- cross-posted to:
- [email protected]
- cross-posted to:
- [email protected]
You must log in or register to comment.
32bit does need to go but we’re at a time where a lot of people are switching their main PCs or gaming PCs over to Fedora to get ahead of Windows 10 EoS. The timing in this change couldn’t be worse (even if it’s two versions ahead.)
It’s bad PR to break Steam and gaming at this time. Valve needs to sort this out on their end but the Fedora Project needs to check in with their users to see what they’re using on Fedora.
Also loved the gaslighting at the end. Very Linux dev.
Steam is kind of junk if you look into how it actually works
Value needs to get its stuff together
Go on
It is 32bit and depends on a runtime from a really old version of Ubuntu. It is also proprietary but that is a different issue.
It would be nice if Value worked with the Fedora project to build something open and modern so that devs could easily package games for Linux. The problem is that Value wants everyone to go though Steam as Steam is where they actually make money. People like to see Value as the good guy but it isn’t all sunshine and rainbows. I can’t really blame them in the end but it is important to realize they have a interest that may not always align with the community.
Tyty
I fully agree with what you wrote. Someone email Gabe!!!
Get Gabe and Linus on the discord chat so that they can hash this out right quick!
What gaslight?
Containerizing the 2038 problem won’t fix 32 bit issues that need recompilations.
Have you set your devices to 2038 and reported the issues your distro needs fixed?
I’m using multilib for gaming compatibility, I’m not saying we need full _time32 support.
If I pull 32bit libs for specific applications, why would that affect the kernel or the user spaces apps that are all 64 bit and using 64 bit function calls??
2038 will break games that make specific calls to 32bit time functions, sure. Thats gonna suck in 12 years from now. But what if instead of breaking compatibility for gaming now, we work towards a solution in the next 12 years?
It’s the other way around: your _time32 libs will fail, especially user apps that call Fedora{Linux _time64. The kernel as of right now is safe for year 2038.
You have a decade to test everything, report, and fix. Recompilation projects for everything. I want to see year2038_bazzite_safedb.org with all the games and apps you use marked ☑ 2039+ compliant. You need to start now.
You should have started on 2004. So time to catch up! So yes, let’s database the solutions an app at a time, before the decade is up!
Wow that was fast, but as the developer said himself, clickbaity YouTubers and news sites will have a co responsibility for that. The drop of 32bit us inevitable, and in my opinion it would be smarter to think about how to do this the good way better earlier than later. But probably these kind of discussions have to be moved from public to internal I guess.
No, that’s how you lose users. Private, nontransparent decision taking makes projects get dropped immediately.
The timing just sucked. 32 bit has to go, but it can’t be this year or next year. And it can’t be a blanket drop as the dev wanted. Alternatives are not ready yet to keep gaming working, and gaming was the number 1 factor holding back desktop adoption.
He is also falling for the internet fundamental attribution error: “If I hate or love something and everybody on the internet agrees with me, it’s because I’m always right and we are all intelligent individuals. If I love/hate something, and everybody on the internet disagrees, they were lied to, manipulated, astroturfed, are ignorant, misinformed, etc.”
It could be true. But it could also be that your proposal is very unpopular and you’re wrong.
Lol, how many people do you think are using fedora specifically for gaming? You think fedora is some last year project that’ll die off once it stops supporting this specific “thing”? That’s some delusion of grandeur.
May be Steam and gamers who want to keep things running without problems should step up to maintain those packages required to keep their games running. Then we’ll see how fast the toxicity towards the developers die down.
Why do you think “alternatives are not ready”? Is it maybe because people who work on opensource project on their free time aren’t bothered by what internet strangers think? Why aren’t you working on one yourself?
People shitting on opensource devs to work beyond their comfortability are fucking shitty human beings.
If you plan it 5 years in advance, yeah, otherwise no.
Also if it means destroying Bazzite, they should really think hard about it, as it’s a great “product” with big potential, even for profit in the long run if it takes off.We’ve known about problem 2038 since 2003. You’ve had decades to fix it, and one more. Get to it!
How much 32bit only hardware is still out there in the wild and not still running on windows XP?
Sounds like a problem for a purpose built distro, not a mainline one.
EDIT: I stand corrected. I thought it was hardware 32bit support being discussed. It’s premature to discontinue 32-bit libraries.
The support isn’t exclusivity for native 32 bit cpus, it’s for 32bit libraries that compatibility applications like wine/proton depend on to run 32bit windows executables
Thank you for the correction.
Even further: The support is exclusively for the 32bit libraries. The 32bit kernel and therefore cpu support was dropped a long time ago in Fedora. Fedora 31 in 2019.
