Skip to content

Commit

Permalink
Add IndexOfAnyValues.Contains (dotnet#78996)
Browse files Browse the repository at this point in the history 
* Add IndexOfAnyValues.Contains

* Unsafe => unsafe
  • Loading branch information
@MihaZupan
MihaZupan authored Nov 30, 2022
1 parent 1076cf5 commit b85e152
Show file tree
Hide file tree
Showing 21 changed files with 563 additions and 421 deletions.
5 changes: 2 additions & 3 deletions src/coreclr/System.Private.CoreLib/src/System/TypeNameParser.cs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -89,8 +89,7 @@ internal sealed partial class TypeNameParser : IDisposable

#region Private Data Members
private readonly SafeTypeNameParserHandle m_NativeParser;
private const string SpecialChars = ",[]&*+\\"; // see typeparse.h
private static readonly IndexOfAnyValues<char> s_specialChars = IndexOfAnyValues.Create(SpecialChars);
private static readonly IndexOfAnyValues<char> s_specialChars = IndexOfAnyValues.Create(",[]&*+\\"); // see typeparse.h
#endregion

#region Constructor and Disposer
Expand Down Expand Up @@ -289,7 +288,7 @@ private static string EscapeTypeName(string name)

foreach (char c in name.AsSpan(specialCharIndex))
{
if (SpecialChars.Contains(c))
if (s_specialChars.Contains(c))
sb.Append('\\');

sb.Append(c);
Expand Down
67 changes: 0 additions & 67 deletions src/libraries/System.Memory/tests/Span/IndexOfAny.byte.cs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -530,73 +530,6 @@ public static void MakeSureNoChecksGoOutOfRangeMany_Byte()
}
}

[Fact]
[OuterLoop("Takes about a second to execute")]
public static void TestIndexOfAny_RandomInputs_Byte()
{
IndexOfAnyCharTestHelper.TestRandomInputs(
expected: IndexOfAnyReferenceImpl,
indexOfAny: (searchSpace, values) => searchSpace.IndexOfAny(values),
indexOfAnyValues: (searchSpace, values) => searchSpace.IndexOfAny(values));

static int IndexOfAnyReferenceImpl(ReadOnlySpan<byte> searchSpace, ReadOnlySpan<byte> values)
{
for (int i = 0; i < searchSpace.Length; i++)
{
if (values.Contains(searchSpace[i]))
{
return i;
}
}

return -1;
}
}

[Theory]
[InlineData(true)]
[InlineData(false)]
public static void AsciiNeedle_ProperlyHandlesEdgeCases_Byte(bool needleContainsZero)
{
// There is some special handling we have to do for ASCII needles to properly filter out non-ASCII results
ReadOnlySpan<byte> needleValues = needleContainsZero ? "AEIOU\0"u8 : "AEIOU!"u8;
IndexOfAnyValues<byte> needle = IndexOfAnyValues.Create(needleValues);

ReadOnlySpan<byte> repeatingHaystack = "AaAaAaAaAaAa"u8;
Assert.Equal(0, repeatingHaystack.IndexOfAny(needle));
Assert.Equal(1, repeatingHaystack.IndexOfAnyExcept(needle));
Assert.Equal(10, repeatingHaystack.LastIndexOfAny(needle));
Assert.Equal(11, repeatingHaystack.LastIndexOfAnyExcept(needle));

ReadOnlySpan<byte> haystackWithZeroes = "Aa\0Aa\0Aa\0"u8;
Assert.Equal(0, haystackWithZeroes.IndexOfAny(needle));
Assert.Equal(1, haystackWithZeroes.IndexOfAnyExcept(needle));
Assert.Equal(needleContainsZero ? 8 : 6, haystackWithZeroes.LastIndexOfAny(needle));
Assert.Equal(needleContainsZero ? 7 : 8, haystackWithZeroes.LastIndexOfAnyExcept(needle));

Span<byte> haystackWithOffsetNeedle = new byte[100];
for (int i = 0; i < haystackWithOffsetNeedle.Length; i++)
{
haystackWithOffsetNeedle[i] = (byte)(128 + needleValues[i % needleValues.Length]);
}

