by

Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • 21:58
    BiocondaBot labeled #22406
  • 21:58
    LiterallyUniqueLogin commented #22406
  • 21:10
    BiocondaBot labeled #22413
  • 21:10
    BiocondaBot labeled #22413
  • 21:10
    BiocondaBot opened #22413
  • 21:06
    BiocondaBot labeled #22412
  • 21:06
    BiocondaBot labeled #22412
  • 21:06
    BiocondaBot opened #22412
  • 20:57
    sjanssen2 synchronize #22411
  • 20:41
    sjanssen2 synchronize #22411
  • 20:23
    sjanssen2 synchronize #22411
  • 20:00
    sjanssen2 synchronize #22411
  • 19:39
    LiterallyUniqueLogin commented #22406
  • 19:38
    LiterallyUniqueLogin synchronize #22406
  • 19:17
    sjanssen2 synchronize #22411
  • 19:05
    sjanssen2 synchronize #22411
  • 18:37
    sjanssen2 synchronize #22411
  • 13:45
    sjanssen2 synchronize #22411
  • 13:18
    sjanssen2 synchronize #22411
  • 13:05
    sjanssen2 synchronize #22411
LiterallyUniqueLogin
@LiterallyUniqueLogin

Hey there,

I opened a bioconda/bioconda-recipes#22387 whose lint check is failing, but when I click on "View more details on CircleCI Checks" under the checks tab, I get sent to a page with a spinning wheel that never loads and I don't see more details. Can someone help me fix this/tell me how my recipe failed the lint check?

Devon Ryan
@dpryan79
@bwlang can you send me the output of conda list?
It seems that you're getting an old pysam with a broken openssl pinning for some reason
Manuel Holtgrewe
@holtgrewe
Off-Topic: is there any good journal out there that accepts extended tutorials for using data analysis software? F1000 Research is the only thing that comes to my mind.
Devon Ryan
@dpryan79
@holtgrewe I suppose Nature Methods too, but F1000 would be my first choice unless the software is extremely popular
Manuel Holtgrewe
@holtgrewe
@dpryan79 thanks for the feedback
Bioconda Bot
@BiocondaBot
PR ready for review: bioconda/bioconda-recipes#22375
Bioconda Bot
@BiocondaBot
PR ready for review: bioconda/bioconda-recipes#22379
Jacques Dainat
@Juke34
Is it normal I don't see the new version of the tool introduced by this Merged PR 2 days ago in bioconda? bioconda/bioconda-recipes#22343
Devon Ryan
@dpryan79
@Juke34 No, that's not normal, it must have failed for random reasons when it got merged into the master branch.
I'm surprised the nightly cron job didn't upload it then.
Pablo Moreno
@pcm32

Hi, our monocl3-cli upload failed to quay.io (https://circleci.com/gh/bioconda/bioconda-recipes/110097?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-checks-link) due to the following error:

10:49:39 BIOCONDA INFO (ERR) [May 29 10:49:39] ERRO Task processing failed: API error (500): {"message":"Get https://quay.io/v2/dpryan79/mulled_container/manifests/latest: error parsing HTTP 429 response body: invalid character 'T' looking for beginning of value: \"Too many login attempts. \\nPlease reset your Quay password and try again.\”"}

should I just retry or is there a password reset as it is advising? Thanks!

I guess @dpryan79 would know since he is explicitly mentioned in the error? Also, the conda package didn’t get uploaded either (I guess related to that failure for consistency?).
Devon Ryan
@dpryan79
@pcm32 That happens on occassion due to an unknown error on quay.io's side. The only solution is to hit rerun and hope that it works.
@bgruening Apparently quay.io is still not functioning correctly.
Pablo Moreno
@pcm32
hehe, thanks. Yes, I was going to do it and Bjoern beat me to it.
Nicola Soranzo
@nsoranzo
Can't beat the bjoernbot :D
Devon Ryan
@dpryan79
:+1:
Brad Langhorst
@bwlang
@dpryan79 : here are the conda lists outputs from the not-working. 3.4.3 and working 3.3.2 envs:
https://gist.github.com/bwlang/c33e1cc505ca45c85426ebc1f8f09a3e
let me know if it’s better to write up as an issue in github, i thought maybe this was just my setup...
Devon Ryan
@dpryan79
@bwlang Your setup is right, this is the umpteenth example of conda solving things incorrectly. Please run conda update pysam in the broken environment and you should get a modern build of pysam that is correctly pinned to openssl.
@mbargull If we need more examples of current repodata issues then I expect what's above is an example of that, since that pysam build is from almost 2 years ago.
Marcel Bargull
@mbargull

Yes, I don't know yet what the actual cause is but that's current_repodata.json related. If you do

CONDA_REPODATA_FNS=repodata.json conda create --name=some-env -c conda-forge -c bioconda deeptools=3.4.3

you get the latest pysam.
(To make that current_repodata.json omission permanent, you can run

conda config --prepend repodata_fns repodata.json
# if the repodata_fns key is already there (unlikely), also run
conda config --remove repodata_fns current_repodata.json

. The conda issue for that is conda/conda#9905 .)

