I am currently looking for a way to execute visual regression tests with storybook, without having to use a service like Browserstack or SauceLabs. This docker-image seems to be exactly the thing that I am looking for, except for one little drawback:
When I run this image as "service" in Gitlab-CI, I cannot access the server on the build-container. The only feasible solution seems to be a reverse tunnel, similar to what Saucelabs does: If there was a running sshd in the image, the test-runner could do ssh -R 8080:127.0.0.1:8080 docker-selenium before starting the tests, and then tell the browser to load localhost:8080.
Setting up ssh in the docker-container, running as seluser was not so easy, so in my PoC, I used chisel to create the tunnel.
Would you be interested in adding this to the docker-image? Do you have another idea how to achieve my goal? Or should I rather continue extending my PoC to an image that can be used in production?
I'd rather have a way of doing this without building and maintaining my own image.
I am currently looking for a way to execute visual regression tests with storybook, without having to use a service like Browserstack or SauceLabs. This docker-image seems to be exactly the thing that I am looking for, except for one little drawback:
When I run this image as "service" in Gitlab-CI, I cannot access the server on the build-container. The only feasible solution seems to be a reverse tunnel, similar to what Saucelabs does: If there was a running sshd in the image, the test-runner could do
ssh -R 8080:127.0.0.1:8080 docker-seleniumbefore starting the tests, and then tell the browser to loadlocalhost:8080.Setting up ssh in the docker-container, running as seluser was not so easy, so in my PoC, I used chisel to create the tunnel.
Would you be interested in adding this to the docker-image? Do you have another idea how to achieve my goal? Or should I rather continue extending my PoC to an image that can be used in production?
I'd rather have a way of doing this without building and maintaining my own image.