public ReStr(TokensGen tks, string str)
{
this.m_str = str;
for (int index = 0; index < str.Length; ++index)
{
tks.m_tokens.UsingChar(str[index]);
}
}
public ReStr(TokensGen tks, string str)
{
m_str = str;
for (int i = 0; i < str.Length; i++)
{
tks.m_tokens.UsingChar(str[i]);
}
}
public ReUStr(TokensGen tks, string str)
{
this.m_str = str;
for (int index = 0; index < str.Length; ++index)
{
tks.m_tokens.UsingChar(char.ToLower(str[index]));
tks.m_tokens.UsingChar(char.ToUpper(str[index]));
}
}
public ReUStr(TokensGen tks, string str)
{
m_str = str;
for (int i = 0; i < str.Length; i++)
{
tks.m_tokens.UsingChar(Char.ToLower(str[i]));
tks.m_tokens.UsingChar(Char.ToUpper(str[i]));
}
}
public TokClassDef(GenBase gbs, string name, string bas)
{
if (gbs is TokensGen)
{
TokensGen tks = (TokensGen)gbs;
m_name = name;
tks.m_tokens.tokens[name] = this;
m_refToken = bas;
}
}
public static ResWds New(TokensGen tks, string str)
{
ResWds resWds = new ResWds();
str = str.Trim();
if (str[0] == 'U')
{
resWds.m_upper = true;
str = str.Substring(1).Trim();
}
if (str[0] == '{' && str[str.Length - 1] == '}')
{
str = str.Substring(1, str.Length - 2).Trim();
string str1 = str;
char[] chArray = new char[1] {
','
};
foreach (string str2 in str1.Split(chArray))
{
string str3 = str2.Trim();
string name = str3;
int num = str3.IndexOf(' ');
if (num > 0)
{
name = str3.Substring(num).Trim();
str3 = str3.Substring(0, num);
}
resWds.m_wds[(object)str3] = (object)name;
if (tks.m_tokens.tokens[(object)name] == null)
{
TokClassDef tokClassDef = new TokClassDef((GenBase)tks, name, "TOKEN");
tks.m_outFile.WriteLine("//%{0}+{1}", (object)name, (object)tokClassDef.m_yynum);
tks.m_outFile.Write("public class {0} : TOKEN", (object)name);
tks.m_outFile.WriteLine("{ public override string yyname { get { return \"" + name + "\";}}");
tks.m_outFile.WriteLine("public override int yynum { get { return " + (object)tokClassDef.m_yynum + "; }}");
tks.m_outFile.WriteLine(" public " + name + "(Lexer yyl):base(yyl) {}}");
}
}
return(resWds);
}
tks.m_tokens.erh.Error(new CSToolsException(47, "bad ResWds element"));
return((ResWds)null);
}
public static ResWds New(TokensGen tks, string str)
{
ResWds r = new ResWds();
str = str.Trim();
if (str[0] == 'U')
{
r.m_upper = true;
str = str.Substring(1).Trim();
}
if (str[0] != '{' || str[str.Length - 1] != '}')
{
goto bad;
}
str = str.Substring(1, str.Length - 2).Trim();
string[] wds = str.Split(',');
for (int j = 0; j < wds.Length; j++)
{
string w = wds[j].Trim();
string a = w;
int i = w.IndexOf(' ');
if (i > 0)
{
a = w.Substring(i).Trim();
w = w.Substring(0, i);
}
r.m_wds[w] = a;
if (tks.m_tokens.tokens[a] == null)
{
TokClassDef t = new TokClassDef(tks, a, "TOKEN");
tks.m_outFile.WriteLine("//%{0}+{1}", a, t.m_yynum);
tks.m_outFile.Write("public class {0} : TOKEN", a);
tks.m_outFile.WriteLine("{ public override string yyname { get { return \"" + a + "\";}}");
tks.m_outFile.WriteLine("public override int yynum { get { return " + t.m_yynum + "; }}");
tks.m_outFile.WriteLine(" public " + a + "(Lexer yyl):base(yyl) {}}");
}
}
return(r);
bad:
tks.m_tokens.erh.Error(new CSToolsException(47, "bad ResWds element"));
return(null);
}
public Regex(TokensGen tks, int p, string str)
{
int length = str.Length;
int num1 = 0;
int num2 = 0;
int num3 = 0;
this.m_sub = (Regex)null;
if (length == 0)
{
return;
}
int startIndex;
if (str[0] == '(')
{
int index;
for (index = 1; index < length; ++index)
{
if (str[index] == '\\')
