[dependabot]

Bumps the go-dependencies group with 6 updates in the / directory:

Package	From	To
github.com/riverqueue/river	0.50.0	0.51.0
github.com/riverqueue/river/riverdriver	0.50.0	0.51.0
github.com/riverqueue/river/riverdriver/riverpgxv5	0.50.0	0.51.0
github.com/riverqueue/river/riverdriver/riversqlite	0.50.0	0.51.0
github.com/riverqueue/river/rivershared	0.50.0	0.51.0
github.com/riverqueue/river/rivertype	0.50.0	0.51.0

Bumps the go-dependencies group with 4 updates in the /riverproui directory: github.com/riverqueue/river, github.com/riverqueue/river/riverdriver, github.com/riverqueue/river/rivershared and github.com/riverqueue/river/rivertype.

Updates github.com/riverqueue/river from 0.50.0 to 0.51.0

Release notes

Sourced from github.com/riverqueue/river's releases.

v0.51.0

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Changelog

Sourced from github.com/riverqueue/river's changelog.

[0.51.0] - 2026-10-31

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Commits

• 69c0b45 [go] Prepare v0.51.0 (#1571)
• 737146d Add a Where escape hatch to JobDeleteManyParams (#1570)
• e898054 [go, java, js, rust] Alternative implementation for reconciling jobs against ...
• 4f79f7b Bump crass from 1.0.7 to 1.0.8 in /ruby/rails/riverqueue-rails (#1562)
• f995760 [ruby] Prepare v0.15.0 (#1561)
• b330cfb [ruby] Add river migrate-get/validate/version subcommands (#1560)
• 5c2d873 [go] Fix notifier hanging after a listen or unlisten times out (#1559)
• 53ce383 [go] Fix InsertManyFast and QueueUpdate panics on drivers without listene...
• fc6c1d5 [java, js, ruby] bump compatible dependencies (#1553)
• 012b72c Fix intermittently failing test TestJobCleaner/CustomizableInterval (#1531)
• Additional commits viewable in compare view

Updates github.com/riverqueue/river/riverdriver from 0.50.0 to 0.51.0

Release notes

Sourced from github.com/riverqueue/river/riverdriver's releases.

v0.51.0

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Changelog

Sourced from github.com/riverqueue/river/riverdriver's changelog.

[0.51.0] - 2026-10-31

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Commits

• 69c0b45 [go] Prepare v0.51.0 (#1571)
• 737146d Add a Where escape hatch to JobDeleteManyParams (#1570)
• e898054 [go, java, js, rust] Alternative implementation for reconciling jobs against ...
• 4f79f7b Bump crass from 1.0.7 to 1.0.8 in /ruby/rails/riverqueue-rails (#1562)
• f995760 [ruby] Prepare v0.15.0 (#1561)
• b330cfb [ruby] Add river migrate-get/validate/version subcommands (#1560)
• 5c2d873 [go] Fix notifier hanging after a listen or unlisten times out (#1559)
• 53ce383 [go] Fix InsertManyFast and QueueUpdate panics on drivers without listene...
• fc6c1d5 [java, js, ruby] bump compatible dependencies (#1553)
• 012b72c Fix intermittently failing test TestJobCleaner/CustomizableInterval (#1531)
• Additional commits viewable in compare view

Updates github.com/riverqueue/river/riverdriver/riverpgxv5 from 0.50.0 to 0.51.0

Release notes

Sourced from github.com/riverqueue/river/riverdriver/riverpgxv5's releases.

v0.51.0

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Changelog

Sourced from github.com/riverqueue/river/riverdriver/riverpgxv5's changelog.

[0.51.0] - 2026-10-31

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Commits

• 69c0b45 [go] Prepare v0.51.0 (#1571)
• 737146d Add a Where escape hatch to JobDeleteManyParams (#1570)
• e898054 [go, java, js, rust] Alternative implementation for reconciling jobs against ...
• 4f79f7b Bump crass from 1.0.7 to 1.0.8 in /ruby/rails/riverqueue-rails (#1562)
• f995760 [ruby] Prepare v0.15.0 (#1561)
• b330cfb [ruby] Add river migrate-get/validate/version subcommands (#1560)
• 5c2d873 [go] Fix notifier hanging after a listen or unlisten times out (#1559)
• 53ce383 [go] Fix InsertManyFast and QueueUpdate panics on drivers without listene...
• fc6c1d5 [java, js, ruby] bump compatible dependencies (#1553)
• 012b72c Fix intermittently failing test TestJobCleaner/CustomizableInterval (#1531)
• Additional commits viewable in compare view

Updates github.com/riverqueue/river/riverdriver/riversqlite from 0.50.0 to 0.51.0

Release notes

Sourced from github.com/riverqueue/river/riverdriver/riversqlite's releases.

v0.51.0

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Changelog

Sourced from github.com/riverqueue/river/riverdriver/riversqlite's changelog.

