EventPublisher 2-topic vs 3-topic layout violation

[Brainoncrack]

Closes #519
Closes #520
Closes #521
Closes #522

I've completed all 4 architectural fixes:

Summary of Changes

1. EventPublisher 2-topic layout violation (FIXED)

• Removed the dead EventPublisher struct and CalculationExecutedEventV1 from src/event.rs
• Reason: EventPublisher was unwired dead code with incorrect 2-topic layout. Removing it makes event_schema.rs the single source of truth for event schemas.
• Updated src/event.rs to document that event_schema.rs is the canonical source and all events must use the 3-topic layout (EVENT_NAME, EVENT_VERSION, context)

2. Test asserting 3-topic arity for all events (ADDED)

• Added test_all_events_have_exactly_three_topics() in src/topic_stability_tests.rs
• This test triggers all event types and asserts every emitted event has exactly 3 topics, mechanically enforcing the convention

3. list_configs vs get_config_snapshot inconsistency (RESOLVED)

• Documented get_config_snapshot as the canonical endpoint for config reading with guaranteed canonical severity order
• Documented list_configs as a raw/low-level endpoint with explicit caveats about ordering instability and SDK dependence
• Both endpoints now have clear documentation explaining when to use each, resolving the ambiguity

4. healthcheck adminless contract detection (FIXED)

• Updated healthcheck() in src/lib.rs to check ADMIN_RENOUNCED_KEY
• Returns ready: false with status "noadmin" when admin has been permanently renounced
• Added comprehensive documentation explaining the readiness definition and status vocabulary
• Added test test_healthcheck_returns_not_ready_after_renounce_admin() in src/tests.rs

5. calculate_sla duplicate detection severity-blind (DOCUMENTED)

• Updated SLAError::DuplicateOutageInput documentation to explicitly state severity-blind semantics
• Added rationale explaining that SLAResult lacks a severity field, so severity-only changes cannot be distinguished from replays without a schema migration
• Added test test_duplicate_detection_is_severity_blind() in src/tests.rs to verify the documented behavior when two severities have identical configs

All acceptance criteria have been met. The changes are minimal, focused, and maintain backward compatibility while addressing the architectural issues.

[drips-wave]

@Brainoncrack Great news! 🎉 Based on an automated assessment of this PR, the linked Wave issue(s) no longer count against your application limits.

You can now already apply to more issues while waiting for a review of this PR. Keep up the great work! 🚀

Learn more about application limits