FTR I have the mesa maintainers permission to push mesa.
Also I don’t consider closed source steam a reliable indicator of issues, it has no debug symbols.
6 Likes
This is not reasonable.
Linus Torvalds pointed it out years ago https://www.youtube.com/watch?v=7SofmXIYvGM&t=510s
3 Likes
The people who reported Steam issues on Bodhi during testing weren’t told this.
They were thanked for their feedback and asked to take the time to provide further detail, which they did.
2 Likes
Maybe you should vent your anger at the mesa maintainer who thought it was a good idea to add untested mesa merge requests to a stable release.
Gaming literally pushes platforms, and is one of the biggest pushes of Linux on the desktop. This is frankly madness to say. And, gaming is clearly a good indicator when Mesa is broken such as this exact case.
1 Like
Well, if you ask, from my perspective of professionally developing software for many years, it was not reasonable as well 
Hi folks,
Let’s please all take a minute and calm down here. Please do not resort to blaming/screaming—these are not welcome on community platforms since these actions are not in line with the CoC: it’s not “being excellent to each other”.
So, unless you are writing a constructive comment, please refrain from doing so. Bugs will happen, updates with bugs will get through, and while we understand people feeling frustrated, we also expect people to not let their frustrated feelings result in frustrated actions towards community members.
Thanks,
9 Likes
Blaming and screaming are indeed not appropriate, but what’s the means by which we can have a calm, constructive discussion about the lessons learned from this, that leads to better stability in future?
Having worked in corporate tech and having done my share of bad code pushes, what I’m curious about is what happens after issues like this when they happen in Fedora. And this particular issue does seem to lie at the “avoidable” end of the spectrum.
Over the last few months of using Fedora, I’ve seen a number of updates that clearly caused significant disruption to users, but what I’ve not seen in any of those cases is the output of a “lessons learned” process where folks get together to understand what went wrong to cause breakage of production systems that users rely on, and what needs to change to prevent recurrences.
3 Likes
Dave Airlie: fedora 43: bad mesa update oopsie
need to probably give changes like this more karma thresholds
That would probably be a good start. Again, this had multiple reported issues and IMO there’s no way it should have been allowed to be pushed by anyone.
2 Likes
@leigh123linux, perhaps, we can, instead, work to institute a formal evaluation policy, if none exists? I imagine that lessons can be learnt to prevent this, and those lessons can be codified.
Some parts, like Fossilize, appear to be FOSS, so, perhaps RPMFusion could include symbols for it [1] in their new DebugInfoD repository, at least. If that’s useful.
1 Like
Just to note - autopush Karma was disabled for this update so regardless of the Karma it will have always had to have been a manual push to stable.
2 Likes
Oeer, still getting up to speed on the way Fedora works. Any info on why that was allowed for such a critical package??
Please see response #20 - the person that pushed the update provided their reasoning within this thread.
2 Likes
Then with all due respect the system is not working. Yes, issues happen. But this highlights room for needed improvement. It shouldn’t have been possible for one person to just push an update that had various reports of it being broken. I can’t be crazy in thinking that right?
4 Likes
@leigh123linux , you know that there’s no need for that. You are an important member of the community, and we are extremely grateful for the time and effort you put in with your Fedora and RPM Fusion packages.
@gtb : can we not go down the road where when a mistake is made, “someone has to pay”. This is really not how the community works.
Everyone here: there was an issue, it was very quickly fixed. it’s time to move on. Please stop trying to lay blame, that’s NOT how we work.
The lesson to learn here is that bugs will get through our systems for a variety of reasons (human and tech)—whatever systems are in place because it is impossible to have 100% coverage—but as soon as they are identified, a fix will also be released. Of course, we will all be more careful, as we always try to be.
9 Likes
No need for any blood or name calling or anything sure. But brushing it under the rug and trying to shut down conversation on getting improvements doesn’t help anything either.
3 Likes
@liamdawe, perhaps, better yet, instead of philosophical generalisations, we solely comment if we’ve something specific, helpful, and actionable. (Remember that the moderators must be more generalised in their criticisms, in order to not single-out people unnecessarily, so I’m not commenting on @ankursinha’s reasonable comment.)
1 Like
Of course it’s true that bugs happen and we won’t 100% prevent them, but we also shouldn’t fall into a defeatist mindset where we believe nothing can be done to reduce the current rate of breakage.
For me, this along with a number of other recent issues (especially around the F43 release) strengthens my view that Fedora needs to give greater weight to production stability, probably by putting a better defined process around it.
For example, as a sketch of an action plan - and it’s just a sketch because I’m not organisationally familiar enough with Fedora to know what would best achieve these aims:
- Establish a Production Stability SIG
- Formalise incident management processes which set out:
- What constitutes a significant incident in production systems
- How an incident is tracked, managed and communicated
- How an incident is followed up to learn lessons and generate continuous improvement
4 Likes
That sounds like a great idea!. Would you please start a new discussion to set up the SIG, which hopefully you can lead/organise?
We’re not brushing anything under the rug or shutting down the conversation. This conversation is happening on open channels, for a start, and we haven’t closed the topic or even put it in “slow mode” yet.
What we’re saying is: say what you have to say while being excellent to each other. This is required by our Code of Conduct that everyone agrees to before they participate in the community. So, all we’re doing is reminding people of this.
4 Likes