| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
This reverts commit dd779c9686f49f5ed6dda8ad5a1cb3b0788e1dd4.
|
|
|
|
|
|
| |
ActiveSupport::TestCase now"
This reverts commit 98d0f7ebd34b858f12a12dcf37ae54fdbb5cab64.
|
|
|
|
| |
This reverts commit 8d2866bb80fbe81acb04f5b0c44f152f571fb29f.
|
| |
|
|
|
|
|
|
|
|
| |
This hack prevails everywhere in the codebase by being copy & pasted, and it's actually not a negative thing but a necessary thing for framework implementors,
so it should better have a name and be a thing.
And with this commit, activesupport/test/abstract_unit.rb now doesn't silently autoload AS::TestCase,
so we're ready to establish clearner environment for running AS tests (probably in later commits)
|
|
|
|
| |
It's used everywhere, clean and mature enough
|
|
|
|
| |
and Git taught me that this crap was added via this commit... https://github.com/rails/rails/commit/68db6bc431fbff0b2291f1f60ccf974b4eece596
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- `AM::Error#to_h` was kind of broken before and would return in the
hash values a single error message.
```ruby
person = Person.new
person.errors.add(:name, "cannot be blank")
person.errors.add(:name, "too long")
puts person.errors.to_h # {name: 'too long'}
```
Since an attribute can have different errors, the previous behavior
didn't make much sense.
Now, `ActiveModel::Errors#to_hash` correctly returns an array of
error messages containing all the errors for an attribute.
However, one can easily be surprised by this change, so let's
deprecated it first.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- In ef4d3215b1198c456780b8d18aa62be7795b9b8c I made a change to
pass `AM::Error` object in case the arity of the block passed to
`each` accepted less than 2 arguments.
This is causing one issue for `to_h` as it expects the argument
passed to the block to be an Array (and were are passing it an
instance of `AM::Error`).
There is no real reason to use `to_h` anymore since `to_hash` exists
Deprecating `to_h` inf favor of `to_hash`
Co-Authored-By: Rafael França <rafael@franca.dev>
|
|
|
|
|
|
|
|
|
|
|
|
| |
- `AM::Errors#each` is implemented for the `Enumerator` module and
get called indirectly by a bunch of method in the ruby land
(map, first, select ...)
These methods have a `-1` arity as they are written in C and they
wrongly trigger a deprecation warning.
This commit fixes that and correctectly return a `AM::Error` object
when `each` is called with a negative arity.
|
|
|
|
|
|
|
|
|
| |
- Since `ActiveModel::Error` can now be inherited by
`ActiveModel::NestedError`, when the latter generates a
`full_message`, the `i18n_customize_full_message` accessor set in
the parent class is not set.
This commit fixes that by using a `class_attribute` instead.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- 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.
|
|\
| |
| | |
Returns `nil` when `AM::Errors#delete` doesn't delete anything:
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
- `AM::Errors#delete` currently returns an empty array when trying
to delete an error that doesn't exist.
This behaviour is surprising and I think it would be better
to no return a truthy value but instead return nil like
`Hash#delete` does.
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- When a ActiveRecord record get saved and validated as part of a
collection association, the errors attribute are changed to reflect
the children names. You end up with an error attribute that will
look like this:
`author.errors # {:'books.title' => [:blank]}`
https://github.com/rails/rails/blob/2fe20cb55c76e6e50ec3a4ec5b03bbb65adba290/activerecord/lib/active_record/autosave_association.rb#L331-L340
We then can't check if the `books.title` errors was added using
`ActiveModel::Errors#added?` because it tries to generate a message
to make the match and end up calling the "books.title" method
on the Author.
```
author.errors.added?(:'books.title', :blank) => NoMethodError: undefined method `books.title'
```
This patch modify the behaviour of `strict_match?` to not generate
a message to make the comparison but instead make a strict
comparison with the `options` from the error.
|
|
|
|
|
|
|
|
|
|
|
| |
We sometimes say "✂️ newline after `private`" in a code review (e.g.
https://github.com/rails/rails/pull/18546#discussion_r23188776,
https://github.com/rails/rails/pull/34832#discussion_r244847195).
Now `Layout/EmptyLinesAroundAccessModifier` cop have new enforced style
`EnforcedStyle: only_before` (https://github.com/rubocop-hq/rubocop/pull/7059).
That cop and enforced style will reduce the our code review cost.
|
|
|
|
|
|
|
|
|
| |
This fixes the following warnings.
