Confusing behavior of pipes with empty elements

Describe the bug

Pipelines are allowed to have empty parts. However behavior of these empty parts and where they are allowed to appear is inconsistent.

How to reproduce

1. In the middle of pipeline empty element is equivalent to $in

   ls
       |
       | print
   

   is equivalent to

   ls
       | $in
       | print
   

2. In the beginning of pipeline empty element is equivalent to null

    | describe
   

   is equivalent to

   null | describe
   

3. At the end of pipeline empty elements are not allowed. For example nu parser is not considering this a complete statement:
   

Expected behavior

Behavior 3. is expected. However behavior 1. and 2. are not. I expect them to be incorrect syntax, beacuse 1. is simply useless, I see no need for pipeline element that does nothing, and 2. is just confusing and also kinda useless (writing explicit null is much more clear)

Screenshots

No response

Configuration

Additional context

This issue arises when writing grammar for tree-sitter in PR https://github.com/nushell/tree-sitter-nu/pull/28 . Parsing rules around handling pipelines and newlines become significantly harder if empty pipeline elements are involved

Start with the pipeline and newline parsing rules discussed in tree-sitter-nu PR #28, comparing empty elements in the middle, beginning, and end. Reproduce the three examples in Nushell 0.84.0 to confirm the current behavior; done means empty pipeline elements are rejected consistently while explicit syntax remains valid.