Bumped ICU4N to 60.1.0-alpha.440 (#1353)

Author: NightOwl888

.build/dependencies.props

@@ -34,7 +34,7 @@
         https://github.com/apache/lucene-solr/tree/31d7ec7bbfdcd2c4cc61d9d35e962165410b65fe/lucene/analysis/icu/src/data/utr30
         Just make sure they are adjusted to the right version of ICU/Lucene.
     <ICU4NPackageVersion>[60.1,60.2)</ICU4NPackageVersion> -->
-    <ICU4NPackageVersion>[60.1.0-alpha.438,60.1.0-alpha.446)</ICU4NPackageVersion>
+    <ICU4NPackageVersion>[60.1.0-alpha.440,60.1.0-alpha.446)</ICU4NPackageVersion>
     <IKVMPackageVersion>8.7.5</IKVMPackageVersion>
     <IKVMMavenSdkPackageVersion>1.6.7</IKVMMavenSdkPackageVersion>
     <!-- J2N will break binary compatibility in 3.0.0 to fix the APIs of collection types -->

src/Lucene.Net.Analysis.Common/Analysis/Th/ThaiTokenizer.cs

@@ -1,18 +1,15 @@
 // Lucene version compatibility level 4.8.1
 #if FEATURE_BREAKITERATOR
-using ICU4N.Support.Text;
 using ICU4N.Text;
 using J2N;
 using Lucene.Net.Analysis.TokenAttributes;
 using Lucene.Net.Analysis.Util;
 using Lucene.Net.Support;
-using Lucene.Net.Support.Threading;
 using Lucene.Net.Util;
 using System;
 using System.Collections.Generic;
 using System.Globalization;
 using System.IO;
-using System.Text.RegularExpressions;
 
 namespace Lucene.Net.Analysis.Th
 {
@@ -44,6 +41,19 @@ namespace Lucene.Net.Analysis.Th
     /// <summary>
     /// Tokenizer that use <see cref="BreakIterator"/> to tokenize Thai text.
     /// </summary>
+    /// <remarks>
+    /// This is an attempt to mimic the behavior of the JDK's <c>java.Text.BreakIterator</c> approach
+    /// to tokenizing Thai text. While it passes the Lucene tests, there may be innumerable differences
+    /// between this implementation and the one in the JDK.
+    /// <para/>
+    /// Unlike the JDK, this implementation is guaranteed to be stable across all supported target frameworks.
+    /// While it does use ICU4N's <see cref="RuleBasedBreakIterator"/>, this implementation doesn't follow
+    /// the UAX #29 specification (http://unicode.org/reports/tr29) and is not guaranteed to behave the same as
+    /// either the one in the JDK or in ICU4J.
+    /// <para/>
+    /// This implementation is provided primarily for API compatibility with Lucene. If strict Unicode compliance
+    /// is desired, it is highly recommended to use the <see cref="Icu.Segmentation.ICUTokenizer"/> instead.
+    /// </remarks>
     public class ThaiTokenizer : SegmentingTokenizerBase
     {
         // LUCENENET specific - DBBI_AVAILABLE removed because ICU always has a dictionary-based BreakIterator