fix(testing): emit StopBenchmark command even when benches failed

[not-matthias]

No description provided.

[Copilot]

Pull Request Overview

This PR fixes an issue where the StopBenchmark command was not being emitted when benchmarks fail, which could result in excessive sampling of irrelevant performance data and incorrect marker pairs (having SampleStart without SampleStop).

Key changes:

• Added ensureBenchmarkIsStopped() function to safely stop benchmarks on failure
• Calls this function in two failure scenarios: in run1() when the initial benchmark run fails, and in processBench() when the main benchmark run fails
• Updated the build process to apply the new patch

Reviewed Changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
testing/testing/codspeed.go	Adds ensureBenchmarkIsStopped() helper function to ensure StopBenchmark is called when benchmarks fail
testing/testing/benchmark.go	Calls ensureBenchmarkIsStopped() in two failure paths: after run1() fails and after doBench() fails in processBench()
testing/patches/benchmark_stopbenchmark_fail.patch	Patch file containing the changes to benchmark.go for the fork process
testing/fork.sh	Applies the new patch during the build process to modify the forked Go testing package

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[codspeed-hq]

CodSpeed Performance Report

Merging #40 will degrade performances by 17.34%

_{Comparing cod-1703-codspeed-go-run-fails-due-to-multiple-start-markers (ed16991) with main (140fee0)}

Summary

❌ 1 regression
✅ 41 untouched

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Benchmark	BASE	HEAD	Change
❌	bench_collect_results[10000000, 25]	5 s	6 s	-17.34%

[GuillaumeLagrange]

lgtm