feat(exchange): scope pending-exchange cancellation by origin + ladder plumbing (L3)

[cristim]

What

Work item L3 of the auto-laddering full-automation plan (closes gap G10 / issue #1348). Prevents the standalone ri_exchange_reshape task and ladder runs from silently cancelling each other's pending exchanges, and adds the plumbing the ladder write path needs.

Today RunAutoExchange begins by cancelling all pending exchanges store-wide. Once laddering creates pending reshapes, the standalone task (its own 6h schedule) would silently cancel a ladder run's awaiting-approval reshapes.

Changes

• Scoped cancellation: new CancelPendingExchangesByOrigin(ctx, origin common.ExchangeOrigin) replaces the store-wide cancel. ExchangeOriginStandalone cancels only ladder_run_id IS NULL rows; ExchangeOriginLadder cancels only IS NOT NULL. The origin is a validated typed enum (pkg/common, fails loud on an unknown value) — not a bare bool.
• LadderRunID plumbing: added to the exchange record type and the ri_exchange_history.ladder_run_id column is now written/read (INSERT is 21 columns/args, all 5 SELECT/RETURNING + Scan aligned; an alignment regression test guards it). Ladder-originated exchanges are tagged; standalone rows stay NULL.
• True dry-run: DryRun=true performs zero mutations (no cancel, no save, no Execute, no token generation) and returns Simulated=true.
• Seam correctness: exchangeRunner.RunAutoExchange is widened to (ctx, cfg, ladderRunID *string, dryRun bool) so the concrete runner (wired in L16) is structurally forced to thread LadderRunID — it cannot compile while defaulting to nil and re-introducing the cross-task-cancel bug.

Scope / current state

Standalone-side scoping is active now (a single-source deployment behaves exactly as before). Ladder-side scoping activates when the write path's concrete runner is wired (L16); until then the ladder reshape path returns errWriteNotWired. The IS NOT NULL partition is deliberately coarse (is-ladder, not per-run); a per-run/per-config scoping key is tracked in #1367 for before ladder reshapes create pendings.

Verification (exit 0)

• go build ./..., go vet ./... clean.
• pkg go test ./exchange/... 100 pass; go test ./internal/server/... ./internal/config/... 1010 pass; integration (real PG/pgxmock) 733 pass; providers/aws ladder 115 pass; gocyclo -over 10 clean.
• Tests prove: standalone cancel does NOT touch ladder pendings and vice-versa (fail-before/pass-after on a WHERE-clause swap); dry-run performs zero mutations (asserted via mock.AssertExpectations, no sleeps); LadderRunID round-trips (nil forwarded as nil, no fabricated value); unknown origin fails loud with no query issued; 21-arg INSERT column alignment.

Part of #1336. Tracker #1333. Addresses #1348.

[coderabbitai]

Warning

Review limit reached

You’ve reached a temporary PR review limit under our Fair Usage Limits Policy.

Your recent review volume is higher than typical usage, so adaptive limits are currently applied.

Next review available in: 49 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 4b6aeae5-fa34-4c26-b831-c902c69af397

📥 Commits

Reviewing files that changed from the base of the PR and between 7adacd7 and 5b9a8e5.

⛔ Files ignored due to path filters (1)

• pkg/go.sum is excluded by !**/*.sum

📒 Files selected for processing (18)

• internal/analytics/collector_test.go
• internal/config/interfaces.go
• internal/config/store_postgres.go
• internal/config/store_postgres_nil_db_test.go
• internal/config/store_postgres_pgxmock_test.go
• internal/config/types.go
• internal/mocks/stores.go
• internal/server/handler_ri_exchange.go
• internal/server/handler_ri_exchange_test.go
• internal/server/test_helpers_test.go
• pkg/common/types.go
• pkg/exchange/auto.go
• pkg/exchange/auto_test.go
• pkg/go.mod
• pkg/ladder/capability.go
• providers/aws/ladder/interfaces.go
• providers/aws/ladder/reshape.go
• providers/aws/ladder/reshape_test.go

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch feat/ladder-reshape-scoped-cancel

---

_{Comment @coderabbitai help to get the list of available commands.}

[cristim]

@coderabbitai review

[coderabbitai]

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[cristim]

Rebased onto post-#1362 main (7adacd715), resolving the 4 reported textual conflicts. All were identical, trivial import-block collisions (this branch adds pkg/common, #1362 added pkg/ladder); resolved by keeping both import lines (union) in:

• internal/config/interfaces.go
• internal/mocks/stores.go
• internal/server/test_helpers_test.go
• internal/analytics/collector_test.go

No semantic conflict: main's factory.go WithWriteSide is satisfied by the widened RunAutoExchange seam, and all StoreInterface implementers satisfy both widenings.

Union gates on the rebased head 5b9a8e525 (all exit 0):

Gate	Result
go build ./... (root)	exit 0
go vet ./... (root)	exit 0
cd pkg && go test ./exchange/...	100 passed
go test ./internal/server/... ./internal/config/...	1041 passed
cd providers/aws && go build && go vet && go test ./ladder/...	115 passed
go test -tags=integration ./internal/config/...	749 passed

[cristim]

@coderabbitai review

[coderabbitai]

✅ Action performed

Review finished.

Note: CodeRabbit is an incremental review system and does not re-review already reviewed commits. This command is applicable only when automatic reviews are paused.

[cristim]

Merged to main after: two Fable review rounds (typed ExchangeOrigin enum, correct-by-construction ladder seam forcing LadderRunID/DryRun through, pgxmock WHERE-clause + 21-arg INSERT alignment tests), CodeRabbit clean, rebase onto post-#1362 main (4 import-union conflicts), and a local union re-verify against post-#1365 main (build/vet/ladder tests all exit 0) since both PRs touched providers/aws/ladder/interfaces.go. Standalone and ladder pending-exchange cancellation are now origin-scoped and cannot cross-cancel; per-run scoping before ladder reshapes create pendings is tracked in #1367 (needed with L16).