arch4edu / arch4edu Goto Github PK
View Code? Open in Web Editor NEWArch Linux Repository for Education
Home Page: https://arch4edu.org
Arch Linux Repository for Education
Home Page: https://arch4edu.org
From 2020-01-01, we will start to drop all the python 2 related packages.
Hi,
mxnet-cuda is compiled with cuda 10.1 while the official repos have already updated to cuda 10.2. Can you rebuild the package?
Thanks for your awesome work here!
After I added the keybase.pub mirror to my /etc/pacman.conf following these steps, I am getting "error: could not open file /var/lib/pacman/sync/arch4edu.db: Unrecognized archive format" when running pacman:
$ sudo pacman -Syyu
:: Synchronizing package databases...
core 133.4 KiB 494K/s 00:00 [###] 100%
extra 1646.2 KiB 5.36M/s 00:00 [###] 100%
community 4.9 MiB 8.22M/s 00:01 [###] 100%
arch4edu 7.3 KiB 0.00B/s 00:00 [###] 100%
:: Starting full system upgrade...
error: could not open file /var/lib/pacman/sync/arch4edu.db: Unrecognized archive format
This is the section of my /etc/pacman.conf
[arch4edu]
Server = https://keybase.pub/arch4edu/$arch
And I am not getting the error if I use the mirror https://mirrors.tuna.tsinghua.edu.cn/arch4edu/$arch
Is there a problem with the keybase mirror or did I do something wrong? Thanks!
最新archlinux更新把libhdf5.so.100升级成了libhdf5.so.101,openmpi里面libmpi.so.12升级成了libmpi.so.20,导致这两个包对libhdf5.so.100和libmpi.so.12的依赖被破坏。bot端能否重新编译一下?
以后依赖于openmpi的包要格外留意,因为升级比较频繁,而很多软件包是直接链到libmpi.so.XX而不是libmpi.so上的。
r-mkl-3.6.1-1-x86_64.pkg.tar.xz currently on the repository was built with mkl 2019.4.243. However, when installing r-mkl with pacman, it will install intel-mkl-2019.5.281 and intel-common-libs-2019.5.281 as dependencies because they are the latest version. This causes installing certain R packages to fail since options in /etc/R/Makeconf still point to /opt/intel/compilers_and_libraries_2019.4.243/linux/mkl/include, which doesn't exist.
Can you please rebuild r-mkl with the latest intel-mkl and intel-common-libs to fix this?
Thanks!
SigLevel = Never
is always considered insecure, please add a package arch4edu-keyring
just as the one in archlinuxcn
mxnet 1.4.0-1.1, cuda 10.1.105-6. ldd /usr/lib/python3.7/site-packages/mxnet/libmxnet.so|grep found
gives me:
libcudart.so.10.0 => not found
libcublas.so.10.0 => not found
libcurand.so.10.0 => not found
libcusolver.so.10.0 => not found
libcufft.so.10.0 => not found
libnvrtc.so.10.0 => not found
It seems that cuda is updated, and mxnet needs to be rebuilt.
I think openblas-lapack (and possibly many other packages) need to be rebuilt using GCC8 since Arch switched, libgfortran was updated to libgfortran.so.5. These packages depend on the older libgfortran.so.4 that was in GCC7.
I think r-mkl needs recompiling, following the update of package icu to v65 on Nov. 14th, 2019. Otherwise, r-mkl does not find libicu.so, because it's looking for the previous version.