{
++index;
}
else if (str[index] == ']' && num2 > 0)
{
num2 = 0;
}
else if (num2 <= 0)
{
if (str[index] == '"' || str[index] == '\'')
{
if (num3 == (int)str[index])
{
num3 = 0;
}
else if (num3 == 0)
{
num3 = (int)str[index];
}
}
else if (num3 <= 0)
{
if (str[index] == '[')
{
++num2;
}
else if (str[index] == '(')
{
++num1;
}
else if (str[index] == ')' && num1-- == 0)
{
break;
}
}
}
}
if (index != length)
{
this.m_sub = new Regex(tks, p + 1, str.Substring(1, index - 1));
startIndex = index + 1;
}
else
{
goto label_99;
}
}
else if (str[0] == '[')
{
int index;
for (index = 1; index < length && str[index] != ']'; ++index)
{
if (str[index] == '\\')
{
++index;
}
}
if (index != length)
{
this.m_sub = (Regex) new ReRange(tks, str.Substring(0, index + 1));
startIndex = index + 1;
}
else
{
goto label_99;
}
}
else if (str[0] == '\'' || str[0] == '"')
{
StringBuilder stringBuilder = new StringBuilder();
int index;
for (index = 1; index < length && (int)str[index] != (int)str[0]; ++index)
{
if (str[index] == '\\')
{
char ch = str[++index];
switch (ch)
{
case 'r':
stringBuilder.Append('\r');
continue;
case 't':
stringBuilder.Append('\t');
continue;
case 'v':
stringBuilder.Append('\v');
continue;
default:
switch (ch)
{
case '\n':
continue;
case '"':
stringBuilder.Append('"');
continue;
case '\'':
stringBuilder.Append('\'');
continue;
case '0':
stringBuilder.Append(char.MinValue);
continue;
case '\\':
stringBuilder.Append('\\');
continue;
case 'n':
stringBuilder.Append('\n');
continue;
default:
stringBuilder.Append(str[index]);
continue;
}
}
}
else
{
stringBuilder.Append(str[index]);
}
}
if (index != length)
{
startIndex = index + 1;
this.m_sub = (Regex) new ReStr(tks, stringBuilder.ToString());
}
else
{
goto label_99;
}
}
else if (str.StartsWith("U\"") || str.StartsWith("U'"))
{
StringBuilder stringBuilder = new StringBuilder();
int index;
for (index = 2; index < length && (int)str[index] != (int)str[1]; ++index)
{
if (str[index] == '\\')
{
char ch = str[++index];
switch (ch)
{
case 'r':
stringBuilder.Append('\r');
continue;
case 't':
stringBuilder.Append('\t');
continue;
case 'v':
stringBuilder.Append('\v');
continue;
default:
switch (ch)
{
case '\n':
continue;
case '"':
stringBuilder.Append('"');
continue;
case '\'':
stringBuilder.Append('\'');
continue;
case '\\':
stringBuilder.Append('\\');
continue;
case 'n':
stringBuilder.Append('\n');
continue;
default:
stringBuilder.Append(str[index]);
continue;
}
}
}
else
{
stringBuilder.Append(str[index]);
}
}
if (index != length)
{
startIndex = index + 1;
this.m_sub = (Regex) new ReUStr(tks, stringBuilder.ToString());
}
else
{
goto label_99;
}
}
else if (str[0] == '\\')
{
char ch1;
char ch2 = ch1 = str[1];
switch (ch2)
{
case 'r':
ch1 = '\r';
break;
case 't':
ch1 = '\t';
break;
case 'v':
ch1 = '\v';
break;
default:
if (ch2 == 'n')
{
ch1 = '\n';
break;
}
break;
}
this.m_sub = (Regex) new ReStr(tks, ch1);
startIndex = 2;
}
else if (str[0] == '{')
{
int index = 1;
while (index < length && str[index] != '}')
{
++index;
}
if (index != length)
{
string str1 = str.Substring(1, index - 1);
string define = (string)tks.defines[(object)str1];
this.m_sub = define != null ? new Regex(tks, p + 1, define) : (Regex) new ReCategory(tks, str1);
startIndex = index + 1;
}
else
{
goto label_99;
}
}
else
{
this.m_sub = str[0] != '.' ? (Regex) new ReStr(tks, str[0]) : (Regex) new ReRange(tks, "[^\n]");
startIndex = 1;
}
if (startIndex >= length)
{
return;
}
if (str[startIndex] == '?')
