#testing #snapshot-testing

bin+lib runt

A lightweight and parallel snapshot testing framework

14 releases

0.2.8 Dec 7, 2020
0.2.7 Jul 24, 2020
0.2.4 Apr 18, 2020
0.2.1 Mar 30, 2020
0.1.4 Mar 30, 2020

#88 in Testing

Download history 16/week @ 2021-02-21 7/week @ 2021-02-28 21/week @ 2021-03-07 2/week @ 2021-03-14 30/week @ 2021-03-21 30/week @ 2021-03-28 28/week @ 2021-04-04 20/week @ 2021-04-11 2/week @ 2021-04-18 4/week @ 2021-04-25 20/week @ 2021-05-02 15/week @ 2021-05-16 3/week @ 2021-05-23 16/week @ 2021-06-06

75 downloads per month

Custom license

672 lines

RUN Test (RUNT)   latest

Runt is a lightweight, parallel snapshot testing framework. It aims to enable snapshot testing with minimal configuration.

Install the latest version of runt using:

cargo install runt

Runt is most useful when you have the following test setup:

  • One command that needs to run on many input files.
  • Test suites grouped by commands run on the files.
  • Test outputs are sent to IO streams (stdout and stderr).
  • Test and test suites are unordered.

Runt is not useful when you want to do:

  • Rich introspective testing of data structures.
  • Test suites with complex setups, dependencies, and teardowns.

Snapshot testing with runt is extremely flexible. For example, the tests under runt-cli-test test the outputs of the runt CLI.

Building & Developing

  • Install Rust.
  • Run cargo build --release. The runt executable is generated under target/release/runt.
  • Runt is tested using runt. Run runt runt-cli-test to test runt.


Runt is configured using a single runt.toml file:

# Version of runt to be used with this configuration.
ver = "0.2.2"

# Configuration for each test suite. File paths are relative to the folder
# containing runt.toml.
# Optional name for this test suite.
name = "Cat tests"
# Test paths can be globs or exact.
paths = [ "cat-test/*.txt" ]
# Command to run on each test file. {} is replaced with input name.
cmd = "cat {}"
# (Optional) Directory to store the generated .expect files.
expect_dir = "cat-out/"

name = "Ls test"
paths = [ "ls-test/input.txt" ]
cmd = "cat {} | ls"

name = "Error test"
paths = ["error-test/input.txt"]
cmd = "echo error message 1>&2 && exit 1"

Run runt <dir> to execute all the tests. <dir> defaults to the current directory.


Showing diffs: By default, runt does not show diffs between the new output and the expect file. Use --diff to show the diffs.

Saving changes: The --save flag overwrites the expect files to save the updated outputs.

Suppress specific outputs: The --only flag can be used to focus on only failing, missing, or correct tests. It composes with the diff and save flags.


  • Runt has a minimal configuration example under cli-tools. The runt.toml file contains all the configuration and explanation for various options.


  • When executing a large test suite, I get Too many open files (os error 24). Runt tries to spawn as many processes in parallel as possible and might hit the system limit on open file descriptors. Use ulimit -n 4096 to increase the number of file descriptors that can be opened at the same time.

Other options

  • Turnt is a testing framework that allows for more complex snapshot comparisons. It's particularly powerful when you have several intermediate files you'd like to compare. runt forgoes the flexibility of turnt for faster execution and built-in output diffing.
  • insta enables snapshot testing of inline rust programs. Useful when the testing intrinsic structure of Rust programs. runt operators on arbitrary shell commands which enables testing CLI programs.
  • jest is a JavaScript snapshot testing framework that allow formulation of complex expectation queries.


~144K SLoC