Crates.io | testdir |
lib.rs | testdir |
version | 0.9.1 |
source | src |
created_at | 2020-12-06 21:07:10.961428 |
updated_at | 2023-12-14 17:12:46.64529 |
description | Semi-persistent, scoped test directories |
homepage | |
repository | https://github.com/flub/testdir |
max_upload_size | |
id | 320251 |
size | 61,444 |
This crate aims to make it easier to use temporary directories in
tests, primarily by calling the testdir!()
macro somewhere in a test
function. The direcetories are structured per-scope and per-test and
will be available for inspection after the test has finished. On
subsequent test runs older generations of test directories will be
cleaned up.
If you've ever used pytest's tmp_path
or tmpdir
fixtures
this pattern should be similar.
By default test directories are created in cargo's target directory:
target/testdir-$N/module/path/test_name
Here $N
is an integer generation number increasing with each cargo test
invocation. Generations older than the 8 most recent ones are
removed on the next cargo test
run.
There is also a symlink pointing to the most recent generation:
target/testdir-current -> testdir-$N`
Note however that on windows sometimes this can not be updated due to permissions, this symlink is a best-effort on windows.
Even when executing this with cargo test --jobs=1
these tests will
pass as each gets their own unique directory:
// E.g. in lib.rs
mod tests {
use std::path::PathBuf;
use testdir::testdir;
#[test]
fn test_write() {
let dir: PathBuf = testdir!();
let path = dir.join("hello.txt");
std::fs::write(&path, "hi there").ok();
assert!(path.exists());
}
#[test]
fn test_read() {
let dir: PathBuf = testdir!();
let path = dir.join("hello.txt");
assert!(!path.exists());
}
}
Afterwards you can inspect the directories and they should look something like this:
$ tree target/
target/
+- testdir-0
| +- cratename
| +- tests
| +- test_read
| +- test_write
| +- hello.txt
+- testdir-current -> testdir-0
The code lives in a git repository at https://github.com/flub/testdir from which you can create issues, clone the repository and create pull requests.