{
this.m_sub = (Regex) new ReOpt(this.m_sub);
++startIndex;
}
else if (str[startIndex] == '*')
{
this.m_sub = (Regex) new ReStar(this.m_sub);
++startIndex;
}
else if (str[startIndex] == '+')
{
this.m_sub = (Regex) new RePlus(this.m_sub);
++startIndex;
}
if (startIndex >= length)
{
return;
}
if (str[startIndex] == '|')
{
this.m_sub = (Regex) new ReAlt(tks, this.m_sub, p + startIndex + 1, str.Substring(startIndex + 1, length - startIndex - 1));
return;
}
if (startIndex >= length)
{
return;
}
this.m_sub = (Regex) new ReCat(tks, this.m_sub, p + startIndex, str.Substring(startIndex, length - startIndex));
return;
label_99:
tks.erh.Error((CSToolsException) new CSToolsFatalException(1, tks.sourceLineInfo(p), str, "ill-formed regular expression " + str));
}
public ReRange(TokensGen tks, string str)
{
StringBuilder stringBuilder = new StringBuilder();
int num1 = str.Length - 1;
for (int index = 1; index < num1; ++index)
{
if (str[index] == '\\')
{
if (index + 1 < num1)
{
++index;
}
if (str[index] >= '0' && str[index] <= '7')
{
int num2;
for (num2 = (int)str[index++] - 48; index < num1 && str[index] >= '0' && str[index] <= '7'; ++index)
{
num2 = num2 * 8 + (int)str[index] - 48;
}
stringBuilder.Append((char)num2);
}
else
{
char ch = str[index];
switch (ch)
{
case 'r':
stringBuilder.Append('\r');
continue;
case 't':
stringBuilder.Append('\t');
continue;
case 'v':
stringBuilder.Append('\v');
continue;
default:
if (ch == 'n')
{
stringBuilder.Append('\n');
continue;
}
stringBuilder.Append(str[index]);
continue;
}
}
}
else
{
stringBuilder.Append(str[index]);
}
}
int length = stringBuilder.Length;
if (length > 0 && stringBuilder[0] == '^')
{
this.m_invert = true;
stringBuilder.Remove(0, 1).Append(char.MinValue).Append(char.MaxValue);
}
for (int index1 = 0; index1 < length; ++index1)
{
if (index1 + 1 < length && stringBuilder[index1 + 1] == '-')
{
for (int index2 = (int)stringBuilder[index1]; index2 <= (int)stringBuilder[index1 + 2]; ++index2)
{
this.Set(tks, (char)index2);
}
index1 += 2;
}
else
{
this.Set(tks, stringBuilder[index1]);
}
}
}
public ReStr(TokensGen tks, char ch)
{
m_str = new string(ch, 1);
tks.m_tokens.UsingChar(ch);
}
public ReCategory(TokensGen tks, string str)
{
m_str = str;
m_test = tks.m_tokens.GetTest(str);
}
public string m_sTerminal = ""; // or something for the Lexer
#endregion Fields
#region Constructors
/// <exclude/>
public NfaNode(TokensGen tks)
: base(tks)
{
}
public ReAlt(TokensGen tks,Regex sub,int p,string str)
{
m_sub = sub;
m_alt = new Regex(tks,p,str);
}
public ReCategory(TokensGen tks,string str)
{
m_str = str;
m_test = tks.m_tokens.GetTest(str);
}
void Set(TokensGen tks,char ch)
{
m_map[ch] = true;
tks.m_tokens.UsingChar(ch);
}
/// <exclude/>
public ReStr(TokensGen tks,char ch)
{
m_str = new string(ch,1);
tks.m_tokens.UsingChar(ch);
}
public Nfa(TokensGen tks) : base(tks)
{
m_end = new NfaNode(m_tks);
}
// build an NFA for a given regular expression
public Nfa(TokensGen tks, Regex re) : base(tks)
{
m_end = new NfaNode(tks);
re.Build(this);
}
public ObjectList m_eps = new ObjectList(); // of NfaNode for unlabelled arcs
public NfaNode(TokensGen tks) : base(tks)
{
}
void Set(TokensGen tks, char ch)
{
m_map[ch] = true;
tks.m_tokens.UsingChar(ch);
}
public ReRange(TokensGen tks, string str)
{
StringBuilder ns = new StringBuilder();
int n = str.Length - 1, v;
int p;
for (p = 1; p < n; p++) // fix \ escapes
{
if (str[p] == '\\')
{
if (p + 1 < n)
{
p++;
}
if (str[p] >= '0' && str[p] <= '7')
{
for (v = str[p++] - '0'; p < n && str[p] >= '0' && str[p] <= '7'; p++)
{
v = v * 8 + str[p] - '0';
}
ns.Append((char)v);
}
else
{
switch (str[p])
{
case 'n': ns.Append('\n'); break;
case 't': ns.Append('\t'); break;
case 'r': ns.Append('\r'); break;
case 'v': ns.Append('\v'); break;
default: ns.Append(str[p]); break;
}
}
}
else
{
ns.Append(str[p]);
}
}
n = ns.Length;
if (n > 0 && ns[0] == '^')
