#metabrainz

/

      • samj1912
        umm, later versions of pyqt introduce some changes in the network module
      • 2017-04-03 09322, 2017

      • ruaok
        meh, there are still several containers running.
      • 2017-04-03 09333, 2017

      • samj1912
        so it might not run until I submit the PR for py3
      • 2017-04-03 09302, 2017

      • Slurpee joined the channel
      • 2017-04-03 09320, 2017

      • samj1912
        I know its for arch, that's why I wanted to let you know it doesn't work on the latest pyqt :P
      • 2017-04-03 09300, 2017

      • samj1912
        or it might? not sure if those changes were specifically for py3/qt5 since it was related to bytes and str problems
      • 2017-04-03 09325, 2017

      • iliekcomputers
        ruaok: the problem is probably related to something like that, maybe the old influx container had some old listens that got exported also
      • 2017-04-03 09330, 2017

      • ruaok
        but the integration tests should use non-permanent volumes that should be discarded after each test run.
      • 2017-04-03 09334, 2017

      • ruaok
        why is that not happening?
      • 2017-04-03 09329, 2017

      • zas
        it doesn't for sure, and is the reason we were running out of disk space on serge
      • 2017-04-03 09353, 2017

      • zas
        i manually deleted dangling volumes
      • 2017-04-03 09313, 2017

      • ruaok
        got an easy command for that?
      • 2017-04-03 09313, 2017

      • zas
        same goes for images
      • 2017-04-03 09343, 2017

      • antlarr
        LOL, I didn't notice PICARD-1041 until now :)
      • 2017-04-03 09343, 2017

      • BrainzBot
        PICARD-1041: Replace Ok button text by Make It So! in Options dialog https://tickets.metabrainz.org/browse/PICARD-1041
      • 2017-04-03 09353, 2017

      • zas
        yes, easy but can be dangerous, docker version 1.13 (we are still on 1.12) has a prune command
      • 2017-04-03 09302, 2017

      • ruaok
        back to how it was meant to be!
      • 2017-04-03 09308, 2017

      • zas
        yes !
      • 2017-04-03 09310, 2017

      • samj1912
        antlarr: we shifted to qt5 on master :D
      • 2017-04-03 09321, 2017

      • samj1912
        you are free to file tickets and submit PRs :P
      • 2017-04-03 09327, 2017

      • antlarr
        cool :)
      • 2017-04-03 09311, 2017

      • zas
        docker rmi $(docker images -f dangling=true -q)
      • 2017-04-03 09325, 2017

      • zas
        or using xargs if too many
      • 2017-04-03 09348, 2017

      • colbydray joined the channel
      • 2017-04-03 09314, 2017

      • Freso
        samj1912: I made the package against 5.8 and at least it ran fine (with a bunch of plugin errors).
      • 2017-04-03 09317, 2017

      • zas
        docker volume rm $(docker volume ls -qf dangling=true)
      • 2017-04-03 09324, 2017

      • ruaok
        thanks!
      • 2017-04-03 09327, 2017

      • samj1912
        Freso: +1:
      • 2017-04-03 09330, 2017

      • Freso
        Didn't try loading releases etc.
      • 2017-04-03 09339, 2017

      • zas
        the first one is for images, the second for volumes, but use with care...
      • 2017-04-03 09345, 2017

      • samj1912
        ugh, zas , 1 problem while rebasing, added the ok button twice
      • 2017-04-03 09324, 2017

      • zas
        samj1912: np, just make a fix PR, i didn't expect no issue with so big PR ;)
      • 2017-04-03 09326, 2017

      • Freso
        Loaded a release one just fine too. Will push this PKGBUILD.
      • 2017-04-03 09307, 2017

      • zas
        Freso: we'll discover some issues for sure, but at least things are moving
      • 2017-04-03 09309, 2017

      • Freso
        zas: Yes. But this was specifically with regard to the picard-git Arch/AUR package, which samj1912 said might not work on Qt>5.7. :)
      • 2017-04-03 09315, 2017

      • Gentlecat
        zas: should we dedicate a server just for tests, builds, and other workflow things?
      • 2017-04-03 09345, 2017