```
/rails/activemodel/test/cases/nested_error_test.rb:9: warning: method redefined; discarding old test_initialize
/rails/activemodel/test/cases/error_test.rb:29: warning: previous definition of test_initialize was here
```
|
|
|
|
| |
Otherwise we get deprecation warnings in the generated scaffold template files
|
|\
| |
| | |
Model error as object
|
| | |
|
| |
| |
| |
| | |
maintaining behavior errors.details[:foo].any?
|
| | |
|
| | |
|
| | |
|
| |
| |
| |
| | |
Revert some tests to ensure back compatibility
|
| | |
|
| |
| |
| |
| |
| | |
autosave duplicate errors can be removed
See SHA 7550f0a016ee6647aaa76c0c0ae30bebc3867288
|
| |
| |
| |
| | |
This is because we try to accommodate old hash behavior, so `first` and `last` now does not return Error object.
|
| | |
|
| |
| |
| |
| |
| |
| | |
Many operations need grouping of errors by attributes, e.g. ActiveRecord::AutosaveAssociation#association_valid?
Refactor other methods using group_by_attribute
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Allow `each` to behave in new way if block arity is 1
Ensure dumped marshal from Rails 5 can be loaded
Make errors compatible with marshal and YAML dumps from previous versions of Rails
Add deprecation warnings
Ensure each behave like the past, sorted by attribute
|
| |
| |
| |
| |
| |
| |
| | |
Add initialize_dup to deep dup.
Move proc eval and flexible message position out to Errors,
because proc eval is needed for Errors#added? and Errors#delete
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This adds `.attribute_names` and `#attribute_names` to
`ActiveModel::Attributes` along the same lines as the corresponding
methods in `ActiveRecord::AttributeMethods` (see
[`.attribute_names`][class_method] and
[`#attribute_names`][instance_method].
While I was here I also added documentation for '#attributes', which I
added in 043ce35b186. The whole class is still `#:nodoc:` so I don't
think this will have any effect for now.
[class_method]: https://github.com/rails/rails/blob/cc834db1d0815744cfa173813c05d928e008e167/activerecord/lib/active_record/attribute_methods.rb#L154-L160
[instance_method]: https://github.com/rails/rails/blob/cc834db1d0815744cfa173813c05d928e008e167/activerecord/lib/active_record/attribute_methods.rb#L299-L301
|
| | |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Each acceptance validator applied to a model class includes an instance
of a module builder (LazilyDefineAttributes) into that class. In tests,
if the original model class is not subclassed, these modules pile up and
cannot be removed, potentially leading to flakey specs and false
positive/negatives.
To avoid this, always use subclasses in tests whose names (constants)
can be removed when the test is done.
|
| | |
|
| |
| |
| |
| |
| |
| | |
Follow up of #26764 and #35700.
And add test case for #35700.
|
|/ |
|
|\
| |
| | |
Type cast falsy boolean symbols on boolean attribute as false
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Before 34cc301, type casting by boolean attribute when querying is a
no-op, so finding by truthy boolean string (i.e.
`where(value: "true") # => value = 'true'`) didn't work as expected
(matches it to FALSE in MySQL #32624). By type casting is ensured, a
value on boolean attribute is always serialized to TRUE or FALSE.
In PostgreSQL, `where(value: :false) # => value = 'false'` was a valid
SQL, so 34cc301 is a regresson for PostgreSQL since all symbol values
are serialized as TRUE.
I'd say using `:false` is mostly a developer's mistake (user's input
basically comes as a string), but `:false` on boolean attribute is
serialized as TRUE is not a desirable behavior for anybody.
This allows falsy boolean symbols as false, i.e.
`klass.create(value: :false).value? # => false` and
`where(value: :false) # => value = FALSE`.
Fixes #35676.
|
| |
| |
| |
| | |
#35789
|
|/
|
|
|
|
| |
- I feel `i18n_customize_full_messages` explains the meaning of the
config better.
- Followup of https://github.com/rails/rails/pull/32956
|
| |
|
| |
|
|\
| |
| |
| |
| | |
kamipo/dont_allow_non_numeric_string_matches_to_zero
Don't allow `where` with non numeric string matches to 0 values
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This is a follow-up of #35310.
Currently `Topic.find_by(id: "not-a-number")` matches to a `id = 0`
record. That is considered as silently leaking information.
If non numeric string is given to find by an integer column, it should
not be matched to any record.
Related #12793.
|
|/
|
|
|
|
|
|
|
|
|
| |
This reverts commit 52fddcc653458456f98b3683dffd781cf00b35fe.
52fddcc was to short-circuit `ensure_in_range` in `cast_value`. But that
caused a regression for empty string deserialization.
Since 7c6f393, `ensure_in_range` is moved into `serialize`. As 52fddcc
said, the absolute gain is quite small. So I've reverted that commit to
fix the regression.
|
|\
| |
| |
| | |
Return correct date in ActiveModel for time to date conversions
|
| |
| |
| |
| |
| |
| |
| | |
time.to_date conversion happens considering leap years
so a conversion of "Day.new({'day(1i)'=>'1', 'day(2i)'=>'1', 'day(3i)'=>'1'})" results in saving the date as Mon, 03 Jan 0001
which might seem weird on the user level, hence falling back to parsing on string level resolves this data mismatch
Fixes #28521
|