{ // invert range
m_invert = true;
ns.Remove(0, 1).Append((char)0).Append((char)0xFFFF);
}
for (p = 0; p < n; p++)
{
if (p + 1 < n && ns[p + 1] == '-')
{
for (v = ns[p]; v <= ns[p + 2]; v++)
{
Set(tks, (char)v);
}
p += 2;
}
else
{
Set(tks, ns[p]);
}
}
}
public ReRange(TokensGen tks, string str)
{
string ns = "";
int n = str.Length - 1, v;
int p;
for (p = 1; p < n; p++) // fix \ escapes
{
if (str[p] == '\\')
{
if (p + 1 < n)
{
p++;
}
if (str[p] >= '0' && str[p] <= '7')
{
for (v = str[p++] - '0'; p < n && str[p] >= '0' && str[p] <= '7'; p++)
{
v = v * 8 + str[p] - '0';
}
ns += (char)v;
}
else
{
switch (str[p])
{
case 'n': ns += '\n'; break;
case 't': ns += '\t'; break;
case 'r': ns += '\r'; break;
default: ns += str[p]; break;
}
}
}
else
{
ns += str[p];
}
}
n = ns.Length;
if (ns[0] == '^')
{ // invert range
m_invert = true;
ns = ns.Substring(1) + (char)0 + (char)0xFFFF;
}
for (p = 0; p < n; p++)
{
if (p + 1 < n && ns[p + 1] == '-')
{
for (v = ns[p]; v <= ns[p + 2]; v++)
{
Set(tks, (char)v);
}
p += 2;
}
else
{
Set(tks, ns[p]);
}
}
}
/// <exclude/>
public ReUStr(TokensGen tks,char ch)
{
m_str = new string(ch,1);
tks.m_tokens.UsingChar(Char.ToLower(ch));
tks.m_tokens.UsingChar(Char.ToUpper(ch));
}
public ReUStr(TokensGen tks, char ch)
{
this.m_str = new string(ch, 1);
tks.m_tokens.UsingChar(char.ToLower(ch));
tks.m_tokens.UsingChar(char.ToUpper(ch));
}
/// <exclude/>
public Dfa(TokensGen tks)
: base(tks)
{
m_tokens = tks.m_tokens;
}
/// <exclude/>
public ReCat(TokensGen tks,Regex sub, int p, string str)
{
m_sub = sub;
m_next = new Regex(tks,p,str);
}
/// <exclude/>
public Nfa(TokensGen tks)
: base(tks)
{
m_end = new NfaNode(m_tks);
}
public Hashtable m_map = new Hashtable(); // char->bool
#endregion Fields
#region Constructors
public ReRange(TokensGen tks,string str)
{
StringBuilder ns = new StringBuilder();
int n = str.Length-1,v;
int p;
for (p=1;p<n;p++) // fix \ escapes
if (str[p] == '\\')
{
if (p+1<n)
p++;
if (str[p]>='0' && str[p]<='7')
{
for (v = str[p++]-'0';p<n && str[p]>='0' && str[p]<='7';p++)
v=v*8+str[p]-'0';
ns.Append((char)v);
}
else
switch(str[p])
{
case 'n' : ns.Append('\n'); break;
case 't' : ns.Append('\t'); break;
case 'r' : ns.Append('\r'); break;
default: ns.Append(str[p]); break;
}
}
else
ns.Append(str[p]);
n = ns.Length;
if (ns[0] == '^')
{// invert range
m_invert = true;
ns.Remove(0,1).Append((char)0).Append((char)0xFFFF);
}
for (p=0;p<n;p++)
if (p+1<n && ns[p+1]=='-')
{
for (v=ns[p];v<=ns[p+2];v++)
Set(tks,(char)v);
p += 2;
}
else
Set(tks,ns[p]);
}
public ReRange(TokensGen tks,string str)
{
string ns = "";
int n = str.Length-1,v;
int p;
for (p=1;p<n;p++) // fix \ escapes
if (str[p] == '\\')
{
if (p+1<n)
p++;
if (str[p]>='0' && str[p]<='7')
{
for (v = str[p++]-'0';p<n && str[p]>='0' && str[p]<='7';p++)
v=v*8+str[p]-'0';
ns += (char)v;
}
else
switch(str[p])
{
case 'n' : ns += '\n'; break;
case 't' : ns += '\t'; break;
case 'r' : ns += '\r'; break;
default: ns += str[p]; break;
}
}
else
ns += str[p];
n = ns.Length;
if (ns[0] == '^')
{// invert range
m_invert = true;
ns = ns.Substring(1)+(char)0 +(char)0xFFFF;
}
for (p=0;p<n;p++)
if (p+1<n && ns[p+1]=='-')
{
for (v=ns[p];v<=ns[p+2];v++)
Set(tks,(char)v);
p += 2;
}
else
Set(tks,ns[p]);
}
/// <exclude/>
public ReStr(TokensGen tks,string str)
{
m_str = str;
for (int i=0;i<str.Length;i++)
tks.m_tokens.UsingChar(str[i]);
}
/*
* Construct a Regex from a given string
*
* 1. First examine the given string.
* If it is empty, there is nothing to do, so return (having cleared m_sub as a precaution).
* 2. Look to see if the string begins with a bracket ( . If so, find the matching ) .
* This is not as simple as it might be because )s inside quotes or [] or escaped will not count.
* Recursively call the constructor for the regular expression between the () s.
* Mark everything up to the ) as used, and go to step 9.
