via: faker-ruby#2297

* 🐛 Fix indentation ( `ja.faker.name` to `ja.faker.color.name` )

via #2315

* 🎨 Convert a huge one-line array to multiple lines

Because the file cannot be opened with a weak editor...

* 🐛 Japanese postcode has a hyphen.

via: #2297

* Update test/test_ja_locale.rb

Co-authored-by: Koichi ITO <koic.ito@gmail.com>

Co-authored-by: Koichi ITO <koic.ito@gmail.com>

…ormat

Removes the full list of japanese postcodes in favor or a simple 7-digit format:
"###-####".

The reason is for performance, as generating a postcode in this locale was very expensive.

Japanese postal codes were changed from  a pattern to a list of real postal codes, introduced in #2297.

The problem is that when running Faker on `locale = 'ja'`, about 60% of the time to load and call `Address#postalcode` is spent reading and parsing the [ja/address.yml](https://github.com/faker-ruby/faker/blob/main/lib/locales/ja/address.yml) file, which has 2.2mb.

Benchmark:
```
Postcode - Smaller Address File:  1.2 i/s
Postcode:                         0.7 i/s - 1.65x  slower

1.34 s/i vs 811.55 ms/i
```

…ormat

Replaces the full list of japanese postcodes in favor of a simple 7-digit format:
"###-####".

The reason is for performance, as generating a postcode in this locale
is too expensive.

Japanese postal codes were changed from a simple pattern to a list of real postal codes, introduced in #2297.

The problem is that when running Faker on `locale = 'ja'`, about 60% of the time to load and call `Address#postalcode` is spent reading and parsing the [ja/address.yml](https://github.com/faker-ruby/faker/blob/main/lib/locales/ja/address.yml) file, which has 2.2mb.

Benchmark:
```
Postcode - Smaller Address File:  1.2 i/s
Postcode:                         0.7 i/s - 1.65x  slower

1.34 s/i vs 811.55 ms/i
```

This has the potential to improve the performance of generating
postcodes in japanese by 65%.

…ormat

Replaces the full list of japanese postcodes in favor of a simple 7-digit format:
"###-####".

The reason is for performance, as generating a postcode in this locale
is too expensive.

Japanese postal codes were changed from a simple pattern to a list of real postal codes, introduced in #2297.

The problem is that when running Faker on `locale = 'ja'`, about 60% of the time to load and call `Faker::Address.postalcode` is spent reading and parsing the [ja/address.yml](https://github.com/faker-ruby/faker/blob/main/lib/locales/ja/address.yml) file, which has 2.2mb.

Benchmark:
```
Postcode - Smaller Address File:  1.2 i/s
Postcode:                         0.7 i/s - 1.65x  slower

1.34 s/i vs 811.55 ms/i
```

This has the potential to improve the performance of generating
postcodes in japanese by 65%.

…ormat (#3201)

* perf: replaces list of postcodes in `ja/address.yml` with a 7-digit format

Replaces the full list of japanese postcodes in favor of a simple 7-digit format:
"###-####".

The reason is for performance, as generating a postcode in this locale
is too expensive.

Japanese postal codes were changed from a simple pattern to a list of real postal codes, introduced in #2297.

The problem is that when running Faker on `locale = 'ja'`, about 60% of the time to load and call `Faker::Address.postalcode` is spent reading and parsing the [ja/address.yml](https://github.com/faker-ruby/faker/blob/main/lib/locales/ja/address.yml) file, which has 2.2mb.

Benchmark:
```
Postcode - Smaller Address File:  1.2 i/s
Postcode:                         0.7 i/s - 1.65x  slower

1.34 s/i vs 811.55 ms/i
```

This has the potential to improve the performance of generating
postcodes in japanese by 65%.

* fix: remove flaky specs