youmuliu/stgsharp
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

regex interpreter now can build partial correct AST (i forget groups), but in bad shape. building an optimizer.

Browse Source
This commit is contained in:
Youmu Liu
2026-02-06 17:07:03 +08:00
parent 1bfb8b67c4
commit 3a441b270c
13 changed files with 236 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files
StgSharp.RegularAnalysis/TextRegex/RegexInterpreter.cs
+62 -1
View File
@@ -53,13 +53,18 @@ namespace StgSharp.RegularAnalysis.TextRegex
{
if ((token.Flag & RegexElementLabel.SEQUENCE) != 0)
{
Console.WriteLine($"Unit {token.Value}");
// char or charset
_stack.PushOperand(new RegexAstNode(token));
} else if ((token.Flag & RegexElementLabel.GROUP_BEGIN) != 0)
{
Console.WriteLine($"Group begin {token.Value}");
_stack.IncreaseDepth(0);
} else if ((token.Flag & RegexElementLabel.GROUP_END) != 0)
{
Console.WriteLine($"Group end {token.Value}");
// TODO: close all nodes here
while (_stack.TryPopOperator(out RegexAstNode? op))
{
@@ -69,14 +74,61 @@ namespace StgSharp.RegularAnalysis.TextRegex
throw new InvalidOperationException();
}
_stack.DecreaseDepth();
} else if ((token.Flag & RegexElementLabel.OPERATOR) != 0) { }
} else if ((token.Flag & RegexElementLabel.OPERATOR) != 0)
{
Console.WriteLine($"Operator {token.Value} with flag {token.Flag}");
if (_stack.OperatorInDepthCount == 0)
{
// the first operator in stack
_stack.PushOperator(new RegexAstNode(token));
} else
{
// process operator by precedence
while (_stack.TryPeekOperator(out RegexAstNode topOp))
{
(bool Valid, int Value) cmp = RegexAstNode.ComparePrecedence(token.Flag, topOp.Source.Flag);
if (!cmp.Valid) {
throw new InvalidOperationException("Operators with same precedence are not supported.");
}
if (cmp.Value < 0)
{
_stack.PopOperator();
ProcessOperator(topOp);
continue;
} else
{
break;
}
}
_stack.PushOperator(new RegexAstNode(token));
}
}
}
Console.WriteLine(1);
if (_stack.Depth == 1)
{
if (_stack.OperandInDepthCount != 1 || _stack.OperatorInDepthCount != 0)
{
while (_stack.TryPopOperator(out RegexAstNode? op)) {
ProcessOperator(op);
}
}
if (_stack.OperandInDepthCount == 1 && _stack.OperatorInDepthCount == 0)
{
_tree.Root = _stack.PopOperand();
return _tree;
}
throw new InvalidOperationException("Invalid regular expression syntax.");
}
throw new InvalidOperationException("Invalid regular expression syntax.");
}
private void ProcessOperator(
RegexAstNode op
)
{
Console.WriteLine($"Processing a {op.Source.Flag} operator");
if (op.EqualityTypeConvert == RegexElementLabel.CONCAT)
{
if (_stack.TryPopOperand(out RegexAstNode? _1) &&
@@ -85,11 +137,15 @@ namespace StgSharp.RegularAnalysis.TextRegex
if (RegexAstNode.TryMergeUnit(_2, _1, out RegexAstNode? _3))
{
_stack.PushOperand(_3);
_tree.AddNode(_3);
} else
{
op.Left = _2;
op.Right = _1;
_stack.PushOperand(op);
_tree.AddNode(op);
_tree.AddNode(_1);
_tree.AddNode(_2);
}
} else
{
@@ -101,6 +157,8 @@ namespace StgSharp.RegularAnalysis.TextRegex
{
op.Left = _1;
_stack.PushOperand(op);
_tree.AddNode(_1);
_tree.AddNode(op);
} else
{
throw new InvalidOperationException("Insufficient operands for CONCAT operator.");
@@ -113,6 +171,9 @@ namespace StgSharp.RegularAnalysis.TextRegex
op.Left = _1;
op.Right = _2;
_stack.PushOperand(op);
_tree.AddNode(op);
_tree.AddNode(_1);
_tree.AddNode(_2);
} else
{
throw new InvalidOperationException("Insufficient operands for CONCAT operator.");