Crash when upgrading to Ecto 3.12

[guess]

Environment

• Oban Version: 2.18.0
• PostgreSQL image: imresamu/postgis-arm64:15-3.4-alpine3.19
• Elixir 1.16.2 (compiled with Erlang/OTP 25)

Current Behavior

Since upgrading to Ecto 3.12.1, I have been getting this crash, and I'm not sure if this is an Ecto issue or if Oban needs to change anything related to the changes made from Ecto 3.12, so just wanted to bring to your attention.

[error] GenServer {Oban.Registry, {Oban, Oban.Stager}} terminating
** (FunctionClauseError) no function clause matching in Ecto.Adapters.Postgres.Connection.distinct/3
    (ecto_sql 3.11.3) lib/ecto/adapters/postgres/connection.ex:482: Ecto.Adapters.Postgres.Connection.distinct(%Ecto.Query.ByExpr{expr: true, file: "/app/deps/oban/lib/oban/engines/basic.ex", line: 178, params: nil, subqueries: []}, {{[[34, "public", 34], 46, [34, "oban_jobs", 34]], [111 | "0"], Oban.Job}, []}, #Ecto.Query<from j0 in Oban.Job, where: j0.state == "available", where: not is_nil(j0.queue), distinct: true, select: j0.queue>)
    (ecto_sql 3.11.3) lib/ecto/adapters/postgres/connection.ex:162: Ecto.Adapters.Postgres.Connection.all/2
    (ecto_sql 3.11.3) lib/ecto/adapters/postgres.ex:132: Ecto.Adapters.Postgres.prepare/2
    (ecto 3.12.1) lib/ecto/query/planner.ex:202: Ecto.Query.Planner.query_without_cache/4
    (ecto 3.12.1) lib/ecto/query/planner.ex:170: Ecto.Query.Planner.query_prepare/6
    (ecto 3.12.1) lib/ecto/query/planner.ex:143: Ecto.Query.Planner.query_with_cache/8
    (ecto 3.12.1) lib/ecto/repo/queryable.ex:214: Ecto.Repo.Queryable.execute/4
    (ecto 3.12.1) lib/ecto/repo/queryable.ex:19: Ecto.Repo.Queryable.all/3
    (oban 2.18.0) lib/oban/engines/basic.ex:180: Oban.Engines.Basic.check_available/1
    (oban 2.18.0) lib/oban/engine.ex:264: anonymous fn/2 in Oban.Engine.check_available/1
    (oban 2.18.0) lib/oban/engine.ex:339: anonymous fn/3 in Oban.Engine.with_span/4
    (telemetry 1.2.1) /app/deps/telemetry/src/telemetry.erl:321: :telemetry.span/3
    (oban 2.18.0) lib/oban/stager.ex:88: Oban.Stager.notify_queues/1
    (oban 2.18.0) lib/oban/stager.ex:73: anonymous fn/1 in Oban.Stager.stage_and_notify/2
    (ecto_sql 3.11.3) lib/ecto/adapters/sql.ex:1358: anonymous fn/3 in Ecto.Adapters.SQL.checkout_or_transaction/4
    (db_connection 2.7.0) lib/db_connection.ex:1756: DBConnection.run_transaction/4
    (oban 2.18.0) lib/oban/stager.ex:70: Oban.Stager.stage_and_notify/2
    (oban 2.18.0) lib/oban/stager.ex:57: anonymous fn/2 in Oban.Stager.handle_info/2
    (telemetry 1.2.1) /app/deps/telemetry/src/telemetry.erl:321: :telemetry.span/3
    (oban 2.18.0) lib/oban/stager.ex:56: Oban.Stager.handle_info/2
Last message: :stage
State: %Oban.Stager{conf: %Oban.Config{dispatch_cooldown: 5, engine: Oban.Engines.Basic, get_dynamic_repo: nil, insert_trigger: true, log: false, name: Oban, node: "45f857b40976", notifier: {Oban.Notifiers.Postgres, []}, peer: {Oban.Peers.Postgres, []}, plugins: [{Oban.Plugins.Cron, [crontab: [{"@hourly", Qultr.EventLinkScraperWorker}]]}, {Oban.Plugins.Pruner, [max_age: 604800]}], prefix: "public", queues: [default: [limit: 5], event_link_watcher: [limit: 1]], repo: Qultr.Repo, shutdown_grace_period: 15000, stage_interval: 1000, testing: :disabled}, timer: #Reference<0.690515242.1288962049.43907>, interval: 1000, limit: 5000, mode: :global}

Possibly related to one of these changes:

• Adapter changes: distinct, group_by, order_by and window expressions use the new Ecto.Query.ByExpr struct rather than the old Ecto.Query.QueryExpr struct
• [Ecto.Query] Subqueries are now supported in distinct, group_by, order_by and window expressions
• [Ecto.Query] Allow select_merge to be used in more insert_all and subquery operations by merging distinct fields

Expected Behavior

Not crash

[sorentwo]

Please try forcing a recompile, and/or clearing your build and recompiling. Our tests all failed in CI after the upgrade until we cleared the cache.

[sorentwo]

Closing this for now. If clearing the cache doesn't fix it, let us know.