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

Optimizer is almost ok. Will go on source generate

Browse Source
This commit is contained in:
Youmu Liu
2026-02-13 18:36:59 +08:00
parent 6c78373375
commit 27add7c6e8
5 changed files with 163 additions and 53 deletions
Download Patch File Download Diff File Expand all files Collapse all files
StgSharp.RegularAnalysis/TextRegex/RegexInterpreter.Optimize.cs
+89 -19
View File
@@ -35,14 +35,15 @@ using System.Runtime.CompilerServices;
using System.Security.Permissions;
using System.Text;
using System.Threading.Tasks;
using System.Xml.Linq;
namespace StgSharp.RegularAnalysis.TextRegex
{
public partial class RegexInterpreter
{
private static List<RegexAstNode> FlattenAlt(
RegexAstNode root
private static RegexAstNode FlattenAlt(
RegexAstNode root
)
{
TreeEnumerator<RegexAstNode, RegexElementLabel> enumerator = new(root);
@@ -59,7 +60,13 @@ namespace StgSharp.RegularAnalysis.TextRegex
}
current.Parent.RemoveChild(current);
}
return cases;
for (int i = 0; i < cases.Count - 1; i++)
{
cases[i].AppendNode(cases[i + 1]);
cases[i].Parent = root;
}
cases[^1].Parent = root;
return cases[0];
}
private static bool MergeConcat(
@@ -69,6 +76,25 @@ namespace StgSharp.RegularAnalysis.TextRegex
if (root.EqualityTypeConvert != RegexElementLabel.CONCAT) {
return false;
}
/*
* concat → 'a'
* ╱
* 'a'
*/
if ((root.Left.EqualityTypeConvert & RegexElementLabel.SEQUENCE) != 0 &&
RegexAstNode.IsNullOrEmpty(root.Right))
{
root.RemoveAndAsChild(true);
return true;
}
if ((root.Right.EqualityTypeConvert & RegexElementLabel.SEQUENCE) != 0 &&
RegexAstNode.IsNullOrEmpty(root.Left))
{
root.RemoveAndAsChild(false);
return true;
}
/*
* concat → concat
* ╱ ╲ /
@@ -184,7 +210,9 @@ namespace StgSharp.RegularAnalysis.TextRegex
)
{
UnionAlt union = new(0, cases);
return union.AsNormal();
RegexAstNode root = union.AsNormal();
return root;
}
private class UnionAlt : RegexAstNode
@@ -196,11 +224,21 @@ namespace StgSharp.RegularAnalysis.TextRegex
int level,
RegexAstNode node
)
: base(new Token<RegexElementLabel>(string.Empty, 0, 0, 0))
: base(node.Source)
{
Level = level;
Cases = new List<RegexAstNode>();
Cases.Add(node.Right ?? Empty);
RegexAstNode right = IsNullOrEmpty(node.Right) ?
new RegexAstNode(
new Token<RegexElementLabel>(node.Value, 0, node.Source.Column,
RegexElementLabel.UNIT)
) :
node.Right;
Cases.Add(right);
node.PrependNode(Empty);
node.AppendNode(Empty);
UnionBegin = node;
UnionEnd = node;
}
@@ -209,7 +247,7 @@ namespace StgSharp.RegularAnalysis.TextRegex
int level,
List<RegexAstNode> c
)
: base(new Token<RegexElementLabel>(string.Empty, 0, 0, 0))
: base(c[0].Source)
{
Cases = c;
Level = level;
@@ -233,6 +271,17 @@ namespace StgSharp.RegularAnalysis.TextRegex
List<RegexAstNode> buffer = [];
for (int i = 0; i < Cases.Count; i++)
{
if ((Cases[i].EqualityTypeConvert & RegexElementLabel.SEQUENCE) != 0)
{
RegexAstNode concat = new RegexAstNode(new Token<RegexElementLabel>(string.Empty,
0,
Cases[i].Source.Column,
RegexElementLabel.CONCAT));
concat.Left = Cases[i];
Cases[i].Parent = concat;
concat.Right = Empty;
Cases[i] = concat;
}
RegexAstNode current = Cases[i];
int j = 0;
for (; j < buffer.Count; j++)
@@ -258,11 +307,11 @@ namespace StgSharp.RegularAnalysis.TextRegex
buffer.Add(current);
}
}
Cases = buffer;
if (Cases.Count != buffer.Count) {
Cases = buffer;
}
// TODO problem1 没处理case链表
// TODO problem2 没有处理立刻返回策略
if (Cases.Count == 1)
if (buffer.Count == 1)
{
// if there is only one case,
// then we suppose that its contained cases has a same beginning,
@@ -276,6 +325,7 @@ namespace StgSharp.RegularAnalysis.TextRegex
{
alt.UnionBegin.Parent = UnionEnd;
UnionEnd.Right = alt.UnionBegin;
UnionEnd = alt.UnionBegin;
}
Cases = alt.Cases;
@@ -291,18 +341,34 @@ namespace StgSharp.RegularAnalysis.TextRegex
Cases[i] = union.AsNormal();
}
}
RegexAstNode branch = new RegexAstNode(new Token<RegexElementLabel>(string.Empty, 0,
UnionEnd.Source.Column,
RegexElementLabel.ALT));
UnionEnd.Right = branch;
branch.Parent = UnionEnd;
UnionEnd = branch;
RegexAstNode branch;
if (IsNullOrEmpty(UnionEnd))
{
branch = new RegexAstNode(new Token<RegexElementLabel>(string.Empty, 0,
Source.Column,
RegexElementLabel.ALT));
UnionBegin = branch;
UnionEnd = branch;
} else
{
branch = new RegexAstNode(new Token<RegexElementLabel>(string.Empty, 0,
UnionEnd.Source.Column,
RegexElementLabel.ALT));
UnionEnd.Right = branch;
branch.Parent = UnionEnd;
UnionEnd = branch;
}
for (int i = 0; i < Cases.Count - 1; i++)
{
Cases[i] ??= NewEmptyNode();
Cases[i].AppendNode(Cases[i + 1]);
Cases[i].Parent = UnionEnd;
}
UnionEnd.Right = Cases[0];
TreeEnumerator<RegexAstNode, RegexElementLabel> enumerator = new TreeEnumerator<RegexAstNode, RegexElementLabel>(UnionBegin);
return UnionBegin;
}
@@ -310,8 +376,12 @@ namespace StgSharp.RegularAnalysis.TextRegex
RegexAstNode _case
)
{
RegexAstNode right = _case.Right;
right ??= Empty;
RegexAstNode right = IsNullOrEmpty(_case.Right) ?
new RegexAstNode(
new Token<RegexElementLabel>(_case.Value, 0, _case.Source.Column,
RegexElementLabel.UNIT)
) :
_case.Right;
Cases.Add(right);
}