* 3. Look to see if the string begins with a bracket [ . If so, find the matching ] , watching for escapes.
* Construct a ReRange for everything between the []s.
* Mark everything up to the ] as used, and go to step 9.
* 4. Look to see if the string begins with a ' or " . If so, build the contents interpreting
* escaped special characters correctly, until the matching quote is reached.
* Construct a ReStr for the contents, mark everything up to the final quote as used, and go to step 9.
* 4a. Look to see if the string begins with a U' or U" . If so, build the contents interpreting
* escaped special characters correctly, until the matching quote is reached.
* Construct a ReUStr for the contents, mark everything up to the final quote as used, and go to step 9.
* 5. Look to see if the string begins with a \ .
* If so, build a ReStr for the next character (special action for ntr),
* mark it as used, and go to step 9.
* 6. Look to see if the string begins with a { .
* If so, find the matching }, lookup the symbolic name in the definitions table,
* recursively call this constructor on the contents,
* mark everything up to the } as used, and go to step 9.
* 7. Look to see if the string begins with a dot.
* If so, construct a ReRange("^\n"), mark the . as used, and go to step 9.
* 8. At this point we conclude that there is a simple character at the start of the regular expression.
* Construct a ReStr for it, mark it as used, and go to step 9.
* 9. If the string is exhausted, return.
* We have a simple Regex whose m_sub contains what we can constructed.
* 10. If the next character is a ? , *, or +, construct a ReOpt, ReStart, or RePlus respectively
* out of m_sub, and make m_sub point to this new class instead. Mark the character as used.
* 11. If the string is exhausted, return.
* 12. If the next character is a | , build a ReAlt using the m_sub we have and the rest of the string.
* 13. Otherwise build a ReCat using the m_sub we have and the rest of the string.
*/
public Regex(TokensGen tks, int p, string str)
{
int n = str.Length;
int nlp = 0;
int lbrack = 0;
int quote = 0;
int j;
char ch;
//1. First examine the given string.
// If it is empty, there is nothing to do, so return (having cleared m_sub as a precaution).
m_sub = null;
if (n == 0)
{
return;
}
//2. Look to see if the string begins with a bracket ( . If so, find the matching ) .
// This is not as simple as it might be because )s inside quotes or [] or escaped will not count.
// Recursively call the constructor for the regular expression between the () s.
// Mark everything up to the ) as used, and go to step 9.
else if (str[0] == '(')
{ // identify a bracketed expression
for (j = 1; j < n; j++)
{
if (str[j] == '\\')
{
j++;
}
else if (str[j] == ']' && lbrack > 0)
{
lbrack = 0;
}
else if (lbrack > 0)
{
continue;
}
else if (str[j] == '"' || str[j] == '\'')
{
if (quote == str[j])
{
quote = 0;
}
else if (quote == 0)
{
quote = str[j];
}
}
else if (quote > 0)
{
continue;
}
else if (str[j] == '[')
{
lbrack++;
}
else if (str[j] == '(')
{
nlp++;
}
else if (str[j] == ')' && nlp-- == 0)
{
break;
}
}
if (j == n)
{
goto bad;
}
m_sub = new Regex(tks, p + 1, str.Substring(1, j - 1));
j++;
//3. Look to see if the string begins with a bracket [ . If so, find the matching ] , watching for escapes.
// Construct a ReRange for everything between the []s.
// Mark everything up to the ] as used, and go to step 9.
}
else if (str[0] == '[')
{ // range of characters
for (j = 1; j < n && str[j] != ']'; j++)
{
if (str[j] == '\\')
{
j++;
}
}
if (j == n)
{
goto bad;
}
m_sub = new ReRange(tks, str.Substring(0, j + 1));
j++;
}
//4. Look to see if the string begins with a ' or " . If so, build the contents interpreting
// escaped special characters correctly, until the matching quote is reached.
// Construct a CReStr for the contents, mark everything up to the final quote as used, and go to step 9.
else if (str[0] == '\'' || str[0] == '"')
{ // quoted string needs special treatment
StringBuilder qs = new StringBuilder();
for (j = 1; j < n && str[j] != str[0]; j++)
{
if (str[j] == '\\')
{
switch (str[++j])
{
case 'n': qs.Append('\n'); break;
case 'r': qs.Append('\r'); break;
case 't': qs.Append('\t'); break;
case 'v': qs.Append('\v'); break;
case '\\': qs.Append('\\'); break;
case '\'': qs.Append('\''); break;
case '0': qs.Append((char)0); break; // 4.7f
case '"': qs.Append('"'); break;
case '\n': break;
default: qs.Append(str[j]); break;
}
}
else
{
qs.Append(str[j]);
}
}
if (j == n)
{
goto bad;
}
j++;
m_sub = new ReStr(tks, qs.ToString());
}
//4a. Look to see if the string begins with a U' or U" . If so, build the contents interpreting
// escaped special characters correctly, until the matching quote is reached.
