Skip to content

Frozen/String/Hashing: replace unsafe pointer arithmetic with safe span indexing #125547

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
Copilot wants to merge 2 commits into from
Closed

Frozen/String/Hashing: replace unsafe pointer arithmetic with safe span indexing #125547

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b4de2ee
Initial plan
Copilot Mar 14, 2026
0035c1f
Remove unsafe/fixed/pointer code from Hashing.cs, replace with safe s…
Copilot Mar 14, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
188 changes: 86 additions & 102 deletions src/libraries/System.Collections.Immutable/src/System/Collections/Frozen/String/Hashing.cs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,6 @@
using System.Buffers;
using System.Diagnostics;
using System.Numerics;
using System.Runtime.InteropServices;
using System.Text;

namespace System.Collections.Frozen
Expand All @@ -20,63 +19,56 @@ internal static class Hashing
private const uint Hash1Start = (5381 << 16) + 5381;
private const uint Factor = 1_566_083_941;

public static unsafe int GetHashCodeOrdinal(ReadOnlySpan<char> s)
public static int GetHashCodeOrdinal(ReadOnlySpan<char> s)
{
int length = s.Length;
fixed (char* src = &MemoryMarshal.GetReference(s))
uint hash1, hash2;
switch (s.Length)
{
uint hash1, hash2;
switch (length)
{
case 0:
return (int)(Hash1Start + unchecked(Hash1Start * Factor));

case 1:
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ src[0];
return (int)(Hash1Start + (hash2 * Factor));

case 2:
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ src[0];
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ src[1];
return (int)(Hash1Start + (hash2 * Factor));

case 3:
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ src[0];
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ src[1];
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ src[2];
return (int)(Hash1Start + (hash2 * Factor));

case 4:
hash1 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ ((uint*)src)[0];
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ ((uint*)src)[1];
return (int)(hash1 + (hash2 * Factor));

default:
hash1 = Hash1Start;
hash2 = hash1;

uint* ptrUInt32 = (uint*)src;
while (length >= 4)
{
hash1 = (BitOperations.RotateLeft(hash1, 5) + hash1) ^ ptrUInt32[0];
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ ptrUInt32[1];
ptrUInt32 += 2;
length -= 4;
}

char* ptrChar = (char*)ptrUInt32;
while (length-- > 0)
{
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ *ptrChar++;
}

return (int)(hash1 + (hash2 * Factor));
}
case 0:
return (int)(Hash1Start + unchecked(Hash1Start * Factor));

case 1:
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ s[0];
return (int)(Hash1Start + (hash2 * Factor));

case 2:
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ s[0];
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ s[1];
return (int)(Hash1Start + (hash2 * Factor));

case 3:
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ s[0];
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ s[1];
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ s[2];
return (int)(Hash1Start + (hash2 * Factor));

case 4:
hash1 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ ((uint)s[0] | ((uint)s[1] << 16));
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ ((uint)s[2] | ((uint)s[3] << 16));
return (int)(hash1 + (hash2 * Factor));
Comment thread
EgorBo marked this conversation as resolved.

default:
hash1 = Hash1Start;
hash2 = hash1;

int i = 0;
for (; i <= s.Length - 4; i += 4)
{
hash1 = (BitOperations.RotateLeft(hash1, 5) + hash1) ^ ((uint)s[i] | ((uint)s[i + 1] << 16));
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ ((uint)s[i + 2] | ((uint)s[i + 3] << 16));
}

for (; i < s.Length; i++)
{
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ s[i];
}
Comment thread
EgorBo marked this conversation as resolved.

return (int)(hash1 + (hash2 * Factor));
}
}

// useful if the string only contains ASCII characters
public static unsafe int GetHashCodeOrdinalIgnoreCaseAscii(ReadOnlySpan<char> s)
public static int GetHashCodeOrdinalIgnoreCaseAscii(ReadOnlySpan<char> s)
{
Debug.Assert(Ascii.IsValid(s));

Expand All @@ -86,57 +78,49 @@ public static unsafe int GetHashCodeOrdinalIgnoreCaseAscii(ReadOnlySpan<char> s)
const uint LowercaseChar = 0x20u;
const uint LowercaseUInt32 = 0x0020_0020u;

int length = s.Length;
fixed (char* src = &MemoryMarshal.GetReference(s))
uint hash1, hash2;
switch (s.Length)
{
uint hash1, hash2;
switch (length)
{
case 0:
return (int)(Hash1Start + unchecked(Hash1Start * Factor));

case 1:
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ (src[0] | LowercaseChar);
return (int)(Hash1Start + (hash2 * Factor));

case 2:
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ (src[0] | LowercaseChar);
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ (src[1] | LowercaseChar);
return (int)(Hash1Start + (hash2 * Factor));

case 3:
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ (src[0] | LowercaseChar);
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ (src[1] | LowercaseChar);
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ (src[2] | LowercaseChar);
return (int)(Hash1Start + (hash2 * Factor));

case 4:
hash1 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ (((uint*)src)[0] | LowercaseUInt32);
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ (((uint*)src)[1] | LowercaseUInt32);
return (int)(hash1 + (hash2 * Factor));

default:
hash1 = Hash1Start;
hash2 = hash1;

uint* ptrUInt32 = (uint*)src;
while (length >= 4)
{
hash1 = (BitOperations.RotateLeft(hash1, 5) + hash1) ^ (ptrUInt32[0] | LowercaseUInt32);
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ (ptrUInt32[1] | LowercaseUInt32);
ptrUInt32 += 2;
length -= 4;
}

char* ptrChar = (char*)ptrUInt32;
while (length-- > 0)
{
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ (*ptrChar | LowercaseUInt32);
Copy link
Copy Markdown
Member

@EgorBo EgorBo Mar 15, 2026 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

a copy-paste issue in the previous code - the single-char loop uses LowercaseUInt32 (for 2 chars at a time) instead of LowercaseChar. Not a correctness issue

ptrChar++;
}

return (int)(hash1 + (hash2 * Factor));
}
case 0:
return (int)(Hash1Start + unchecked(Hash1Start * Factor));

case 1:
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ (s[0] | LowercaseChar);
return (int)(Hash1Start + (hash2 * Factor));

case 2:
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ (s[0] | LowercaseChar);
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ (s[1] | LowercaseChar);
return (int)(Hash1Start + (hash2 * Factor));

case 3:
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ (s[0] | LowercaseChar);
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ (s[1] | LowercaseChar);
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ (s[2] | LowercaseChar);
return (int)(Hash1Start + (hash2 * Factor));

case 4:
hash1 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ (((uint)s[0] | ((uint)s[1] << 16)) | LowercaseUInt32);
hash2 = (BitOperations.RotateLeft(Hash1Start, 5) + Hash1Start) ^ (((uint)s[2] | ((uint)s[3] << 16)) | LowercaseUInt32);
return (int)(hash1 + (hash2 * Factor));
Comment thread
EgorBo marked this conversation as resolved.

default:
hash1 = Hash1Start;
hash2 = hash1;

int i = 0;
for (; i <= s.Length - 4; i += 4)
{
hash1 = (BitOperations.RotateLeft(hash1, 5) + hash1) ^ (((uint)s[i] | ((uint)s[i + 1] << 16)) | LowercaseUInt32);
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ (((uint)s[i + 2] | ((uint)s[i + 3] << 16)) | LowercaseUInt32);
}

for (; i < s.Length; i++)
{
hash2 = (BitOperations.RotateLeft(hash2, 5) + hash2) ^ (s[i] | LowercaseChar);
}

return (int)(hash1 + (hash2 * Factor));
}
}

Expand Down
Loading