This should probably be something like:

I don't really like the way the code looks / readability at this point, but that's more of a comment on the overall state of this normalization method... If you can clean up this little section some it would be appreciated.

We should decide what we want the project name to be in workspace.json because if I am not mistaked, the projectName is what is used to add the project to NX.

So I see a couple of cases that could be NX pathScheme and dotnet.
1- nx g @nx-dotnet/core:library SomeCompany.Foobar --directory=somefolder ...
2- nx g @nx-dotnet/core:library SomeCompany.Foobar ...

In other words, with and without directory

NX Scheme with directory: somefolder-some-company.foobar, -
NX Scheme without directory: some-company.foobar,

And I would assume the following for dotnet scheme
Dotnet Scheme with directory: Somefolder.SomeCompany.Foobar
Dotnet Scheme without directory: SomeCompany.Foobar

What is the reason for adding the directory to the front of the name? Just to namespace it?

Feel free to message me directory to discuss. I am not sure if I was clear here. I want to get clarity on all cases before I rip this methods apart and try make it cleaner. That includes making the changes for projectName.

Yeah, adding the directory is mainly to namespace it, but also to maintain consistency w/ the other Nx generators. The csproj name matches the generated project name, and w/ the nx first party generators when you specify --directory, it is prepended to the project name.

This will be kinda ugly too, but we should default to Test rather than test for the suffix when using dotnet path scheme.

Perhaps:

  const suffix = schema.testProjectNameSuffix || (pathScheme === 'nx' ? 'test' : 'Test');

I don't really like having the || and a ternary here though, that's a lot of complexity represented by the single line. A decent amount of stuff changes based on that too, so it may be easier if we changed to something like:

let suffix: string;
let separator: string;
if (schema.pathScheme === 'nx') {
  separator = '-'
  suffix = schema.testProjectNameSuffix || 'test'
} else {
  separator = '.'
  suffix = schema.testProjectNameSuffix || 'Test'
}

I'm not really a fan of that either though... It'd look better if abstracted into a method.

// ....
const { suffix, separator } = getPathPartsFromSchema(schema)
// ...

function getPathPartsFromSchema(schema) {
  if (schema.pathScheme === 'nx') {
   return {
    separator: '-'
    suffix: schema.testProjectNameSuffix || 'test'
    }
  } else {
    return {
      separator: '.'
      suffix: schema.testProjectNameSuffix || 'Test'
    }
  }
}

This doesn't need to be exported