ci: overlap corpus and tool setup

[gaborbernat]

The test jobs waited for the corpus submodules before installing Node and uv. The conformance job installed its Node oracle dependencies before setting up uv.

Fetch the corpus in the background and wait before starting tox. Install the conformance oracle dependencies alongside uv setup after Node is ready. Keep the CodSpeed workflows unchanged.

In the PR run, corpus fetching overlapped a median 3.5s of Node and uv setup across 22 jobs (range 1 to 10s). The conformance job took 357s versus 344s in the baseline run; we have not demonstrated a job-level reduction there. The timing snapshot lacks a completed whole-workflow comparison. These are single-run comparisons. Workflow wall time runs from workflow start to the last job completion and includes runner queues; cache differences and variation in unchanged jobs prevent attributing the whole-workflow difference to this change.

[codspeed]

Merging this PR will degrade performance by 20.1%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

❌ 1 regressed benchmark
✅ 580 untouched benchmarks
⏩ 32 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
❌	test_feature[shadow-slot-comments]	110.3 µs	138 µs	-20.1%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing ci-parallel-steps (88a06c6) with main (9ff1626)}

Footnotes

1. 32 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