Where communities thrive

  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
Repo info
  • 20:10
    rainwoodman commented #7227
  • 19:34
    thermokarst commented #2980
  • 18:31
    KiberDron edited #10535
  • 18:31
    KiberDron opened #10535
  • 18:24
    lguangyu commented #2980
  • 16:10
    Vichoko opened #10534
  • 15:01
    LunarLanding commented #7753
  • 14:51
    LunarLanding commented #7753
  • 14:45
    spikerobinsonDLG opened #10533
  • 10:27
    anoldmaninthesea closed #10513
  • 08:12
    BerginJay commented #10513
  • 08:08
    gqqnbig commented #7227
  • 04:46
    duytrung opened #10532
  • 03:45
    mingyen1202 commented #9746
  • 02:27
    ecardenas1123 commented #344
  • Mar 01 21:40
    giladmaya synchronize #10530
  • Mar 01 20:44
    anaconda-issue-bot labeled #10531
  • Mar 01 20:44
    felker review_requested #10531
  • Mar 01 20:44
    felker opened #10531
  • Mar 01 19:39
    chenghlee commented #10118
Wolf Vollprecht
then you can use myfile as a command
so whatever ends up in the /bin folder inside your conda package is executable with a single command. Usually people install these files with a setup.py script.
you can make python executables quite easily using the console scripts entry point
I hope with this information you can find some additional help online

@JasonYangShadow , sorry I have never worked with those packages you mentioned... :)
But this looks like working:
$ conda create -n for_trinity -c bioconda trinity # create environment "for_trinity" for the "trinity" package
$ conda create -n for_picard -c bioconda picard # create environment "for_picard" for the "picard" package

Then you will have 2 environments, where different openjdk versions are present. To run picard and trinity you need to activate their corresponding environments first (e.g. in separate terminals if you want to use them at the same time). Or use "conda run" command.
I'm not sure, maybe this was what you wanted to sort out...

Uwe L. Korn
It seems like currently all repodata.json requests to labels end in a 500 error on anaconda.org:
requests.exceptions.HTTPError: 500 Server Error: INTERNAL SERVER ERROR for url: https://conda.anaconda.org/conda-forge/label/broken/noarch/repodata.json
Did also test rust_dev as an alternative label to ensure it is not a problem specific to broken.
Wolf Vollprecht
the same problem on other non-cdn channels (e.g. robostack)
Wolf Vollprecht

@DerKevinRiehl the first step is to learn how to write a setup.py and a proper python package (that is completely orthogonal to conda packaging).
If you want to hack something, then you need to learn a little bit of bash or cmd.exe scripting and just copy your script into $PREFIX/bin (inside your conda meta.yaml).


script: cp myscript.py $PREFIX/bin/myscript.py  # [unix]
script: COPY myscript.py %PREFIX%\Library\bin\myscript.py  # [win]

but this is really hacky and should be avoided. But at this point you're asking really general questions and you should read some tutorials, e.g. the python packaging ones

Ben Mares
Hello @xhochy and @wolfv , the HTTP 500 is causing problems for lots of users, e.g. pytorch/pytorch#51029 . Do you know if this has been reported to the people at anaconda.org? The best I could come up with was sending an e-mail to info@anaconda.com .
Uwe L. Korn
We typically use this channel as a way of communication but people typically wake up around "now".
Ben Mares
Ah, cool. I know you from github effectively as "Mr. conda-forge". ;) But I had the impression that that was completely separate from Anaconda Inc.
Uwe L. Korn
conda-forge is also hosted on Anaconda Inc. server, but I'm completely unaffiliated with them, so all I have too is this channel ;)
conda-forge is on a separate CDN mirror in front of anaconda.org. That's why it is still working.
Ben Mares
Seems to be working for me again!
Uwe L. Korn
Can confirm :+1:
Cheng H. Lee
Re: the 500 errors - one of the worker nodes that backs all of the channels (except conda-forge and main) ran into a storage problem late last night. Anaconda IT has now resolved the situation.
Wolf Vollprecht
Leo Fang
Wolf Vollprecht
you're on windows, right? you need a bld.bat hten
anyways, I actually looked at your repo. you need a source section. pleaes try some existing recipes first. and try to learn the basics
e.g. git, shell, ...
Kaleb Barrett
I am having issues embedding Python from an Miniconda install on Windows 10. I've tracked it down to a problem with the PATH. {install_dir}\Library\bin isn't on the PATH after loading into an embedded Python, but {install_dir}\Library\usr\bin is on the PATH twice. The PATH is correct when sitting on the command line and running Python directly. So I'm wondering how PATH is built and/or modified by the conda environment manager and/or the shipped python executable (or libpython package? good lord...). I'm not sure how all the pieces of conda work together and who I should be reporting this bug to.
4 replies
Ian Stokes-Rees
Any tips for how to debug a very slow conda activate operation? For me it is taking 90 seconds. I looked at etc/conda/activate.d/*.sh scripts, and all they do is check file/directory existence and set environment variables. I can’t figure out what is taking so long.
1 reply
Faustin Carter
Is there a way to do basic http auth for conda channels in any way other than including the creds directly in the channel URL?
I have a protected channel, and currently my condarc has a line like:
- https://MY_USER:MY_LONG_API_KEY@myserver.com/mychannel
But when I build packages, that channel URL is getting included in the package metadata, which is not cool
Patrick Sodré
@FaustinCarter, you can put your secrets in a ~/.netrc location and conda will look for it there
Faustin Carter
Ooh, where is documentation on that?
2 replies
Matti Picus
Over at conda-forge/triqs-feedstock#18, the admin would like to disable pypy until they are convinced it is properly supported upstream. How could they best do that?
3 replies
Faustin Carter
Anyone know of a way to force environment creation from a yaml file to respect packages set in create_default_packages
I'm currently doing conda env create -f envfile.yaml
And I get a new env, but it is missing the "default" packages I specified in .condarc via the create_default_packages key
Faustin Carter
Nevermind, I found an old github issue. Apparently this is a requested feature that is not yet implemented.
Will Furnass
Hi all. Any gotchas to be wary of if wanting to update a conda install on a multi-tenant system from conda-4.3.17 to the latest version? The symlinks to per user envs and the hardlinks within the central install both make me a little nervous.
2 replies
Suhail Shergill
Hi, the current version of tensorflow-gpu for linux is at 2.2 (latest at 2.4). tensorflow is at 2.3 (latest at 2.4). Is there a timeline for when these packages get updated in conda?
Isuru Fernando
Is there a way to avoid user level config file?
4 replies
Is there a status page for the Python 3.9 migration? I'm having trouble building a package and I'm not sure if it's because of missing dependencies etc
Wolf Vollprecht
@mattchan-tencent I believe you're referring to conda-forge? That's actually better to ask on teh conda-forge gitter
@wolfv I mean for defaults
I'm trying to figure out if I can build something on defaults without using conda-forge
since it's a bit more complicated using them...
and no MKL etc...
Wolf Vollprecht
ok, then I am sorry!
(and I don't know)