Fedora already won’t run on 32bits hardware. That is not the issue being discussed here.
I stand corrected. Thank you.
Yet Apple somehow got away with removing 32-bit apps and changing architectures multiple times.
And they broke A LOT of software along the way.
I don’t think Apple is concerned with making every app and game work on their systems to the same degree as a Linux distro. They have a niche they seem satisfied with and that niche isn’t really Steam games.
With that being said, Valve made a 64 bit client for Mac so whichever major distro is first will probably push Valve to finally make a 64 bit version for Linux.
Is dropping support for 32bit hardware more important than being able to run on everything?
Because it has always seemed like one of Linux’s core strengths is that no matter what your hardware is, you can run Linux on it.
Being able to run on everything is nice, but what you’re actually asking for here is for a highly skilled (and rare) distro maintainer to dedicate substantial amounts of time to maintaining and testing thousands of packages on 32 bit systems in case someone wants to use it. It’s not just like users get to click a button and voila, they can run more games than before.
If you really need to use 32 bit software, eventually you’ll just have to manage it yourself because no one else is able to do it for you. That capability will almost never go away.
There are and will always be distros optimized for running on everything. Fedora is a “move fast” distro, it’s hard to move fast with a lot of baggage.
Run fast is fine, until it’s run away fast…
Is dropping support for 32bit hardware more important than being able to run on everything?
Yes evidently, because they dropped that hardware support in 2019. Specifically they dropped 32-bit x86 kernels in Fedora 31
To be a little more precise, Linux is still available for 32-bit x86, just not from the Fedora distro. The Linux project is just now dropping support for 486 CPUs, because the maintenance burden for a virtually unused system type is too high for the mainline. That still leaves 32-bit Pentiums and newer though.
Fedora is usually the first to pioneer something
I think the last time I had a 32 bit CPU was around 2005 but I could be remembering that incorrectly. Supporting 20 year old hardware isn’t always easy.
It isn’t easy, but this isn’t about the hardware. It’s about the software packages. Tons of software meant to run on 32-bit hasn’t been updated to run on 64-bit natively. Thus the burden of keeping a lot of packages that serve as backwards compatibility.
Tons of software meant to run on 32-bit hasn’t been updated to run on 64-bit natively.
32bit only Linux apps are basically non-existent, anything with the source available and maintainers would have been ported at some point in the last 2 decades, otherwise they have very specific technical reasons for being 32bit only (like OBS iiuc), the source has been lost somehow, or it’s a proprietary program where the company has no interest (e.g. Valve with Steam)
In fact I think Steam might really be it.
True, yeah I read that too. Started as a hardware thing but now it’s a “this is the state of things as a result of things that were hard to change” thing.
Well, that makes Fedora a distro I can marginally trust for now.
at least for another 6 months…
So we’re just gonna keep chugging along until 2038? Please.
No. Valve (the biggest offender) will have to make native 64-bit Steam before then, as will the remaining holdouts, so Linux distros will be able to remove 32-bit packages in a timely manner.
Removing then now will break too much to be worth doing.
I don’t think Fedora users care about breakages ngl.
I am a fedora user. I care. Why wouldn’t I? …
Same. If most of my games stopped working, I would be very annoyed, especially because it was entirely preventable.
Thankfully, the Fedora project and community agree.
Yeah. I’m not saying the process went perfectly, but I think it’s good they proposed this and then nixed it. Gotta do it someday though.
Because Fedora is a testing distro which gets rid of legacy stuff the first. I wouldn’t recommend it if you care tbh.
I mean, I have more than one machine. Some can be closer to guinea pigs than others. In this case, it’s a laptop that I don’t keep anything unbacked-up on. Had Fedora on it for about 6 months and I cannot remember an update breaking anything for me so far. The previous machine I had it on was used less but I had the same experience. If you’re mainly just web browsing on a machine, bleeding edge is good imo
Don’t break userspace. 32-bit support should never be removed.
Easy thing to say when you’re not the one donating your time for free.
I love what Valve is doing for Linux, but longer term, the onus is on them to solve the 32-bit compatibility layer issue (a-la Proton for win-to-linux, as well as their upcoming x86-to-ARM layer).
Expecting all distros (who again, are staffed mostly by volunteers) to do this work separately (i.e. duplicating all that work), for all time, is a big ask.
Don’t break userspace.
That’s a kernel saying. A bit unfitting to repeat it for the distro that builds said userspace.
By that logic we should never remove anything legacy ever. It don’t work like that.
Userspace is allowed to break themselves
Isn’t the context of that quote around the kernel and kernel space vs user space? I don’t see how that thought really extends to distros that simply implement the kernel as one of their packages.