Assert.Equal(-1, haystackWithOffsetNeedle.IndexOfAny(needle));
Assert.Equal(0, haystackWithOffsetNeedle.IndexOfAnyExcept(needle));
Assert.Equal(-1, haystackWithOffsetNeedle.LastIndexOfAny(needle));
Assert.Equal(haystackWithOffsetNeedle.Length - 1, haystackWithOffsetNeedle.LastIndexOfAnyExcept(needle));

// Mix matching characters back in
for (int i = 0; i < haystackWithOffsetNeedle.Length; i += 3)
{
haystackWithOffsetNeedle[i] = needleValues[i % needleValues.Length];
}

Assert.Equal(0, haystackWithOffsetNeedle.IndexOfAny(needle));
Assert.Equal(1, haystackWithOffsetNeedle.IndexOfAnyExcept(needle));
Assert.Equal(haystackWithOffsetNeedle.Length - 1, haystackWithOffsetNeedle.LastIndexOfAny(needle));
Assert.Equal(haystackWithOffsetNeedle.Length - 2, haystackWithOffsetNeedle.LastIndexOfAnyExcept(needle));
}

private static int IndexOf(Span<byte> span, byte value)
{
int index = span.IndexOf(value);
Expand Down
203 changes: 0 additions & 203 deletions src/libraries/System.Memory/tests/Span/IndexOfAny.char.cs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -4,9 +4,6 @@
using System.Buffers;
using System.Linq;
using System.Numerics;
using System.Runtime.CompilerServices;
using System.Runtime.InteropServices;
using System.Text;
using Xunit;

namespace System.SpanTests
Expand Down Expand Up @@ -776,95 +773,6 @@ public static void MakeSureNoChecksGoOutOfRangeMany_Char()
}
}

[Fact]
[OuterLoop("Takes about a second to execute")]
public static void TestIndexOfAny_RandomInputs_Char()
{
IndexOfAnyCharTestHelper.TestRandomInputs(
expected: IndexOfAnyReferenceImpl,
indexOfAny: (searchSpace, values) => searchSpace.IndexOfAny(values),
indexOfAnyValues: (searchSpace, values) => searchSpace.IndexOfAny(values));

static int IndexOfAnyReferenceImpl(ReadOnlySpan<char> searchSpace, ReadOnlySpan<char> values)
{
for (int i = 0; i < searchSpace.Length; i++)
{
if (values.Contains(searchSpace[i]))
{
return i;
}
}

return -1;
}
}

[Theory]
[InlineData(true)]
[InlineData(false)]
public static void AsciiNeedle_ProperlyHandlesEdgeCases_Char(bool needleContainsZero)
{
// There is some special handling we have to do for ASCII needles to properly filter out non-ASCII results
ReadOnlySpan<char> needleValues = needleContainsZero ? "AEIOU\0" : "AEIOU!";
IndexOfAnyValues<char> needle = IndexOfAnyValues.Create(needleValues);

ReadOnlySpan<char> repeatingHaystack = "AaAaAaAaAaAa";
Assert.Equal(0, repeatingHaystack.IndexOfAny(needle));
Assert.Equal(1, repeatingHaystack.IndexOfAnyExcept(needle));
Assert.Equal(10, repeatingHaystack.LastIndexOfAny(needle));
Assert.Equal(11, repeatingHaystack.LastIndexOfAnyExcept(needle));

ReadOnlySpan<char> haystackWithZeroes = "Aa\0Aa\0Aa\0";
Assert.Equal(0, haystackWithZeroes.IndexOfAny(needle));
Assert.Equal(1, haystackWithZeroes.IndexOfAnyExcept(needle));
Assert.Equal(needleContainsZero ? 8 : 6, haystackWithZeroes.LastIndexOfAny(needle));
Assert.Equal(needleContainsZero ? 7 : 8, haystackWithZeroes.LastIndexOfAnyExcept(needle));

Span<char> haystackWithOffsetNeedle = new char[100];
for (int i = 0; i < haystackWithOffsetNeedle.Length; i++)
{
haystackWithOffsetNeedle[i] = (char)(128 + needleValues[i % needleValues.Length]);
}

