Mention error handling as a potential reason to use a loop

[ItsHarper]

When you're doing validation on iterable data and want to report problems using error make, it's much better to use for than each, as each will throw an nu::shell::eval_block_with_input instead of the actual error you made.

Example

for

> for x in [0] { error make { msg: "Custom error" } }
Error:   × Custom error
   ╭─[entry #635:1:16]
 1 │ for x in [0] { error make { msg: "Custom error" } }
   ·                ─────┬────
   ·                     ╰── originates from here
   ╰────

each

> [0] | each { error make { msg: "Custom error" } }
Error: nu::shell::eval_block_with_input

  × Eval block failed with pipeline input
   ╭─[entry #629:1:2]
 1 │ [0] | each { error make { msg: "Custom error" } }
   ·  ┬
   ·  ╰── source value
   ╰────

Error:
  × Custom error
   ╭─[entry #629:1:14]
 1 │ [0] | each { error make { msg: "Custom error" } }
   ·              ─────┬────
   ·                   ╰── originates from here
   ╰────

The fact that you get a nu::shell::eval_block_with_input instead of the error you made of course impacts catch blocks as well.

[Kissaki]

Where would this be mentioned? The Loop Advantages sections seems fitting. Will that be enough?

I created #2020 for it. I took the liberty of using your example code.

I don't see dedicated error handling/error raising documentation, so I'll consider this ticket done with the mention on loop advantages.