chore(deps): update hex dependencies (patch)

[renovate]

This PR contains the following updates:

Package	Type	Update	Change
bandit (source)	prod	patch	1.10.1 → == 1.10.2
credo (source)	dev	patch	1.7.13 → == 1.7.16
ecto_sql (source)	prod	patch	3.13.2 → == 3.13.4
ex_aws_s3 (source)	prod	patch	2.5.8 → == 2.5.9
oban (source)	prod	patch	2.20.1 → == 2.20.3
phoenix_live_reload (source)	dev	patch	1.6.1 → == 1.6.2
phoenix_live_view (source)	prod	patch	1.1.17 → == 1.1.22

---

Release Notes

mtrudel/bandit (bandit)

v1.10.2

Compare Source

Enhancements

• Distinguish client disconnects from genuine body read timeouts (#​564, thanks @​pepicrft!)

rrrene/credo (credo)

v1.7.16

Compare Source

• Fix compatibility & compiler warnings with Elixir 1.20.0-rc.1
• Credo.Check.Refactor.PassAsyncInTestCases add new param :force_comment_on_explicit_false (defaults to false)
• Credo.Check.Warning.Dbg add new param :allow_captures (defaults to false)
• New Check: Credo.Check.Warning.UnusedMapOperation
• New Check: Credo.Check.Warning.UnusedOperation

v1.7.15

Compare Source

• Improve performance on large projects
• Parse token_metadata for source files
• Credo.Check.Warning.ExpensiveEmptyEnumCheck have better issue messages
• Credo.Check.Refactor.MatchInCondition add new param :allow_operators
• Credo.Check.Refactor.MatchInCondition fix false positive
• Credo.Check.Readability.AliasOrder fix false positive
• Credo.Check.Readability.FunctionNames fix false positive
• Credo.Check.Readability.SinglePipe add new param :allow_blocks (defaults to true)
• Credo.Check.Refactor.ModuleDependencies fix false positive

v1.7.14

Compare Source

• Fix compatibility & compiler warnings with Elixir 1.20.0-rc.0
• Credo.Check.Refactor.PassAsyncInTestCases add new param :force_comment_on_explicit_false (defaults to false)
• New Check: Credo.Check.Warning.UnusedMapOperation
• New Check: Credo.Check.Warning.UnusedOperation

elixir-ecto/ecto_sql (ecto_sql)

v3.13.4

Compare Source

Bug fixes

• [mysql] Do not crash mix ecto.load with large dumped databases

v3.13.3

Compare Source

Enhancements

• [sql] Tag generated functions as :generated
• [sql] Add :wrap_in_transaction option to explain

Bug fixes

• [mysql] Fix structure_load/2 for MySQL 9.4+

ex-aws/ex_aws_s3 (ex_aws_s3)

v2.5.9

Compare Source

oban-bg/oban (oban)

v2.20.3

Compare Source

Enhancements

• [Worker] Allow snoozing jobs by returning a tuple period

  It's now possible to snooze jobs with a period such as {1, :minute} instead of just a raw
  number of seconds.

• [Oban] Validate no duplicate options are passed to config on init

  Because options are provided as a keyword list, duplicate options were accepted while only one
  was actually used.

Bug Fixes

• [Oban] Fix starting a queue on a specific node

  The :node option was incorrectly preserved when starting a queue, which would crash it. Now
  the :node is dropped after scoping the start signal.

• [Reindexer] Fix dropping invalid indexes from the reindexer

  The reindexer would fail sliently with "DROP INDEX CONCURRENTLY" cannot be executed from a
  function message because the deindex operation used a DO block to loop through and drop
  invalid indexes, but PostgreSQL prohibits CONCURRENTLY operations inside functions or DO
  blocks. Now invalid indexes are fetched first, then dropped as individual queries.

• [Installer] Prevent installer crash with unsupported adapters

  The installer would crash with a CaseClauseError when a project had an Ecto repo using an
  unsupported adapter like Ecto.Adapters.Tds.

  Now the installer filters repos to find one with a supported adapter, skipping unsupported ones
  automatically. Without a compatible repo, it displays an error message listing the found repos
  and their adapters, along with guidance on how to specify a repo explicitly.

• [Pruner] Better sqlite timestamp default and pruning query

  The CURRENT_TIMESTAMP type lacks a trailing z, which causes it to be compared incorrectly
  against UTC datetimes. Any jobs inserted without a scheduled_at, where the default is used,
  could be returned in queries that compare against a DateTime.

  This prevents the issue in the future two ways:

  1. Switch the default inserted_at/scheduled_at timestamp to a format that can be queried
     properly.
  2. Change the pruning check to use completed_at rather than scheduled_at for existing
     databases. This is a more accurate query that was avoided before because it didn't match the
     Basic engine.

v2.20.2

Compare Source

Enhancements

• [Telemetry] Add domain to attach_default_logger metadata

  Although this information is present in the source attribute, it’s not part of the metadata,
  it’s part of the attributes. And when encode = true, it’s generated as a string, which makes
  it less convenient to work with.

• [Cron] Add unique cron entry identifiers to job meta

  A id value is generated for each cron entry and stored in the job's meta to distinguish between
  different cron jobs with the same expression.

Bug Fixes

• [Stager] Order staging query to maximize compound index usage

  The core compound index couldn't be utilized by staging queries when the planner estimated a
  large number of hits. Changing the query to order by scheduled_at and id, it becomes an
  index scan.

  It would still be more efficient to use a dedicated index, but OSS doesn't have any mechanisms
  for automatic concurrent index creation and we have to save it for later.

• [Executor] Take measurements from the executing process

  When Producer starts jobs, it calls Executor.new in its own process, before handing it off to
  Task.Supervisor.async_nolink to call Executor.call. That means it used the Producer pid,
  and took measurements from that process.

• [Oban] Set scheduled state correctly when updating with update_job/3

  When the scheduled_at timestamp is set during update_job, the state is automatically set to
  scheduled. This mirrors the functionality of insert_job.

  Note that the value of the timestamp isn't considered. Setting a timestamp in the past will set
  the job as scheduled. This isn't a problem in practice because the stager will change the
  state to available on the next cycle anyhow.

• [Repo] Include :deadlock_detected in expected errors

  Deadlocks are automatically resolved by Postgres because it aborts one transaction. At that
  point, the retry should succeed quickly since the blocking transaction is gone.

phoenixframework/phoenix_live_reload (phoenix_live_reload)

v1.6.2

Compare Source

• Bug fixes
  • Properly deal with Unicode when forwarding logs

phoenixframework/phoenix_live_view (phoenix_live_view)

v1.1.22

Compare Source

v1.1.21

Compare Source

v1.1.20

Compare Source

v1.1.19

Compare Source

v1.1.18

Compare Source

---

Configuration

📅 Schedule: Branch creation - Between 12:00 AM and 03:59 AM ( * 0-3 * * * ) in timezone America/New_York, Automerge - At any time (no schedule defined).

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

👻 Immortal: This PR will be recreated if closed unmerged. Get config help if that's undesired.

---

• If you want to rebase/retry this PR, check this box

---

This PR was generated by Mend Renovate. View the repository job log.