The new option allows any Ruby namespace to be registered and set
up for eager load. We are effectively exposing the structure existing
in Rails since v3.0 for all developers in order to make their applications
thread-safe and CoW friendly.
The flag was mainly used to add a Rack::Lock middleware to
the stack, but the only scenario the lock is desired is in
development.
If you are deploying on a not-threaded server, the Rack::Lock
does not provide any benefit since you don't have concurrent
accesses. On the other hand, if you are on a threaded server,
you don't want the lock, since it defeats the purpose of using
a threaded server.
If there is someone out there, running on a thread server
and does want a lock, it can be added to your environment
as easy as: `use Rack::Lock`
Previously, the eager load behavior was mostly coupled to
config.cache_classes, however this was suboptimal since in
some environments a developer may want to cache classes but
not necessarily load them all on boot (for example, test env).
This pull request also promotes the use of config.eager_load
set to true by default in production. In the majority of the
cases, this is the behavior you want since it will copy most
of your app into memory on boot (which was also the previous
behavior).
Finally, this fix a long standing Rails bug where it was
impossible to access a model in a rake task when Rails was
set as thread safe.
When you mount your application at a path, for example /myapp, server
should set SCRIPT_NAME to /myapp. With such information, rails
application knows that it's mounted at /myapp path and it should generate
routes relative to that path.
Before this patch, rails handled SCRIPT_NAME correctly only for regular
apps, but it failed to do it for mounted engines. The solution was to
hardcode default_url_options[:script_name], which is not the best answer
- it will work only when application is mounted at a fixed path.
This patch fixes the situation by respecting original value of
SCRIPT_NAME when generating application's routes from engine and the
other way round - when you generate engine's routes from application.
This is done by using one of 2 pieces of information in env - current
SCRIPT_NAME or SCRIPT_NAME for a corresponding router. This is because
we have 2 cases to handle:
- generating engine's route from application: in this situation
SCRIPT_NAME is basically SCRIPT_NAME set by the server and it
indicates the place where application is mounted, so we can just pass
it as :original_script_name in url_options. :original_script_name is
used because if we use :script_name, router will ignore generating
prefix for engine
- generating application's route from engine: in this situation we
already lost information about the SCRIPT_NAME that server used. For
example if application is mounted at /myapp and engine is mounted at
/blog, at this point SCRIPT_NAME is equal /myapp/blog. Because of that
we need to keep reference to /myapp SCRIPT_NAME by binding it to the
current router. Later on we can extract it and use when generating url
Please note that starting from now you *should not* use
default_url_options[:script_name] explicitly if your server already
passes correct SCRIPT_NAME to rack env.
(closes#6933)
It's not really a good idea to have this as a global config option. We
should allow people to specify the behaviour per association.
There will now be two new values:
* :dependent => :restrict_with_exception implements the current
behaviour of :restrict. :restrict itself is deprecated in favour of
:restrict_with_exception.
* :dependent => :restrict_with_error implements the new behaviour - it
adds an error to the owner if there are dependent records present
See #4727 for the original discussion of this.
This allows us to do:
In your configuration:
Rails.queue[:image_queue] = SomeQueue.new
Rails.queue[:mail_queue] = SomeQueue.new
In your app code:
Rails.queue[:mail_queue].push MailJob.new
Both jobs pushed to the same default queue
Rails.queue.push DefaultJob.new
Rails.queue[:default].push DefaultJob.new
For instance, running
rails g migration CreateMediaJoinTable artists musics:uniq
will create a migration with
create_join_table :artists, :musics do |t|
# t.index [:artist_id, :music_id]
t.index [:music_id, :artist_id], unique: true
end
AddXXXToYYY/RemoveXXXFromYYY migrations are produced with references
statements, for instance
rails g migration AddReferencesToProducts user:references
supplier:references{polymorphic}
will generate the migration with:
add_reference :products, :user, index: true
add_reference :products, :supplier, polymorphic: true, index: true
Jobs pushed to the queue should not contain a reference to it. As the queue
itself cannot be marshalled, and as a consequence of checking the
marshallability of all jobs in the test environment, we can now guarantee this
to be the case in the test environment when using the default TestQueue
implementation.
By marshalling and unmarshalling jobs when adding them to the test queue, we
can ensure that jobs created during test runs are valid candidates for
marshalling, and, thus, that they can be used with queueing backends other than
the default simple in-process implementation.
This will also be used in a subsequent commit to ensure that jobs pushed to the
queue do not contain a reference to the queue itself.
Using an anonymous class prevented marshalling: we're not doing that yet, but
the next commit will introduce this. This also provided an opportunity to
improve the expressivity of the tests and to make the assertion failure
messages clearer.
When Rails::Paths::Root's []= is used to replace a path it should persist the previous path's glob. Without passing the glob along we get gnarly bugs when trying to wire up things like engines.
module FooEngine
class Engine < ::Rails::Engine
isolate_namespace FooEngine
config.paths['config/initializers'] = "lib/foo_engine/initializers"
end
end
## Example of behaviour before this commit.
#
# Before the initializer override:
>> FooEngine::Engine.config.paths["config/initializers"].glob
=> "**/*.rb"
# After the initializer override:
>> FooEngine::Engine.config.paths["config/initializers"].glob
=> nil
## Example of behaviour after this commit.
#
# Before the initializer override:
>> FooEngine::Engine.config.paths["config/initializers"].glob
=> "**/*.rb"
# After the initializer override:
>> FooEngine::Engine.config.paths["config/initializers"].glob
=> "**/*.rb"
This reverts commit 6acebb38bc0637bc05c19d87f8767f16ce79189b.
Usage of this feature did not reveal any improvement in existing apps.
Conflicts:
actionpack/lib/action_dispatch/routing/mapper.rb
guides/source/routing.textile
railties/lib/rails/engine.rb
railties/lib/rails/paths.rb
railties/test/paths_test.rb
Since plugins were removed, we can clean up a few methods in engines.
We also use this opportunity to move `load_console`, `load_tasks` and
`load_runner` to Rails::Engine. This means that, if someone wants to
improve script/rails for engines to support console or runner commands,
part of the work is already done.
Add polymorphic option to model generator
For instance,
$ rails g model Product supplier:references{polymorphic}
generate model with `belongs_to :supplier, polymorphic: true` association and appropriate migration.
For instance,
$ rails g model Product supplier:references{polymorphic}
generate model with `belongs_to :supplier, polymorphic: true` association and appropriate migration.
Also fix model_generator_test.rb#L196 and #L201
