Match blank text with a blank-text regex under ByWordAny [patch]

TextFilter.Test/ReadmeTests.cs

@@ -4,6 +4,7 @@ namespace TextFilter.Test;
 
 using System;
 using System.Collections.Generic;
+using System.ComponentModel;
 using System.IO;
 using System.Linq;
 using System.Reflection;
@@ -89,7 +90,10 @@ public void EveryMethodInTheReadmeApiReferenceExistsWithItsDocumentedDefaults()
 		foreach (Match row in rows)
 		{
 			string methodName = row.Groups["method"].Value;
-			MethodInfo[] overloads = [.. methods.Where(m => m.Name == methodName)];
+			// The binary-compatibility overloads are hidden from IntelliSense and deliberately carry no
+			// defaults, since defaults on both sets would make calls ambiguous. The README documents the
+			// visible overloads, so only those are held to its defaults.
+			MethodInfo[] overloads = [.. methods.Where(m => m.Name == methodName && !IsHiddenFromCallers(m))];
 			if (overloads.Length == 0)
 			{
 				problems.Add($"{methodName} does not exist");
@@ -115,6 +119,9 @@ public void EveryMethodInTheReadmeApiReferenceExistsWithItsDocumentedDefaults()
 		Assert.IsEmpty(problems, string.Join("; ", problems));
 	}
 
+	private static bool IsHiddenFromCallers(MethodInfo method) =>
+		method.GetCustomAttribute<EditorBrowsableAttribute>()?.State == EditorBrowsableState.Never;
+
 	[TestMethod]
 	public void ReadmeMatchOptionsSampleGivesTheDocumentedResults()
 	{

TextFilter.Test/TextFilterTests.cs

@@ -1169,8 +1169,8 @@
 	public void FilterWithAClassFollowedByALiteralBracketKeepsAndExcludesTheRightItems()
 	{
 		List<string> items = ["items[3]", "items[3", "notes.txt"];
 		CollectionAssert.AreEqual(new List<string> { "items[3]" }, TextFilter.Filter(items, "*[[][0-9]]").ToList());
 		CollectionAssert.AreEqual(new List<string> { "items[3", "notes.txt" }, TextFilter.Filter(items, "-*[[][0-9]]").ToList());
 	}
 
 	[TestMethod]
@@ -1205,6 +1205,47 @@
 			TextFilter.Filter([text], "zzz", filterType, TextFilterMatchOptions.ByWordAll).ToList());
 	}
 
+	[TestMethod]
+	[DataRow("   ", @"\s")]
+	[DataRow("   ", @"^\s*$")]
+	[DataRow("   ", @"\s+")]
+	[DataRow("", "^$")]
+	[DataRow("", @"^\s*$")]
+	[DataRow(" \t ", @"^\s*$")]
+	public void RegexByWordAnyMatchesBlankTextWithAPatternForBlankText(string text, string pattern)
+	{
+		// Blank text splits into no words, and the no-words guard used to return before the whole-text
+		// attempt, so a filter written to find blank entries missed exactly those under the default.
+		Assert.IsTrue(TextFilter.IsMatch(text, pattern, TextFilterType.Regex), $"'{pattern}' should match '{text}' under ByWordAny.");
+		Assert.IsTrue(TextFilter.IsMatch(text, pattern, TextFilterType.Regex, TextFilterMatchOptions.ByWholeString), $"'{pattern}' should match '{text}' under ByWholeString.");
+	}
+
+	[TestMethod]
+	[DataRow("   ", @"\s")]
+	[DataRow("", "^$")]
+	[DataRow("   ", @"^\s*$")]
+	public void RegexByWordAllStillMatchesNoBlankText(string text, string pattern)
+	{
+		Assert.IsFalse(TextFilter.IsMatch(text, pattern, TextFilterType.Regex, TextFilterMatchOptions.ByWordAll));
+	}
+
+	[TestMethod]
+	[DataRow("")]
+	[DataRow("   ")]
+	public void RegexByWordAnyDoesNotMatchBlankTextWithAPatternThatNeedsAWord(string text)
+	{
+		Assert.IsFalse(TextFilter.IsMatch(text, "zzz", TextFilterType.Regex));
+		Assert.IsFalse(TextFilter.IsMatch(text, @"\w", TextFilterType.Regex));
+	}
+
+	[TestMethod]
+	public void RegexByWordAnyFilterKeepsTheBlankEntries()
+	{
+		CollectionAssert.AreEqual(
+			new List<string> { "", "   " },
+			TextFilter.Filter(["", "hello", "   ", "a b"], @"^\s*$", TextFilterType.Regex).ToList());
+	}
+
 	[TestMethod]
 	public void RegexByWordAllFilterDropsBlankEntries()
 	{
@@ -1326,7 +1367,7 @@
 	{
 		List<string> result = [.. TextFilter.Filter(["readme.md", "src/a.cs", "docs/", "a b"], filter)];
 
 		Assert.AreEqual(1, result.Count);
 		Assert.AreEqual("docs/", result[0]);
 	}
 
@@ -1337,7 +1378,7 @@
 	{
 		List<string> result = [.. TextFilter.Filter(["readme.md", "src/a.cs", "docs/", "a b"], filter)];
 
 		Assert.AreEqual(3, result.Count);
 		Assert.AreEqual("readme.md", result[0]);
 		Assert.AreEqual("src/a.cs", result[1]);
 		Assert.AreEqual("a b", result[2]);
@@ -1348,7 +1389,7 @@
 	{
 		List<string> result = [.. TextFilter.Filter(["readme.md", "src/a.cs", "docs/"], "+**/")];
 
