Who contributed to Rails on July 21, 2026?

8 developers shipped this update, including fxn, Ryuta Kamizono, hachi8833, Dinesh Saini, Andy Pfister, andyjeffries, bubiche, and alex.

What were the notable Rails updates?

Avoid reloading job rows after dispatch and schedule, Let dump_schema_migrations_sort_by be :reverse by default, and Merge pull request #58166 from rails/fxn/change-default.

Jul 21, 2026

$ the-wire · showcase

SOLID QUEUE CUTS JOB DISPATCH OVERHEAD, RAILS FIXES SCHEMA MIGRATION DEFAULTS

By RepoJournal · Filed 06:55 UTC on July 21, 2026 · About Rails · Composed from the cited sources · methodology

8 people shipped this

fxn @fxn 2 cited

Dinesh Saini @ImDineshSaini 1 cited

Andy Pfister @andyundso 1 cited

andyjeffries @andyjeffries 1 cited

bubiche @bubiche 1 cited

Ryuta Kamizono @kamipo

hachi8833 @hachi8833

alex @ghiculescu

Solid Queue eliminated redundant database queries in its hottest path, cutting unnecessary round-trips on every bulk job enqueue while Rails corrected a schema.rb default it shipped with reservations.

In brief

Solid Queue now skips reloading job rows after dispatch. one less roundtrip to the DB per job. tiny win that compounds across millions of jobs.
Rails schema dumps default to :reverse migration sort order. stops the annoying reversal dance when you're working with lots of migrations at once.
Solid Queue's DISTINCT queries on Postgres got a recursive CTE bump. better query plans for jobs with lots of duplicates.
Solid Queue now auto-picks Rails' timezone config for job schedules. one less thing to wire up when you're setting up a worker.

The Solid Queue team shipped a performance fix [1] that removes wasteful row reloading after job dispatch and schedule operations. The library was issuing two database statements per execution table to re-read jobs already in memory. On high-volume job queues, this doubles the overhead of `perform_all_later`. Paired with a new recursive CTE optimization [4] for PostgreSQL's DISTINCT queries on large job tables, Solid Queue is getting leaner. The CTE approach sidesteps MySQL's Loose Index Scan limitation by walking the B-tree index directly on distinct queue names. Rails itself course-corrected on schema migration handling [2] [3], changing the default sort order for dumped schema migrations to reverse chronological instead of lexicographic. The maintainers called the original default "painful" and expected merge conflicts to surface the issue in production apps. Windows users on Solid Queue will see signal handling simplified [5] with QUIT trapping removed from that platform. The queue will also now respect Rails application timezones for scheduled tasks [6] instead of forcing server-configured time, eliminating a common source of confusion in multi-timezone deployments.

Action items

→ Review and reset schema.rb sort defaults if you have merge conflicts in db/schema.rb after next Rails merge rails/rails [plan]
→ Update Solid Queue to the latest version for job dispatch performance gains rails/solid_queue [monitor]
→ Verify recurring job timing if using Solid Queue on Rails with non-UTC timezones rails/solid_queue [plan]

References

[1] Avoid reloading job rows after dispatch and schedule ↗ rails/solid_queue
[2] Let dump_schema_migrations_sort_by be :reverse by default ↗ rails/rails
[3] Merge pull request #58166 from rails/fxn/change-default rails/rails ↗
[4] Improve DISTINCT queries on PostgreSQL via recursive CTE ↗ rails/solid_queue
[5] Remove trapping QUIT on Windows ↗ rails/solid_queue
[6] Automatically use Rails or a default timezone for the schedule ↗ rails/solid_queue