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

add COUNT and CONCAT ir emitter

Browse Source
This commit is contained in:
Youmu Liu
2026-06-17 18:32:55 +08:00
parent a7bb3209dd
commit 5570c6f338
6 changed files with 409 additions and 580 deletions
Download Patch File Download Diff File Expand all files Collapse all files
StgSharp.RegularAnalysis/TextRegex/RegexInterpreter.IRGenerate.cs
+172 -30
View File
@@ -28,8 +28,12 @@
using StgSharp.RegularAnalysis.Abstraction;
using System;
using System.Collections.Generic;
using System.ComponentModel.Design;
using System.Data;
using System.Diagnostics.Metrics;
using System.Linq;
using System.Reflection.Emit;
using System.Runtime.CompilerServices;
using System.Text;
using System.Threading.Tasks;
using System.Xml;
@@ -45,27 +49,68 @@ namespace StgSharp.RegularAnalysis.TextRegex
AbstractSyntaxTree<RegexAstNode, RegexElementLabel> tree
)
{
return GenerateIRRecursion(tree.Root, 0);
return GenerateIRRecursion(tree.Root, 0).ToList();
}
private List<RegexIR> GenerateIRRecursion(
RegexAstNode node,
int depth
[MethodImpl(MethodImplOptions.AggressiveInlining)]
private void EmitCount(
int curDepth,
RegexIRGenerator ir,
int min,
int max,
bool isGreedy,
RegexAstNode right
)
{
if (depth > 1024) {
bool isRightOperator = !RegexAstNode.IsNullOrEmpty(right) &&
(right.Source.Flag & RegexElementLabel.OPERATOR) != 0;
if (isRightOperator) // seq to be count is complex
{
ir.EmitCount(min, max, isGreedy, 1);
RegexIRGenerator rightIR = GenerateIRRecursion(right, curDepth + 1);
ir.EmitTry(rightIR.Count);
ir.EmitIRStream(rightIR);
} else
{
RegexElementLabel count_source_mask = right.Source.Flag;
switch (count_source_mask)
{
case RegexElementLabel.UNIT:
ir.EmitCountSeq(min, max, isGreedy, right.Source.Value);
break;
case RegexElementLabel.UNIT_SET:
ir.EmitCountSet(min, max, isGreedy, right.Source.Value);
break;
case RegexElementLabel.UNIT_SPAN:
ir.EmitCountSet(min, max, isGreedy, right.Source.Value);
break;
default:
throw new InvalidOperationException($"Unexpected regex element type for count operator: {count_source_mask}");
}
}
}
private RegexIRGenerator GenerateIRRecursion(
RegexAstNode node,
int depth
)
{
#pragma warning disable IDE0010
if (depth > 1024)
{
throw new OverflowException("Regex parse exceeds maximum recursion depth");
}
int curDepth = depth;
RegexAstNode left = node.Left;
RegexAstNode right = node.Right;
List<RegexIR> leftIR = EmptyIR;
List<RegexIR> rightIR = EmptyIR;
RegexIRGenerator leftIR;
RegexIRGenerator rightIR;
bool isLeftOperator = !RegexAstNode.IsNullOrEmpty(left) &&
(left.Source.Flag & RegexElementLabel.OPERATOR) != 0;
(left.Source.Flag & RegexElementLabel.VAST_OPERATOR) != 0;
bool isRightOperator = !RegexAstNode.IsNullOrEmpty(right) &&
(right.Source.Flag & RegexElementLabel.OPERATOR) != 0;
(right.Source.Flag & RegexElementLabel.VAST_OPERATOR) != 0;
Token<RegexElementLabel> token = node.Source;
@@ -73,30 +118,127 @@ namespace StgSharp.RegularAnalysis.TextRegex
{
case RegexElementLabel.GROUP_BEGIN:
string group_source = token.Value;
bool is_group = group_source.Length == 1;
string group_name = group_source.StartsWith(@"(?<", StringComparison.Ordinal) ?
group_source[3..^2] :
string.Empty;
List<RegexIR> groupIR = [];
// process group type
bool is_group = group_source.Length == 1; // simple group: (pattern)
string group_name =
group_source.StartsWith(@"(?<", StringComparison.Ordinal) ?