Assert.Equal(-1, haystackWithOffsetNeedle.IndexOfAny(needle));
Assert.Equal(0, haystackWithOffsetNeedle.IndexOfAnyExcept(needle));
Assert.Equal(-1, haystackWithOffsetNeedle.LastIndexOfAny(needle));
Assert.Equal(haystackWithOffsetNeedle.Length - 1, haystackWithOffsetNeedle.LastIndexOfAnyExcept(needle));

// Mix matching characters back in
for (int i = 0; i < haystackWithOffsetNeedle.Length; i += 3)
{
haystackWithOffsetNeedle[i] = needleValues[i % needleValues.Length];
}

Assert.Equal(0, haystackWithOffsetNeedle.IndexOfAny(needle));
Assert.Equal(1, haystackWithOffsetNeedle.IndexOfAnyExcept(needle));
Assert.Equal(haystackWithOffsetNeedle.Length - 1, haystackWithOffsetNeedle.LastIndexOfAny(needle));
Assert.Equal(haystackWithOffsetNeedle.Length - 2, haystackWithOffsetNeedle.LastIndexOfAnyExcept(needle));

// With chars, the lower byte could be matching, but we have to check that the higher byte is also 0
for (int i = 0; i < haystackWithOffsetNeedle.Length; i++)
{
haystackWithOffsetNeedle[i] = (char)(((i + 1) * 256) + needleValues[i % needleValues.Length]);
}

Assert.Equal(-1, haystackWithOffsetNeedle.IndexOfAny(needle));
Assert.Equal(0, haystackWithOffsetNeedle.IndexOfAnyExcept(needle));
Assert.Equal(-1, haystackWithOffsetNeedle.LastIndexOfAny(needle));
Assert.Equal(haystackWithOffsetNeedle.Length - 1, haystackWithOffsetNeedle.LastIndexOfAnyExcept(needle));

// Mix matching characters back in
for (int i = 0; i < haystackWithOffsetNeedle.Length; i += 3)
{
haystackWithOffsetNeedle[i] = needleValues[i % needleValues.Length];
}

Assert.Equal(0, haystackWithOffsetNeedle.IndexOfAny(needle));
Assert.Equal(1, haystackWithOffsetNeedle.IndexOfAnyExcept(needle));
Assert.Equal(haystackWithOffsetNeedle.Length - 1, haystackWithOffsetNeedle.LastIndexOfAny(needle));
Assert.Equal(haystackWithOffsetNeedle.Length - 2, haystackWithOffsetNeedle.LastIndexOfAnyExcept(needle));
}

private static int IndexOf(Span<char> span, char value)
{
int index = span.IndexOf(value);
Expand Down Expand Up @@ -893,115 +801,4 @@ private static int IndexOfAny(Span<char> span, ReadOnlySpan<char> values)
return index;
}
}

public static class IndexOfAnyCharTestHelper
{
private const int MaxNeedleLength = 10;
private const int MaxHaystackLength = 40;

private static readonly char[] s_randomAsciiChars;
private static readonly char[] s_randomLatin1Chars;
private static readonly char[] s_randomChars;
private static readonly byte[] s_randomAsciiBytes;
private static readonly byte[] s_randomBytes;

static IndexOfAnyCharTestHelper()
{
s_randomAsciiChars = new char[100 * 1024];
s_randomLatin1Chars = new char[100 * 1024];
s_randomChars = new char[1024 * 1024];
s_randomBytes = new byte[100 * 1024];

var rng = new Random(42);

for (int i = 0; i < s_randomAsciiChars.Length; i++)
{
s_randomAsciiChars[i] = (char)rng.Next(0, 128);
}

for (int i = 0; i < s_randomLatin1Chars.Length; i++)
{
s_randomLatin1Chars[i] = (char)rng.Next(0, 256);
}

rng.NextBytes(MemoryMarshal.Cast<char, byte>(s_randomChars));

s_randomAsciiBytes = Encoding.ASCII.GetBytes(s_randomAsciiChars);

rng.NextBytes(s_randomBytes);
}

public delegate int IndexOfAnySearchDelegate<T>(ReadOnlySpan<T> searchSpace, ReadOnlySpan<T> values) where T : IEquatable<T>?;