I think loadcaffe needs to be rebuilt against the new protobuf:
luajit: /usr/share/lua/5.1/trepl/init.lua:389: /usr/share/lua/5.1/loadcaffe/ffi.lua:10: libprotobuf.so.15: cannot open shared object file: No such file or directory
stack traceback:
[C]: in function 'error'
/usr/share/lua/5.1/trepl/init.lua:389: in function 'require'
neural_style.lua:6: in main chunk
[C]: in function 'dofile'
/usr/bin/th:150: in main chunk
[C]: at 0x00404680
root 是欧洲核子中心(CERN)开发的基于LGPL2.1协议的数据分析软件。广泛应用于高能物理、加速器物理、同步辐射实验或仿真的数据处理。
AUR中将root拆分成root 和root-extra两个包。由于编译量有点大,依赖有点多。。。。如果方便的话,请收录这两个包。
万分感谢。
PS. ROOT-Mirror on github and ROOT Homepage.
opencv is upgrade, and libopencv_gapi.so.4.0
is not found, libopencv_gapi.so.4.1
is availabel.
I wonder why an AUR helper is included, according to the wiki:
- Warning: AUR helpers are not supported by Arch Linux.
- Since AUR helpers are unsupported, they are not present in the official repositories.
I know that this is not an official (yet) repository but this is almost a similar case.
brackets 找来的
从 intel-compiler-base=2019.2.187-1
升级到 intel-compiler-base=2019.3.199-3
发现新的包依赖 intel-common-libs=2019.3.199
intel-openmp=2019.3.199
而他们在旧版本中并不存在,在源目录 (tuna mirror arch4edu) 中也没有找到。貌似是 lilac 的 auto update 85034d1 引起的。谢谢!
hdfview及hdfview-beta 可用于查看高性能科学计算hdf输出文件。由于网络问题经常编译失败。故请求收录。
AUR分别为hdfview和hdfview-beta。
万分感谢arch4edu。
Hello,
First off thank you for hosting these packages.
I would like to request a package and it's dependencies for this repository.
ros-melodic-desktop-full
is the latest version of the Robot Operating System. (https://www.ros.org/) It is very prominent in the robotics community and it would definitely be a nice addition to have for this repository. There are a couple packages like gazebo
, ogre-1.9
, and pcl
which take forever to compile and installers of ROS-melodic on Arch, all feel the pain. As such having binaries for these PKGBUILDs would be lovely.
Thank you
请检查报错邮件
这个配置文件里我配置了镜像,于是安装 nvidia-docker 的时候它就冲突了。另外它没有写在 backup 里,是不是意味着我卸载 nvidia-docker 的时候我自己的配置也没了?
如果 docker 无法把这个配置分文件处理的话,建议使用安装脚本来处理这个配置文件的更改。
以下几个包的pkgbuild更新了,但新的binary一直没有发布:
请检查一下编译状态?
mxnet 1.4.0 was released, see https://github.com/apache/incubator-mxnet/releases/tag/1.4.0.
I followed the instructions and got the error when I run pacman -Syy. How can i solve?, Sorry if here isn't the right place for the poblem.
感谢你们提供的方便的二进制源!可以节省不少编译时间。
我在使用你们的paraview包。paraview包依赖protobuf,protobuf包更新较为频繁,但paraview包的更新经常几个月不更新,这就导致更名后的动态连接库无法被正常使用:
$ paraview
/usr/lib/paraview-5.4/paraview: error while loading shared libraries: libprotobuf.so.14: cannot open shared object file: No such file or directory
我检查了一下你们的PKGBUILD,与AUR一致。应该只要重新编译就行。
为了永久的解决这个问题,建议你们做一个脚本,定时自动检查官方仓库更新,并对依赖有更新的包重新编译发布。
These two package can help to creating 3D graphics by python, it often be used as visualization in scientific computing by python.
Intel(R) Math Kernel Library for Deep Neural Networks (Intel(R) MKL-DNN) is an open-source performance library for deep-learning applications. The library accelerates deep-learning applications and frameworks on Intel architecture. Intel MKL-DNN contains vectorized and threaded building blocks that you can use to implement deep neural networks (DNN) with C and C++ interfaces.
Given that most modern Intel processors support MKL acceleration, and python-numpy-mkl
and python-scipy-mkl
already exist in arch4edu, it should be reasonable to add 'mkl-dnn' as well.
