GitHub

Practice Algorithms with NAUX Learn

Tutorial index · First program · Language reference

Learners may solve a problem in their own style. The corpus solutions are examples, not mandatory templates; a different source algorithm is valid when it stays inside the admitted language profile and produces the required output.

Deterministic input and output

  • read_int() reads one signed integer token.
  • read_token() reads one whitespace-delimited text token.
  • read_line() reads through the next line ending.
  • All input functions share one deterministic cursor.
  • !say value writes the value followed by exactly one newline.

Example: sum n integers

Save as sum.nx:

~ rite
    $n = read_int()
    $sum = 0
    ~ loop $n
        $sum = $sum + read_int()
    ~ end
    !say $sum
~ end

Save as input.txt:

5
10 -3 5 30 8

Linux:

"$NAUX" run sum.nx < input.txt

Windows PowerShell (source-built toolchain only; there is no 0.1.4 Windows binary):

Get-Content -Raw input.txt | & $Naux run sum.nx

Expected output is 50.

Example: list maximum

~ rite
    $n = read_int()
    $values = []
    ~ loop $n
        $values = queue_push($values, read_int())
    ~ end

    $maximum = $values[0]
    $i = 1
    ~ while $i < len($values)
        ~ if $values[$i] > $maximum
            $maximum = $values[$i]
        ~ end
        $i = $i + 1
    ~ end
    !say $maximum
~ end

Lists use zero-based indexing. queue_push returns a new list, so assign its result. NAUX Learn v0.1 admits ~ loop and ~ while, but no for, break, continue, or switch spelling.

Useful commands

naux check solution.nx
naux run solution.nx < input.txt
naux run solution.nx --engine interp < input.txt
naux run solution.nx --max-work 100000 --max-call-depth 64

Normal run uses the bounded VM. --engine interp selects the reference interpreter; both must agree on learner-visible output within this profile.

The 30-exercise corpus

The public corpus includes:

  • basics and math;
  • linear search, binary search, and prefix sums;
  • bubble, selection, insertion, and counting sort;
  • BFS, DFS, shortest paths, Dijkstra, connected components, and topological sort;
  • greedy algorithms;
  • dynamic programming.

Browse the source solutions and input/output fixtures. Download any .nx source and matching .in fixture and execute it through ordinary naux run.

Resource limits

Normal learner execution is bounded to reduce accidental runaway work. The defaults are 1,000,000 semantic work checkpoints and 128 active user calls; positive CLI overrides have hard ceilings. These limits are not a security sandbox or an operating-system memory limit. See the execution envelope.