As AlmaLinux chairperson benny Vasquez explained, the precise goal is "ABI compatibility [which] in our case means working to ensure that applications built to run on RHEL (or RHEL clones) can run without issue on AlmaLinux. Adjusting to this expectation removes our need to ensure that everything we release is an exact copy of the source code that you would get with RHEL."
To do that, AlmaLinux will use the CentOS Stream source code. In return, Vasquez added, "We'll continue to contribute upstream in Fedora and CentOS Stream and to the greater Enterprise Linux ecosystem, just as we have been doing since our inception, and we invite our community to do the same!"
Officially, Red Hat had nothing to say. But, I'm told by Red Hatters that this is exactly "the approach that we've suggested that RHEL-like distributions take - working with the broader community in CentOS Stream."
So, what's the problem? Well, KnownHost CTO and AlmaLinux Infrastructure Team Leader Jonathan Wright recently posted a CentOS Stream fix for CVE-2023-38403, a memory overflow problem in iperf3. Iperf3 is a popular open-source network performance test. This security hole is an important one, but not a huge problem. Still, it's better by far to fix it than let it linger and see it eventually used to crash a server.
That's what I and others felt anyway. But, then, a senior Red Hat software engineer replied, "Thanks for the contribution. At this time, we don't plan to address this in RHEL, but we will keep it open for evaluation based on customer feedback."
AlmaLinux: "Is customer demand really necessary to fix CVEs?"
Red Hat: "We commit to addressing Red Hat defined Critical and Important security issues. Security vulnerabilities with Low or Moderate severity will be addressed on demand when [a] customer or other business requirements exist to do so."
AlmaLinux: "I can even understand that, but why reject the fix when the work is already done and just has to be merged?"
At this point, Mike McGrath, Red Hat's VP of Core Platforms, AKA RHEL, stepped in. He explained, "We should probably create a 'what to expect when you're submitting' doc. Getting the code written is only the first step in what Red Hat does with it. We'd have to make sure there aren't regressions, QA, etc. … So thank you for the contribution, it looks like the Fedora side of it is going well, so it'll end up in RHEL at some point."
One user wrote, "You want customer demand? Here is customer demand. FIX IT, or I will NEVER touch RHEL EVER." While another, snarked, "Red Hat: We're going totally commercial because Alma never pushes fixes upstream! Also, Red Hat: We don't want your fixes, Alma!"
So, the immediate problem has been fixed. Still, bad feelings have been left behind. As Wright wrote, "The worst part of this for me is feeling that I wasted my time by even submitting a PR [Pull Request] here." That's the last reaction you want from developers in an open-source community.
Looking ahead, though, Vasquez is optimistic. In an interview, she said, "This is uncharted territory for all of us, and they appear to be willing to make things better. If we go back to our true goal (improve the ecosystem for everyone), this interaction is a learning opportunity for everyone. They have processes and practices for accepting stuff from the SIGs [CentOS Stream Special Interest Groups] already, but I'm hoping they'll get better about accepting PRs outside of the SIGs."