twilightcoders / active_record-mti Goto Github PK

Bundler could not find compatible versions for gem "activerecord":
In snapshot (Gemfile.lock):
activerecord (= 5.1.5)

In Gemfile:
active_record-mti (~> 0.2.1) was resolved to 0.2.1, which depends on
activerecord (< 5, >= 4.1)

rails (= 5.1.5) was resolved to 5.1.5, which depends on
  activerecord (= 5.1.5)

Hi @voltechs, I'm interested in using this in a Rails 6 project. Wondering if this is still maintained.

Thanks!

Hi,
I just want to use active_record-mti with Rails 5.2 but it's not supported. The following errors appears when running bundle install:

Bundler could not find compatible versions for gem "pg":
  In snapshot (Gemfile.lock):
    pg (= 1.1.3)

  In Gemfile:
    pg

    active_record-mti (~> 0.3.2) was resolved to 0.3.2, which depends on
      pg (~> 0)

Running bundle update will rebuild your snapshot from scratch, using only
the gems in your Gemfile, which may resolve the conflict.

Should I make a project fork changing gemspec file with pg pointing to '~> 1'?

Postgresql does not allow for foreign key constraints from a base table to some other table. See Caveats section in https://www.postgresql.org/docs/9.5/static/ddl-inherit.html

Scenario:

• You have a parent table called persons with a primary key id. You have students and teachers tables that inherit from persons.

• You have another table called orders and it has an attribute person_id.

• You'd like to add a foreign_key constraint from persons.id to orders.id. Such that anytime you inserted a student or teacher record it would maintain this constraint. However postgresql will not allow this as the record must exist in the base class.

Request:

• Optionally allow a way to maintain this foreign_key constaint at the application layer (MTI)

• For instance, if you tried to insert an orders record and the person did exist in the persons table then MTI would throw a foreign constraint exception.

Unfortunately my AdminUser class is not recognized as a MTI class. :(
I'm running Rails 4.2.10 and this is my setup:

ActiveRecord::Migration.create_table :users do |t|
  t.string :first_name, :last_name
  t.timestamps null: false
end

ActiveRecord::Migration.create_table :admin_users, inherits: :users do |t|
  t.text :roles, array: true, default: [], null: false
end

# db/structure.rb
CREATE TABLE users (
    id integer NOT NULL,
    first_name character varying,
    last_name character varying,
    created_at timestamp without time zone NOT NULL,
    updated_at timestamp without time zone NOT NULL,
);

CREATE TABLE admin_users (
    roles text[] DEFAULT '{}'::text[] NOT NULL
)
INHERITS (users);

# app/models/application_record.rb
class ApplicationRecord < ActiveRecord::Base
  self.abstract_class = true
end

# app/models/user.rb
class User < ApplicationRecord
end

class AdminUser < User
end

When I check AdminUser.mti_type_column and AdminUser.tableoid_column on the console, they are both nil, which they shouldn't be according to the specs.
What can I do to further help debugging this problem?

Issue:
When mti adds indexes to children tables it can fail if the generated index name is too long. Rails' add_index function generates the index name by default using the table name and the column your indexing (or array of columns if it's a compound index). If this constructed table name is > 64 characters then add_index will fail.

One thing to note is the :name attribute isn't being passed to add_index when constructing the child index in "active_record-mti/lib/active_record/mti/connection_adapters/postgresql/schema_statements.rb #create_table"

Steps to Reproduce:
Add the following to the spec/schema.rb file and try to run tests.

create_table :vehicles, force: true do |t|
    t.string :color
    t.string :type # Inheritance column
    t.timestamps null: false
  end

  add_index :vehicles, [:color, :type]

  create_table "vehicles/trucks", force: true, inherits: :vehicles do |t|
    t.integer :bed_size
  end

  # different schema of parent
  create_table "vehicles/some_really_really_really_really_long_table_name", force: true, inherits: :vehicles do |t|
    t.integer :long_size
  end

  # same schema of parent
  create_table "vehicles/some_other_really_really_really_long_table_name", force: true, inherits: :vehicles do |t|
end

Expected Result:
It creates the compound indexes on the children tables.

Actual Result:
The following error is thrown:
Index name 'index_vehicles/some_really_really_really_really_long_table_name_on_color_and_type' on table 'vehicles/some_really_really_really_really_long_table_name' is too long; the limit is 63 characters (ArgumentError)

Hi guys,
I have an issue with test fixture. The fixture related to extended model is not loaded into the test DB.

My environment: OSX, Rails 5.1.4, Ruby 2.4.2, active_record-mti 0.3.0.pre.rc4

My code (extract)
Migrations:

create_table :rules do |t|
      t.string :regex,             null: false
     ...
      t.timestamps
    end

create_table :rule_products, inherits: :rules do |t|
      t.xxx xxx ....
end

Models:

class Rule < ApplicationRecord
  validates :regex, presence: true
  ...
end

class RuleProduct < Rule
  self.table_name = 'rule_products'
  ...
end

Fixtures:

--- rules.yml
one:
  regex: rule_one.*

two:
  regex: rule_two.*

--- rule_products.yml
one:
  regex: one.*
  otherfield: value1

two:
  regex: two.*
  otherfield: value2

Now in the test, if I use:

rule = rules(:one)

all works fine, but, if I use:

rule = rule_products(:one)

I got the following error:

Minitest::UnexpectedError: ActiveRecord::RecordNotFound: Couldn't find RuleProduct with 'id'=980190962

and, if I check the test DB, the related table is empty.

I'm doing something wrong?
Thanks for the help.
Best Regards,
Nicola