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 18:02:00 +08:00
parent 3a441b270c
commit ed7972f1da
5 changed files with 66 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files
StgSharp.RegularAnalysis/Abstraction/AbstractSyntaxTree.cs
+5
View File
@@ -43,6 +43,11 @@ namespace StgSharp.RegularAnalysis.Abstraction
private TNode _root;
public HashSet<TNode> AllNodes
{
get { return _allToken; }
}
public int Count
{
get { return _allToken.Count; }
StgSharp.RegularAnalysis/Abstraction/CompileStack.cs
+1 -1
View File
@@ -59,7 +59,7 @@ namespace StgSharp.RegularAnalysis.Abstraction
public int OperatorInDepthCount
{
[MethodImpl(MethodImplOptions.AggressiveInlining)]
get => _operators.Count - _depthMarks.Peek().OperandPos;
get => _operators.Count - _depthMarks.Peek().OperatorPos;
}
public int Depth
StgSharp.RegularAnalysis/Abstraction/ISyntaxNode.cs
+2
View File
@@ -53,6 +53,8 @@ namespace StgSharp.RegularAnalysis.Abstraction
TNode Right { get; set; }
TNode Parent { get; set; }
static abstract TNode Empty { get; }
Token<TLabel> Source { get; }
StgSharp.RegularAnalysis/TextRegex/RegexAstNode.cs
+2
View File
@@ -66,6 +66,7 @@ namespace StgSharp.RegularAnalysis.TextRegex
get;
set
{
field?.Parent = null;
field = value;
value.Parent = this;
}
@@ -76,6 +77,7 @@ namespace StgSharp.RegularAnalysis.TextRegex
get;
set
{
field?.Parent = null;
field = value;
value.Parent = this;
}
StgSharp.RegularAnalysis/TextRegex/RegexInterpreter.cs
+56 -24
View File
@@ -41,42 +41,59 @@ namespace StgSharp.RegularAnalysis.TextRegex
private AbstractSyntaxTree<RegexAstNode, RegexElementLabel> _tree = new();
private CompileStack<RegexAstNode, RegexElementLabel> _stack = new();
private string _source;
public AbstractSyntaxTree<RegexAstNode, RegexElementLabel> Analyze(
string regex
public static AbstractSyntaxTree<RegexAstNode, RegexElementLabel> Interpret(
string regex
)
{
RegexTokenReader reader = new RegexTokenReader(regex);
RegexInterpreter interpreter = new RegexInterpreter();
interpreter._source = regex;
interpreter.Analyze();
interpreter.OptimizeTree();
return interpreter._tree;
}
private void Analyze()
{
RegexTokenReader reader = new RegexTokenReader(_source);
TokenParser<RegexElementLabel, RegexElementLabel> lexer = reader.Pipe(() => new RegexTokenParser());
while (lexer.TryReadToken(out Token<RegexElementLabel> token))
{
if ((token.Flag & RegexElementLabel.SEQUENCE) != 0)
{
Console.WriteLine($"Unit {token.Value}");
// char or charset
Console.WriteLine($"Unit {token.Value}");
_stack.PushOperand(new RegexAstNode(token));
} else if ((token.Flag & RegexElementLabel.GROUP_BEGIN) != 0)
{
Console.WriteLine($"Group begin {token.Value}");
_stack.PushOperator(new RegexAstNode(token));
_stack.IncreaseDepth(0);
Console.WriteLine(/**/_stack.OperatorInDepthCount/**/);
} else if ((token.Flag & RegexElementLabel.GROUP_END) != 0)
{
// close all nodes here
Console.WriteLine($"Group end {token.Value}");
// TODO: close all nodes here
while (_stack.TryPopOperator(out RegexAstNode? op))
{
RegexAstNode op;
while (_stack.TryPopOperator(out op)) {
ProcessOperator(op);
}
if (_stack.OperandInDepthCount != 1) {
throw new InvalidOperationException();
}
_stack.DecreaseDepth();
op = _stack.PopOperator();
if (op.Source.Flag != RegexElementLabel.GROUP_BEGIN) {
throw new InvalidOperationException("Unmatched group begin and end symbol");
}
op.Right = _stack.PopOperand();
_stack.PushOperand(op);
} else if ((token.Flag & RegexElementLabel.OPERATOR) != 0)
{
Console.WriteLine($"Operator {token.Value} with flag {token.Flag}");
Console.Write($"Operator {token.Value} with flag {token.Flag}, ");
Console.WriteLine(/**/_stack.OperatorInDepthCount/**/);
if (_stack.OperatorInDepthCount == 0)
{
// the first operator in stack
@@ -117,13 +134,35 @@ namespace StgSharp.RegularAnalysis.TextRegex
if (_stack.OperandInDepthCount == 1 && _stack.OperatorInDepthCount == 0)
{
_tree.Root = _stack.PopOperand();
return _tree;
return;
}
throw new InvalidOperationException("Invalid regular expression syntax.");
}
throw new InvalidOperationException("Invalid regular expression syntax.");
}
private void OptimizeTree(
)
{
// phase0P: ALT flatten
int opCount = 0;
do
{
opCount = 1;
} while (opCount != 0);
// phase1: ALT merge
opCount = 0;
do
{
opCount = 1;
} while (opCount != 0);
// phase2: CONCAT merge
// phase3: scan container and remove no parent nodes
_tree.AllNodes.RemoveWhere(node => node.Parent == null && node != _tree.Root);
}
private void ProcessOperator(
RegexAstNode op
)
@@ -134,19 +173,12 @@ namespace StgSharp.RegularAnalysis.TextRegex
if (_stack.TryPopOperand(out RegexAstNode? _1) &&
_stack.TryPopOperand(out RegexAstNode? _2))
{
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);
}
op.Left = _2;
op.Right = _1;
_stack.PushOperand(op);
_tree.AddNode(op);
_tree.AddNode(_1);
_tree.AddNode(_2);
} else
{
throw new InvalidOperationException("Insufficient operands for CONCAT operator.");