fix: unparse Int8 casts as SIGNED for MySQL dialect - #26883

…sh_collisions (apache#27034)

## Which issue does this PR close?

- Closes apache#27033.

## Rationale for this change

`completed_groups_release_first_value_bitmap` (added in apache#26782) takes
about 10 minutes in the `cargo test hash collisions (amd64)` merge-queue
job. It aggregates 131,072 distinct keys in one batch. Under
`force_hash_collisions`, every key has the hash 0, so the aggregate is
quadratic. The test checks that `FIRST_VALUE(v ORDER BY v)` releases its
per-batch bitmap, which does not depend on hashing.

The long job gives spot runner shutdowns more time to hit it. Evidence
from merge-queue runs:

- In the passing rerun for apache#26926
([job](https://github.com/apache/datafusion/actions/runs/41571721394/job/125710730373)),
the physical-plan unit test binary took 609.89s. All other tests in it
finished by 03:40:31. This test ran from about 03:40:07 to 03:50:17.
- The median job time went from 6.2 minutes (230 jobs, Oct 21 to Oct 30)
to 12.0 minutes (70 jobs after apache#26782 merged).
- All 5 failed attempts of the job since then were runner shutdowns
while this test was running: apache#26929
([job](https://github.com/apache/datafusion/actions/runs/41114163010/job/124263887540)),
apache#26898
([job](https://github.com/apache/datafusion/actions/runs/41115099549/job/124267135386)),
apache#26883
([job](https://github.com/apache/datafusion/actions/runs/41120437208/job/124285496814)),
apache#26974
([job](https://github.com/apache/datafusion/actions/runs/41559552415/job/125666576696)),
apache#26926
([job](https://github.com/apache/datafusion/actions/runs/41571721394/job/125705896210)).

## What changes are included in this PR?

Skip the test under `force_hash_collisions`, as apache#26594 did for another
test. The imports that only this test uses move into its body, so they
are not unused under the feature.

## What is the testing strategy for this PR?

Test config change only. The regular `cargo test` job still runs the
test. Locally (`--profile ci`), the test takes 0.92s by default and 309s
with `force_hash_collisions`. With the feature, it is now skipped, and
Clippy passes with and without the feature.

## Are there any user-facing changes?

No.

## Which issue does this PR close?

  - Closes apache#26887.

  ## Rationale for this change

MySQL's `CAST` does not accept `SMALLINT` or the `... UNSIGNED` integer
types as target types, but `arrow_dtype_to_ast_dtype` hard-codes them
for `Int16` and `UInt8`..`UInt64`. As a result, `MySqlDialect` cannot
override them and the generated SQL fails with `ERROR 1064`.

See
https://dev.mysql.com/doc/refman/8.0/en/cast-functions.html#function_cast

Checked against MySQL 8.0.46 and 8.4.13 with `SELECT CAST('1' AS
<type>)`: every type in the issue table fails with `ERROR 1064`, while
`SIGNED` and `UNSIGNED` succeed.

  ## What changes are included in this PR?

- New `Dialect` hooks `int16_cast_dtype` and
`uint8`/`uint16`/`uint32`/`uint64_cast_dtype`. Their defaults return the
previous types.
- `MySqlDialect` returns `SIGNED` for `Int16` and `UNSIGNED` for the
unsigned types.

  This builds on apache#26883.

  ## What is the testing strategy for this PR?

  New unit tests in `datafusion/sql/src/unparser/expr.rs`:
  - `mysql_int16_and_unsigned_casts`: the new MySQL output
- `default_dialect_int16_and_unsigned_casts`: the default output is
unchanged
  - `custom_dialect_int16_and_unsigned_casts`: the builder setters

  ## Are there any user-facing changes?

With `MySqlDialect`, `Int16` casts now unparse as `SIGNED` and unsigned
integer casts as `UNSIGNED`.