Bioconda Bot
@BiocondaBot
PR ready for review: bioconda/bioconda-recipes#22398
Bioconda Bot
@BiocondaBot
PR ready for review: bioconda/bioconda-recipes#22383
Pablo Moreno
@pcm32
Has anyone had any luck in having an upload-macos or upload-linux working today? Mine have all failed once and again….
LiterallyUniqueLogin
@LiterallyUniqueLogin

Hey there, can anyone help with an issue with my PR bioconda/bioconda-recipes#22387

It seems that the build system is caching and using an old version of my test script (which is called during build) instead of the one that's up to date in the package distribution on PyPI, but I don't know how this is happening/how to debug it. Or maybe this is an error on my end and I'm not seeing it.

Any help appreciated - I've put more details on the PR.

Devon Ryan
@dpryan79
@pcm32 I'll have a look. I expect it's the quay.io issue, since I don't think I added the explicit docker pull in the upload-linux step (shouldn't be needed on OSX)
OK, so at least most of the failures are due to quay.io. That's something I can likely fix real quick.
Devon Ryan
@dpryan79
That should now hopefully be fixed, including for the nightly cron job. The OSX uploads have mostly been working except for monocle3-cli, which keeps timing out for reasons that I don't understand.
LiterallyUniqueLogin
@LiterallyUniqueLogin

Bump.

I'm pretty positive my issue is with CircleCI caching the wrong version of trtools. I noticed that CircleCI's checks are pulling trtools-2.0.17 even though the branch now says version = 2.0.18. Any ideas? How can I flush the CircleCI cache and make it run checks on the latest version of my package? bioconda/bioconda-recipes#22387

LiterallyUniqueLogin
@LiterallyUniqueLogin
Okay, I solved that issue by just creating a new PR. New issue:

Here's the PR https://github.com/bioconda/bioconda-recipes/pull/22406/checks?check_run_id=721932723 . Basically, my tests pass the first time on the the configured Docker container, but then the mulled-build fails.

The mulled-build fails because my script doesn't have access to git (it uses git to download files needed for a few of the tests). However, my recipe specifically declares a dependency on git in the test.requires section of meta.yaml. Is this a bug with mulled-build, or an issue with my meta.yaml?

Thanks for any help!

Pablo Moreno
@pcm32

@dpryan79 when you have time, unfortunately the linux upload (which is what I really need) is not working for monocle3-cli and now the error is different:

From ssh://github.com/bioconda/bioconda-recipes
 * [new branch]          master     -> __upstream__xdxlpY8euA/master
Your bioconda-recipes CI configuration is out of date.
Please update it to the latest version of the upstream master branch.

Have @BiocondaBot attempt to fix this by creating a comment on your PR:

   @BiocondaBot update

I’m guessing that since that PR is merged now, there is not much point in doing the bot update command, right? Thanks!

Devon Ryan
@dpryan79
@pcm2 I'll make a new PR. Yeah that's the downside to changing the CI after merging.
Devon Ryan
@dpryan79
@LiterallyUniqueLogin The mulled test ignores the requires: section since the point of it is to test that only the package and its run time dependencies are really needed. The normal way we handle this is to put tests with extra requirements in a run_test.sh file. Since that's not installed with the package it's then not run in the mulled test (only those tests still in the meta.yaml are run).
Devon Ryan
@dpryan79
@pcm2 Is monocle3-cli still restricted to monocle3 <=0.2.0?
Pablo Moreno
@pcm32
Yes, very much I would say.
I don’t want to spoil your Sat, I can do the PR if needed.
Devon Ryan
@dpryan79
OK, I'll have it built for R 3.6 then
Already have the PR made :)
Pablo Moreno
@pcm32
(I just wasn't sure what was the best way to proceed)
Devon Ryan
@dpryan79
Can you go ahead and approve this PR then: bioconda/bioconda-recipes#22407 ?
Pablo Moreno
@pcm32
Done
Devon Ryan
@dpryan79
Thanks!
Devon Ryan
@dpryan79
@pcm32 I'll build and push this locally to the anaconda cloud. quay.io won't allow containers to get uploaded from circleci for some reason.
Pablo Moreno
@pcm32
ahh, ok, thanks… I had sent it to rebuild just now, should I stop it?
Devon Ryan
@dpryan79
No need, failing again won't hurt anything
Bioconda Bot
@BiocondaBot
PR ready for review: bioconda/bioconda-recipes#22406
LiterallyUniqueLogin
@LiterallyUniqueLogin
@dpryan79 Thanks! That worked. I don't know how closely involved you are with the documentation, but that facet of the mulled build was really unintuitive, and I couldn't find that information anywhere online. Maybe add that information to this bit of the docs?
Björn Grüning
@bgruening
@LiterallyUniqueLogin you can change the documentation here as well: https://github.com/bioconda/bioconda-utils/edit/master/docs/source/contributor/building-locally.rst