scanning min scan length

2026-07-08 18:25:05 +08:00
parent 439128798e
commit 29b24191e1
8 changed files with 125 additions and 93 deletions
@@ -0,0 +1,101 @@
+// -----------------------------------------------------------------------------
+// file="RegexAnalyzer.InfoScan"
+// Project: StgSharp
+// Copyright (c) Nitload.
+// SPDX-License-Identifier: MIT
+// -----------------------------------------------------------------------------
+
+using System;
+using System.Collections.Generic;
+using System.Linq;
+using System.Runtime.CompilerServices;
+using System.Runtime.InteropServices;
+using System.Text;
+using System.Threading.Tasks;
+
+namespace StgSharp.RegularAnalysis.Text
+{
+    public partial class RegexAnalyzer
+    {
+
+        internal static RegexInfo ScanInfo(
+                                             List<RegexIR> ir
+        )
+        {
+            RegexInfo info = new();
+            int process_count = 0;
+
+            ReadOnlySpan<RegexIR> irSpan = CollectionsMarshal.AsSpan(ir);
+
+            int max_region = 0;
+            int min_length = 0;
+            int single_line_count = 0;
+
+            ScanSpan(irSpan);
+            return info;
+
+            void ScanSpan(
+                 ReadOnlySpan<RegexIR> ir_span
+            )
+            {
+                do
+                {
+                    int count = int.Min(8, ir_span.Length - process_count);
+                    ReadOnlySpan<RegexIR> window = ir_span.Slice(process_count, count);
+                    RegexIRPrefix prefix;
+                    if (window[0].CommandMask == RegexIRCommand.Match_Seq)
+                    {
+                        RegexMatchSeqIR match_seq = Unsafe.As<RegexMatchSeqIR>(window[0]);
+                        min_length += match_seq.Pattern.Length;
+                        prefix = match_seq.Prefix;
+                        max_region = int.Max(max_region, prefix.WorkingRegion);
+                        if ((prefix.Mask & RegexIRPrefixMask.ObserveMethod) == RegexIRPrefixMask.Observe_Line)
+                        {
+                            int line_idx = prefix.ObserveSource;
+                            RegexIR source = ir[line_idx];
+                            if (!source.SingleLineResultRequired)
+                            {
+                                source.SingleLineResultIndex = line_idx;
+                                single_line_count++;
+                            }
+                        }
+                        process_count++;
+                    } else if (window[0] is RegexFindSeqIR find_seq)
+                    {
+                        min_length += find_seq.Pattern.Length;
+                        prefix = find_seq.Prefix;
+                        max_region = int.Max(max_region, prefix.WorkingRegion);
+                        if ((prefix.Mask & RegexIRPrefixMask.ObserveMethod) == RegexIRPrefixMask.Observe_Line)
+                        {
+                            int line_idx = prefix.ObserveSource;
+                            RegexIR source = ir[line_idx];
+                            if (!source.SingleLineResultRequired)
+                            {
+                                source.SingleLineResultIndex = line_idx;
+                                single_line_count++;
+                            }
+                        }
+                        process_count++;
+                    } else if (window[0] is RegexCountSeqIR count_seq)
+                    {
+                        min_length += count_seq.Pattern.Length * count_seq.Min;
+                        prefix = count_seq.Prefix;
+                        max_region = int.Max(max_region, prefix.WorkingRegion);
+                        if ((prefix.Mask & RegexIRPrefixMask.ObserveMethod) == RegexIRPrefixMask.Observe_Line)
+                        {
+                            int line_idx = prefix.ObserveSource;
+                            RegexIR source = ir[line_idx];
+                            if (!source.SingleLineResultRequired)
+                            {
+                                source.SingleLineResultIndex = line_idx;
+                                single_line_count++;
+                            }
+                        }
+                        process_count++;
+                    } else if (window[0] is RegexMatchSetIR) { }
+                } while (process_count < ir.Count);
+            }
+        }
+
+    }
+}