - q);
if (m_tokens.toupper)
{
trgx = trgx.ToUpper();
}
Regex rgx = new Regex(this, q, trgx);
Nfa nfa1 = new Nfa(this, rgx);
if (!m_startstates.Contains(startsym))
{
m_startstates[startsym] = new Nfa(this);
}
nfa = (Nfa)m_startstates[startsym];
nfa.AddEps(nfa1);
White(buf, ref p, max);
m_actions[nfa1.m_end.m_state] = nfa1.m_end;
// handle multiline actions enclosed in {}
nfa1.m_end.m_sTerminal = ToBraceIfFound(ref buf, ref p, ref max, m_inFile);
// examine action string
if (nfa1.m_end.m_sTerminal.Length > 0 && nfa1.m_end.m_sTerminal[0] == '%')
{
string tokClass, b = nfa1.m_end.m_sTerminal;
q = 1;
max = b.Length;
int n;
for (n = 0; q < max && b[q] != ' ' && b[q] != '\t' && b[q] != '\n' && b[q] != '{' && b[q] != ':'; q++, n++) // extract the class name
{
;
}
tokClass = b.Substring(1, n); // new-style auto token construction
object ob = m_tokens.tokens[tokClass];
TokClassDef t = (TokClassDef)ob;
bool isNew = (t == null);
// check for initialisation action following %name
string init = b.Substring(n + 1, b.Length - n - 1);
string bas1 = "TOKEN";
bool haveInit = false;
for (int j = 0; j < init.Length; j++)
{
if (init[j] == '{')
{
haveInit = true;
break;
}
else if (init[j] == ':')
{
bas1 = "";
for (; init[j] == ' ' || init[j] == '\r'; j++)
{
;
}
for (; init[j] != ' ' && init[j] != '\t' && init[j] != '{' && init[j] != '\n'; j++)
{
bas1 += init[j];
}
break;
}
}
if (isNew && tokClass != "TOKEN")
{ // this token class has not been declared. Do so now
bool isNode = (m_tokens.tokens[bas1] != null);
t = new TokClassDef(this, tokClass, bas1); // updates TOKEN.tokens
m_outFile.WriteLine("//%{0}+{1}", tokClass, t.m_yynum);
m_outFile.Write("public class {0} : {1}", tokClass, bas1);
m_outFile.WriteLine("{ public override string yyname { get { return \"" + tokClass + "\";}}");
m_outFile.WriteLine("public override int yynum { get { return " + t.m_yynum + "; }}");
m_outFile.WriteLine(" public " + tokClass + "(Lexer yyl):base(yyl) {}}");
}
if (haveInit && init.IndexOf("%except") < 0)
{
init = init.Replace("yylval", "yymval"); // yuk: avoid the next line munging yylval
init = FixActions(init);
init = init.Replace("yymval", "yylval");
nfa1.m_end.m_sTerminal = "%" + NewConstructor(t, init);
}
}
}
if (!m_lexerseen)
{
m_tokens.erh.Error(new CSToolsException(26, "No %lexer directive detected: possibly incorrect text encoding?"));
}
Console.WriteLine("Constructing DFAs");
foreach (string s in m_startstates.Keys)
{
Dfa d = new Dfa((Nfa)m_startstates[s]);
m_tokens.starts[s] = d;
if (d.m_actions != null)
{
Console.WriteLine("Warning: This lexer script generates an infinite token stream on bad input");
}
}
Console.WriteLine("Output phase");
Emit(m_actions, m_actvars, m_namespace, m_showDfa);
Console.WriteLine("End of Create");
object o = m_tokens.starts["YYINITIAL"];
if (o == null)
{
Console.WriteLine("Warning: No lexer");
}
else if (((Dfa)o).m_actions != null) // repeat the above warning
{
Console.WriteLine("Warning: This lexer script generates an infinite token stream on bad input");
}
}