test: fill coverage gaps with 87 new integration tests across bot + dispatch + events

[devin-ai-integration]

Summary

Autonomous coverage fill-in from an audit over scripts/audit_teamtalk_coverage.py plus a manual walk of crates/teamtalk/src/ and crates/teamtalk/tests/. Adds 87 new integration tests across six focused files; no changes to src/ and no public API shifts.

Test count on this branch: 287 → 374 (+87), all green under cargo test --workspace --all-features.

Files added

• crates/teamtalk/tests/bot_fsm_extra_tests.rs — 19 tests. DialogState defaults/expiry-boundary/timeout-policy round-trip, metadata replace+remove, legacy pipe decode & malformed rejections; DialogFlow next/previous/start/terminal navigation; DialogMachine session id uniqueness, advance reactivates paused, Clear vs Pause timeout policy on live expired, set/clear timeout, metadata persistence across reloads, custom prefix namespacing, is_in ignores paused, restart_flow rotates session id.
• crates/teamtalk/tests/dispatcher_extra_tests.rs — 11 tests. Empty-queue step, event-specific filtering, wildcard for-all, Stop flow from any slot, shortcut handlers (connect_success/connection_lost/connect_failed/command_error/user_joined/text_message), multi-specific ordering, specific+wildcard insertion order merge, Stop semantics (surfaces via step() while siblings still run).
• crates/teamtalk/tests/events_error_extra_tests.rs — 14 tests. Event::is_reconnect_needed and is_reconnect_needed_with (discriminant match, payload irrelevance, empty extras); ConnectionState::default() and Joining(ChannelId) inequality; Error Display for each variant; IoError/ClientError message fields; FfiError Display per variant; From<FfiError> cascade through Error::Ffi; SdkError Clone payload round-trip; Event::Unknown(ClientEvent) payload.
• crates/teamtalk/tests/bot_state_v2_extra_tests.rs — 13 tests. set overwrite without growth, set_with_ttl replaces permanent, plain set clears TTL, remove on expired entry still evicts (documents current implementation: returns stale value but frees the slot), remove_prefix zero-match, empty-prefix removes all, keys with empty prefix, exists for expired, get_many preserves requested order, set_many empty-batch and overwrite, get for missing/expired, remove_prefix with mixed TTL expiry.
• crates/teamtalk/tests/scheduler_tests.rs — 11 tests. Empty-scheduler sentinel delay, every_named + every (unique auto-names), remove idempotence, set_enabled toggle, set_enabled on missing is noop, is_enabled → None on missing, recurring job fires across ticks, after job is one-shot and fully removed after running, disabled job skipped, next_run_delay only considers enabled jobs.
• crates/teamtalk/tests/bot_command_extra_tests.rs — 19 tests. parse_command multi-prefix / empty+prefix-only rejection / argument order / whitespace collapsing; CommandPattern::parse rejects empty/required-after-optional/variadic-not-last/command-token-after-args; min_args/max_args for fixed and variadic tails; accepts arity validation; descriptor flags (required/variadic); multi-word command names; usage bracketing + prefix insertion; Args raw/rest/get-parse-error/missing/all.

What was NOT changed

• No src/ files touched.
• No public API surface shifts; no new items re-exported; no trait methods added.
• All new tests compile under existing feature flags (mock / bot) already used by sibling tests.

Observations from audit that did NOT become test changes

• audit_teamtalk_coverage.py reports 203 "missing_tests" symbols, but the audit's own README notes the metric is strict direct-symbol grep — it flags tests that don't literally reference TT_* names. Most high-level wrappers are exercised through the safe API, so expanding that metric by parroting TT_* identifiers into assertions would not increase real behavioural coverage. The gaps I did fill are areas where the high-level Rust behaviour itself was lightly covered.
• Router-group / unknown-command-policy / auto-help coverage is a legitimate next fill-in, but it requires a mock harness that wires a Bot through a full message flow; deferred to a follow-up rather than stubbed shallowly here.

Review & Testing Checklist for Human

• cargo test --workspace --all-features — should report 374 passed on this branch.
• Spot-check one non-obvious test assumption: bot_state_v2_extra_tests::remove_evicts_expired_entry_even_though_it_returns_the_stale_value documents that MemoryStateStore::remove does not consult the TTL — it returns the stored value even if expired, then evicts the slot. If you want strict TTL-aware remove, that's an src/ change (not in this PR).
• Spot-check dispatcher_extra_tests::dispatcher_stop_from_first_handler_still_surfaces_stop_and_lets_siblings_run — confirms current semantics: Stop from one handler still lets siblings on the same event fire; Stop is reported at step() return so the outer loop exits after the event. If you'd rather have "short-circuit within event", that's also an src/ change.

Notes

This PR is the coverage-fill follow-up the user explicitly approved ("сам почитаеш и посмотриш как там что ну точнее если быть точнее куда где что ещё можно тестами в 58 и сделаеш"). Two new tests were written as documenting tests rather than bug reports — they record existing behaviour that looked surprising but is internally consistent. Call out in review if either should change behaviour instead; happy to follow up with the src/ change in a separate PR.

Link to Devin session: https://app.devin.ai/sessions/71fdd6196cb74723a2e277bb81993a9c
Requested by: @BlindMaster24

---

[devin-ai-integration]

🤖 Devin AI Engineer

I'll be helping with this pull request! Here's what you should know:

✅ I will automatically:

• Address comments on this PR. Add '(aside)' to your comment to have me ignore it.
• Look at CI failures and help fix them

Note: I can only respond to comments from users who have write access to this repository.

⚙️ Control Options:

• Disable automatic comment and CI monitoring

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no potential bugs to report.

View in Devin Review to see 5 additional findings.