 		Assert.AreEqual(1, result.Count);
 		Assert.AreEqual("docs/", result[0]);
 	}
 
@@ -1368,9 +1409,9 @@
 		// Splitting only on ' ' left a chunk of other whitespace that trimmed to an empty token, and
 		// the glob tokenizer then read its first character. A pasted trailing tab or newline is enough.
 		Assert.IsTrue(TextFilter.IsMatch("foo", filter));
 		CollectionAssert.AreEqual(
 			new List<string> { "foo" },
 			TextFilter.Filter(["foo"], filter).ToList());
 	}
 
 	[TestMethod]
@@ -1559,11 +1600,11 @@
 	[TestMethod]
 	public void Pre17GenericFilterSignatureStillExists()
 	{
 		Assert.IsTrue(typeof(TextFilter).GetMethods().Any(m =>
 			m.Name == nameof(TextFilter.Filter)
 			&& m.IsGenericMethodDefinition
 			&& m.GetParameters().Length == 5
 			&& m.GetParameters()[4].ParameterType == typeof(TextFilterMatchOptions)));
 	}
 
 	[TestMethod]

TextFilter/TextFilter.cs

@@ -303,7 +303,7 @@
 
 			return ExcludedTokenPrefixes.Contains(prefix)
 				? TextFilterTokenType.Excluded
 				: RequiredTokenPrefixes.Contains(prefix)
 				? TextFilterTokenType.Required
 				: TextFilterTokenType.Optional;
 		})
@@ -334,7 +334,7 @@
 	/// matches nor rejects anything; as a required token it matches everything; and as an excluded
 	/// token it excludes nothing.
 	/// </remarks>
 	public static bool DoesMatchGlob(string text, string filter, TextFilterMatchOptions textFilterMatchOptions, TextFilterCaseSensitivity caseSensitivity = TextFilterCaseSensitivity.CaseSensitive)
 	{
 		Ensure.NotNull(text);
 		Ensure.NotNull(filter);
@@ -408,7 +408,7 @@
 	/// <param name="textTokens">The set of text tokens to match against.</param>
 	/// <param name="caseSensitivity">Whether the match distinguishes uppercase from lowercase.</param>
 	/// <returns><c>true</c> if any token matches the glob filter token; otherwise, <c>false</c>.</returns>
 	public static bool AnyTokenMatchesGlobFilter(string filterToken, HashSet<string> textTokens, TextFilterCaseSensitivity caseSensitivity = TextFilterCaseSensitivity.CaseSensitive)
 	{
 		Ensure.NotNull(filterToken);
 		Ensure.NotNull(textTokens);
@@ -425,7 +425,7 @@
 	/// <param name="textTokens">The set of text tokens to match against.</param>
 	/// <param name="caseSensitivity">Whether the match distinguishes uppercase from lowercase.</param>
 	/// <returns><c>true</c> if all tokens match the glob filter token; otherwise, <c>false</c>.</returns>
 	public static bool AllTokensMatchGlobFilter(string filterToken, HashSet<string> textTokens, TextFilterCaseSensitivity caseSensitivity = TextFilterCaseSensitivity.CaseSensitive)
 	{
 		Ensure.NotNull(filterToken);
 		Ensure.NotNull(textTokens);
@@ -509,7 +509,7 @@
 	// MaxGlobClassMembers characters, or that has no members at all because every range is reversed.
 	private const int MaxGlobClassMembers = 4096;
 
 	internal static string? ExpandGlobClasses(string token)
 	{
 		if (!token.Contains(']'))
 		{
@@ -712,7 +712,7 @@
 	/// <see cref="TextFilterMatchOptions.ByWholeString"/> for those.
 	/// </para>
 	/// </remarks>
 	public static bool DoesMatchRegex(string text, string filter, TextFilterMatchOptions textFilterMatchOptions, TextFilterCaseSensitivity caseSensitivity = TextFilterCaseSensitivity.CaseSensitive)
 	{
 		Ensure.NotNull(text);
 		Ensure.NotNull(filter);
@@ -729,6 +729,16 @@
 		// Swapped for RegexMatchNothing if the pattern times out partway through
 		Regex activeRegex = regex;
 
+		// No single word contains whitespace, so a pattern like "hello world" or "o\sw" can only ever
+		// match the whole text. Under ByWordAny, try that first; the per-word pass below still runs,
+		// so a word-anchored pattern like "^world" keeps matching "hello world". It comes before the
+		// no-words check below because blank text has no words at all, so a pattern written to find
+		// blank entries, such as "^\s*$" or "\s", can only ever match the whole text.
+		if (textFilterMatchOptions is TextFilterMatchOptions.ByWordAny && IsMatchWithinTimeout(ref activeRegex, cacheKey, text))
+		{
+			return true;
+		}
+
 		// Text with no words -- "" or "   " split by word -- has nothing for the pattern to match.
 		// Without this, Enumerable.All over the empty set is vacuously true, so under ByWordAll blank
 		// text matched every pattern while the glob path reported no match for the same input.
@@ -737,14 +747,6 @@
 			return false;
 		}
 
-		// No single word contains whitespace, so a pattern like "hello world" or "o\sw" can only ever
-		// match the whole text. Under ByWordAny, try that first; the per-word pass below still runs,
-		// so a word-anchored pattern like "^world" keeps matching "hello world".
-		if (textFilterMatchOptions is TextFilterMatchOptions.ByWordAny && IsMatchWithinTimeout(ref activeRegex, cacheKey, text))
-		{
-			return true;
-		}
-
 		Func<IEnumerable<string>, Func<string, bool>, bool> matchFunc = textFilterMatchOptions is TextFilterMatchOptions.ByWordAny
 			? Enumerable.Any
 			: Enumerable.All;