coveralls-public VS thinkdeep

cpan_coverage: This calculates the coverage of your test suite and reports the results. It also uploads the results to coveralls.io

I will normally use GitHub Actions to automatically run my test suite on each push, on every major version of Perl I support. One of the test runs will load Devel::Cover and use it to upload test coverage data to Codecov and Coveralls.

Several years ago I got into Travis CI and set up lots of my GitHub repos so they automatically ran the tests each time I committed to the repo. Later on, I also worked out how to tie those test runs into Coveralls.io so I got pretty graphs of how my test coverage was looking. I gave a talk about what I had done.

This approach will create two json coverage files, which will be merged together by NYC. Therefore the results will be purely local. If You don't mind using online tools like Codecov or Coveralls for merging data from different tests, then go ahead and use them. They will probably also be more accurate. But if You still want to learn how to get coverage from E2E, then please read through

Ideally, software can quickly go from development to production. Continuous deployment and delivery are some processes that make this possible. Continuous deployment means establishing an automated pipeline from development to production while continuous delivery means maintaining the main branch in a deployable state so that a deployment can be requested at any time. Predecos uses these tools. When a commit goes into master, the code is pushed directly to the public environment. Deployment also occurs when a push is made to a development branch enabling local/e2e testing before push to master. In this manner the master branch can be kept clean and ready for deployment most of the time. Problems that surface resulting from changes are visible before reaching master. Additional automated tools are used. Docker images are built for each microservice on commit to a development or master branch, a static code analysis is performed by SonarCloud revealing quality and security problems, Snyk provides vulnerability analysis and CodeClimate provides feedback on code quality while Coveralls provides test coverage. Finally, a CircleCI build is done. Each of these components use badges which give a heads-up display of the health of the system being developed. Incorporating each of these tools into the development process will keep the code on a trajectory of stability. For example, eliminating code smells, security vulnerabilities, and broken tests before merging a pull-request (PR) into master. Using Husky on development machines to ensure that code is well linted and locally tested before it is allowed to be pushed to source-control management (SCM). Applying additional processes such as writing tests around bugs meaning reintroduction of a given bug would cause a test to fail. The automated tools would then require that test to be fixed before push to SCM meaning fewer bugs will be reintroduced. Proper development processes and automation have a strong synergy.

There is https://coveralls.io/ and https://github.com/marketplace/codecov , but they are both priced for commercial usage. Do you know some free alternatives or approaches to have something similiar?

Add a Code Coverage CI step using Coveralls.io Add Dependency monitoring using Snyk

Since there is no need to reinvent the wheel, I will take advantage of an existing github action in the Continuous integration workflows category: Node.js. With this action I will set up this action in one of my public repositories. I will set up Node.js action for automating my unit test and also integrate with coveralls.io for getting a badge of how much my tests covers relevant lines.

Looks like https://github.com/lemurheavy/coveralls-public/issues/632 this issue based on the log. Try going through their solutions, maybe?

A complex codebase needs to be kept as simple as possible. In Predecos, one way this is done is through use of full-stack JavaScript, meaning developers only need experience with one language to maintain the codebase. Some differences exist in NodeJS when compared with browser-based JavaScript but those are minor compared with having to work in multiple languages. Front-end and back-end dependencies are also kept as uniform as possible. For example, Mocha and Chai are used in both the front-end and back-end meaning experience with said tools enables development in either. Developer setup and deployment should also be straightforward. Predecos uses Helm v3.8.2 to trigger one-click deployment of all the microservices. After doing this and setting a couple environment variables, one can easily start the front-end by using yarn run start then they are ready for development. This is also valuable when doing a deployment. When CircleCI went down, the impact was slight because automated deployment and manual deployment were very similar. All that was required was modifying the CircleCI config for the terminal, building the necessary docker images manually, pushing those images and running helm upgrade. A simple and fast mitigation that saved a lot of time. Diverging from simplicity was sometimes necessary though. Some may argue that use of Kubernetes makes the codebase difficult to understand by virtue of its high learning curve. However, the stability, reliability, documentation and wide-spread adoption of Kubernetes made it hard to resist. Google Cloud Platform (GCP), Amazon Web Services (AWS), Microsoft Azure and DigitalOcean all have Kubernetes offerings. This means applications developed on Kubernetes are largely portable from one cloud provider to another. Predecos uses DigitalOcean’s managed Kubernetes product because it is low-cost when compared to the other options. Should the desire for features such as, for example, encryption at rest arise, one option would be to migrate to Google Kubernetes Engine (GKE). Some modifications would be necessary but because Predecos is in a proof-of-concept (PoC) state and minimal data has been stored that migration would likely be simple.