[0.51.0] - 2026-10-31

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Commits

• 69c0b45 [go] Prepare v0.51.0 (#1571)
• 737146d Add a Where escape hatch to JobDeleteManyParams (#1570)
• e898054 [go, java, js, rust] Alternative implementation for reconciling jobs against ...
• 4f79f7b Bump crass from 1.0.7 to 1.0.8 in /ruby/rails/riverqueue-rails (#1562)
• f995760 [ruby] Prepare v0.15.0 (#1561)
• b330cfb [ruby] Add river migrate-get/validate/version subcommands (#1560)
• 5c2d873 [go] Fix notifier hanging after a listen or unlisten times out (#1559)
• 53ce383 [go] Fix InsertManyFast and QueueUpdate panics on drivers without listene...
• fc6c1d5 [java, js, ruby] bump compatible dependencies (#1553)
• 012b72c Fix intermittently failing test TestJobCleaner/CustomizableInterval (#1531)
• Additional commits viewable in compare view

Updates github.com/riverqueue/river/rivershared from 0.50.0 to 0.51.0

Release notes

Sourced from github.com/riverqueue/river/rivershared's releases.

v0.51.0

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Changelog

Sourced from github.com/riverqueue/river/rivershared's changelog.

[0.51.0] - 2026-10-31

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Commits

• 69c0b45 [go] Prepare v0.51.0 (#1571)
• 737146d Add a Where escape hatch to JobDeleteManyParams (#1570)
• e898054 [go, java, js, rust] Alternative implementation for reconciling jobs against ...
• 4f79f7b Bump crass from 1.0.7 to 1.0.8 in /ruby/rails/riverqueue-rails (#1562)
• f995760 [ruby] Prepare v0.15.0 (#1561)
• b330cfb [ruby] Add river migrate-get/validate/version subcommands (#1560)
• 5c2d873 [go] Fix notifier hanging after a listen or unlisten times out (#1559)
• 53ce383 [go] Fix InsertManyFast and QueueUpdate panics on drivers without listene...
• fc6c1d5 [java, js, ruby] bump compatible dependencies (#1553)
• 012b72c Fix intermittently failing test TestJobCleaner/CustomizableInterval (#1531)
• Additional commits viewable in compare view

Updates github.com/riverqueue/river/rivertype from 0.50.0 to 0.51.0

Release notes

Sourced from github.com/riverqueue/river/rivertype's releases.

v0.51.0

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Changelog

Sourced from github.com/riverqueue/river/rivertype's changelog.

[0.51.0] - 2026-10-31

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Commits

• 69c0b45 [go] Prepare v0.51.0 (#1571)
• 737146d Add a Where escape hatch to JobDeleteManyParams (#1570)
• e898054 [go, java, js, rust] Alternative implementation for reconciling jobs against ...
• 4f79f7b Bump crass from 1.0.7 to 1.0.8 in /ruby/rails/riverqueue-rails (#1562)
• f995760 [ruby] Prepare v0.15.0 (#1561)
• b330cfb [ruby] Add river migrate-get/validate/version subcommands (#1560)
• 5c2d873 [go] Fix notifier hanging after a listen or unlisten times out (#1559)
• 53ce383 [go] Fix InsertManyFast and QueueUpdate panics on drivers without listene...
• fc6c1d5 [java, js, ruby] bump compatible dependencies (#1553)
• 012b72c Fix intermittently failing test TestJobCleaner/CustomizableInterval (#1531)
• Additional commits viewable in compare view

Updates github.com/riverqueue/river from 0.50.0 to 0.51.0

Release notes

Sourced from github.com/riverqueue/river's releases.

v0.51.0

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Changelog

Sourced from github.com/riverqueue/river's changelog.

