sh.test is a test runner that can be used to write functional tests as single executable scripts. sh.test will find and execute these tests and report on the test status. In addition, sh.test is able to export the test results as JUnit compatible XML, which makes it easy to integrate the test runs in your CI server.
You can install sh.test from pip
pip install sh.test
If you want to install the latest version directory from the github repository, checkout the repository and run:
python setup.py install
After sh.test is installed, a single executable sh.test will be available in your path.
If you want to install sh.test in user space or you do not have root privileges to install sh.test globally, append the --user option to the installation command. This will install sh.test in $HOME/.local. Note that you might have to put $HOME/.local/bin in your PATH.
You can see a few very limited examples in the test_scripts folder. sh.test searched for files that contain the word test in their name and that are executable. Those file are executed by sh.test and if they fail to execute (exit status != 0) the test is marked as failed.
By default sh.test searches recursively in the current working directory, but you can specify directories to search for script or test script as command line options. For example:
sh.test test_script
This will search for all test script in the test_sript folder and execute them.
By default, sh.test will execute all test in the current working directory. You can use the --cwd option to tell sh.test to switch to the tests script parent folder before execution. This can be helpful if you want to store test data next to your script.
In order to allow you to create temporary output during a test run, sh.script creates and exposes a dedicated temporary folder for each test. The path to the folder is stored in the TEST_DIR environment variable and the folder will be automatically removed after the test execution. Here is a simple example of a test script that creates an output file and then diffs the file against some expected output. The folder structure looks like this:
test_script/
test_use_dir.sh
use_dir.sh
Here is the content of the test script test_use_dir.sh:
#!/bin/bash echo "TEST" > $TEST_DIR/testfile diff -q use_dir.txt $TEST_DIR/testfile
The script writes its results to a file in $TEST_DIR and compares it against a use_dir.txt file located next to the script. The test can be successfully executed with:
sh.test --cwd test_script/test_use_dir.sh
Please note the --cwd option that is necessary to run the test as the test depends on files located relative to the test script.