RPM Lands Support For x86_64 Microarchitecture Feature Levels

[brucethemoose]

It took them only 7 hours to shoot it down, giving a six word long explanation.

Oof.

I love Arch and continue to run CachyOS for lack of a better alternative (as maintaining Nvidia and some other stuff on Clear Linux is too much of a pain), but I've recently been disheartened by the opaqueness of the maintainers when, for instance, trying to figure out the status of/blockers for Python 3.11, or figuring out whats holding up x86-64-v3.

[ms178]

Oof.

I love Arch and continue to run CachyOS for lack of a better alternative (as maintaining Nvidia and some other stuff on Clear Linux is too much of a pain), but I've recently been disheartened by the opaqueness of the maintainers when, for instance, trying to figure out the status of/blockers for Python 3.11, or figuring out whats holding up x86-64-v3.

Yeah, me, too. A constructive way forward would be to communicate roadmaps or obstacles for delaying certain important packages or changes. From what I remember when debating the Arch toolchain debacle, there are some flaws in their infrastructure and procedures that need to be adressed to keep up with the demands of the present. It really shows some bad sides of being a community effort that isn't backed by a corporation or cohesive organizational structure. I fear that they are not up for the task in the long term and slowly but surely will be fading away as a project with other distributions taking its place.

[brucethemoose]

Yeah, me, too. A constructive way forward would be to communicate roadmaps or obstacles for delaying certain important packages or changes. From what I remember when debating the Arch toolchain debacle, there are some flaws in their infrastructure and procedures that need to be adressed to keep up with the demands of the present. It really shows some bad sides of being a community effort that isn't backed by a corporation or cohesive organizational structure. I fear that they are not up for the task in the long term and slowly but surely will be fading away as a project with other distributions taking its place.

On the other hand, Arch seems to be gaining and holding users. The maintainer base declining with a growing userbase seems... problematic.

And here I am willing to test some Python 3.11 and CUDA stuff for the testing repos, but I feel almost shut out from those efforts.

[ms178]

On the other hand, Arch seems to be gaining and holding users. The maintainer base declining with a growing userbase seems... problematic.

Exactly my thoughts as well, that's why I was wondering back in the day why Valve stepped away from Debian/Ubuntu as their base as that was better maintained in a more stable way. I would have hoped that Valve would put more ressources into the Arch project to help them in that transition, but I haven't read anything yet if that is actually happening. Let's wait and see.

[TemplarGR]

On the other hand, Arch seems to be gaining and holding users. The maintainer base declining with a growing userbase seems... problematic.

And here I am willing to test some Python 3.11 and CUDA stuff for the testing repos, but I feel almost shut out from those efforts.

Even if the whole world switches to Arch, it won't make a difference in terms of workload. This is not problematic. What IS problematic, is transparency. They need to become more "professional" about it. This is not a hobbyist pet project anymore, Arch is a serious operating system that is used by many, thus they need to be more responsible. I am pretty sure they have a good work ethic and are good people, but they need to organize the project better. In another thread i was told that many changes regarding "new blood" have happened, this is in the good direction. They need to draw in more people, provide clear recruitment "procedures" for people wishing to join. Many would be willing, i am sure. I am looking forward for my work schedule to lighten a little so i will look into joining myself.

[Svyatko]

It is about time... Basic x64 is almost 20 years old. 20 freaking years.... Modern bleeding edge distros need to upgrade to at least v2. Anyone with that old hardware can keep using older distros, or use distros tailored to basic x64 machines.

openSUSE developers tried to rebase to V2, but had collided with strong opposition.
I'd suggested upgrade to AMD K10-compatible part of V2 https://code.opensuse.org/leap/features/issue/79
x86-64 + {CMPXCHG16B, LAHF-SAHF, POPCNT, SSE3} or just to x86-64 + SSE3.
At first it was accepted because openSUSE decided to use V2, but then it was rejected, because openSUSE decided to use V1.

Discussion about architecture levels: https://en.opensuse.org/openSUSE:X86...tecture-Levels