But it was closed without any real discussion. I’d be happy to ACTUALLY discuss the issue I’m trying to resolve here.
FreeBSD compatibility is broken is many places within CheckMK’s agent code right now. PRs for fixes either languish for months on end, or apparently get closed without the ability to discuss WHY a change is being proposed.
Okay, sure, I get it, you want “consistency” in the code base. Problem is, that already doesn’t exist. So let’s set a universal, COMPATIBLE standard for shell scripts that work on multiple OSes, adopt that, and use it. The suggested change in the PR is exactly that. Was the PR for the ENTIRETY of the code base to change in one giant leap? No, but it was a start. It was something I’m personally using and validated. But if you want more files updated w/ the same? Okay, cool, let’s have that discussion and do it rather than just closing out PRs without the actual chance for discussion.
Otherwise, this just looks bad for CheckMK, denying non-Linux OSes the ability to have functional code.
I have checked with my colleagues about the pull request. We understand the importance of making Checkmk compatible with FreeBSD, but redoing the system to support this would be a complex change that requires a lot of careful planning and research. For now, we cannot commit to this.
We really appreciate your feedback, as it helps us understand what is important to our users. The best way for us to consider this change for a future release is for you to add it to our Ideas portal. Other users can then vote on it, which shows our product team the level of community interest.
In the meantime, many of our users are already running Checkmk on FreeBSD successfully. If you have any specific issues, you could share them on the forum, where someone from the community or the team might be able to help you.
The PR is literally a one-line change to the shebang to make it more compatible with more OSes, not just FreeBSD, but also any other OS that has a different path to where BASH is stored.
Its one line. That’s it. Very small change.
I don’t need help finding a solution. I found it, and gave it in a PR. It was that simple. A single line change.
The only push back is “consistency”, which already doesn’t exist, as pointed out with examples of existing inconsistency in the PR.
So my point here was that if inconsistency already exists, the that shouldn’t be the one and only blocker of the PR. Instead, agree upon a compatible solution that helps the most ecosystems out (ex: the one listed in the PR) and start rolling it out. It doesn’t need to be one massive PR that updates literally every single instance in the code base all at once (that’s never how things are done in CheckMK based on other inconsistent updates I’ve followed w/ agents), so roll it out slowly a section or file at a time. That’s perfectly fine, and I’m more than willing to help in that process.
But just entirely shutting me and an entire section of the community out with a “wont fix” like I said isn’t a good look for CheckMK.
Yes, there are plenty of us using CheckMK on FreeBSD, but look at the other PRs. We’re all running out-of-tree modifications due to the lack of willingness to fix the issues we’re running into. They’re all small and trivial changes, but the PRs either get flat out rejected, or ignored for months upon months at a time. I’m just trying to figure out “why”