'mkl-dnn':
https://github.com/intel/mkl-dnn
https://aur.archlinux.org/packages/mkl-dnn
mkl-dnn 打包的时候使用了 mklml,而 mklml 是 mkl 的一个子集,里面包含了 libiomp5.so
,也就是造成 openmp
文件冲突的源头.我看 mkl-dnn 并没有说一定要在链接的时候使用 mklml 的,可以直接用 gcc 或者 openmp 的 OpenMP 实现.这样就不会有文件冲突了.
>>> import mxnet
Traceback (most recent call last):
File "<input>", line 1, in <module>
import mxnet
File "/usr/lib/python3.7/site-packages/mxnet/__init__.py", line 24, in <module
>
from .context import Context, current_context, cpu, gpu, cpu_pinned
File "/usr/lib/python3.7/site-packages/mxnet/context.py", line 24, in <module>
from .base import classproperty, with_metaclass, _MXClassPropertyMetaClass
File "/usr/lib/python3.7/site-packages/mxnet/base.py", line 213, in <module>
_LIB = _load_lib()
File "/usr/lib/python3.7/site-packages/mxnet/base.py", line 204, in _load_lib
lib = ctypes.CDLL(lib_path[0], ctypes.RTLD_LOCAL)
File "/usr/lib/python3.7/ctypes/__init__.py", line 356, in __init__
self._handle = _dlopen(self._name, mode)
OSError: libopenblas.so.0: 无法打开共享对象文件: 没有那个文件或目录
It works if I install openblas-lapack
from arch4edu, but not openblas
from community repo.
请问能包含scilab 的beta版本: Scilab 6.0.0 beta 2 吗? 可以将其命名为scilab-beta,与scilab5.5.2区分。
>>> sudo pacman -S nvidia-docker
[sudo] lilydjwg 的密码:
正在解析依赖关系...
正在查找软件包冲突...
错误:检测到未解决的软件包冲突
错误:无法准备事务处理 (有冲突的依赖关系)
:: nvidia-container-toolkit 与 nvidia-container-runtime-hook 有冲突
具体的容器可以用 docker,也可以用 LXC/LXD.
警告:无法解决 "llvm-libs<8.1.0","python-llvmlite" 的一个依赖关系
警告:无法解决 "python-llvmlite>=0.28.0","python-numba" 的一个依赖关系
警告:无法解决 "python-numba","python-librosa" 的一个依赖关系
警告:无法解决 "llvm-libs<8.1.0","python-llvmlite" 的一个依赖关系
警告:无法解决 "python-llvmlite>=0.28.0","python-numba" 的一个依赖关系
警告:无法解决 "python-numba","python-resampy" 的一个依赖关系
警告:无法解决 "python-resampy","python-librosa" 的一个依赖关系
:: 因为无法解决依赖关系,以下软件包无法进行更新:
python-librosa
想打包 horovod 才发现,mxnet 不包含头文件.官方的 wheel 是有的.
LAMMPS, or the Large-scale Atomic/Molecular Massively Parallel Simulator is a software package that lets you perform molecular dynamics simulations, and various other related analyses.
