Before Rails 4.0, `config.cache_classes` determined whether application
code was eager loaded. The `config.eager_load` option was introduced to
allow the two behaviours to be configured independently, but this
documentation was never updated to reflect that change.
Puma gets bundler's info from `Bundler::ORIGINAL_ENV` for restart.
f6f3892f4d/lib/puma/launcher.rb (L168)
So, specified `BUNDLE_GEMFILE` env for use same Gemfile in the restart.
Fixes#31351
This pull request handles `FrozenError` introduced by Ruby 2.5.
Refer https://bugs.ruby-lang.org/projects/ruby-trunk/repository/revisions/61131
Since `FrozenError` is a subclass of `RuntimeError` minitest used by master
branch can handle it, though it would be better to handle `FrozenError`
explicitly if possible.
`FrozenError` does not exist in Ruby 2.4 or lower, `frozen_error_class`
handles which exception is expected to be raised.
This pull request is intended to be merged to master,
then backported to `5-1-stable` to address #31508
The `app:update` rake task will regenerate `development.rb` so that it
contains this option; that means we're currently adding it to existing
apps in two places, which is unnecessary and confusing.
Also:
- Remove inaccurate comment about which stack frames are ignored
- Clarify that the feature uses `caller_locations`, not `caller`
- Remove unused return value in `extract_callstack`
`bootsnap` is a useful gem normally. However, `bootsnap` is unnecessary
when generating a Rails application to be used only for testing.
So I want to control whether use this or not by option.
I've noticed during pair/mob programming sessions with peers that
despite the speed boosts provided by Bootsnap and Spring, there is a
noticeable latency between firing a bin/rails server command and any
feedback being provided to the console. Depending on the size of the
application this lack of feedback can make it seem like something is
wrong when Rails is simply busy initializing.
This change may seem gratuitous but by just printing one line to STDOUT
we're giving a clear signal to the Rails user that their command has
been received and that Rails is indeed booting. It almost imperciptibly
makes Rails feel more responsive.
Sure the code doesn't look very fancy but there's no other appropriate
place I could think of putting it than boot.rb.
Compare these two GIFs of booting without and with this change:
Before:
After:
This new ActiveRecord configuration option allows you to easily
pinpoint what line of application code is triggering SQL queries in the
development log by appending below each SQL statement log the line of
Ruby code that triggered it.
It’s useful with N+1 issues, and to locate stray queries.
By default this new option ignores Rails and Ruby code in order to
surface only callers from your application Ruby code or your gems.
It is enabled on newly generated Rails 5.2 applications and can be
enabled on existing Rails applications:
```ruby
Rails.application.configure do
# ...
config.active_record.verbose_query_logs = true
end
```
The `rails app:upgrade` task will also add it to
`config/development.rb`.
This feature purposely avoids coupling with
ActiveSupport::BacktraceCleaner since ActiveRecord can be used without
ActiveRecord. This decision can be reverted in the future to allow more
configurable backtraces (the exclusion of gem callers for example).
- The namespacing should be `ActiveRecord::FixtureSet`
- I might be missing something but I'm not sure why `create_fixtures` is useful for nowaday (unless for testing rails internal /shrug) and since it's been that long it wasn't working I think it should be fine to just fire it
Option parsing happens too late to have any impact on the Rails
environment. Rails accesses the environment name and memoizes it too
early in the boot process for a commandline option to have any impact on
the database connection, so we'll change this test to set the
environment from an environment variable (and ensure it still works when
running tests with `ruby`)
When tests are run with just `ruby`, the RAILS_ENV is set to
`development` too early, and we connect to the development database
rather than the test database.
I frequently run tests with `ruby`, not with a runner like `rake` or
`rails`. When running the test with just `ruby` the `RAILS_ENV`
environment variable did not get set to "test", and this would cause the
tests to fail (and even mutate the development database!)
This commit adds integration tests for running tests with just `ruby`
and ensures the environment gets defaulted to "test". I also added a
test to ensure that passing an environment to `-e` actually works (and
fixed that case too).
An interesting / annoying thing is that Minitest picks up it's plugins
by asking RubyGems for a list of files:
ca6a71ca90/lib/minitest.rb (L92-L100)
This means that RubyGems needs to somehow know about the file before it
can return it to Minitest. Since we are not packaging Rails as a Gem
before running the integration tests on it (duh, why would you do
that?), RubyGems doesn't know about the file, so it can't tell Minitest,
so Minitest doesn't automatically require it. This means I had to
manually require and insert the plugin in our integration test. I've
left comments about that in the test as well.
Ugh.
With the current code, a failing test shows this error, which is missing
the number of times called and has two periods at the end.
```
/railties$ be ruby -Itest test/generators/app_generator_test.rb -n test_active_storage_install
Failure:
AppGeneratorTest#test_active_storage_install [test/generators/app_generator_test.rb:313]:
active_storage:install expected to be called once, but was called times..
Expected: 1
Actual: 2
```
After the fix, the error message looks correct:
```
/railties$ be ruby -Itest test/generators/app_generator_test.rb -n test_active_storage_install
Failure:
AppGeneratorTest#test_active_storage_install [test/generators/app_generator_test.rb:313]:
active_storage:install expected to be called once, but was called 2 times.
Expected: 1
Actual: 2
```
