| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| | |/ /
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Transactional-fixture using tests with racing threads and inter-thread
synchronisation inside transaction blocks will now deadlock... but
without this, they would just crash.
In 5.0, the threads didn't share a connection at all, so it would've
worked... but with the main thread inside the fixture transaction, they
wouldn't've been able to see each other.
So: as far as I can tell, the set of operations this "breaks" never had
a compelling use case. Meanwhile, it provides an increased level of
coherency to the operational feel of transactional fixtures.
If this does cause anyone problems, they're probably best off disabling
transactional fixtures on the affected tests, and managing transactions
themselves.
|
|\ \ \ \
| | | | |
| | | | | |
Tweaks #28678
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
* Fix the comment on `formatted_version`
* Extract `define_params`
* Remove duplicated guard clause for `@version`
|
| |/ / /
|/| | |
| | | |
| | | |
| | | |
| | | | |
Also, explicitly apply the order: generate_subscripts is unlikely to
start returning values out of order, but we should still be clear about
what we want.
|
| | | | |
|
|\ \ \ \
| |/ / /
|/| / /
| |/ / |
Expose `queries` for `AssociationQueryValue` and `PolymorphicArrayValue`
|
| | | |
|
| | |
| | |
| | |
| | | |
It should be handled by `PolymorphicArrayHandler` if polymorphic association.
|
|\ \ \
| | | |
| | | | |
use formatted number as schema version
|
| | |/
| |/| |
|
| | |
| | |
| | |
| | | |
To ease to customize a relation for `exists?`.
|
|/ /
| |
| |
| | |
fixtures, not an empty array.
|
|\ \
| | |
| | | |
Clear active connections after initialization
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Any connections that were checked out during initialization should be
checked back in before the first request is processed, for two reasons:
- Returning the connection to the pool allows it to be health checked
before it's used again. If the connection dies before the first
request arrives, the health check will replace it with a new one.
- If the thread that initialized Rails is not the same thread that will
be performing work, checking in the connection will allow it to be
reused instead of being stuck to the initialization thread forever.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This reverts commit a680a5814184e2f37c4686aa53d0ad3c7fb6b1ee, reversing
changes made to 842f67dd242e738419f27e752ea7dcd0bbe87b6d.
Reason: I can't resist to the joke, so better to keep it there
https://github.com/rails/rails/pull/28598#issuecomment-290945339.
|
| | | |
|
| | |
| | |
| | | |
`ActiveRecord::Base.establish_connection` accepts a single symbol argument to specify a named connection; a single string argument appears to be interpreted as a connection URI
|
|\ \ \
| | | |
| | | |
| | | |
| | | | |
htanata/ar_relation_inspect_should_not_load_all_records
Load only needed records on ActiveRecord::Relation#inspect
|
| | |/
| |/|
| | |
| | |
| | |
| | |
| | | |
Instead of loading all records and returning only a subset of those,
just load the records as needed.
Fixes #25537.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
silly method gets a silly doc fix,
or I'm missing an even sillier joke and I'm about to get schooled.
BUT I'm pretty sure this is some serious Beaudrillard simulacrum, though.
I'm just doing my part to spread the gospel of Douglas Adams.
|
|\ \ \
| | | |
| | | | |
Fix `primary_keys` across multiple schemas
|
| | | |
| | | |
| | | |
| | | | |
Fixes #28470.
|
|\ \ \ \
| |_|/ /
|/| | | |
Make internal methods to private
|
| | | | |
|
|\ \ \ \
| | | | |
| | | | | |
Preprocess association query handling in predicate builder
|
| | | | |
| | | | |
| | | | |
| | | | | |
Because `RelationHandler` uses `value.arel`.
|
| | |/ /
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Currently association query is handled as a postprocess. This has two
problems.
1. When `value` is a `Hash`, we need to skip the postprocess using
`next`.
2. `can_be_bound?` should return false if
`table.associated_with?(column_name)` is true (pass to the postprocess).
