/*
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));
}
// 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);
}
/// <exclude/>
public void AddArcEx(Regex re,NfaNode next)
{
m_arcs.Add(new ArcEx(re,next));
}
public RePlus(Regex sub)
{
m_sub = sub;
}
public ReStar(Regex sub)
{
m_sub = sub;
}
public ReOpt(Regex sub)
{
m_sub = sub;
}
/// <exclude/>
public ReCat(TokensGen tks,Regex sub, int p, string str)
{
m_sub = sub;
m_next = new Regex(tks,p,str);
}
public ReAlt(TokensGen tks,Regex sub,int p,string str)
{
m_sub = sub;
m_alt = new Regex(tks,p,str);
}
public ArcEx(Regex re,NfaNode next)
{
m_ref=re; m_next=next;
}
/// <summary>
/// 是否为时间格式
/// </summary>
/// <returns>是则返加true 不是则返回 false</returns>
public static bool IsTime(string timeval)
{
return(Regex.IsMatch(timeval, @"^((([0-1]?[0-9])|(2[0-3])):([0-5]?[0-9])(:[0-5]?[0-9])?)$"));
}