// Construct a ReUStr for the contents, mark everything up to the final quote as used, and go to step 9.
else if (str.StartsWith("U\"") || str.StartsWith("U'"))
{ // quoted string needs special treatment
StringBuilder qs = new StringBuilder();
for (j = 2; j < n && str[j] != str[1]; j++)
{
if (str[j] == '\\')
{
switch (str[++j])
{
case 'n': qs.Append('\n'); break;
case 'r': qs.Append('\r'); break;
case 't': qs.Append('\t'); break;
case 'v': qs.Append('\v'); break;
case '\\': qs.Append('\\'); break;
case '\'': qs.Append('\''); break;
case '"': qs.Append('"'); break;
case '\n': break;
default: qs.Append(str[j]); break;
}
}
else
{
qs.Append(str[j]);
}
}
if (j == n)
{
goto bad;
}
j++;
m_sub = new ReUStr(tks, qs.ToString());
}
//5. Look to see if the string begins with a \ .
// If so, build a ReStr for the next character (special action for ntr),
// mark it as used, and go to step 9.
else if (str[0] == '\\')
{
switch (ch = str[1])
{
case 'n': ch = '\n'; break;
case 't': ch = '\t'; break;
case 'r': ch = '\r'; break;
case 'v': ch = '\v'; break;
}
m_sub = new ReStr(tks, ch);
j = 2;
//6. Look to see if the string begins with a { .
// If so, find the matching }, lookup the symbolic name in the definitions table,
// recursively call this constructor on the contents,
// mark everything up to the } as used, and go to step 9.
}
else if (str[0] == '{')
{
for (j = 1; j < n && str[j] != '}'; j++)
{
;
}
if (j == n)
{
goto bad;
}
string ds = str.Substring(1, j - 1);
string s = (string)tks.defines[ds];
if (s == null)
{
m_sub = new ReCategory(tks, ds);
}
else
{
m_sub = new Regex(tks, p + 1, s);
}
j++;
}
else
{ // simple character at start of regular expression
//7. Look to see if the string begins with a dot.
// If so, construct a CReDot, mark the . as used, and go to step 9.
if (str[0] == '.')
{
m_sub = new ReRange(tks, "[^\n]");
}
//8. At this point we conclude that there is a simple character at the start of the regular expression.
// Construct a ReStr for it, mark it as used, and go to step 9.
else
{
m_sub = new ReStr(tks, str[0]);
}
j = 1;
}
//9. If the string is exhausted, return.
// We have a simple Regex whose m_sub contains what we can constructed.
if (j >= n)
{
return;
}
//10. If the next character is a ? , *, or +, construct a CReOpt, CReStart, or CRePlus respectively
// out of m_sub, and make m_sub point to this new class instead. Mark the character as used.
if (str[j] == '?')
{
m_sub = new ReOpt(m_sub);
j++;
}
else if (str[j] == '*')
{
m_sub = new ReStar(m_sub);
j++;
}
else if (str[j] == '+')
{
m_sub = new RePlus(m_sub);
j++;
}
// 11. If the string is exhausted, return.
if (j >= n)
{
return;
}
// 12. If the next character is a | , build a ReAlt using the m_sub we have and the rest of the string.
if (str[j] == '|')
{
m_sub = new ReAlt(tks, m_sub, p + j + 1, str.Substring(j + 1, n - j - 1));
}
// 13. Otherwise build a ReCat using the m_sub we have and the rest of the string.
else if (j < n)
{
m_sub = new ReCat(tks, m_sub, p + j, str.Substring(j, n - j));
}
return;
bad:
tks.erh.Error(new CSToolsFatalException(1, tks.sourceLineInfo(p), str, "ill-formed regular expression " + str));
}
/// <exclude/>
public ReUStr(TokensGen tks,string str)
{
m_str = str;
for (int i=0;i<str.Length;i++)
{
tks.m_tokens.UsingChar(Char.ToLower(str[i]));
tks.m_tokens.UsingChar(Char.ToUpper(str[i]));
}
}
public Dfa(TokensGen tks) : base(tks)
{
m_tokens = tks.m_tokens;
}
/// <exclude/>
public LNode(TokensGen tks)
{
m_tks = tks;
m_state = tks.NewState();
}
public ReAlt(TokensGen tks, Regex sub, int p, string str)
{
m_sub = sub;
m_alt = new Regex(tks, p, str);
}
/// <exclude/>
public static ResWds New(TokensGen tks,string str)
{
ResWds r = new ResWds();
str = str.Trim();
if (str[0]=='U')
{
r.m_upper = true;
str = str.Substring(1).Trim();
}
if (str[0]!='{' || str[str.Length-1]!='}')
goto bad;
str = str.Substring(1,str.Length-2).Trim();
string[] wds = str.Split(',');
for (int j=0;j<wds.Length;j++)
{
string w = wds[j].Trim();
string a = w;
int i = w.IndexOf(' ');
if (i>0)
{
a = w.Substring(i).Trim();
w = w.Substring(0,i);
}
r.m_wds[w] = a;
if (tks.m_tokens.tokens[a]==null)
{
TokClassDef t = new TokClassDef(tks,a,"TOKEN");
tks.m_outFile.WriteLine("//%{0}+{1}",a,t.m_yynum);
tks.m_outFile.Write("public class {0} : TOKEN",a);
tks.m_outFile.WriteLine("{ public override string yyname { get { return \""+a+"\";}}");
tks.m_outFile.WriteLine("public override int yynum { get { return "+t.m_yynum+"; }}");
tks.m_outFile.WriteLine(" public "+a+"(Lexer yyl):base(yyl) {}}");
}
}
return r;
bad:
tks.m_tokens.erh.Error(new CSToolsException(47,"bad ResWds element"));
return null;
}
public ReCat(TokensGen tks, Regex sub, int p, string str)
{
m_sub = sub;
m_next = new Regex(tks, p, str);
}
// build an NFA for a given regular expression
/// <exclude/>
public Nfa(TokensGen tks,Regex re)
: base(tks)
{
m_end = new NfaNode(tks);
re.Build(this);
}
public LNode(TokensGen tks)
{
m_tks = tks;
m_state = tks.NewState();
