From 235b3a03f27f752a00b82448a7fc128e4dd802c5 Mon Sep 17 00:00:00 2001
From: Curtis Wensley <curtis.wensley@gmail.com>
Date: Sun, 23 May 2021 16:01:12 -0700
Subject: [PATCH] Fix memory issue with large character set ranges

Fixes #48
---
 .../Samples/LargeCharcterSetRange.cs          | 17 ++++++++++++
 Eto.Parse/Grammar.cs                          | 26 +++++++++++++++----
 2 files changed, 38 insertions(+), 5 deletions(-)
 create mode 100644 Eto.Parse.Tests/Samples/LargeCharcterSetRange.cs

diff --git a/Eto.Parse.Tests/Samples/LargeCharcterSetRange.cs b/Eto.Parse.Tests/Samples/LargeCharcterSetRange.cs
new file mode 100644
index 0000000..54ea68a
--- /dev/null
+++ b/Eto.Parse.Tests/Samples/LargeCharcterSetRange.cs
@@ -0,0 +1,17 @@
+using Eto.Parse.Grammars;
+using NUnit.Framework;
+
+namespace Eto.Parse.Tests.Samples
+{
+	[TestFixture]
+	public class LargeCharcterSetRange
+	{
+		[Test]
+		public void LargeRangeShouldntCauseMemoryException()
+		{
+			var _grammar = new EbnfGrammar(EbnfStyle.W3c).Build($"id ::= [a-zA-Z\u0100-\uffff_][0-9a-zA-Z\u0100-\uffff_]*", "id");
+			var _match = _grammar.Match("张三李四");
+		}
+
+	}
+}
\ No newline at end of file
diff --git a/Eto.Parse/Grammar.cs b/Eto.Parse/Grammar.cs
index e2ccff0..2f82acd 100644
--- a/Eto.Parse/Grammar.cs
+++ b/Eto.Parse/Grammar.cs
@@ -80,6 +80,12 @@ public class Grammar : UnaryParser
 		public bool AllowPartialMatch { get; set; }
 
 		public bool Trace { get; set; }
+		
+		/// <summary>
+		/// Sets the maximum character set range when <see cref="GrammarOptimizations.CharacterSetAlternations"/> is enabled.
+		/// </summary>
+		/// <value></value>
+		public int MaxCharacterSetRangeOptimization { get; set; } = 100;
 
 		public GrammarOptimizations Optimizations { get; set; }
 
@@ -264,6 +270,7 @@ void OptimizeCharacterSets()
 				{
 					var chars = new List<char>();
 					var inverse = new List<char>();
+					var additionalParsers = new List<Parser>();
                     for (int i1 = 0; i1 < alt.Items.Count; i1++)
 					{
                         Parser item = alt.Items[i1];
@@ -288,12 +295,19 @@ void OptimizeCharacterSets()
 						var charRange = item as CharRangeTerminal;
 						if (charRange != null)
 						{
-							for (char i = charRange.Start; i <= charRange.End; i++)
+							if (charRange.End - charRange.Start > MaxCharacterSetRangeOptimization)
+							{
+								additionalParsers.Add(charRange);
+							}
+							else
 							{
-								if (charRange.Inverse)
-									inverse.Add(i);
-								else
-									chars.Add(i);
+								for (char i = charRange.Start; i <= charRange.End; i++)
+								{
+									if (charRange.Inverse)
+										inverse.Add(i);
+									else
+										chars.Add(i);
+								}
 							}
 							continue;
 						}
@@ -313,6 +327,8 @@ void OptimizeCharacterSets()
 						{
 							Inverse = true
 						});
+					if (additionalParsers.Count > 0)
+						alt.Items.AddRange(additionalParsers);
 				}
 			}
 		}