URL: https://lammps.sandia.gov/
I currently maintain its stable, beta and git packges in the aur. It would be nice if you folks could build and host the binaries of the stable version:
https://aur.archlinux.org/packages/lammps/
This is the beta version, in case you want to take a look:
https://aur.archlinux.org/packages/lammps-beta/
在编译openfoam3.0-git的时候,由于foamyquadmesh,foamyhexmesh依赖于libboost_thread.so而编译文件内没有指定,可能会编译出错。虽然软件包能正确生成,但是依赖于boost的部分可能会编译失败。可以在log里面找一下Error。
用这个patch可以修复这个错误。另外,openfoam2.4-git可能也会有这个问题,可以在log里面找一下然后修改相关options文件修复。
--- src/OpenFOAM-3.0.x/applications/utilities/mesh/generation/foamyMesh/foamyHexMesh/Make/options 2016-04-23 10:46:06.634726444 -0400
+++ src/OpenFOAM-3.0.x/applications/utilities/mesh/generation/foamyMesh/foamyHexMesh/Make/options 2016-04-23 10:50:35.496222009 -0400
@@ -35,4 +35,5 @@ EXE_LIBS = \
-lfileFormats \
-ltriSurface \
-ldynamicMesh \
- -lsampling
+ -lsampling \
+ -lboost_thread
--- src/OpenFOAM-3.0.x/applications/utilities/mesh/generation/foamyMesh/foamyQuadMesh/Make/options 2016-04-23 10:46:06.634726444 -0400
+++ src/OpenFOAM-3.0.x/applications/utilities/mesh/generation/foamyMesh/foamyQuadMesh/Make/options 2016-04-23 10:48:25.558854579 -0400
@@ -38,4 +38,5 @@ EXE_LIBS = \
-ldecompositionMethods \
-L$(FOAM_LIBBIN)/dummy -lptscotchDecomp \
-lsampling \
- -lfileFormats
+ -lfileFormats \
+ -lboost_thread
python-scipy-mkl has a literal $pkgver in its provides. This happens because you have enclosed these variables in strong quotes.
sudo pacman -S cntk
resolving dependencies...
warning: cannot resolve "nccl", a dependency of "cntk"
:: The following package cannot be upgraded due to unresolvable dependencies:
cntk
:: Do you want to skip the above package for this upgrade? [y/N] n
error: failed to prepare transaction (could not satisfy dependencies)
:: unable to satisfy dependency 'nccl' required by cntk
The latest release is 0.10.1, https://github.com/xflr6/graphviz/releases/tag/0.10.1.
libnvidia-container-tools 的版本应该是和 libnvidia-container 保持一致的才对.考虑一下从源码构建?AUR 上已经改成从源码构建的了.
openfoam
The open source CFD toolbox,
http://www.openfoam.org
https://github.com/OpenFOAM
The AUR links:
https://aur.archlinux.org/packages/openfoam/
https://aur.archlinux.org/packages/openfoam2.4-git/
https://aur.archlinux.org/packages/openfoam3.0-git/
openfoam开源的计算流体力学软件,有个非官方软件源
home_Minerva_W_Science_Arch_Extra
Maintainer:
Description: OpenFOAM packages.
[home_Minerva_W_Science_Arch_Extra]
Server = http://download.opensuse.org/repositories/home:/Minerva_W:/Science/Arch_Extra/$arch
gmsh
A free 3D finite element grid generator with a build-in CAD engine and post-processor
计算流体力学的网格划分软件
http://gmsh.info/
AUR:https://aur.archlinux.org/packages/gmsh-bin/
I have a plan to add R packages into arch4edu such as r-devtools
, r-jsonlite
.
But:
So I post this issue as a vote or survey before starting to add them.
If there are enough votes that requring the R packages, we will start to do it.
Jabref is a citation manager, and bibtex editor. It eases the work of creating bib
file for publication. All you need is to provide doi
number and it can generates bib
file automatically.
The official website is https://www.jabref.org. There is already an AUR package https://aur.archlinux.org/packages/jabref/.
The software is licensed with MIT. So theoretically, we should be able to do anything with it?
numpy/scipi在计算机科学领域广泛使用,与官方源的numpy相比,mkl加持的numpy并行能力有肉眼可见的巨大提升,但intel的工具链体积太大,makepkg非常煎熬,因此十分希望能用上prebuilt版本。
AUR:
https://aur.archlinux.org/packages/python-scipy-mkl/
https://aur.archlinux.org/packages/python-numpy-mkl/
Introduction: Extracts data points from images of graphs
Homepage: http://markummitchell.github.io/engauge-digitizer/
License: GPL 2.0
AUR link: https://aur.archlinux.org/packages/engauge
This is a neat tool when you need to extract data from figures in papers.
The cudnn package is in this repository, but can not be found on the server.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.