[0.51.0] - 2026-10-31

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleting jobs by an arbitrary SQL predicate, like JobListParams.Where does for lists. For example, it can select queued jobs whose args name a given customer. A Where predicate counts as a filter, so it doesn't need UnsafeAll, and each predicate is wrapped in parentheses so that an OR in it can't widen the deletion past other filters or to running jobs. The existing batch limit (100 jobs by default) and Postgres behavior of skipping locked jobs still apply. [PR #1570](riverqueue/river#1570).

Fixed

• Fixed a running job not being cancelled after Client.JobCancel when the cancellation notification was lost, as happens if it's sent while the client's notifier is reconnecting. Clients with a notifier now also check their running jobs for cancellation requests once a minute. Client.JobRetry now clears an earlier attempt's cancellation request so that the retried job isn't cancelled again. [PR #1564](riverqueue/river#1564).
• Fixed Client.InsertManyFast panicking after its transaction committed on a client configured with queues when the driver doesn't support a listener, such as a riverdatabasesql driver created with New instead of NewWithPgxListener. It now wakes producers for the inserted jobs' queues right away, like Client.InsertMany. [PR #1558](riverqueue/river#1558).
• Fixed Client.QueueUpdate without metadata crashing the program when the client works that queue and the driver doesn't support a listener. [PR #1558](riverqueue/river#1558).
• Fixed the notifier's run loop blocking forever when a topic LISTEN or UNLISTEN handed to it finished after its caller had stopped waiting (after 5 seconds, with timed out waiting for interrupt operation), as can happen on a slow or unresponsive connection. Client.Stop then never completed, and neither did a Client.Start that failed this way. [PR #1559](riverqueue/river#1559).

Commits

• 69c0b45 [go] Prepare v0.51.0 (#1571)
• 737146d Add a Where escape hatch to JobDeleteManyParams (#1570)
• e898054 [go, java, js, rust] Alternative implementation for reconciling jobs against ...
• 4f79f7b Bump crass from 1.0.7 to 1.0.8 in /ruby/rails/riverqueue-rails (#1562)
• f995760 [ruby] Prepare v0.15.0 (#1561)
• b330cfb [ruby] Add river migrate-get/validate/version subcommands (#1560)
• 5c2d873 [go] Fix notifier hanging after a listen or unlisten times out (#1559)
• 53ce383 [go] Fix InsertManyFast and QueueUpdate panics on drivers without listene...
• fc6c1d5 [java, js, ruby] bump compatible dependencies (#1553)
• 012b72c Fix intermittently failing test TestJobCleaner/CustomizableInterval (#1531)
• Additional commits viewable in compare view

Updates github.com/riverqueue/river/riverdriver from 0.50.0 to 0.51.0

Release notes

Sourced from github.com/riverqueue/river/riverdriver's releases.

v0.51.0

Added

• Added JobDeleteManyParams.Where, an escape hatch for deleti...

  Description has been truncated

[bgentry]

🤖 Codex review: Dependency security and local compatibility review are clear; merge remains blocked by required Pro image CI.

Upgrade

• Six public River modules advance v0.50.0 → v0.51.0: github.com/riverqueue/river, riverdriver, riverdriver/riverpgxv5, riverdriver/riversqlite, rivershared, and rivertype. The two driver modules also advance as indirect dependencies in the Pro module.
• Reviewed head: a7959494897ea156db6eec9370496a95fe2b05cc; base: 310c675a1dfe6eb9d22b0018c8c2e91aa0837042.

Security Review

• Downloaded and diffed all six actual published module pairs. All twelve module ZIP and go.mod sums match the PR and public Go checksum verification. Canonical source remains github.com/riverqueue/river; old module origins resolve to 3901b3d6fa6fa700c87551caec0d84fe211e0c98, new origins to 69c0b4502af6676797c18b5226e6fe2912ff9a3a.
• No external application dependency upgrades, newly introduced modules, same-version checksum rewrites, replacement/fork changes, toolchain/go directive changes, vendoring, generators, or new native/binary artifacts. Existing private dependencies are unchanged. Application dependency overrides are unaffected.
• Runtime changes match the documented cancellation polling fallback, cancellation metadata cleanup on retry, notifier buffering, and no-listener panic fixes. The opt-in JobDeleteManyParams.Where API uses existing named-argument handling and parenthesizes predicates with a newline guard for trailing SQL comments; this does not automatically expose raw SQL through existing UI callers.
• No unexpected process execution, environment/credential reads, new external network destinations, filesystem effects, unsafe/cgo or initialization side effects in the changed production source. Exact-version OSV queries return no advisories for all twelve identities. This is a feature/correctness update, not a confirmed security update.

