Session: Upstream code quality: Making open source work out of the box
Many open source projects are used as building blocks in products. Full software stacks for commercial products using them tend to be heavily tested. They typically include downstream changes and fixes. Ideally, those eventually get sent upstream although this can easily become a second priority compared to releasing products.
This talk will look into the implications of enabling automated testing upstream: What are the benefits? How can it be done? Which projects are already doing it?
A strong example is the Linux kernel, with all its various applications from tiny embedded devices to smartphones, laptops, servers and super-computers. Many companies run kernel tests all the time, but the mainline or stable releases do not depend on any particular mandatory tests to pass. KernelCI is an attempt at solving this problem by gathering results from product companies and running dedicated tests upstream. This will save time downstream, in the same way that sharing code avoids duplication of efforts. The same principles can be applied to any open source project, large or small, with some appropriate adjustments. What about the open source projects you care about?