group_source[3..^2] : /* named group (?<name>) */
string.Empty; /* unnamed group (?:) */
RegexIRGenerator groupIR = new();
rightIR = GenerateIRRecursion(right, curDepth + 1);
groupIR.Add(RegexIR.Try(rightIR.Count));
groupIR.AddRange(rightIR);
groupIR.Add(RegexIR.Condition(2, 1));
groupIR.Add(RegexIR.PopLevel(false));
groupIR.Add(RegexIR.PopLevel(true));
break;
case RegexElementLabel.COUNT:
break;
case RegexElementLabel.CONCAT:
break;
case RegexElementLabel.ALT:
break;
default:
break;
}
groupIR.EmitTry(rightIR.Count);
groupIR.EmitIRStream(rightIR);
groupIR.EmitCondition(2, 1);
groupIR.EmitPopLevel(false);
return [];
// a grouping op needed here, before popping
groupIR.EmitPopLevel(true);
return groupIR;
case RegexElementLabel.COUNT:
string count_source = token.Value;
bool isGreedy = !(count_source.Length > 1 && !count_source.EndsWith('?'));
ReadOnlySpan<char> count_source_span = count_source;
RegexIRGenerator countIR = new();
if (count_source_span[0] == '*') // infinite count
{
EmitCount(curDepth, countIR, 0, int.MaxValue, isGreedy, right);
} else if (count_source_span[0] == '+') // at least one
{
EmitCount(curDepth, countIR, 1, int.MaxValue, isGreedy, right);
} else if (count_source_span[0] == '?') // zero or one
{
EmitCount(curDepth, countIR, 0, 1, isGreedy, right);
} else if (count_source_span[0] == '{') // zero or one, non-greedy
{
int comma_pos = count_source_span.IndexOf(',');
if (comma_pos == -1)
{
if (int.TryParse(count_source[1..(isGreedy ? ^3 : ^2)], out int count))
{
EmitCount(curDepth, countIR, count, count, isGreedy, right);
} else
{
throw new InvalidCastException($"Invalid count format: {count_source}");
}
} else
{
if (int.TryParse(count_source[1..comma_pos], out int min) &&
int.TryParse(count_source[(comma_pos + 1)..(isGreedy ? ^3 : ^2)], out int max))
{
EmitCount(curDepth, countIR, min, max, isGreedy, right);
} else
{
throw new InvalidCastException($"Invalid count format: {count_source}");
}
}
} else
{
throw new InvalidOperationException();
} // unknown count format
return countIR;
case RegexElementLabel.CONCAT:
RegexIRGenerator concat = new();
// process left node
if (RegexAstNode.IsNullOrEmpty(left))
{
// do nothing
} else if ((left.Source.Flag & RegexElementLabel.SEQUENCE) != 0)
{
switch (left.Source.Flag)
{
case RegexElementLabel.UNIT_SPAN:
goto case RegexElementLabel.UNIT_SPAN;
case RegexElementLabel.UNIT:
concat.EmitMatchSeq(left.Source.Value);
break;
case RegexElementLabel.UNIT_SET:
concat.EmitMatchSet(left.Source.Value);
break;
default:
break;
}
} else if ((left.Source.Flag & RegexElementLabel.VAST_OPERATOR) != 0)
{
leftIR = GenerateIRRecursion(left, curDepth + 1);
concat.EmitIRStream(leftIR);
}
// process right node
if (RegexAstNode.IsNullOrEmpty(right))
{
// do nothing
} else if ((right.Source.Flag & RegexElementLabel.SEQUENCE) != 0)
{
switch (right.Source.Flag)
{
case RegexElementLabel.UNIT_SPAN:
goto case RegexElementLabel.UNIT;
case RegexElementLabel.UNIT:
concat.EmitMatchSeq(right.Source.Value);
break;
case RegexElementLabel.UNIT_SET:
concat.EmitMatchSet(right.Source.Value);
break;
default:
break;
}
} else if ((right.Source.Flag & RegexElementLabel.VAST_OPERATOR) != 0)
{
rightIR = GenerateIRRecursion(right, curDepth + 1);
concat.EmitIRStream(rightIR);
}
return concat;
case RegexElementLabel.ALT:
return null!;
default:
return null!;
}
#pragma warning restore IDE0010
}
}