}
/*
Construct a Regex from a given string
1. First examine the given string.
If it is empty, there is nothing to do, so return (having cleared m_sub as a precaution).
2. Look to see if the string begins with a bracket ( . If so, find the matching ) .
This is not as simple as it might be because )s inside quotes or [] or escaped will not count.
Recursively call the constructor for the regular expression between the () s.
Mark everything up to the ) as used, and go to step 9.
3. Look to see if the string begins with a bracket [ . If so, find the matching ] , watching for escapes.
Construct a ReRange for everything between the []s.
Mark everything up to the ] as used, and go to step 9.
4. Look to see if the string begins with a ' or " . If so, build the contents interpreting
escaped special characters correctly, until the matching quote is reached.
Construct a ReStr for the contents, mark everything up to the final quote as used, and go to step 9.
4a. Look to see if the string begins with a U' or U" . If so, build the contents interpreting
escaped special characters correctly, until the matching quote is reached.
Construct a ReUStr for the contents, mark everything up to the final quote as used, and go to step 9.
5. Look to see if the string begins with a \ .
If so, build a ReStr for the next character (special action for ntr),
mark it as used, and go to step 9.
6. Look to see if the string begins with a { .
If so, find the matching }, lookup the symbolic name in the definitions table,
recursively call this constructor on the contents,
mark everything up to the } as used, and go to step 9.
7. Look to see if the string begins with a dot.
If so, construct a ReRange("^\n"), mark the . as used, and go to step 9.
8. At this point we conclude that there is a simple character at the start of the regular expression.
Construct a ReStr for it, mark it as used, and go to step 9.
9. If the string is exhausted, return.
We have a simple Regex whose m_sub contains what we can constructed.
10. If the next character is a ? , *, or +, construct a ReOpt, ReStart, or RePlus respectively
out of m_sub, and make m_sub point to this new class instead. Mark the character as used.
11. If the string is exhausted, return.
12. If the next character is a | , build a ReAlt using the m_sub we have and the rest of the string.
13. Otherwise build a ReCat using the m_sub we have and the rest of the string.
*/
/// <exclude/>
public Regex(TokensGen tks,int p,string str)
{
int n = str.Length;
int nlp = 0;
int lbrack = 0;
int quote = 0;
int j;
char ch;
//1. First examine the given string.
// If it is empty, there is nothing to do, so return (having cleared m_sub as a precaution).
m_sub = null;
if (n==0)
return;
//2. Look to see if the string begins with a bracket ( . If so, find the matching ) .
// This is not as simple as it might be because )s inside quotes or [] or escaped will not count.
// Recursively call the constructor for the regular expression between the () s.
// Mark everything up to the ) as used, and go to step 9.
else if (str[0]=='(')
{ // identify a bracketed expression
for (j=1;j<n;j++)
if (str[j]=='\\')
j++;
else if (str[j]=='[' && quote==0 && lbrack==0)
lbrack++;
else if (str[j]==']' && lbrack>0)
lbrack = 0;
else if (str[j]=='"' || str[j]=='\'')
{
if (quote==str[j])
quote = 0;
else if (quote==0)
quote = str[j];
}
else if (str[j]=='(' && quote==0 && lbrack==0)
nlp++;
else if (str[j]==')' && quote==0 && lbrack==0 && nlp--==0)
break;
if (j==n)
goto bad;
m_sub = new Regex (tks,p+1,str.Substring(1,j-1));
j++;
//3. Look to see if the string begins with a bracket [ . If so, find the matching ] , watching for escapes.
// Construct a ReRange for everything between the []s.
// Mark everything up to the ] as used, and go to step 9.
