feat: Add support for log bucket configuration

[matty-rose]

Fixes #88

[comment-bot-dev]

Thanks for the PR! 🚀
✅ Lint checks have passed.

[morgante]

We should use this same logic for the top-level module as well.

[matty-rose]

I'd prefer to keep the top-level module logic the same, as I believe it aligns with the raw bucket resource inputs better. For example this is a valid value for the logging variable in the top-level module:

{
  first = {
    log_bucket = "test-bucket"
  }
  second = {
    log_bucket = "test-bucket"
    log_object_prefix = "test-prefix"
  }
}

and in this case, GCS itself handles making the prefix the same as the bucket name for first.
The only reason I implemented the ternary logic in the simple_bucket module is that I set the type of the input variable to be an object with two strings, and the optional constraint is currently experimental in terraform 0.14. This means that the following is NOT a valid value for logging in the simple_bucket module, and so to obtain the default GCS behaviour I implemented the bucket name/prefix logic manually:

{
  log_bucket = "test-bucket"
}

There are two solutions that I'd choose over re-implementing the logic in the root module:

1. Change the type of the logging variable in the simple_bucket module to any and document the input format better, similar to the nested objects in lifecycle_rules - this would allow passing in the log_object_prefix field to actually be optional
2. Leave it as-is, and have a small difference in logic between the modules

What are your thoughts @morgante (Also thanks for taking a look at this so quick!)

[morgante]

In that case, I'd like to actually update the simple_bucket submodule to accept log_bucket and log_object_prefix as top-level variables. I don't think the layer of indirection in an object is helpful.

[matty-rose]

makes sense, have made this change @morgante