The main goal of the repo is to show how to marry Mocha with Docker in order to make your end-to-end tests even greater. There are quite a few benefits of test dockerization. It:
- reduces test execution time, which is especially visible in scale.
- elimates one of the stages - the building - which makes the tests more reliable in general
- gives more control over the running environment which increases even more the reliability of the whole thing
- lets your colleagues run the tests on their local machines, no matter what OS they're using or in what tech they're working on in daily basis
While developing high-level tests, like for UI, it's pretty often desired to run only a subset of tests or to inject a variable, like environment under test, while running the scenarios. In general, mocha allows to filter tests in two ways: based on path and by --grep parameter.
Therefore, for presentation purposes, I've decided to create repo that contains tests of more than one suite. By suite I mean a directory inside /tests, and there are three of them: end-to-end, foo, bar. Let's say that both foo and bar contain tests of some service, and I believe that end-to-end is self-explanatory.
Also, I've put it() blocks inside nested describe()s, which only include tags like @smoke or @regression. This way we can use the grep feature to cherry pick tests.
In result, we gained ability of running only those tests that we want.
mocha ./tests/end-to-end/*.test.js --grep "@smoke"
and we can still easily set env variables:
ENV="staging" mocha ./tests/end-to-end/*.test.js --grep "@smoke"
To make it work with Docker I've decided to go with a bash script, which is an entrypoint for Docker image. Nothing fancy there, but it gives us a lot of space for future development & improvements. To make my life easier, I defined a script for each suite in package.json
"scripts": {
"test": "mocha ./tests/**/*.test.js",
"test:e2e": "mocha ./tests/end-to-end/*.test.js",
"test:bar": "mocha ./tests/bar/*.test.js",
"test:foo": "mocha ./tests/foo/*.test.js"
},
Mainly, I did it because of two reasons:
- we need to run the tests without docker as well, so having something like
test:e2eis simply useful - I wanted to reduce complexity of bash script
This way, to make it work inside Docker container we must pass two or three variables into docker run commend:
ENV="$ENV" npm run test:"${DOMAIN}"
ENV="$ENV" npm run test:"${DOMAIN}" -- -g "$EXPRESSION"
which can be done like this:
docker run -e ENV="staging" -e DOMAIN="foo" -e EXPRESSION="@smoke" mocha-with-docker:latest
To build an image, execute:
docker build -t mocha-with-docker .
Depending on the needs, you'll be probably required to install the necessary library dependecies into Dockerfile. It's always a good idea to search for 'how to' manuals in the official documentation of the particular tool. Here are some of them:
To copy test reports from non-running container you can use docker cp command, like this:
docker cp ${DOCKER_CONTAINER_NAME}:/usr/src/app/${REPOSITORY_NAME}/${TEST_OUTPUT_FILE} ./