Are Boolean operators lazy?

[numero-744]

Hello,

TL;DR: Are the and and or Boolean operators guaranteed to be lazy, so we can document it?

---

I encountered this bug nushell/nushell#10237 which makes last error out on an empty input instead of returning a null if I understand correctly.

So I have written this workaround:

  let filler = if ($text != "") and ($text | split chars | last) == "%" { "%" } else { " " }

This code works without errors for me, so the RHS of the and Boolean operator must not be evaluated / be lazily evaluated.
To put it short, the and operator seems lazy (nu version 0.108.0)

I cannot find it documented in Boolean | Nushell and looking for "lazy" in the documentation does not yield anything related to Booleans.

Is and guaranteed to be lazy in future versions? What about or?
Should we document this behavior?

[Kissaki]

Rather than "lazy", the concept is typically called operator precedence, which ranks the syntax operators into a precedence ranking. (where "lazy" would be lower precedence.) (Example: C#)

Looks like there's nothing about precedence either, though.

The "Language Reference Guide" has a page for Operators, that that only lists operators, categorized.

[numero-744]

I think precedence is only a syntax thing allowing to remove some braces while still having a defined evaluation order. Here the question is not to know when parts of the expression are evaluated, it is to know if they are evaluated on not. In the documentation you sent, it seems to relate more to the "operand evaluation" section than the "operator precedence" one. They mention this as conditional evaluation:

Typically, all operator operands are evaluated. However, some operators evaluate operands conditionally. That is, the value of the leftmost operand of such an operator defines if (or which) other operands should be evaluated. These operators are the conditional logical AND (&&) and OR (||) operators, […]

Wikipedia page for Lazy Evaluation redirects to Short-circuit Evaluation for Boolean operators: https://en.wikipedia.org/wiki/Short-circuit_evaluation So maybe the "short-circuit evaluation" name better fits? Le 26 novembre 2025 17:51:23 UTC, Jan Klass ***@***.***> a écrit :

…

Kissaki left a comment (nushell/nushell.github.io#2072) Rather than "lazy", the concept is typically called operator precedence, which ranks the syntax operators into a precedence ranking. (where "lazy" would be lower precedence.) *([Example: C#](https://learn.microsoft.com/en-us/dotnet/csharp/language-reference/operators/#operator-precedence))* Looks like there's nothing about *precedence* either, though. The "Language Reference Guide" has a page for [Operators](https://www.nushell.sh/lang-guide/chapters/operators.html), that that only lists operators, categorized. -- Reply to this email directly or view it on GitHub: #2072 (comment) You are receiving this because you authored the thread. Message ID: ***@***.***>

[Kissaki]

Ah, yeah, now I get what you meant.

I remember reading relating to it within Nushell, but possibly within past release notes.

if/else does not mention and/or. Operators mentions and/or without information or references. if command help/docs don't mention it.

My expectation is that Nushell uses short-circuit/lazy evaluation.