      • samj1912
        Freso: I think it might be a py3 only issue :P not sure with so many upgrades :P
      • 2017-04-03 09308, 2017

      • Freso
        samj1912: Mhm. :)
      • 2017-04-03 09308, 2017

      • zas
        Gentlecat: you mean to replace serge ? well, not much an issue for now, but if it becomes one we'll see
      • 2017-04-03 09334, 2017

      • Freso
        Anyway, updated dependencies have been pushed: https://aur.archlinux.org/packages/picard-git/
      • 2017-04-03 09301, 2017

      • github joined the channel
      • 2017-04-03 09301, 2017

      • github
        [picard] samj1912 opened pull request #671: Fix duplicate Ok button (master...master) https://git.io/vS4f1
      • 2017-04-03 09301, 2017

      • github has left the channel
      • 2017-04-03 09315, 2017

      • samj1912
        !m Freso :D
      • 2017-04-03 09315, 2017

      • BrainzBot
        You're doing good work, Freso :D!
      • 2017-04-03 09327, 2017

      • Gentlecat
        zas: not necessarily to replace, but rather move already deployed things around to make it more organized
      • 2017-04-03 09337, 2017

      • samj1912
        We will have to wait a while for plugins to be functional
      • 2017-04-03 09308, 2017

      • Freso
        samj1912: I guess at least until we have some progress on the versioned plugin repository API thing, right?
      • 2017-04-03 09326, 2017

      • Freso
        Huh.
      • 2017-04-03 09327, 2017

      • samj1912
        Freso: actually I have implemented it and submitted a PR :P
      • 2017-04-03 09334, 2017

      • Freso
        Trying to open settings crashed picard.
      • 2017-04-03 09343, 2017

      • Freso tries with "clean" settings
      • 2017-04-03 09347, 2017

      • samj1912
        it shouldnt?
      • 2017-04-03 09303, 2017

      • samj1912
      • 2017-04-03 09306, 2017

      • samj1912
        with "unclean" settings
      • 2017-04-03 09310, 2017

      • Freso
        samj1912: I would think not. :)
      • 2017-04-03 09317, 2017

      • Freso
        samj1912: I doubt we have the exact same settings.
      • 2017-04-03 09331, 2017

      • samj1912
        hehe okay :P
      • 2017-04-03 09353, 2017

      • samj1912
        I expect to be kept busy with JIRA bugs the next few days :P
      • 2017-04-03 09359, 2017

      • samj1912
        and zas with reviewing PRs :P
      • 2017-04-03 09308, 2017

      • Freso
        Seems to be the no_release.py plugin.
      • 2017-04-03 09325, 2017

      • samj1912
        ah, can you delete your plugins folder?
      • 2017-04-03 09335, 2017

      • Freso
      • 2017-04-03 09321, 2017

      • UmkaDK_ has quit
      • 2017-04-03 09333, 2017

      • samj1912
        Freso: you can find the latest plugins @ https://github.com/metabrainz/picard-plugins/pull…
      • 2017-04-03 09334, 2017

      • UmkaDK joined the channel
      • 2017-04-03 09316, 2017

      • Freso
        samj1912: 👍
      • 2017-04-03 09336, 2017

      • zas
        samj1912: we'll need to handle auto download of new versions of plugins when one is upgrading from 1.4.x to 2.x, prolly at start
      • 2017-04-03 09300, 2017

      • samj1912
        zas: I made a PR
      • 2017-04-03 09330, 2017

      • zas
        damn ... you are fixing issues before we report them... ;)
      • 2017-04-03 09339, 2017

      • samj1912
      • 2017-04-03 09347, 2017

      • samj1912
      • 2017-04-03 09305, 2017

      • samj1912
      • 2017-04-03 09330, 2017

      • samj1912
        PW tests are failing currently because gh:PP#79 is not yet merged
      • 2017-04-03 09330, 2017

      • BrainzBot
        Build plugin files for specific api versions: https://github.com/metabrainz/picard-plugins/pull…
      • 2017-04-03 09350, 2017

