conda-forge/conda-forge.github.io

General discussion / chat for conda-forge

Anthony Scopatz

@scopatz

specifically systemd

Ray Donnelly

@mingwandroid

Ha systemd!

Anthony Scopatz

@scopatz

Normally I would :wink:

Ray Donnelly

@mingwandroid

Well, are you making a conda-forge linux distro here?

To be honest, a systemd package doesn't make a lot of sense to me.

Anthony Scopatz

@scopatz

I need libudev to get bluetooth support, which is now part of systemd

I agree most parts of systemd are not relevant

And the build instructions for systemd say in needs to be built into /usr

Ray Donnelly

@mingwandroid

well, udev existed in centos6 and if you run something that links to that on centos7 it will be running a systemd-aware udev and should integrate into your system OK.

(in theory)

Anthony Scopatz

@scopatz

Yes, so the problem there is that centos6 udev is v147 and bluez needs at least libudev v172

Ray Donnelly

@mingwandroid

but from our perspective, we should try to avoid systemd since it's not the only init system on Linux. Gentoo do not even use it by default. I'm not a systemd hater though so I'm not arguing from that perspective at all.

Anthony Scopatz

@scopatz

Yeah, I hear you... It doesn't really matter to me

I just wish they hadn't merged like this

It is kind of a mess for us

Ray Donnelly

@mingwandroid

Gentoo forked udev AFAIK, don't know if that's any use?

Anthony Scopatz

@scopatz

hahah oh Gentoo... my long lost love

Ray Donnelly

@mingwandroid

I started using it again a month ago for first time in about 10 years.

Ryan May

@dopplershift

I'm surprised it's still around

Ray Donnelly

@mingwandroid

It's on my raspi4 serving all my things now .. haven't tried conda on it though.

Francesco Biscani

@bluescarni

$ head /var/log/emerge.log
1309433336: Started emerge on: Jun 30, 2011 11:28:56
1309433336:  *** emerge  unmerge dev-util/ccache
1309433336:  *** exiting successfully.
1309433336:  *** terminating.
1309995063: Started emerge on: Jul 06, 2011 23:31:03
1309995063:  *** emerge  sync
1309995063:  === sync
1309995063: >>> Starting rsync with rsync://134.68.240.59/gentoo-portage
1309995119: === Sync completed with rsync://134.68.240.59/gentoo-portage
1309995119:  *** terminating.

8 years and going strong :)

Ray Donnelly

@mingwandroid

they fixed 64-bit raspi4 issues before anyone else, that is why I installed it.

OK, I know who to ping when I break the raspi again :-)

Francesco Biscani

@bluescarni

Ray Donnelly

@mingwandroid

but my old reasons to avoid it (too much compiling) aren't as relevant for a minimal server so I'm happy to be back to it. Works a lot better than anything else on raspi4 atm in my experience.

Ryan May

@dopplershift

Gentoo taught me everything I know about Linux. Just got tired of trying to figure out why my browser couldn't play flash videos, or open a PDF, etc.

Francesco Biscani

@bluescarni

I used to install it everywhere, now it's only on my desktop. This installation has gone through 2 country moves, countless updates, and the only thing surviving from the original hardware is the case. It has sentimental value at this point :)

Anthony Scopatz

@scopatz

conda skeleton rpm --distro=centos6 --arch=x86_64 systemd isn't working for me at all

I don't seem to be able to build any systemd CDT package at all

Isuru Fernando

@isuruf

@mingwandroid, https://github.com/isuruf/OpenMacOSSDK

WIP script

what's the fastest way to check the type of a file in python?

Ray Donnelly

@mingwandroid

There's no systemd for centos6 for us to repackage here.

Anthony Scopatz

@scopatz

Yeah, it also doesn't work for centos7

Ray Donnelly

@mingwandroid

centos7 is only tested on ppc64le and aarch64 ATM, there are hardcoded tables in cran.py (could be moved to a file if anyone cared enough, I don't see us distro hopping enough for that to be worthwhile personally)

I didn't add entries for centos7 intel as we don't target that as our base on intel as things stand.

teseoch

@teseoch

Hi, i am getting a wierd error when i try to rerender:

I tried to re-render for you, but it looks like I wasn't able to push to the v0.3a branch of libigl/igl-feedstock. Did you check the "Allow edits from maintainers" box?

any idea why?

teseoch

@teseoch

(the checkbox is checked and i tried uncheck -> check)

teseoch

@teseoch

no idea? I believe that missing re-render is causing the problems on windows

Mark Harfouche

@hmaarrfk

@mingwandroid the autogenerated ones don't really work for Conda forge. The split devel packages don't get pushed due to the dependency on an other package that is being built at the same time.

Having them together avoids that dependency circle.

Mark Harfouche

@hmaarrfk

Finally, the skeleton didn't have sha256 autogenerated, nor did it have tests, something that was recently asked. It became important due to the fact that too level directories are stripped sometimes.

Teake Nutma

@teake

There’s a PR on a feedstock I maintain but I can’t merge it (nor make changes to the cloned branch, which I want to do first). I guess I don’t have maintainer rights on my own feedstock repo? The repo in question is https://github.com/conda-forge/astromatic-psfex-feedstock.

Chris Burr

@chrisburr

@teake There is a command in the docs to get the bot to "update team"

Teake Nutma

@teake

Yup that works. Thanks @chrisburr!

Sylvain Corlay

@SylvainCorlay

Hey, do you know how I should use qmake in conda recipes?

since qt is typically a host dependency, and qmake a build tool...

Marcelo Duarte Trevisani

@marcelotrevisani

@SylvainCorlay I think you will need to create a recipe for qmake since this is not present on conda-forge

usually using just cmake is enough , unless you are using something very specific of qmake or there is no config for cmake

Where communities thrive

People

Repo info

Activity