MySQLDatabaseTasks, like AbstractMysqlAdapter, should be able to operate
on any mysql adapter, not just mysql2. Errors having a .error_number
attribute is a mysql2 specific API, which we (Rails) don't control, so
we should instead use connection.error_number(err), which we do.
This also updates tests to better test how this really works, previously
it stubbed create_database to raise Tasks::DatabaseAlreadyExists, which
can never happen.
Previously Rails expected indexes to be an array of columns, but for
PostgreSQL a expression index can just be a string of text. Handle this
by forcing `Index#columns` to be an Array inside `index_exists?`.
Closes#36739
This commit fixes an issue where multi-database configurations were
incompatible with setting a `DATABASE_URL` environment variable.
As part of this work, this commit also includes a light refactor
to make both multi and single database configurations lead into the same
code path so they behave the same.
As mentioned in #36736, this regression was introduced as part of
f2ad69fe7a605b01bb7c37eeac6a9b4e7deb488e
Ruby 2.7 introduces beginless ranges (..value and ...value) and as with
endless ranges we can turn these into inequalities, enabling expressions
such as
Order.where(created_at: ..1.year.ago)
User.where(karma: ...0)
- One regression introduced by the "AM errors as object" features is
about the `full_messages` method.
It's currently impossible to call that method if the `base` object
passed in the constructor of `AM::Errors` doesn't respond to the
`errors` method.
That's because `full_messages` now makes a weird back and forth trip
`AM::Errors#full_messages` -> `AM::Error#full_message` -> `AM::Errors#full_message`
Since `full_message` (singular) isn't needed by AM::Errors, I moved
it to the `AM::Error` (singular) class. This way we don't need to
grab the `AM::Errors` object from the base.
Fixes#36581.
This fixes an issue where validations would return differently when a previously saved invalid association was loaded between calls:
assert_equal true, squeak.valid?
assert_equal true, squeak.mouse.present?
assert_equal true, squeak.valid?
Here the second assert would return
Expected: true
Actual: false
Limiting validations to associations that would be normally saved (using autosave: true) due to changes means that loading invalid associated relations will not change the return value of the parent relations's `valid?` method.
- In 86620cc3aa8e2630bc8d934b1a86453276b9eee9, a change was made
on how we remove error duplication on a record for autosave
association
This fix has one caveat where validation having a `if` / `unless`
options passed as a proc would be considered different.
Example:
```ruby
class Book < ApplicationRecord
has_one :author
validates :title, presence: true, if -> { true }
validates :title, presence: true, if -> { true }
end
Book.new.valid? # false
Book.errors.full_messages # ["title can't be blank", "title can't be blank"]
```
While this example might sound strange, I think it's better to
ignore `AM::Validations` options (if, unless ...) when making the
comparison.
The error happens in PostgreSQL when using `relation.exists?` with
`distinct`, `offset` and `order` for joined table.
However, the error does not happen if either `distinct` or `offset` is
removed. This behavior is confusing.
Fixes#36632
`enum` and `set` are typed cast as `:string`, but currently the
`:string` type is incorrectly reused for schema dumping.
A cast type on columns is not always the same with `sql_type`, this
fixes schema dumping `enum` and `set` columns to use `sql_type` instead
of `type` correctly.
When a enum element contains the prefix 'not_'. I warns to users
to be aware of this new feature.
Example code:
class Foo < ActiveRecord::Base
enum status: [:sent, :not_sent]
end
This PR is to fix#36559 but I also found other issues that haven't been
reported.
The check for `(config.size == 1 && config.values.all? { |v| v.is_a?
String })` was naive. The only reason this passed was because we had
tests that had single hash size configs, but that doesn't mean we don't
want to create a hash config in other cases. So this now checks for
`config["database"] || config["adapter"] || ENV["DATABASE_URL"]`. In the
end for url configs we still get a UrlConfig but we need to pass through
the HashConfig to create the right kind of UrlConfig. The UrlConfig's
are really complex and I don't necessarily understand everything that's
needed in order to act the same as Rails 5.2.
I edited the connection handler test to demonstrate how the previous
implementation was broken when checking config size. Now old and new
tests pass so I think this is closer to 5.2.
Fixes#36559
I think we should change this, but not in 6-0-stable since that's
already in RC and I was trying to only make changes that won't require
any app changes.
This reverts a portion of https://github.com/rails/rails/pull/36439 that
made all schema migration version numbers get dumped as an integer.
While it doesn't _really_ matter it did change behavior. We should bring
this back in 6.1 with a deprecation.
When a record does not have a table name, as in the case for a record
with `self.abstract_class = true` and no `self.table_name` set the error
message raises a cryptic:
"ActiveRecord::StatementInvalid: Could not find table ''" this patch now
raises a new `TableNotSpecified Error`
Fixes: #36274
Co-Authored-By: Eugene Kenny <elkenny@gmail.com>
Before 1340498d2, `order` with no-op value (e.g. `nil`, `""`) had broken
the contract of ordinal methods, which returns a result deterministic
ordered.
When SQLite connects it will silently create a database if the database does not
exist. This behaviour causes different issues because of inconsistent behaviour
between adapters: #36383, #32914. This commit adds a `database_exists?` method
as a way to check the database without creating it. This is a stepping stone to
fully resolving the above issues.
GROUP BY with virtual count attribute is invalid for almost all
databases, but it is valid for PostgreSQL, and it had worked until Rails
5.2.2, so it is a regression for Rails 5.2.3 (caused by 311f001).
I can't find perfectly solution for fixing this for now, but I would not
like to break existing apps, so I decided to allow referencing virtual
count attribute in ORDER BY clause when GROUP BY aggrigation (it partly
revert the effect of 311f001) to fix the regression #36022.
Fixes#36022.
We only use `ToSQL` visitors in the our codebase, do not use
`DepthFirst` and `Dot` visitors.
The `DepthFirst` visitor (which was introduced at c86c37e5f) is used to
traverse an Arel (partial) ast with depth first.
Is there any worth to keep that undocumented feature with much code and
test cases.
This removes that unused `DepthFirst` code and test cases.
If we put the `while_preventing_writes` on the connection then the
middleware that sends reads to the primary and ensures they can't write
will not work. The `while_preventing_writes` will only be applied to the
connection which it's called on - which in the case of the middleware is
Ar::Base.
This worked fine if you called it directly like
`OtherDbConn.connection.while_preventing_writes` but Rails didn't have a
way of knowing you wanted to call it on all the connections.
The change here moves the `while_preventing_writes` method from the
connection to the handler so that it can block writes to all queries for
that handler. This will apply to all the connections associated with
that handler.