Compatibility Verification

• go mod verify and go mod tidy -diff in the root, Pro, and packaging modules — passed with no metadata drift.
• make lint — passed across all three modules with no tracked changes.
• make test/race — passed across all modules against an isolated migrated PostgreSQL 18 database, including SQLite and Pro handler coverage.
• Clean npm ci, npm run lint, npm run test:once (37 files / 278 tests), npm run build, and CGO_ENABLED=0 go build ./cmd/riverui plus the equivalent Pro binary build — passed.
• Local make verify/sqlc encounters only generated-header differences because the installed tool is 1.31.1 while the repository pins 1.31.0. Inspected both PostgreSQL and SQLite output: each difference is solely // sqlc v1.31.0 → // sqlc v1.31.1. Current-head CI with pinned 1.31.0 passes this check, as well as the six Go/PostgreSQL matrix jobs, lint, frontend checks, release builds, and public Docker image builds/manifests.

Required CI Blocker
A maintainer-triggered retry still fails before the Pro Docker build with Could not assume role with OIDC: Not authorized to perform sts:AssumeRoleWithWebIdentity. Required Pro AMD64 image and manifest checks remain failed/cancelled/skipped. Restore AWS authorization and rerun these checks before merge; no branch-protection bypass is appropriate.

Residual Risk
No new blocking supply-chain issue was identified in the dependency delta. Unchanged private dependencies and third-party code retain their existing trust assumptions. Required Pro image validation remains unavailable due to the AWS authorization failure.

[bgentry]

🤖 Codex review: Approved after current-head dependency security and compatibility review.

Upgrade

• Six public River modules: river, riverdriver, riverdriver/riverpgxv5, riverdriver/riversqlite, rivershared, and rivertype, all v0.50.0 → v0.51.0, aligned in the root and Pro graphs.
• Reviewed head: c4f799da15871c6bb0b4c2d41c83c60430489c6a; current base: 36b8f46991115ad75576a682dc86156c0c4c535b.

Security Review
Reused all six published old/new artifact-pair reviews and exact ZIP/go.mod checksum evidence. All twelve identities match both PR sum files and public checksum verification; canonical origins remain 3901b3d6fa6fa700c87551caec0d84fe211e0c98 → 69c0b4502af6676797c18b5226e6fe2912ff9a3a. No external application graph churn, same-version sum rewrites, new hooks/native payloads, fork/replacement/toolchain changes, or unexpected credential/process/network behavior. Source changes match documented cancellation polling, retry metadata cleanup, notifier buffering, no-listener fixes and the opt-in parenthesized SQL predicate API. Exact OSV queries are clear for all twelve identities; this is a feature/correctness update.

Current-Head Delta
Verification-only commit 8f8932b leaves the source tree unchanged. Append-only base integration c4f799d brings in the already reviewed and merged AWS action upgrade from #733. Compared the original reviewed head to this head: only that workflow differs, and all module files/checksums are identical. Reused #733's exact action-source/bundle evidence and reviewed the combined workflow context; no new artifact identity or execution input is introduced.

Compatibility Verification

• Root, Pro and packaging go mod verify / go mod tidy -diff; make lint; and complete make test/race against isolated migrated PostgreSQL 18, including SQLite/Pro handlers — passed.
• Clean frontend installation, lint, all 278 tests, production bundle, public and Pro Go binary builds — passed. Inputs to these local checks are unchanged by the base integration; affected workflow syntax was checked again with actionlint.
• Current-head CI passes the full Go 1.26/1.27 × PostgreSQL 16/17/18 matrix, Go lint, pinned sqlc 1.31.0 verification, frontend checks, release builds, CodeQL, both public/Pro Docker architectures, and both required manifest jobs. The release-only Pro prefetch job correctly skips on a PR.
• Local sqlc 1.31.1 differs from generated code only in version headers; the pinned CI check passes. No tracked drift results from validation.

Residual Risk
No blocking issue is identified in this update. Unchanged private dependencies retain their existing trust assumptions. The merged AWS action carries the unused-path upstream advisory and mutable-tag limitations described in #733; those are unchanged by this PR. Required CI now passes through normal maintainer-triggered workflows without rule, credential, or history bypasses.