public delegate int IndexOfAnyValuesSearchDelegate<T>(ReadOnlySpan<T> searchSpace, IndexOfAnyValues<T> values) where T : IEquatable<T>?;

public static void TestRandomInputs(IndexOfAnySearchDelegate<byte> expected, IndexOfAnySearchDelegate<byte> indexOfAny, IndexOfAnyValuesSearchDelegate<byte> indexOfAnyValues)
{
var rng = new Random(42);

for (int iterations = 0; iterations < 1_000_000; iterations++)
{
// There are more interesting corner cases with ASCII needles, test those more.
Test(rng, s_randomBytes, s_randomAsciiBytes, expected, indexOfAny, indexOfAnyValues);

Test(rng, s_randomBytes, s_randomBytes, expected, indexOfAny, indexOfAnyValues);
}
}

public static void TestRandomInputs(IndexOfAnySearchDelegate<char> expected, IndexOfAnySearchDelegate<char> indexOfAny, IndexOfAnyValuesSearchDelegate<char> indexOfAnyValues)
{
var rng = new Random(42);

for (int iterations = 0; iterations < 1_000_000; iterations++)
{
// There are more interesting corner cases with ASCII needles, test those more.
Test(rng, s_randomChars, s_randomAsciiChars, expected, indexOfAny, indexOfAnyValues);

Test(rng, s_randomChars, s_randomLatin1Chars, expected, indexOfAny, indexOfAnyValues);

Test(rng, s_randomChars, s_randomChars, expected, indexOfAny, indexOfAnyValues);
}
}

private static void Test<T>(Random rng, ReadOnlySpan<T> haystackRandom, ReadOnlySpan<T> needleRandom,
IndexOfAnySearchDelegate<T> expected, IndexOfAnySearchDelegate<T> indexOfAny, IndexOfAnyValuesSearchDelegate<T> indexOfAnyValues)
where T : INumber<T>
{
ReadOnlySpan<T> haystack = GetRandomSlice(rng, haystackRandom, MaxHaystackLength);
ReadOnlySpan<T> needle = GetRandomSlice(rng, needleRandom, MaxNeedleLength);

IndexOfAnyValues<T> indexOfAnyValuesInstance = (IndexOfAnyValues<T>)(object)(typeof(T) == typeof(byte)
? IndexOfAnyValues.Create(MemoryMarshal.CreateReadOnlySpan(ref Unsafe.As<T, byte>(ref MemoryMarshal.GetReference(needle)), needle.Length))
: IndexOfAnyValues.Create(MemoryMarshal.CreateReadOnlySpan(ref Unsafe.As<T, char>(ref MemoryMarshal.GetReference(needle)), needle.Length)));

int expectedIndex = expected(haystack, needle);
int indexOfAnyIndex = indexOfAny(haystack, needle);
int indexOfAnyValuesIndex = indexOfAnyValues(haystack, indexOfAnyValuesInstance);

if (expectedIndex != indexOfAnyIndex)
{
AssertionFailed(haystack, needle, expectedIndex, indexOfAnyIndex, nameof(indexOfAny));
}

if (expectedIndex != indexOfAnyValuesIndex)
{
AssertionFailed(haystack, needle, expectedIndex, indexOfAnyValuesIndex, nameof(indexOfAnyValues));
}
}

private static ReadOnlySpan<T> GetRandomSlice<T>(Random rng, ReadOnlySpan<T> span, int maxLength)
{
ReadOnlySpan<T> slice = span.Slice(rng.Next(span.Length + 1));
return slice.Slice(0, Math.Min(slice.Length, rng.Next(maxLength + 1)));
}

private static void AssertionFailed<T>(ReadOnlySpan<T> haystack, ReadOnlySpan<T> needle, int expected, int actual, string approach)
where T : INumber<T>
{
string readableHaystack = string.Join(", ", haystack.ToString().Select(c => int.CreateChecked(c)));
string readableNeedle = string.Join(", ", needle.ToString().Select(c => int.CreateChecked(c)));

Assert.True(false, $"Expected {expected}, got {approach}={actual} for needle='{readableNeedle}', haystack='{readableHaystack}'");
}
}
}
Loading

0 comments on commit b85e152

Please sign in to comment.