An issue was discovered where certain message payloads were being lost in certain situations. The root cause ended up being overaggressive purge logic in the SQL stored procedures when an orchestration (or entity) does a ContinueAsNew.
This PR updates the ContinueAsNew T-SQL logic to be more selective in what data it deletes to ensure that unprocessed events are not impacted. This PR also adds a new test-case that was previously able to reproduce the issue reliably.
Fixes #146
An issue was discovered where certain message payloads were being lost in certain situations. The root cause ended up being overaggressive purge logic in the SQL stored procedures when an orchestration (or entity) does a ContinueAsNew.
This PR updates the ContinueAsNew T-SQL logic to be more selective in what data it deletes to ensure that unprocessed events are not impacted. This PR also adds a new test-case that was previously able to reproduce the issue reliably.