When `group` is used in combination with any calculation method, the
resulting hash uses the grouping expression as the key. Currently we're
incorrectly always favoring the type reported by the query, instead of
the type known by the class. This causes differing behavior depending on
whether the adaptor actually gives proper types with the query or not.
After this change, the behavior will be the same on all adaptors -- we
see if we know the type from the class, fall back to the type from the
query, and finally fall back to the identity type.

Fixes #25595

The documentation states that parameter values longer than 20 characters
will be truncated by words, but the example shows that a parameter based
on "David Heinemeier Hansson" (with id: 125) becomes "125-david" when
"David Heinemeier".length == 16 so why so short?

The answer lies in the use of the #truncate option omission: nil which
seems to have been intended to mean "nothing", but which actually causes
the default string "..." to be used. This causes #truncate to cleave
words until the "..." can be added and still remain within the requested
size of 20 characters.

The better option is omission: '' (which is probably what was originally
intended).

Furthermore, since the use of #parameterize will remove non-alphanumeric
characters, we can maximize the useful content of the output by calling
parameterize first and then giving truncate a separator: /-/ rather than
a space.

This includes minor grammar fixes throughout the file.

This way, we aren't racing other threads, so we don't need to re-check
the conditional. And we no longer need to hold the lock while calling
remove (which can choose to make a new connection while we wait).

A concurrent thread may have also detected it to be stale, and already
released (or even reassigned) it by now.

Fixes #25585

A thread can only release a connection if it owns it, or it's owned by a
thread that has died.

This nested if checked the same value as the containing case statement.
Moved the code in the if/else into when/else in the containing case.

The current error message only indicates that a touch can fail because the record is new. In practice, we saw cases where touches were failing because the record had been destroyed. `persisted?` checks `new_record?` *and* `destroyed?`. It was confusing to get a message about a new record when in reality we were destroying records.

I also included a helpful tip for users to consider using `persisted?`, `new_record?`, or `destroyed?` before touching.

`create_table_info_cache` is used for sharing `create_table_info` both
`table_options` and `foreign_keys`. But `foreign_keys` no longer uses
`create_table_info_cache` by #25307. No need caching anymore.

This reverts commit 407e0ab5.

Because causing an extra query by `sql_for_insert` for guessing a
primary key.
https://github.com/rails/rails/blob/v5.0.0/activerecord/lib/active_record/connection_adapters/postgresql/database_statements.rb#L121-L125

Do not specal case inspecting associated arrays of over 10 elements, preventing infinite looping in some cases.

Right now with this middleware we are likely always connected, but I'm
hoping to change that soon. This is an easy change to pull out on its
own.

This method appears to have been partially used in connection pool
caching, but it was introduced without much reasoning or any tests. One
edge case test was added later on, but it was focused on implementation
details. This method is no longer used outside of tests, and as such is
removed.

This was added at c7c3f73f but it never raised because MySQL cannot
create text/blob columns with a default value.

As mentioned in 7b86ea67, this is an
internal class.

[ci skip]

r? @sgrif

This reverts commit 7ea502ae, per
internal discussion with @sgrif -- this is documenting the
implementation of a class that isn't intended to be public API.

If you only want the last element of a result set, there's no need to create all
of hash_rows. Also, add a test.

The implementation from abstract/database_statements.rb seems to work just fine.
And with ActiveRecord::Result now implementing an optimized #first method, the
performance concerns previously addressed in
https://github.com/rails/rails/commit/bf79aa4fc14aeb2646331e767038acf0b77e9e7f
should not be an issue.

When you only need the first element of a result set, you shouldn't need to
instantiate all of hash_rows.

`FinderMethods#exists?` should return a boolean rather than raising an
exception.

`UniquenessValidator#build_relation` catches a `RangeError` because it
includes type casting due to a string value truncation. But a string
value truncation was removed at #23523 then type casting in
`build_relation` is no longer necessary. aa062318 removes type casting in
`build_relation` then a `RangeError` moves to `relation.exists?`.

This change will remove the catching a `RangeError`.

Before we enable query caching we check if the connection is
connected. Before this fix we were always checking against the main
connection, and not the model connection.

Fixes #25391

`construct_relation_for_association_calculations` pass a string value to
`construct_join_dependency` when setting a string value in `from`.
It should not pass a string value, but always `joins_values`.

Related #14834, #19452.

Fixes #24193.

[ci skip]

- Followup of https://github.com/rails/rails/commit/1ea6cc11211dc89e3e14b2b641a3cca8a0a91d55.

Previously model file was generated first, which resulted in
inheriting from `ActiveRecord::Base`, but since application_record.rb
is generated as well, it should already be used.