Document know limitation about using `references` in conjunction with custom select clauses [ci skip]

Document know limitation about using `references` in conjunction with eager loading causing it to ignore custom select clauses.

[ci skip]

Fixes #24314

Add link to 'Pry' in debugging guide [ci skip]

update list of rescue_responses default [ci skip]

Follow up to fe859a54

Use `FETCH FIRST` for Oracle12 and test `ROWNUM <=` for Oracle 11g or older version to test sql limit behavior

also test `ROWNUM <=` for Oracle 11g or older version of Oracle and Oracle visitor
Oracle 12c database and Arel Oracle12 visitor supports better top N query.

This code has too much duplication and the rationale for the concatenation
may not be obvious to the reader. You define the ones at class-level, explain
why does the code concatenates there, and then the convenience ones at
instance-level just delegate.

Fix `thread_mattr_accessor` share variable superclass with subclass

A few have been left for aesthetic reasons, but have made a pass
and removed most of them.

Note that if the method `foo` returns an array, `foo << 1`
is a regular push, nothing to do with assignments, so
no self required.

Some case expressions remain, need to think about those ones.

Inserted spaces in the name of Rails components.

Since I was on it, also used PostgreSQL instead of Postgres
because albeit Postgres is an accepted alias, PostgreSQL is
the official name and the actual name of the adapter.

See

    https://wiki.postgresql.org/wiki/ProjectName

with regard to PostgreSQL vs Postgres.

In some test runs `ActionController::Base.enable_fragment_cache_logging` would be false,
based on the test order.

Turns out it was off because we assigned the config to the `@controller` variable, and
not on `ActionController::Base`.

The test failure was reproducible with, and now passes after this:

```
bin/test test/controller/log_subscriber_test.rb --seed 19918
```

The entry was a result of a combination of changes in Action View
and Action Controller.

Better logging of cached partial renders 

Fix Accept header overwritten issue when "xhr: true" in integration test

Implement naive partial caching mechanism.

Add test for LogSubscriber

Use ActionView::Base#log_payload to store log_subscriber's payload, so we can pass cache result into it.

Fixed tests

Remove useless settings

Check if #log_payload exists before calling it. Because other classes also includes CacheHelper but don't have is attribute

Use @log_payload_for_partial_reder instead of #log_payload to carry ActionView's payload.

Update test's hash syntax

Add configuration to enable/disable fragment caching logging

Remove unless test and add new test to ensure cache info won't effect next rendering's log

Move :enable_fragment_cache_logging config from ActionView to ActionPack

Apply new config to tests

Update actionview's changelog

Update configuration guide

Improve actionview's changelog

Refactor PartialRenderer#render and log tests

Mute subscriber's log instead of disabling instrumentation.

Fix typo, remove useless comment and use new hash syntax

Improve actionpack's log_subscriber test

Fix rebase mistake

Apply new config to all caching intstrument actions

In integration test when specify the "Accept" header with "xhr: true"
option, the Accept header is overridden with a default xhr Accept
header. The issue only affects HTTP header "Accept" but not CGI variable
"HTTP_ACCEPT".

For example:

    get '/page', headers: { 'Accept' => 'application/json' }, xhr: true

    # This is WRONG! And the response.content_type is also affected.
    # It should be "application/json"
    assert_equal "text/javascript, text/html, ...", request.accept
    assert_equal 'text/html', response.content_type

The issue is in `ActionDispatch::Integration::RequestHelpers`. When
setting "xhr: true" the helper sets a default HTTP_ACCEPT if blank.
But the code doesn't consider supporting both HTTP header style and
CGI variable style.

For detail see this GitHub issue:
https://github.com/rails/rails/issues/25859

`ActiveRecord::PredicateBuilder#expand` to be private

show error message when error raised in rails runner

Eagerly requiring Rake could put 100ms on to the Rails boot time. Shimmy that
down to 0 by requiring Rake when no native Rails command exists and we try to
run a Rake task. Or when printing all the Rake commands through `rails help`.

Fixes #25029

Add `Style/EmptyLines` in `.rubocop.yml` and remove extra empty lines

This method is not touched from outside.

Revert passing arel node with splat binds for `where`

.to_i is unnecessary for puma threads.