Cod 1815 - Add the full command command that was run to the benchmark uri

[GuillaumeLagrange]

No description provided.

[codspeed-hq]

Merging this PR will not alter performance

✅ 4 untouched benchmarks

---

_{Comparing cod-1815-add-the-command-that-was-run-to-the-upload-payload-and (1ac1eae) with main (0cfb188)}

[not-matthias]

I think we can improve the testing a bit, to avoid re-implementing the same set of shell tests in every new executor (lmk if there are any technical reasons that prevent this for the exec-harness).

Other than that, it looks pretty good. Next review should be quick :)

[not-matthias]

We also have the executor tests, which already include a lot of these edge cases:

runner/src/executor/tests.rs

Lines 14 to 50 in 0cfb188

	const TESTS: [&str; 6] = [
	// Simple echo command
	"echo 'Hello, World!'",
	// Multi-line commands without semicolons
	"echo \"Working\"
	echo \"with\"
	echo \"multiple lines\"",
	// Multi-line commands with semicolons
	"echo \"Working\";
	echo \"with\";
	echo \"multiple lines\";",
	// Directory change and validation
	"cd /tmp
	# Check that the directory is actually changed
	if [ $(basename $(pwd)) != \"tmp\" ]; then
	exit 1
	fi",
	// Quote escaping test
	"#!/bin/bash
	VALUE=\"He said \\\"Hello 'world'\\\" & echo \\$HOME\"
	if [ \"$VALUE\" = \"He said \\\"Hello 'world'\\\" & echo \\$HOME\" ]; then
	echo \"Quote test passed\"
	else
	echo \"ERROR: Quote handling failed\"
	exit 1
	fi",
	// Command substitution test
	"#!/bin/bash
	RESULT=$(echo \"test 'nested' \\\"quotes\\\" here\")
	COUNT=$(echo \"$RESULT\" | wc -w)
	if [ \"$COUNT\" -eq \"4\" ]; then
	echo \"Command substitution test passed\"
	else
	echo \"ERROR: Expected 4 words, got $COUNT\"
	exit 1
	fi",
	];

We should test it there, so that we can reuse the existing test cases that ensure that the same shell scripts work across all executors.

[GuillaumeLagrange]

As discussed, I added a small subset of commands to run with the exec-harness. For now, script-like multi-commands are kinda out of scope, we'll see if we get user demand.

Kept the integration tests in the exec-harness, we'll see if we need more integration tests with pipes and whatnot form the codspeed runner

[not-matthias]

LGTM!