These are unneeded if preprocessing association query handling.
|
|\ \ \ \
| | | | |
| | | | |
| | | | |
| | | | | |
mtsmfm/disable-referential-integrity-without-superuser-privilege-take-2
Use `SET CONSTRAINTS` for `disable_referential_integrity` without superuser privileges (take 2)
|
| | |/ /
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
privileges (take 2)
Re-create https://github.com/rails/rails/pull/21233
eeac6151a5 was reverted (127509c071b4) because it breaks tests.
----------------
ref: 72c1557254
- We must use `authors` fixture with `author_addresses` because of its foreign key constraint.
- Tests require PostgreSQL >= 9.4.2 because it had a bug about `ALTER CONSTRAINTS` and fixed in 9.4.2.
|
|/ / /
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Add stronger assertions to rake migration tasks to make sure the user is providing a numeric VERSION
An empty string was getting converted to version = 0. This would in turn pass the presence check.
Address linting warning
Add test for rake task and refactor code to meet expectations
In particular passing VERSION=0 should not raise an error.
Addressed Comments for PR #28485. Trimmed empty lines + change of wording for error message
Adjust test for change of wording in error message
Change condition to follow rails idioms
|
| | |
| | |
| | |
| | |
| | |
| | | |
Because this comment is not document for `supports_ranges?`
ref: https://github.com/rails/rails/pull/27636#discussion_r107560081
|
| | |
| | |
| | |
| | |
| | | |
Since #28473 `uniq` is delegated to `records`, so `CollectionProxy#uniq`
is unnecessary.
|
|\ \ \
| | | |
| | | | |
Fix `LogSubscriber` to allow legacy `binds`
|
| | | |
| | | |
| | | |
| | | | |
Follow up of #27939.
|
|/ / /
| | |
| | |
| | |
| | | |
The old top level classes PGconn, PGresult and PGError were deprecated
since pg-0.13.0: https://github.com/ged/ruby-pg/blob/master/History.rdoc#v0130-2012-02-09-michael-granger-gedfaeriemudorg
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Fixes a regression introduced in
22ca710f20c3c656811df006cbf1f4dbc359f7a6 where Relation#unscope with a
specific where value (vs unscoping the entire where clause) could result
in the wrong binds being left on the query.
This was caused by an index variable not being incremented properly.
|
|\ \ \
| | | |
| | | | |
Fix extension method with dirty target in has_many associations
|
| |/ /
| | |
| | |
| | |
| | |
| | |
| | | |
Extension methods should not delegate to `scope` to respect dirty
target on `CollectionProxy`.
Fixes #28419.
|
|\ \ \
| | | |
| | | | |
Remove duplicated `columns` definition
|
| |/ / |
|
|\ \ \
| | | |
| | | | |
Delegate `uniq` to `records`
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Since b644964b `ActiveRecord::Relation` includes `Enumerable` so
delegating `collect`, `all?`, and `include?` are also unneeded.
`collect` without block returns `Enumerable` without preloading by that.
We should use `load` rather than `collect` for force loading.
|
|\ \ \ \
| |_|/ /
|/| | | |
Fix `select_all` with legacy `binds`
|
| | | |
| | | |
| | | |
| | | | |
Fixes #27923.
|
| |/ /
|/| |
| | |
| | | |
Follow up of #28453.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This fixes CI failure due to 48f3be8c.
`Enumerable#uniq` was introduced since Ruby 2.4. We should delegate
`uniq` to `records` explicitly.
And since b644964b `ActiveRecord::Relation` includes `Enumerable` so
delegating `map` is unneeded.
|
|\ \ \
| | | |
| | | | |
Allow order to be given expressions as hash keys
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
When `order` is given a hash, the keys are currently assumed to be
attribute names and are quoted as such in the query, which makes it
impossible to pass an expression instead:
Post.order("LENGTH(title)" => :asc).last
# SELECT `posts`.* FROM `posts` ORDER BY `posts`.`LENGTH(title)` DESC LIMIT 1
If the key is an `Arel::Nodes::SqlLiteral`, we now use it directly in
the query. This provides a way to build a relation with a complex order
clause that can still be reversed with `reverse_order` or `last`.
|
|\ \ \ \
| | | | |
| | | | |
| | | | | |
Drop comments from structure.sql in postgresql
|