We want to stay as open as possible during AElfs developement. For this we follow a certain amount rules and guidelines to try and keep the project as accessible as possible. Our project is open source and we publish our code as well as current issues online. It is our responsability to make it as transparent as possible.
AElf is a collaborative project and welcomes outside opinion and requests/discussion for modifications of the code, but since we work in an open environenment all collaborator need to repest a certain standard. We clarify this in the following standard:
collaborative standard: https://github.com/AElfProject/AElf/blob/dev/CODE_OF_CONDUCT.md
We encourage collaborators that want to participate to first read the white paper and the documentations to understand the ideas surrounding AElf. Also a look at our code and architecture and the way current functionality has been implemented. After this if any questions remain, you can open an issues on GitHub stating as clearly as possible what you need to clarify.
Finaly, any collaborator wanting to participate in the developement should open a pull request following our rules. It will be formally reviewed and discussed through GitHub and if validated by core members of AElf, can be merged.
For versionning we use the semver versionning system: https://semver.org
Integrated with github we have cron job that will publish the latest version of devs myget packets.
Testing is one of the most important aspects of software developement. Non tested software is difficult to improve. There are two main types of testing that we perfom: unit testing and performance testing. The unit testing covers functionnality and protocol, which is an essential part of a blockchain system. The performance tests are also very important to show that modifications have not impacted the speed at which our nodes process incoming transactions and blocks.
To ensure the quality of our system and avoid regression, as well as permit safe modifications, we try to cover as much of our functionnality as possible through unit tests. We mostly use the xUnit framework and follow generaly accepted best practices when testing. Our workflow stipulates that for any new functionality, we cover it with tests and make sure other unit tests.
The perfomance testing is crucial to AElf since a strong point of our system is speed.
Server monitoring: zabix monitors instances of aelf metrics like cpu, db...
Chain monitoring: project on github with grafana dashboard from influxdb
Akka monitoring: monitor actors.