1
0
Fork 0
mirror of https://github.com/imjasonh/testscript-rs synced 2026-07-09 01:26:40 +00:00
testscript-rs/README.md
Copilot a69a295cbc
Allow setting environment variables in setup hook (#44)
## Plan: Support passing env vars into test environment in `setup`

- [x] Update `SetupFn` type alias in `params.rs` to accept `&mut
TestEnvironment` instead of `&TestEnvironment`
- [x] Update `setup` method signature in `params.rs` to accept `Fn(&mut
TestEnvironment)` instead of `Fn(&TestEnvironment)`
- [x] Update `setup` method signature in `lib.rs` Builder to accept
`Fn(&mut TestEnvironment)` instead of `Fn(&TestEnvironment)`
- [x] Update execution in `execution.rs` to pass `&mut env` to setup
function
- [x] Update existing test in `setup_hook.rs` to test the new
functionality (make it pass)
- [x] Add a new focused test to validate setting environment variables
in setup hook
- [x] Run tests to validate changes - all tests pass
- [x] Update documentation examples to show the new capability
- [x] Address code review feedback - fix trailing newline inconsistency
- [x] Add comprehensive test for environment variable presence/absence
to ensure proper isolation
- [x] Enhance tests to demonstrate env vars work without sh -c wrapper
(using printenv/set)
- [x] Format assert statement to single line per style guidelines

<!-- START COPILOT ORIGINAL PROMPT -->



<details>

<summary>Original prompt</summary>

> 
> ----
> 
> *This section details on the original issue you should resolve*
> 
> <issue_title>support passing env vars into the test environment in
`setup`</issue_title>
> <issue_description>```
> .setup(|env| => {
>   env.set("FOO", "bar")
> }
> ```
> 
> (or similar)
> 
> this should make `FOO` have the value of `bar` in the test
environment</issue_description>
> 
> ## Comments on the Issue (you are @copilot in this section)
> 
> <comments>
> </comments>
> 


</details>



<!-- START COPILOT CODING AGENT SUFFIX -->

- Fixes imjasonh/testscript-rs#43

<!-- START COPILOT CODING AGENT TIPS -->
---

 Let Copilot coding agent [set things up for
you](https://github.com/imjasonh/testscript-rs/issues/new?title=+Set+up+Copilot+instructions&body=Configure%20instructions%20for%20this%20repository%20as%20documented%20in%20%5BBest%20practices%20for%20Copilot%20coding%20agent%20in%20your%20repository%5D%28https://gh.io/copilot-coding-agent-tips%29%2E%0A%0A%3COnboard%20this%20repo%3E&assignees=copilot)
— coding agent works faster and does higher quality work when set up for
your repo.

---------

Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: imjasonh <210737+imjasonh@users.noreply.github.com>
2025-12-29 21:29:03 -05:00

5.7 KiB

testscript-rs

CI Crates.io

A Rust crate for testing command-line tools using filesystem-based script files.

testscript-rs provides a framework for writing integration tests for CLI applications using a simple DSL using the .txtar format, where test scripts and file contents are combined in a single file.

This crate is inspired by and aims to be compatible with Go's github.com/rogpeppe/go-internal/testscript package, which was itself extracted from internal packages in the Go stdlib, having originally been written by Russ Cox.

Testscript is primarily useful for describing testing scenarios involving executing commands and dealing with files. This makes it a good choice for testing CLI applications in a succinct and human-readable way.

Quick Example

Given a file testdata/simple.txt:

echo "hello"    # print "hello"
stdout "hello"  # check that stdout includes "hello"

You can run this in your tests:

use testscript_rs::testscript;

#[test]
fn test_my_cli() {
    testscript::run("testdata")
        .execute()
        .unwrap();
}

The test will find every testscript file in ./testdata, and execute it.

A more realistic example might look like this:

use testscript_rs::testscript;

#[test]
fn test_my_cli() {
    testscript::run("testdata")
        .setup(|env| {
            // Set environment variables for tests
            env.set_env_var("MY_APP_CONFIG", "/path/to/config");
            
            // Compile your CLI tool
            std::process::Command::new("cargo")
                .args(["build", "--bin", "my-cli"])
                .status()?;

            // Copy binary to test environment
            std::fs::copy("target/debug/my-cli", env.work_dir.join("my-cli"))?;
            Ok(())
        })
        .execute()
        .unwrap();
}

With a test script in testdata/basic.txt:

# Test basic functionality
exec ./my-cli --version
stdout "my-cli 1.0"

exec ./my-cli process input.txt
cmp output.txt expected.txt

-- input.txt --
test content

-- expected.txt --
processed: test content

Running the test will compile the CLI program, make it available to the testscript environment, run the specified commands, and check its output.

Installation

Add testscript-rs to your Cargo.toml:

[dev-dependencies]
testscript-rs = "<release>"

Requires Rust 1.70 or later.

Usage

With Custom Commands

testscript::run("testdata")
    .command("custom-cmd", |env, args| {
        // Your custom command implementation, e.g., the entrypoint of your CLI, so you don't have to `cargo build` it as above.
        println!("Running custom command with args: {:?}", args);
        Ok(())
    })
    .condition("feature-enabled", true)
    .preserve_work_on_failure(true)  // Debug failed tests
    .execute()
    .unwrap();

To call the custom command, in your testscript file:

custom-cmd arg1 arg2 arg3

Test Script Format

Test scripts use the txtar format. For complete format documentation, see the original Go testscript documentation.

Built-in Commands

  • exec - Execute external commands
  • cmp - Compare two files
  • stdout/stderr - Check command output (supports regex)
  • exists - Check file existence
  • mkdir - Create directories
  • cp - Copy files (supports stdout/stderr as source)
  • mv - Move/rename files
  • rm - Remove files/directories
  • chmod - Change file permissions
  • env - Set environment variables
  • cmpenv - Compare files with environment variable substitution
  • stdin - Set stdin for next command
  • cd - Change working directory
  • wait - Wait for background processes
  • kill - Kill background processes
  • skip - Skip test execution
  • stop - Stop test early (pass)
  • unquote - Remove leading > from file lines
  • grep - Search files with regex
  • symlink - Create symbolic links

Commands can be prefixed with conditions ([unix]) or negated (!).

Error Messages

testscript-rs provides detailed error messages with script context to make debugging easy:

Error in testdata/hello.txt at line 6:
  3 | stdout "this works"
  4 |
  5 | # This command will fail
> 6 | exec nonexistent-command arg1 arg2
  7 | stdout "should not get here"
  8 |

Note: Some features of testscript in Go are not supported in this Rust port:

  • [gc] for whether Go was built with gc
  • [gccgo] for whether Go was built with gccgo
  • [go1.x] for whether the Go version is 1.x or later

Examples

See examples/sample-cli/ and its testdata directory for more examples.

There are also more tests in testdata that demonstrate and check this implementations behavior.

UpdateScripts for Easier Test Maintenance

UpdateScripts automatically updates test files with actual command output, making test maintenance easier:

// Enable via API
testscript::run("testdata")
    .update_scripts(true)
    .execute()
    .unwrap();

Or via environment variable:

UPDATE_SCRIPTS=1 cargo test

When enabled, instead of failing on output mismatches, the test files will be updated with actual command output:

Before (failing test):

exec my-tool --version
stdout "my-tool 1.0"

After running with update mode:

exec my-tool --version
stdout "my-tool 2.1.0"

This feature only updates stdout and stderr expectations while preserving file structure and comments.