Hacktoberfest 2026：维护者为十月标记出来的 issue，仍然开放、适合新手。 浏览 Hacktoberfest issue

ByteBuffer - Reduce Allocations for Serialization Path

未关闭
  • 哪里不对？
#141 1 条评论 0 个 reaction 已指派 0 人 在 GitHub 查看

还没有人认领这个 Issue。

评估

难度
4/5
预计耗时
3-5 天
新手友好度
45/100
Issue 类型
重构
描述清晰度
基本清楚
活跃度
停滞
技术栈
csharp
领域
performance

调研方向

Start in Utilities/ByteBuffer.cs by inspecting PutInt, PutLong, GetInt, and GetLong, then review how the buffer manages its internal storage. Compare before and after allocation behavior using benchmark results. Done means the serialization paths avoid the described per-call allocations and supporting before/after benchmarks are provided; the optional Memory or Span refactor is broader scope.

由索引模型根据 Issue 内容生成。

描述

agent enhancement

File: Utilities/ByteBuffer.cs

PutInt and PutLong both call BitConverter.GetBytes(...) which allocates a new byte[] on every call, then immediately Array.Copy it into the buffer. This is avoidable entirely using BinaryPrimitives and MemoryMarshal from System.Buffers:

// BEFORE - allocates byte[] every call
public void PutLong(long value)
{
    var longAsBytes = BitConverter.GetBytes(IPAddress.NetworkToHostOrder(value));
    Array.Copy(longAsBytes, 0, _internalBuffer, Position, longAsBytes.Length);
    Position += longAsBytes.Length;
}

// AFTER - zero allocation

public void PutLong(long value)

{

BinaryPrimitives.WriteInt64BigEndian(_internalBuffer.AsSpan(Position), value);

Position += sizeof(long);

}


public long GetLong()

{

var result = BinaryPrimitives.ReadInt64BigEndian(_internalBuffer.AsSpan(Position));

Position += sizeof(long);

return result;

}

The IPAddress.HostToNetworkOrder / NetworkToHostOrder dance exists solely to handle big-endian byte ordering — BinaryPrimitives.WriteInt64BigEndian does this directly, with no allocation. This applies equally to PutInt/GetInt. On serialization-heavy workloads this is high-impact.

The ByteBuffer itself could also be refactored to work over Memory<byte> or Span<byte> to allow callers to supply pooled or stack-allocated memory rather than always allocating internally.

Any pull requests addressing this change should provide supporting before and after benchmark results.

主要语言
C#
星标
185
派生
31
30 天内没有已合并 PR

环境准备

在 Codespaces 中打开

在浏览器里用你自己的 GitHub 账号启动这个项目的开发容器。

  • 没有 Dockerfile 或 Docker Compose 文件
  • 没有 Pull Request 模板
  • 阅读贡献指南

从这里开始

  1. 先读完整个 Issue，再读项目的贡献指南。
  2. 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
  3. Fork 仓库，在一个分支上完成修改。
  4. 提交 Pull Request，并在描述里引用这个 Issue 编号。

HdrHistogram/HdrHistogram.NET 的其他 Issue

查看 HdrHistogram/HdrHistogram.NET 的全部 Issue

相似的 Issue

更多 C# Issue

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。