}
else if (str[0]=='[')
{ // range of characters
for (j=1;j<n && str[j]!=']';j++)
if (str[j]=='\\')
j++;
if (j==n)
goto bad;
m_sub = new ReRange(tks,str.Substring(0,j+1));
j++;
}
//4. Look to see if the string begins with a ' or " . If so, build the contents interpreting
// escaped special characters correctly, until the matching quote is reached.
// Construct a CReStr for the contents, mark everything up to the final quote as used, and go to step 9.
else if (str[0] == '\'' || str[0] == '"')
{ // quoted string needs special treatment
StringBuilder qs = new StringBuilder();
for (j=1;j<n && str[j]!=str[0];j++)
if (str[j]=='\\')
switch (str[++j])
{
case 'n': qs.Append('\n'); break;
case 'r': qs.Append('\r'); break;
case 't': qs.Append('\t'); break;
case '\\': qs.Append('\\'); break;
case '\'': qs.Append('\''); break;
case '"': qs.Append('"'); break;
case '\n': break;
default: qs.Append(str[j]); break;
}
else
qs.Append(str[j]);
if (j==n)
goto bad;
j++;
m_sub = new ReStr(tks,qs.ToString());
}
//4a. Look to see if the string begins with a U' or U" . If so, build the contents interpreting
// escaped special characters correctly, until the matching quote is reached.
// Construct a ReUStr for the contents, mark everything up to the final quote as used, and go to step 9.
else if (str.StartsWith("U\"")||str.StartsWith("U'"))
{ // quoted string needs special treatment
StringBuilder qs = new StringBuilder();
for (j=2;j<n && str[j]!=str[1];j++)
if (str[j]=='\\')
switch (str[++j])
{
case 'n': qs.Append('\n'); break;
case 'r': qs.Append('\r'); break;
case 't': qs.Append('\t'); break;
case '\\': qs.Append('\\'); break;
case '\'': qs.Append('\''); break;
case '"': qs.Append('"'); break;
case '\n': break;
default: qs.Append(str[j]); break;
}
else
qs.Append(str[j]);
if (j==n)
goto bad;
j++;
m_sub = new ReUStr(tks,qs.ToString());
}
//5. Look to see if the string begins with a \ .
// If so, build a ReStr for the next character (special action for ntr),
// mark it as used, and go to step 9.
else if (str[0]=='\\')
{
switch (ch = str[1])
{
case 'n': ch = '\n'; break;
case 't': ch = '\t'; break;
case 'r': ch = '\r'; break;
}
m_sub = new ReStr(tks,ch);
j = 2;
//6. Look to see if the string begins with a { .
// If so, find the matching }, lookup the symbolic name in the definitions table,
// recursively call this constructor on the contents,
// mark everything up to the } as used, and go to step 9.
}
else if (str[0]=='{')
{
for (j=1;j<n && str[j]!='}';j++)
;
if (j==n)
goto bad;
string ds = str.Substring(1,j-1);
string s = (string)tks.defines[ds];
if (s==null)
m_sub = new ReCategory(tks,ds);
else
m_sub = new Regex(tks,p+1,s);
j++;
}
else
{ // simple character at start of regular expression
//7. Look to see if the string begins with a dot.
// If so, construct a CReDot, mark the . as used, and go to step 9.
if (str[0]=='.')
m_sub = new ReRange(tks,"[^\n]");
//8. At this point we conclude that there is a simple character at the start of the regular expression.
// Construct a ReStr for it, mark it as used, and go to step 9.
else
m_sub = new ReStr(tks,str[0]);
j = 1;
}
//9. If the string is exhausted, return.
// We have a simple Regex whose m_sub contains what we can constructed.
if (j>=n)
return;
//10. If the next character is a ? , *, or +, construct a CReOpt, CReStart, or CRePlus respectively
// out of m_sub, and make m_sub point to this new class instead. Mark the character as used.
if (str[j]=='?')
{
m_sub = new ReOpt(m_sub);
j++;
}
else if (str[j]=='*')
{
m_sub = new ReStar(m_sub);
j++;
}
else if (str[j]=='+')
{
m_sub = new RePlus(m_sub);
j++;
}
// 11. If the string is exhausted, return.
if (j>=n)
return;
// 12. If the next character is a | , build a ReAlt using the m_sub we have and the rest of the string.
if (str[j]=='|')
m_sub = new ReAlt(tks,m_sub,p+j+1,str.Substring(j+1,n-j-1));
// 13. Otherwise build a ReCat using the m_sub we have and the rest of the string.
else if (j<n)
m_sub = new ReCat(tks,m_sub,p+j,str.Substring(j,n-j));
return;
bad:
tks.erh.Error(new CSToolsFatalException(1,tks.sourceLineInfo(p),str,"ill-formed regular expression "+str));
}
private void Set(TokensGen tks, char ch)
{
this.m_map[(object)ch] = (object)true;
tks.m_tokens.UsingChar(ch);
}
