refactor: extract shared cfg_set publish into a helper

[Jagadeeshftw]

Summary

Six single-scalar admin config setters (set_allowlist_enabled, set_min_requests_per_call, set_max_requests_per_call, set_max_requests_per_window, set_rate_window_seconds, set_require_service_registration) each duplicated the identical 3-line env.events().publish((symbol_short!("cfg_set"),), (tag, value)) block. Pure refactor — extracts it into a private publish_cfg_event(env, tag, value) helper and routes all six call sites through it. Behavior is byte-for-byte identical: same topic, same payload shape, same call order relative to the storage write.

While refactoring I noticed the cfg_set topic — despite being the most widely shared event topic in the contract — had no direct payload-correctness test for any of the six setters (only an indirect topic-collision check). Added dedicated tests as regression coverage for this refactor.

Tests added

• One payload-correctness test per setter (6 total): test_cfg_set_payload_set_allowlist_enabled, ..._set_min_requests_per_call, ..._set_max_requests_per_call, ..._set_max_requests_per_window, ..._set_rate_window_seconds, ..._set_require_service_registration
• test_cfg_set_payload_tags_are_mutually_distinct — the six tag values (the field that disambiguates the shared topic) can never collide

Closes #290

Validation

• cargo fmt --all -- --check
• cargo clippy --all-targets -- -D warnings
• cargo test -p escrow — 359 passed (7 new), 1 pre-existing failure unrelated to this change (test_transfer_service_ownership_genuine_transfer_emits_event, fails identically on main)

[mikewheeleer]

nice one @Jagadeeshftw, checks pass. merging 🚀