fix(hints): silence cond{~v} discard-hint on multi-stmt guard bodies

[danieljohnmorris]

Summary

The cond{^"err"} / cond{~v} discarded-result hint added in #300 fires as a false positive on multi-statement guard bodies like

f n:n>R t t;=n 0{prnt "empty";~"ok"};~"done"

Surfaced by interactive-cli rerun4, where every successful run of a tracker emitted the hint. The author has clearly composed the multi-stmt block on purpose; firing there is noise that erodes trust in the hint.

Tightened the walker in collect_hints to require body.len() == 1. Only the original bare single-stmt {^"err"} / {~v} footgun trips the hint now.

Repro

Before (on main):

$ ilo run prog.ilo multi 1
done
{"hints":["hint: `cond{^\"err\"}` and `cond{~v}` discard the body expression..."]}

After:

$ ilo run prog.ilo multi 1
done

Single-stmt shape still fires:

$ ilo run single.ilo g 5    # source: g x:n>R n t;<x 0{^"neg"};~x
5
{"hints":["hint: `cond{^\"err\"}` and `cond{~v}` discard the body expression..."]}

What's in the diff

• src/main.rs: walker requires body.len() == 1 for the discard heuristic; two regression tests (collect_hints_multi_stmt_guard_body_with_trailing_ok_no_hint, collect_hints_multi_stmt_guard_body_with_trailing_err_no_hint).
• examples/cond-multi-stmt-guard-return.ilo: cross-engine pin for the no-hint shape via the examples_engines harness.

Existing tests cover the still-flagged single-stmt shapes (collect_hints_discarded_err_in_guard_body, collect_hints_discarded_ok_in_guard_body) plus nested foreach/match guards (collect_hints_discarded_err_inside_foreach_body, collect_hints_discarded_err_inside_match_arm_body), all of which have single-stmt guard bodies and continue to fire.

Test plan

• cargo test --release --features cranelift (full suite green)
• cargo fmt --check
• cargo clippy --release --features cranelift -- -D warnings
• CLI repro for the false-positive shape: no hint
• CLI repro for single-stmt {^"neg"}: still emits the hint
• examples_engines harness runs the new example across tree + vm engines

Follow-ups

None.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ All tests successful. No failed tests found.

📢 Thoughts on this report? Let us know!