      • TOPIC: MetaBrainz Community and Development channel | MusicBrainz non-development: #musicbrainz | GSoC Getting Started: https://goo.gl/JTNkn0 | MeB meeting agenda: reviews, picard-plugins & Travis (Sophist), DR: MBS-9248 (reo), picard-plugins issue tracking (Freso), Picard v2.0 (samj1912)
      • 2017-04-03 09356, 2017

      • zas
        samj1912: yes, but picard code doesn't automatically download plugins right ? i mean for an existing install, when user is running 1.4.x and upgrade to 2.x
      • 2017-04-03 09321, 2017

      • samj1912
        zas: the new code should query v2 instead of v1 for PW
      • 2017-04-03 09339, 2017

      • samj1912
        i'll make the complete changes with the py3 port
      • 2017-04-03 09308, 2017

      • antfoo has quit
      • 2017-04-03 09330, 2017

      • samj1912
        to ensure it doesn't load old plugins (so that it doesn't crash) , thus we can query for new plugins off of v2 endpoint :)
      • 2017-04-03 09341, 2017

      • zas
        k
      • 2017-04-03 09353, 2017

      • antfoo joined the channel
      • 2017-04-03 09305, 2017

      • samj1912
        I wanted to discuss more about how to handle plugins and the VIDS specification for them
      • 2017-04-03 09311, 2017

      • samj1912
        thus the meeting topic
      • 2017-04-03 09341, 2017

      • samj1912
        meanwhile we should post on discourse that we have updated to 2.0 on master so people should keep to 1.4.X unless they want to break things?
      • 2017-04-03 09350, 2017

      • samj1912
        or update the blog for 1.4.1?
      • 2017-04-03 09301, 2017

      • Freso
        Is there a way to "sandbox" plugins so a broken plugin doesn't crash Picard?
      • 2017-04-03 09312, 2017

      • samj1912
        Freso: yes, I can do that
      • 2017-04-03 09337, 2017

      • samj1912
        1 sec, let me make a PR
      • 2017-04-03 09344, 2017

      • Freso
        Well, people in general shouldn't run from git if they're not able to debug issues and live with oddities.
      • 2017-04-03 09301, 2017

      • Freso
        Let me a make a ticket you can link the PR to. :p
      • 2017-04-03 09340, 2017

      • samj1912
        okay :P
      • 2017-04-03 09349, 2017

      • Freso
      • 2017-04-03 09349, 2017

      • BrainzBot
        PICARD-1046: "Sandbox" plugins so they don't crash Picard
      • 2017-04-03 09315, 2017

      • zas
        samj1912: is there any v1 plugin working without modification with upcoming py3/qt5 ?
      • 2017-04-03 09352, 2017

      • samj1912
        zas: not sure, will have to check
      • 2017-04-03 09303, 2017

      • samj1912
        currently changing the api_version array in picard.__init__
      • 2017-04-03 09311, 2017

      • samj1912
        so that it only has 2.0.0
      • 2017-04-03 09319, 2017

      • samj1912
        thus, none of the plugins will be loaded
      • 2017-04-03 09334, 2017

      • zas
        i'm wondering about how to handle v1/v2 plugins, it can be a branch, or just another directory (meaning a bit of duplication, but may end to be simpler)
      • 2017-04-03 09348, 2017

      • Freso
        zas: I'd say a branch is nicer than a directory.
      • 2017-04-03 09352, 2017

      • D4RK-PH0ENiX joined the channel
      • 2017-04-03 09300, 2017

      • samj1912
        zas: branches
      • 2017-04-03 09317, 2017

      • D4RK-PH0ENiX has quit
      • 2017-04-03 09320, 2017

      • samj1912
        I have handled it using branches in my pr for gh:PP#79
      • 2017-04-03 09320, 2017

      • BrainzBot
        Build plugin files for specific api versions: https://github.com/metabrainz/picard-plugins/pull…
      • 2017-04-03 09350, 2017

      • D4RK-PH0ENiX joined the channel
      • 2017-04-03 09355, 2017

      • Freso
        Branches make it easier to forward/backward port patches for plugins between versions.
      • 2017-04-03 09313, 2017

