galaxyproject/dev

For everybody interested in Galaxy development

Marius van den Beek

@mvdbeek

I suppose this was disabled to not build twice when pushing to branches

A pet-peeve if I ever had one :D

thanks for the heads up @nsoranzo

Nicola Soranzo

@nsoranzo

No worries (and use your fork, Luke!)

Martin Cech

@martenson

Yeah, I did that to most of our projects, I understand it is sometimes annoying but you can always trigger the build if you need it

and the double-build on PR is annoying and consumes our travis quota

Martin Cech

@martenson

@natefoo can we bring the other au nodes up now too?

re jenkins

Nate Coraor

@natefoo

I need to set the keepalive settings on them as well, one sec

Martin Cech

@martenson

(jenkins is being upgraded atm)

Nicola Soranzo

@nsoranzo

@martenson We have added automatic push to PyPI for ephemeris, which is accomplished using a deploy stage on the builds for tags. So we need Travis building on branches, or tags won't be built AFAIK.

Martin Cech

@martenson

sounds good

Martin Cech

@martenson

really @jmchilton ? 🤣

jihaw = JobImportHistoryArchiveWrapper(app, job.id) # yeehaw!

Nate Coraor

@natefoo

yeeeeeeeeeeeeeehaw!

John Chilton

@jmchilton

It is Carl-ism through and through - I don't deserve credit

He had all the best comments

Marius van den Beek

@mvdbeek

anyone ready to approve galaxyproject/ephemeris#144 ? I'll make a new release then

Marius van den Beek

@mvdbeek

thanks @jmchilton !

Martin Cech

@martenson

@nsoranzo re:shellcheck do you have list of ignored rules?

Nicola Soranzo

@nsoranzo

Not really, some need to be carefully checked manually, e.g. when double quotes should not be added around an environment variable like in https://github.com/galaxyproject/galaxy/blob/dev/scripts/common_startup.sh#L193

Martin Cech

@martenson

@nsoranzo Github is down for me, so could you expand here on your comment for the data package pr?

h, up now, nevermind

I'll follow up there

Jennifer Hillman-Jackson

@jennaj

I'm not sure what solution would apply for this data transfer/link update. Help appreciated: https://help.galaxyproject.org/t/launch-a-galaxy-tool-from-an-html-report/1740

Nate Coraor

@natefoo

The Australian Jenkins nodes can be brought back online now.

Martin Cech

@martenson

done

Sergey Golitsynskiy

@ic4f

@nsoranzo , @mvdbeek : question about py2>3 for tools-iuc. This line: https://github.com/galaxyproject/tools-iuc/pull/2032/files#diff-5cfa0d10b441acc12ccbebbcc43c4376R59 - shouldn't it be wrapped in a with stmt? we never close args.output. Is this a special case, or am I missing something else?

Sergey Golitsynskiy

@ic4f

Also, what's the purpose of this try/except? https://github.com/galaxyproject/tools-iuc/blob/master/datatypes/snpeff_datatypes/snpeff.py#L71 The file handle will be left open in case of an exception. Should we fix these? (that's just one example)

Sergey Golitsynskiy

@ic4f

@erasche flake8 is complaining about a 4 year old commit: https://github.com/galaxyproject/tools-iuc/blob/master/tools/progressivemauve/xmfa2gff3.py#L161 . Should the type be argparse.FileType() instead? Or string by default? ('file' is undefined)

Helena Rasche

@erasche

yes

Sergey Golitsynskiy

@ic4f

thank you! (string, so leaving default)

Marius van den Beek

@mvdbeek

Yeah, better to use a with statement. Though in practice it shouldn't be a problem for short lived scripts like these.

Marius van den Beek

@mvdbeek

For the snpeff metadata you'd want to use a with statement as well (always use with, there's no harm in it). The try/except isn't necessary, the framework will handle any exception anyway

Sergey Golitsynskiy

@ic4f

thanks, will change. (also, i'm afraid i made a mess of that PR.. Untangling.. merge/force push accident)

Marius van den Beek

@mvdbeek

Ah, it doesn't for set_metadata, so you do need the try/except

Björn Grüning

@bgruening

For everyone working/interested on data managers, could you please look at this one galaxyproject/tools-iuc#2489 and leave comments.

Sergey Golitsynskiy

@ic4f

except (pun not intended), if that's the except i'm thinking of, it has 'pass' only - so it doesn't handle anything; and without a finally clause, the file stays open. So i'd modify it, i suppose?

Nicola Soranzo

@nsoranzo

@ic4f Here is how I would recommend to read/write JSON files, note the sort_keys=True to make testing of data managers easier.

Nicola Soranzo

@nsoranzo

@jmchilton Can I also get a release of galaxy-containers into PyPI ? It will be needed by Planemo post galaxy-lib (due to from galaxy.containers.docker_model import DockerVolume in planemo/galaxy/config.py)

Martin Cech

@martenson

@ic4f it silences the exception

I think the preferred pattern is to use

finally:
    fh.close()

@jmchilton alternatively write a short release checklist and we can share the load

Sergey Golitsynskiy

@ic4f

@martenson : silences the exception - i know, that was my point (i.e., as opposed to failing 'fast and failing loudly'). And yes, i was going to do just that - adding a finally clause.

Nicola Soranzo

@nsoranzo

I'm happy to be a package maintainer too, if we want to have a group

Marius van den Beek

@mvdbeek

@jmchilton galaxyproject/galaxy#8360 fixes the mulled integration tests for me

Sergey Golitsynskiy

@ic4f

@martenson @nsoranzo for the sake of simplicity, i'll do separate PRs: one to fix python-3 incompatibility issues, the other one (or multiple PRs, if need be), to fix misc issues like unclosed files, etc. Otherwise it's turning into a rather large PR that does many different things.

Marius van den Beek

@mvdbeek

yes, please, that'll be much easier

Nicola Soranzo

@nsoranzo

For stuff related to data managers I'd open a PR against https://github.com/mvdbeek/tools-iuc/tree/py3_data_manager_fixes , if Marius agrees

Sergey Golitsynskiy

@ic4f

@nsoranzo about read/write JSON files: did you mean to provide a link? Or were you only referring to the sort_keys attr?

Where communities thrive

People

Repo info

Activity