Add support for partitioned indexes in PostgreSQL 11+

indexes in a table.

Currently the pg_class catalog is filtered out to retrieve the indexes in a
table by its relkind value. Which in versions lower than 11 of PostgreSQL
is always `i` (lower case). But since version 11, PostgreSQL
supports partitioned indexes referenced with a relkind value of `I`
(upper case). This makes any feature within the current code base to exclude those
partitioned indexes.

The solution proposed is to make use of the `IN` clause to filter those
relkind values of `i` and/or `I` when retrieving a table indexes.

Add a fallback database config when loading schema cache

```ruby
str = "abc"

Benchmark.ips do |x|
  x.report("start_with?") { str.start_with?("a") }
  x.report("first ==")    { str.first == "a" }
end
```

```
Warming up --------------------------------------
         start_with?   282.381k i/100ms
            first ==   207.305k i/100ms
Calculating -------------------------------------
         start_with?     10.239M (± 2.2%) i/s -     51.393M in   5.022151s
            first ==      4.593M (± 4.5%) i/s -     23.011M in   5.021434s
```

I think allowing nested array parameters is almost useless.

This is an alternative of b7e591a5 and a43de738.

Handling `if options.any?` is not preferable solution for kwargs.

Change all "can not"s to the correct "cannot"

It's been 6 years since [the original](https://github.com/rails/rails/pull/13584), so we're overdue for a sequel.

The correct word is "cannot", one word, not "can not", two words.

The schema cache defaults to loading the 'primary' database config.
However, if an app doesn't have a db config with a spec name of
'primary' the filename lookup will blow up.

This adds a fallback for this case.

Update "Eagerly Requiring Helpers" section in testing docs [ci skip]

This will change the path for requiring helper files to be consistent
with previous examples in "Using Separate Files" section

Fix the environment.rb file path in the benchmark generator template

[GitHub Actions] Save few seconds on every build

Install bundler takes about 7-8s and without installing documentations
takes about 3-4s

Correct test helpers paths in test guides

This will correct the paths to the test helpers and make the examples work.

Save a string allocation for each attribute method call

Follow-up c280ae72.

Add support for `if_exists/if_not_exists` on `remove_column/add_column`

This PR adds support for `if_exists` on `remove_column` and
`if_not_exists` on `add_column` to support silently ignoring migrations
if the remove tries to remove a non-existent column or an add tries to
add an already existing column.

We (GitHub) have custom monkey-patched support for these features and
would like to upstream this behavior.

This matches the same behavior that is supported for `create_table` and
`drop_table`. The behavior for sqlite is different from mysql/postgres
and sqlite for remove column and that is reflected in the tests.

Use correct filename for schema cache on load

Cleanup cache expanded_key

Use transform_values to build hash_rows

The initializer that loads the default schema cache on the default
connection doesn't account for the case where an app overrides the
default filename either via ENV["SCHEMA_PATH"], or via the
:schema_cache_path defined in the db config.

Note that as discussed in #34449 this initializer doesn't work for
applications using multiple databases, and this change doesn't fix that.

`lib` isn't autoloaded in most applications and I wouldn't recomment
storing test helpers there even if it was.

This was brought up in #38343. I think the original author meant to
write `test/lib`. I updated the docs to say to use either `test/lib` or
`test/test_helpers` (personally I prefer the latter).

Deduplicate empty `@if` and `@unless` arrays in Active Support callbacks