[claude] fsst: ignored regression test for i32 offset overflow

[connortsui20]

Alternative to #7832 — same regression test for #7833, but placed in the existing encodings/fsst/src/tests.rs module instead of a new tests_large.rs, and gated with #[ignore] rather than the test_with CI env attributes.

The test allocates ~2.5 GiB for the input plus ~2.5 GiB for the compressed output, so it's too expensive to run by default even in CI. With #[ignore] it's skipped on every normal cargo test / cargo nextest run and can be invoked explicitly when needed:

cargo test --release -p vortex-fsst -- --ignored fsst_compress_offsets

When the underlying overflow in fsst_compress_iter (encodings/fsst/src/compress.rs:72, hardcoded VarBinBuilder::<i32>) is fixed, the trailing assert_eq!(compressed.len(), len) becomes the live regression assertion — just drop the #[ignore] (or leave it, since the run cost is unchanged).

Compared to #7832 this drops the test-with dev-dependency, the tests_large.rs module, and the corresponding mod tests_large; in lib.rs.

Tracks #7833.

🤖 Generated with Claude Code

---

Generated by Claude Code

[codspeed-hq]

Merging this PR will improve performance by 30.52%

⚠️ Unknown Walltime execution environment detected

Using the Walltime instrument on standard Hosted Runners will lead to inconsistent data.

For the most accurate results, we recommend using CodSpeed Macro Runners: bare-metal machines fine-tuned for performance measurement consistency.

⚡ 1 improved benchmark
✅ 1205 untouched benchmarks

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	new_alp_prim_test_between[f64, 32768]	250.8 µs	192.2 µs	+30.52%

---

_{Comparing claude/fix-issue-7833-Qunwf (9aa32d9) with develop (eda04e1)}