private readonly List <int> _rules; // negative -> group #, positive -> string #
/// <summary>
/// Since RegexReplacement shares the same parser as Regex,
/// the constructor takes a RegexNode which is a concatenation
/// of constant strings and backreferences.
/// </summary>
public RegexReplacement(string rep, RegexNode concat, Dictionary <int, int> _caps)
{
if (concat.Type() != RegexNode.Concatenate)
{
throw new ArgumentException(SR.ReplacementError);
}
Span <char> buffer = stackalloc char[ReplaceBufferSize];
ValueStringBuilder vsb = new ValueStringBuilder(buffer);
List <string> strings = new List <string>();
List <int> rules = new List <int>();
for (int i = 0; i < concat.ChildCount(); i++)
{
RegexNode child = concat.Child(i);
switch (child.Type())
{
case RegexNode.Multi:
vsb.Append(child.Str);
break;
case RegexNode.One:
vsb.Append(child.Ch);
break;
case RegexNode.Ref:
if (vsb.Length > 0)
{
rules.Add(strings.Count);
strings.Add(vsb.ToString());
vsb.Length = 0;
}
int slot = child.M;
if (_caps != null && slot >= 0)
{
slot = (int)_caps[slot];
}
rules.Add(-Specials - 1 - slot);
break;
default:
throw new ArgumentException(SR.ReplacementError);
}
}
if (vsb.Length > 0)
{
rules.Add(strings.Count);
strings.Add(vsb.ToString());
}
Pattern = rep;
_strings = strings;
_rules = rules;
}
private readonly List <int> _rules; // negative -> group #, positive -> string #
/// <summary>
/// Since RegexReplacement shares the same parser as Regex,
/// the constructor takes a RegexNode which is a concatenation
/// of constant strings and backreferences.
/// </summary>
internal RegexReplacement(string rep, RegexNode concat, Dictionary <int, int> _caps)
{
if (concat.Type() != RegexNode.Concatenate)
{
throw new ArgumentException(SR.ReplacementError);
}
StringBuilder sb = new StringBuilder();
List <string> strings = new List <string>();
List <int> rules = new List <int>();
for (int i = 0; i < concat.ChildCount(); i++)
{
RegexNode child = concat.Child(i);
switch (child.Type())
{
case RegexNode.Multi:
sb.Append(child._str);
break;
case RegexNode.One:
sb.Append(child._ch);
break;
case RegexNode.Ref:
if (sb.Length > 0)
{
rules.Add(strings.Count);
strings.Add(sb.ToString());
sb.Length = 0;
}
int slot = child._m;
if (_caps != null && slot >= 0)
{
slot = (int)_caps[slot];
}
rules.Add(-Specials - 1 - slot);
break;
default:
throw new ArgumentException(SR.ReplacementError);
}
}
if (sb.Length > 0)
{
rules.Add(strings.Count);
strings.Add(sb.ToString());
}
_rep = rep;
_strings = strings;
_rules = rules;
}
/*
* Yet another related computation: it takes a RegexTree and computes the
* leading anchors that it encounters.
*/
internal static int Anchors(RegexTree tree)
{
RegexNode curNode;
RegexNode concatNode = null;
int nextChild = 0;
int result = 0;
curNode = tree._root;
for (; ;)
{
switch (curNode._type)
{
case RegexNode.Concatenate:
if (curNode.ChildCount() > 0)
{
concatNode = curNode;
nextChild = 0;
}
break;
case RegexNode.Greedy:
case RegexNode.Capture:
curNode = curNode.Child(0);
concatNode = null;
continue;
case RegexNode.Bol:
case RegexNode.Eol:
case RegexNode.Boundary:
case RegexNode.ECMABoundary:
case RegexNode.Beginning:
case RegexNode.Start:
case RegexNode.EndZ:
case RegexNode.End:
return(result | AnchorFromType(curNode._type));
case RegexNode.Empty:
case RegexNode.Require:
case RegexNode.Prevent:
break;
default:
return(result);
}
if (concatNode == null || nextChild >= concatNode.ChildCount())
{
return(result);
}
curNode = concatNode.Child(nextChild++);
}
}
/*
* This is a related computation: it takes a RegexTree and computes the
* leading substring if it see one. It's quite trivial and gives up easily.
*/
internal static RegexPrefix Prefix(RegexTree tree)
{
RegexNode curNode;
RegexNode concatNode = null;
int nextChild = 0;
curNode = tree._root;
for (; ;)
{
switch (curNode._type)
{
case RegexNode.Concatenate:
if (curNode.ChildCount() > 0)
{
concatNode = curNode;
nextChild = 0;
}
break;
case RegexNode.Greedy:
case RegexNode.Capture:
curNode = curNode.Child(0);
concatNode = null;
continue;
case RegexNode.Oneloop:
case RegexNode.Onelazy:
if (curNode._m > 0)
{
string pref = string.Empty.PadRight(curNode._m, curNode._ch);
return(new RegexPrefix(pref, 0 != (curNode._options & RegexOptions.IgnoreCase)));
}
else
{
return(RegexPrefix.Empty);
}
case RegexNode.One:
return(new RegexPrefix(curNode._ch.ToString(), 0 != (curNode._options & RegexOptions.IgnoreCase)));
case RegexNode.Multi:
return(new RegexPrefix(curNode._str, 0 != (curNode._options & RegexOptions.IgnoreCase)));
case RegexNode.Bol:
case RegexNode.Eol:
case RegexNode.Boundary:
case RegexNode.ECMABoundary:
case RegexNode.Beginning:
case RegexNode.Start:
case RegexNode.EndZ:
case RegexNode.End:
case RegexNode.Empty:
case RegexNode.Require:
case RegexNode.Prevent:
break;
default:
return(RegexPrefix.Empty);
}
if (concatNode == null || nextChild >= concatNode.ChildCount())
{
return(RegexPrefix.Empty);
}
curNode = concatNode.Child(nextChild++);
}
}