engine: vm.rs is approaching the 6000-line per-file lint cap #588

Problem

After element-cycle-resolution Phase 6, src/simlin-engine/src/vm.rs is 5920 lines — under, but within ~80 lines of, the hard MAX_LINES=6000 per-file threshold that scripts/lint-project.sh (rule at scripts/lint-project.sh:40-47) enforces on Rust sources. That lint runs in the pre-commit hook and CI.

This is tech debt / a maintainability risk, not a current violation: any future task that adds more than ~80 lines to vm.rs will trip the threshold and be forced to do a tangential split mid-stream before it can commit unrelated changes — the same failure mode as #471 (vdf.rs went over and broke pnpm lint on main), caught here proactively the way #518 was for db.rs.

How it got here

vm.rs has already actively managed headroom: the vector-op machinery for VECTOR ELM MAP and VECTOR SORT ORDER was spun out into sibling modules vm_vector_elm_map.rs and vm_vector_sort_order.rs specifically to keep vm.rs under the cap. The element-cycle-resolution work pushed it back up to 5920, so the slack from those extractions is largely consumed.

Why it matters

• The next non-trivial change to vm.rs will fail pnpm lint / pre-commit / CI on an otherwise-unrelated diff.
• The hard threshold exists to force modularization of super-sized files; bumping it ad hoc would undermine the policy.
• vm.rs is the bytecode VM/interpreter hub — a high-traffic file where new opcodes and builtin implementations land — so the squeeze will recur.

Component

• src/simlin-engine/src/vm.rs
• Lint rule in scripts/lint-project.sh

Possible approaches

Behavior-preserving extraction only. Following the existing sibling-module precedent (vm_vector_elm_map.rs, vm_vector_sort_order.rs):

• Extract another self-contained vector/array opcode implementation into its own vm_<op>.rs sibling.
• Or relocate a cohesive VM subsystem (e.g. a group of builtin implementations, or the debug/disassembly helpers such as debug_print_bytecode) into a sibling module.

The fix is mechanical: find a cohesive cluster and relocate it. Doing it now (with headroom to spare) is lower-risk than doing it under duress when a feature commit is blocked.

Context

Surfaced during the element-cycle-resolution Phase 6c code review (docs/implementation-plans/2026-05-18-element-cycle-resolution/). This is the vm.rs companion to #518 (db.rs at the same cap); for reference, db.rs is currently 5895 lines. See the 2026-05-19 update comment on #518.