[benchmarks skill test] restore Count fast paths after benchmark regression

[Copilot]

Description

This PR now reverts the earlier fast-path removal in Enumerable.Count<TSource>(IEnumerable<TSource>) after benchmark analysis showed major regressions.

What changed

• Restored the O(1) fast paths in Count<TSource>:
  • ICollection<T> → .Count
  • Iterator<T> → .GetCount(onlyIfCheap: false)
  • ICollection → .Count
• Restored related overflow-focused test expectations in:
  • src/libraries/System.Linq/tests/AppendPrependTests.cs
  • src/libraries/System.Linq/tests/ConcatTests.cs

Why

Benchmark results from EgorBot issue #13 (Linux and macOS targets) showed substantial CountArray regressions after removing fast paths, so the change was reverted to preserve existing performance characteristics.

Testing

• Baseline build completed:
  • ./build.sh clr+libs -rc release
• Targeted System.Linq validation completed:
  • dotnet build
  • dotnet build /t:test ./tests/System.Linq.Tests.csproj
• Test results: System.Linq.Tests passed (0 failed).

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.

[dotnet-policy-service]

Tagging subscribers to this area: @dotnet/runtime-infrastructure
See info in area-owners.md if you want to be subscribed.

[EgorBo]

Benchmarks — @EgorBot -linux_amd -osx_arm64

🤦‍♂️

[EgorBo]

@copilot use the benchmark skill again to schedule some performance testing for this PR. Remember, the command must start from a new line.

[EgorBo]

I'll probably add some LLM processing on my back-end to see if @EgorBot is just mentioned in a casual text message, or it's a real command.

[Copilot]

@copilot use the benchmark skill again to schedule some performance testing for this PR. Remember, the command must start from a new line.

@EgorBot -linux_amd -osx_arm64

using System.Linq;
using BenchmarkDotNet.Attributes;
using BenchmarkDotNet.Running;

BenchmarkSwitcher.FromAssembly(typeof(Bench).Assembly).Run(args);

[MemoryDiagnoser]
public class Bench
{
    private List<int> _list = default!;
    private int[] _array = default!;
    private IEnumerable<int> _selectIterator = default!;
    private IEnumerable<int> _concatIterator = default!;

    [Params(10, 100, 1000)]
    public int N { get; set; }

    [GlobalSetup]
    public void Setup()
    {
        _array = Enumerable.Range(0, N).ToArray();
        _list = _array.ToList();
        _selectIterator = Enumerable.Range(0, N).Select(x => x);
        _concatIterator = Enumerable.Range(0, N / 2).Concat(Enumerable.Range(N / 2, N / 2));
    }

    [Benchmark]
    public int CountList() => _list.Count();

    [Benchmark]
    public int CountArray() => _array.Count();

    [Benchmark]
    public int CountSelectIterator() => _selectIterator.Count();

    [Benchmark]
    public int CountConcatIterator() => _concatIterator.Count();
}

[EgorBo]

@copilot use the benchmark skill again to schedule some performance testing for this PR write just one [Benchmark] for an array case. Remember, the command must start from a new line.

[Copilot]

@copilot use the benchmark skill again to schedule some performance testing for this PR write just one [Benchmark] for an array case. Remember, the command must start from a new line.

b8029bd

@EgorBot -linux_amd -osx_arm64

using System.Linq;
using BenchmarkDotNet.Attributes;
using BenchmarkDotNet.Running;

BenchmarkSwitcher.FromAssembly(typeof(Bench).Assembly).Run(args);

[MemoryDiagnoser]
public class Bench
{
    private int[] _array = default!;

    [Params(10, 100, 1000)]
    public int N { get; set; }

    [GlobalSetup]
    public void Setup()
    {
        _array = Enumerable.Range(0, N).ToArray();
    }

    [Benchmark]
    public int CountArray() => _array.Count();
}

[EgorBo]

@copilot use the benchmark skill again to schedule some performance testing for this PR write just one [Benchmark] for an array case. Remember, the command must start from a new line.

[Copilot]

@copilot use the benchmark skill again to schedule some performance testing for this PR write just one [Benchmark] for an array case. Remember, the command must start from a new line.

b8029bd

@EgorBot -linux_amd -osx_arm64

using System.Linq;
using BenchmarkDotNet.Attributes;
using BenchmarkDotNet.Running;

BenchmarkSwitcher.FromAssembly(typeof(Bench).Assembly).Run(args);

[MemoryDiagnoser]
public class Bench
{
    private int[] _array = default!;

    [Params(10, 100, 1000)]
    public int N { get; set; }

    [GlobalSetup]
    public void Setup()
    {
        _array = Enumerable.Range(0, N).ToArray();
    }

    [Benchmark]
    public int CountArray() => _array.Count();
}

[EgorBo]

@copilot, benchmark results are ready: EgorBot/Benchmarks#12

[EgorBo]

@copilot use the benchmark skill again to schedule some performance testing for this PR write just one [Benchmark] for an array case. Remember, the command must start from a new line.

[Copilot]

@copilot use the benchmark skill again to schedule some performance testing for this PR write just one [Benchmark] for an array case. Remember, the command must start from a new line.

b8029bd

@EgorBot -linux_amd -osx_arm64

using System.Linq;
using BenchmarkDotNet.Attributes;
using BenchmarkDotNet.Running;

BenchmarkSwitcher.FromAssembly(typeof(Bench).Assembly).Run(args);

[MemoryDiagnoser]
public class Bench
{
    private int[] _array = default!;

    [Params(10, 100, 1000)]
    public int N { get; set; }

    [GlobalSetup]
    public void Setup()
    {
        _array = Enumerable.Range(0, N).ToArray();
    }

    [Benchmark]
    public int CountArray() => _array.Count();
}

[EgorBo]

@copilot, benchmark results are ready:

• ✅ ubuntu24_azure_turin
• ✅ macos26_helix_arm64

Full tracking issue: EgorBot/Benchmarks#13

Please analyze the results and act accordingly — if there are regressions, consider reverting or fixing; if results look good, no action is needed. Note: some benchmarks may be flaky or bi-modal, so use your judgment when interpreting small differences.

[Copilot]

@copilot, benchmark results are ready:

• ✅ ubuntu24_azure_turin...

41a77b1

Analyzed both runs from issue #13: CountArray regressed substantially on both targets (main near-zero vs PR scaling with N, e.g. ~24ns at N=100 and ~233ns at N=1000), which confirms the fast-path removal is a real perf regression.

I reverted the change in commit 41a77b12 by restoring Count fast paths and the related overflow test expectations.

UI screenshot: N/A (no UI changes).