Upstream is dead or nonexistent, therefore unusable & unbuildable. It can be explored again when it’s public. Please refrain from these comments. https://github.com/ghostty-org/ghostty
1 Like
I think this package is from chaotic-aur:
error: failed to prepare transaction (could not satisfy dependencies)
:: installing wine-stable (9.0-15) breaks dependency 'wine>=4.0rc1' required by dxvk-mingw-git
:: installing wine-stable (9.0-15) breaks dependency 'wine' required by wine-gecko
:: installing wine-stable (9.0-15) breaks dependency 'wine' required by wine-installer
:: installing wine-stable (9.0-15) breaks dependency 'wine' required by wine-meta
:: installing wine-stable (9.0-15) breaks dependency 'wine' required by wine-mono
:: installing wine-stable (9.0-15) breaks dependency 'wine' required by wine-nine
:: installing wine-stable (9.0-15) breaks dependency 'wine' required by wineasio
:: installing wine-stable (9.0-15) breaks dependency 'wine' required by winetricks
:: installing wine-stable (9.0-15) breaks dependency 'wine' required by yabridge
1 Like
@z3n this should be resolved by the changes shown here:
2 Likes
Can we get updates in the Choatic-AUR for these packages, please:
python-tokenizers: GitHub - huggingface/tokenizers: 💥 Fast State-of-the-Art Tokenizers optimized for Research and Production
python-transformers: GitHub - huggingface/transformers: 🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
python-safetensors: GitHub - huggingface/safetensors: Simple, safe way to store and distribute tensors
We cannot, since these packages do not belong. They would be considered dangling dependencies, as they aren’t required directly by something we already provide. To quote our guidelines listed on the GitHub repository:
Dependencies without any dependents: Such packages are useless by themselves. Maintaining them wastes effort that is better spent elsewhere.
I understand that python-transformers and others take a long time to install because of slowly resolving the dependency trees, it sucks, and you want it to suck less, that’s fair. The most appropriate way to get them is within a virtual environment, though that doesn’t help speed.
Luckily, the speed issue can be mitigated my using something else for dependency resolution. I’ve been having an EXCELLENT experience with UV recently, and would encourage taking a look at it, especially as you get into the large & messy world of AI dependencies. It’s replaced poetry for me in a few places:
In the very very rare case that something isn’t supported by UV yet, it will definitely work via pyenv:
If you are doing something for work, and cannot use third-party dependencies, here’s the documentation for the built-in (official) module for virtual environments:
Also note, when you research virtual environments, virtualenv should no longer be used at all. There’s still many guides recommending it, but they’re old and wrong, don’t let your time be wasted:
3 Likes
Looks like one more bug we have here.
╭─astro@garuda in ~
╰─λ cpu-x
cpu-x: error while loading shared libraries: libcpuid.so.16: cannot open shared object file: No such file or directory
1 Like
Believe it just needs to be updated. I attempted to run it yesterday and it wouldn’t launch, so I removed it and installed the recently updated cpu-x-git from the AUR and it works perfectly. The Chaotic cpu-x-git errored out during attempted install.
3 Likes
Right, for testing I installed cpu-x and point it up.
The chaotic git version works. 
chaotic-aur/cpu-x-git 5.0.4.r77.gd27244f-1 14,76 MiB 4,68 MiB
4 Likes
This maintainer is known to not bump pkgrel when others normally would, and instead recommends people to rebuild per their pinned comment. That means someone from Chaotic-AUR needs to manually intervene and bump for rebuild any time a shared library updates. Please do report the same in the future since this will definitely continue to occur: AUR (en) - cpu-x
I’ve rebuilt it, thanks: Finished making: cpu-x 5.0.4-1.1 (Tue 03 Sep 2024 11:07:15 PM CEST)
3 Likes
Requesting testers for minq-ananicy-git.
If you are already using ananicy-cpp, you may not be a good candidate to do a “fast test”, since you would likely want to swap back.
Apparently there is a bug on launch which causes a crash, but I can’t reproduce it on my machine. Simply installing then starting the service should be enough to test. The related pull request is here: