Our team has been struggling with a good release process for our browser extensions.
With Chrome Extensions, every build needs to be reviewed by the chrome team. This can sometimes take 5-6 days and has never taken less than 24 hours. If a version is published, there is no way to revert to an old version in the event of bugs/issues that make it through the testing cycle.<p>True CI/CD is out of the question since we have no way to automatically release. Additionally, because of the extension architecture, automated testing is limited almost exclusively to unit level testing the client code.<p>How has your team overcome this? Any advice welcome here.
You just described the situation mobile app developers face every day.<p>1/ Have slower thoughtful release cycles
2/ Construct a test release plan and hire a QA to perform the test every release.
3/ Feature flag releases (have the device check an api to see what is actually allowed to be enabled)<p>> True CI/CD is out of the question since we have no way to automatically release.<p>CircleCI thinks you can[0]? Google has apis for everything.<p>[0] - <a href="https://circleci.com/blog/continuously-deploy-a-chrome-extension/" rel="nofollow">https://circleci.com/blog/continuously-deploy-a-chrome-exten...</a>
Is possible to test a chrome extension with the backend via Sikuli GUI automation.<p>Ended up getting rid of it while making major changes though since the tests are flakey and hard to maintain.