feat(postgresql): message-store reset clears the Postgres transport queue tables - #3
Draft
uniquelau wants to merge 1 commit into
Draft
feat(postgresql): message-store reset clears the Postgres transport queue tables#3uniquelau wants to merge 1 commit into
uniquelau wants to merge 1 commit into
Conversation
ClearAllAsync / RebuildAsync previously truncated only the envelope tables (incoming/outgoing/dead-letter/node/listener), leaving the PostgreSQL queue transport's queue and scheduled-message tables populated. Integration tests over the Postgres queue transport therefore carried rows between runs. Add a per-provider hook, truncateAdditionalTablesAsync(DbTransaction), invoked inside the existing reset transaction so the whole reset stays atomic. The PostgreSQL store overrides it to delete from the transport's own queue + scheduled tables. Scoped to the transport's table types on purpose: AddTable is a general registration path (SQL Server registers a rate-limit table through it that a reset must keep), so a blanket base-class loop would be wrong. Default behaviour for every other provider is unchanged. No public API added. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
uniquelau
force-pushed
the
feat/clear-transport-queue-tables
branch
from
July 20, 2026 08:50
84bbb5e to
eac0c23
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Review copy on our fork — please review before I open this against JasperFx/wolverine.
Problem
ClearAllAsync()/ the message-store reset truncates incoming/outgoing/dead-letter/node/listener tables but not the Postgres transport queue tables. Integration tests over the Postgres queue transport therefore carry queue rows between runs and need a bespoke reset.Change (transactional, additive, Postgres-scoped)
MessageDatabase.Admin.cs— newprotected virtual truncateAdditionalTablesAsync(DbTransaction, CancellationToken)hook (default no-op), invoked inside the existing reset transaction just before commit — so any extra clears run atomically with the envelope truncation.PostgresqlMessageStore.cs— override that clears only the transport'sQueueTable/ScheduledMessageTable(matched by type from the registered_otherTables), reusing the transport's owndelete from {table}pattern.reset_clears_transport_queue_tables— enqueues one immediate + one scheduled message, callsClearAllAsync(), asserts both counts drop to 0. Negative control confirmed: with the hook call disabled the test fails; enabled it passes.Why not a base-class loop: SqlServer registers a
RateLimitTablethrough the sameAddTablepath, so a blanket "truncate every registered table" would wrongly clear it. This keeps the behaviour Postgres-scoped via the virtual hook — SqlServer and other providers keep the no-op default (their build is confirmed green, 0 warnings).No public API added. 43 lines across two source files.
One design point for your call before upstream
The Postgres override identifies the transport tables by filtering
_otherTablesfor the internalQueueTable/ScheduledMessageTabletypes — zero transport changes, no new public surface, but it couples the store to the transport's table classes (same assembly, so acceptable). The alternative is an explicit "register this table for reset-clearing" call from the transport — more decoupled, but adds an API and a call site. I took the minimal no-new-API path; easy to flip if the maintainer prefers the explicit registration.