      • zas
        Freso: yes, i agree
      • 2017-04-03 09334, 2017

      • d4rkie has quit
      • 2017-04-03 09353, 2017

      • Freso
        samj1912: Also, I left a note on the -plugins PR you linked me: https://github.com/metabrainz/picard-plugins/pull…
      • 2017-04-03 09303, 2017

      • github joined the channel
      • 2017-04-03 09303, 2017

      • github
        [picard] samj1912 opened pull request #672: PICARD-1046: Mark plugins with api_versions<2 as incompatible (master...picard1046) https://git.io/vS4Tr
      • 2017-04-03 09303, 2017

      • github has left the channel
      • 2017-04-03 09304, 2017

      • samj1912
        Ah yes, I saw
      • 2017-04-03 09314, 2017

      • samj1912
        ill port the plugins again
      • 2017-04-03 09328, 2017

      • samj1912
        currently failsafes ^
      • 2017-04-03 09300, 2017

      • samj1912
        that should ensure no incompatible plugins are loaded
      • 2017-04-03 09308, 2017

      • D4RK-PH0ENiX has quit
      • 2017-04-03 09351, 2017

      • samj1912
        Ah Freso I saw your comment
      • 2017-04-03 09309, 2017

      • samj1912
        I'll have to think about it properly, how to handle broken plugins
      • 2017-04-03 09322, 2017

      • samj1912
        since plugins can also break while in execution
      • 2017-04-03 09303, 2017

      • Freso
        Yep. I didn't expect it to be an easy, or immediate, fix. :)
      • 2017-04-03 09309, 2017

      • samj1912
        but we provide a common interface for them and they are very much integrated with picard, thus it will be hard to separate them as of now :P
      • 2017-04-03 09314, 2017

      • zas
        i'm not sure we can sandbox plugins easily ...
      • 2017-04-03 09318, 2017

      • samj1912
        yeah
      • 2017-04-03 09328, 2017

      • arbenina_ has quit
      • 2017-04-03 09329, 2017

      • vermileo joined the channel
      • 2017-04-03 09311, 2017

      • zas
        it would require to sandbox an interpreter, and ... well, it will require a lot of changes imho
      • 2017-04-03 09341, 2017

      • samj1912
        yup
      • 2017-04-03 09354, 2017

      • D4RK-PH0ENiX joined the channel
      • 2017-04-03 09354, 2017

      • D4RK-PH0ENiX has quit
      • 2017-04-03 09301, 2017

      • D4RK-PH0ENiX joined the channel
      • 2017-04-03 09308, 2017

      • samj1912
        it would be much easier to have testing for plugins and not merge broken code into the official repo
      • 2017-04-03 09320, 2017

      • zas
        agreed
      • 2017-04-03 09333, 2017

      • Gentlecat
        bitmap, zas: so how difficult would it be to set up another instance of MBS for testing solr indexer?
      • 2017-04-03 09340, 2017

      • samj1912
        we have to be strict about our plugin standards for official repo though
      • 2017-04-03 09348, 2017

      • UmkaDK has quit
      • 2017-04-03 09353, 2017

      • samj1912
        a lot of existing plugins are subpar in terms of code quality
      • 2017-04-03 09326, 2017

      • samj1912
        users can always load unofficial plugins, but I think we are responsible for the official ones since we serve them off of PW
      • 2017-04-03 09330, 2017

      • zas
        that's the main problem, plugins are free to use anything, the API isn't really enforced
      • 2017-04-03 09337, 2017

      • samj1912
        more on that in today's meeting I think
      • 2017-04-03 09357, 2017

      • zas
        Gentlecat: like beta ? i think it can be fast
      • 2017-04-03 09313, 2017

      • Gentlecat
        as long as it's a separate database
      • 2017-04-03 09314, 2017

      • zas
        but bitmap will answer better than me on this
      • 2017-04-03 09334, 2017

      • zas
        ah a separate db ... yes, perhaps not that fast ;)
      • 2017-04-03 09334, 2017

      • Gentlecat
        frontend would be useful too
      • 2017-04-03 09343, 2017

      • zas
        but needed yes
      • 2017-04-03 09356, 2017

      • Gentlecat
        I need to do some tests with triggers and updates