Result: done
Files: crates/nikaia-std/std.contracts tests/compiler/tests/a_set_is_walked_by_iter.nika tests/compiler/tests/a_set_is_walked_by_iter.stdout
Commit: std: every set has iter() (ADR-collections-2 D3)
Body: Only the contract was needed: collections::BTreeSet::iter (Seq[ref(self) T] sync ends sized, smallest first) and collections::HashSet::iter (Seq[ref(self) T] sync sized, no order) in std.contracts; the emitter already lowers both to Rust’s .iter() (no checker or lowering row, lower-std fixpoint without diff), in a program and in a tools file. The orphan comment above BTreeSet::first moved onto BTreeSet::clone. The test’s HashSet sum is bound with let total: i64 first, because println(v.iter().sum()) fails at rustc for a Vec too (see Found).
Fixes: #1006
Tests: a_set_is_walked_by_iter (tests/compiler 358 passed; NK1171 on the unfixed compiler); K3a form lowered to named.iter().any(|name| ...) and run through rustc; tools-file probe under lower-std, restored
Gate: green
Deviation: none
Found: println(v.iter().sum()) on a Vec[i64] reaches rustc: “type annotations needed, Sum
