[12.x] Handle Enum::cases() for enum column type

[iamgergo]

This PR adds a small modification due to the enum method in migrations, to make the development a bit more streamlined.

enum Role: string
{
    case ADMIN = 'admin';
    case MEMBER = 'member';
}

// Before
Schema::create('users', static function (Blueprint $table): void {
    $table->enum('role', ['admin', 'member']);
});

// After
Schema::create('users', static function (Blueprint $table): void {
    $table->enum('role', Role::cases());
});

Normally this PR has no breaking change, since the enum_value works with non-enums as well, returning the original value.

[shaedrich]

Migrations are intended to be non-dependent on other application code that is prone to change as this may break the migration at a later date. So, this is probably not what most devs want. What would work better, though, would be kind of a stub that would prefill the migration with the hard-coded enum values from that point.

[iamgergo]

Migrations are intended to be non-dependent on other application code that is prone to change as this may break the migration at a later date. So, this is probably not what most devs want. What would work better, though, would be kind of a stub that would prefill the migration with the hard-coded enum values from that point.

I see this differently. There are already some methods  – like foreignIdFor(Model::class) – that depend on other part of the application. Also, at the end of the day the original way is still can be used if someone is afraid of breaking the code.

For me, the migrations are part of the same application as the enums or the models.