Beispiel #1
0
        /*
         * 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(CultureInfo.InvariantCulture), 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++);
            }
        }
Beispiel #2
0
        /*
         * FC computation and shortcut cases for each node type
         */
        private void CalculateFC(int NodeType, RegexNode node, int CurIndex)
        {
            bool ci  = false;
            bool rtl = false;

            if (NodeType <= RegexNode.Ref)
            {
                if ((node._options & RegexOptions.IgnoreCase) != 0)
                {
                    ci = true;
                }
                if ((node._options & RegexOptions.RightToLeft) != 0)
                {
                    rtl = true;
                }
            }

            switch (NodeType)
            {
            case RegexNode.Concatenate | BeforeChild:
            case RegexNode.Alternate | BeforeChild:
            case RegexNode.Testref | BeforeChild:
            case RegexNode.Loop | BeforeChild:
            case RegexNode.Lazyloop | BeforeChild:
                break;

            case RegexNode.Testgroup | BeforeChild:
                if (CurIndex == 0)
                {
                    SkipChild();
                }
                break;

            case RegexNode.Empty:
                PushFC(new RegexFC(true));
                break;

            case RegexNode.Concatenate | AfterChild:
                if (CurIndex != 0)
                {
                    RegexFC child = PopFC();
                    RegexFC cumul = TopFC();

                    _failed = !cumul.AddFC(child, true);
                }

                if (!TopFC()._nullable)
                {
                    _skipAllChildren = true;
                }
                break;

            case RegexNode.Testgroup | AfterChild:
                if (CurIndex > 1)
                {
                    RegexFC child = PopFC();
                    RegexFC cumul = TopFC();

                    _failed = !cumul.AddFC(child, false);
                }
                break;

            case RegexNode.Alternate | AfterChild:
            case RegexNode.Testref | AfterChild:
                if (CurIndex != 0)
                {
                    RegexFC child = PopFC();
                    RegexFC cumul = TopFC();

                    _failed = !cumul.AddFC(child, false);
                }
                break;

            case RegexNode.Loop | AfterChild:
            case RegexNode.Lazyloop | AfterChild:
                if (node._m == 0)
                {
                    TopFC()._nullable = true;
                }
                break;

            case RegexNode.Group | BeforeChild:
            case RegexNode.Group | AfterChild:
            case RegexNode.Capture | BeforeChild:
            case RegexNode.Capture | AfterChild:
            case RegexNode.Greedy | BeforeChild:
            case RegexNode.Greedy | AfterChild:
                break;

            case RegexNode.Require | BeforeChild:
            case RegexNode.Prevent | BeforeChild:
                SkipChild();
                PushFC(new RegexFC(true));
                break;

            case RegexNode.Require | AfterChild:
            case RegexNode.Prevent | AfterChild:
                break;

            case RegexNode.One:
            case RegexNode.Notone:
                PushFC(new RegexFC(node._ch, NodeType == RegexNode.Notone, false, ci));
                break;

            case RegexNode.Oneloop:
            case RegexNode.Onelazy:
                PushFC(new RegexFC(node._ch, false, node._m == 0, ci));
                break;

            case RegexNode.Notoneloop:
            case RegexNode.Notonelazy:
                PushFC(new RegexFC(node._ch, true, node._m == 0, ci));
                break;

            case RegexNode.Multi:
                if (node._str.Length == 0)
                {
                    PushFC(new RegexFC(true));
                }
                else if (!rtl)
                {
                    PushFC(new RegexFC(node._str[0], false, false, ci));
                }
                else
                {
                    PushFC(new RegexFC(node._str[node._str.Length - 1], false, false, ci));
                }
                break;

            case RegexNode.Set:
                PushFC(new RegexFC(node._str, false, ci));
                break;

            case RegexNode.Setloop:
            case RegexNode.Setlazy:
                PushFC(new RegexFC(node._str, node._m == 0, ci));
                break;

            case RegexNode.Ref:
                PushFC(new RegexFC(RegexCharClass.AnyClass, true, false));
                break;

            case RegexNode.Nothing:
            case RegexNode.Bol:
            case RegexNode.Eol:
            case RegexNode.Boundary:
            case RegexNode.Nonboundary:
            case RegexNode.ECMABoundary:
            case RegexNode.NonECMABoundary:
            case RegexNode.Beginning:
            case RegexNode.Start:
            case RegexNode.EndZ:
            case RegexNode.End:
                PushFC(new RegexFC(true));
                break;

            default:
                throw new ArgumentException(SR.GetString(SR.UnexpectedOpcode, NodeType.ToString(CultureInfo.CurrentCulture)));
            }
        }
Beispiel #3
0
        internal RegexReplacement(String rep, RegexNode concat, Hashtable _caps)
        {
#endif
            StringBuilder sb;
            List <String> strings;
            List <Int32>  rules;
            int           slot;

            _rep = rep;

            if (concat.Type() != RegexNode.Concatenate)
            {
                throw new ArgumentException(SR.GetString(SR.ReplacementError));
            }

            sb      = new StringBuilder();
            strings = new List <String>();
            rules   = new List <Int32>();

            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;
                    }
                    slot = child._m;

                    if (_caps != null && slot >= 0)
                    {
                        slot = (int)_caps[slot];
                    }

                    rules.Add(-Specials - 1 - slot);
                    break;

                default:
                    throw new ArgumentException(SR.GetString(SR.ReplacementError));
                }
            }

            if (sb.Length > 0)
            {
                rules.Add(strings.Count);
                strings.Add(sb.ToString());
            }

            _strings = strings;
            _rules   = rules;
        }
Beispiel #4
0
 internal RegexTree(RegexNode root, Dictionary <Int32, Int32> caps, Int32[] capnumlist, int captop, Dictionary <String, Int32> capnames, String[] capslist, RegexOptions opts)
Beispiel #5
0
        /*
         * Since RegexReplacement shares the same parser as Regex,
         * the constructor takes a RegexNode which is a concatenation
         * of constant strings and backreferences.
         */
#if SILVERLIGHT
        internal RegexReplacement(String rep, RegexNode concat, Dictionary <Int32, Int32> _caps)
        {