#metabrainz

/

      • _lucifer
        yeah its 3.7-20210115 and https://github.com/metabrainz/docker-python shows its 0,18
      • 2021-02-26 05722, 2021

      • alastairp
        interesting. yvanzo and I hoped that this would fix the error!
      • 2021-02-26 05706, 2021

      • alastairp
        does that ":52:" appear to refer to a line number?
      • 2021-02-26 05707, 2021

      • _lucifer
        i am using the startup branch with only modification to docker-compose.yml
      • 2021-02-26 05747, 2021

      • _lucifer
        let me check
      • 2021-02-26 05748, 2021

      • alastairp
        because this line 52 isn't what I expec the error to be: https://github.com/metabrainz/listenbrainz-server…
      • 2021-02-26 05753, 2021

      • alastairp
        the change was on line 64
      • 2021-02-26 05705, 2021

      • _lucifer
        yeahs its the same line
      • 2021-02-26 05723, 2021

      • alastairp
      • 2021-02-26 05748, 2021

      • yvanzo
        _lucifer: this is caused by this commit: https://github.com/metabrainz/listenbrainz-server…
      • 2021-02-26 05753, 2021

      • alastairp
        can you try this: remove the {{- with }} (and associated end), and replace $rabbitmq_service_name with "rabbitmq"
      • 2021-02-26 05743, 2021

      • yvanzo
      • 2021-02-26 05700, 2021

      • yvanzo
        alastairp: "rabbitmq-prince"?
      • 2021-02-26 05726, 2021

      • alastairp
        yvanzo: we're just running this locally at the moment, and don't even have any services defined in our local consul
      • 2021-02-26 05740, 2021

      • yvanzo
        oh ok :)
      • 2021-02-26 05745, 2021

      • alastairp
        I'm just seeing what it takes to make this render, and then we can look into fixing the actual `with` block
      • 2021-02-26 05708, 2021

      • yvanzo
        I tested with hardcoded "rabbitmq-prince", it was working fine.
      • 2021-02-26 05726, 2021

      • yvanzo
        (for one test container)
      • 2021-02-26 05746, 2021

      • _lucifer
        it proceeds with keyOrDefault
      • 2021-02-26 05753, 2021

      • alastairp
        so it seems that maybe we cannot use our defined KEY template inside a `{{with}}` block (or maybe the `or` block?")
      • 2021-02-26 05714, 2021

      • alastairp
        does `{{- with $rabbitmq_service_name := template "KEY" "rabbitmq_service"}}` work?
      • 2021-02-26 05704, 2021

      • alastairp
        or do you mean that replacing `or(...)` with keyOrDefault(...) works?
      • 2021-02-26 05724, 2021

      • _lucifer
      • 2021-02-26 05739, 2021

      • _lucifer
        I copied this line and removed $svc and it proceeded
      • 2021-02-26 05755, 2021

      • alastairp
        unfortunately that won't work in LB
      • 2021-02-26 05706, 2021

      • alastairp
      • 2021-02-26 05738, 2021

      • alastairp
        this template, called "KEY", takes a value but also includes the value of the DEPLOY_ENV variabliable, to choose the correct config file
      • 2021-02-26 05759, 2021

      • _lucifer
        oh
      • 2021-02-26 05749, 2021

      • alastairp
        this is why I'm thinking we could try `keyOrDefault (template "KEY"....)`
      • 2021-02-26 05757, 2021

      • _lucifer
        {{- with $rabbitmq_service_name := template "KEY" "rabbitmq_service"}} didn't work either
      • 2021-02-26 05718, 2021

      • alastairp
        right, it seems that perhaps we can't use template in the right-hand-side of a with :(
      • 2021-02-26 05737, 2021

      • alastairp
        I don't know enough about consul templates to even know where to start debugging this
      • 2021-02-26 05751, 2021

      • alastairp
        I wonder if we should just revert yvanzo's commit and go back to environment variables for this
      • 2021-02-26 05728, 2021

      • yvanzo
        do you really need this template?
      • 2021-02-26 05735, 2021

      • yvanzo
        (KEY)
      • 2021-02-26 05742, 2021

      • alastairp
        I don't know. this has "always been there"
      • 2021-02-26 05703, 2021

      • alastairp
        I see that it's a helper to include the DEPLOY_ENV in the path to the key in consul
      • 2021-02-26 05743, 2021

      • yvanzo
        alastairp: If you revert this commit in LB repo, you will have to revert the commit in docker server configs too. I would suggest to just use hardcoded "rabbitmq-prince" instead for now.
      • 2021-02-26 05741, 2021

      • alastairp
        I think it's a better idea to go back to the environment variables, because it allows us to redeploy without building a new version of the server
      • 2021-02-26 05712, 2021

      • alastairp
        alternatively, we reproduce the functionality of the KEY template inside keyOrDefault ?
      • 2021-02-26 05721, 2021

      • _lucifer
        on a side note, it seems the latest version is 0.25.2 so why not upgrade directly to it instead of 0.18?
      • 2021-02-26 05742, 2021

      • zas
        ^^yes (if possible)
      • 2021-02-26 05743, 2021

      • yvanzo
        It seems it would be more flexible to have a variable (like $svc in CAA indexer) than a template (like KEY in LB).
      • 2021-02-26 05704, 2021

      • alastairp
        mostly because that's the version in the docker-comsul-template-base repository
      • 2021-02-26 05737, 2021

      • alastairp
        part of whole reason we're doing this upgrade is to be able to upgrade to a new version of consul, and I guess eventually new consul-template
      • 2021-02-26 05721, 2021

      • alastairp
        OK, I'm making an executive decision. In order to be able to do this release, let's roll back to the version of the template and docker-server-configs that uses the env variable to load the rabbitmq host
      • 2021-02-26 05753, 2021

      • alastairp
        and we can open a ticket to see if we can use a variable instead of a template, and after the consul upgrade we look at the steps to allow us to upgrade to a newer consul-template version
      • 2021-02-26 05721, 2021

      • _lucifer
        alastairp: do we need to do it in a with block though? if I understand correctly, with executes the block if the value is defined otherwise not. since we are using keyOrDefault it will always be defined.
      • 2021-02-26 05712, 2021

      • alastairp
        have you found a way to use keyOrDefault and also include DEPLOY_ENV?
      • 2021-02-26 05703, 2021

      • _lucifer
        not yet but trying to find one :)
      • 2021-02-26 05738, 2021

      • _lucifer
        was making sure i had not missed something
      • 2021-02-26 05718, 2021

      • MajorLurker joined the channel
      • 2021-02-26 05750, 2021

      • MajorLurker has quit
      • 2021-02-26 05726, 2021

      • reosarevok
        yvanzo, bitmap: re: https://tickets.metabrainz.org/browse/MBS-9841 - should we add a banner to prod when releasing beta letting people know that they can try to use beta?
      • 2021-02-26 05727, 2021

      • BrainzBot
        MBS-9841: Warn about the new beta version on mb.o
      • 2021-02-26 05743, 2021

      • reosarevok
        I don't think this is an MBS ticket, really - more of a decision about the release process
      • 2021-02-26 05751, 2021

      • Lotheric
        I think that people uninterested by testing beta won't appreciate the banner anf those interested are probably already invested in MusicBrainz enough to know about the beta site, tickets, jira, etc.
      • 2021-02-26 05736, 2021

      • reosarevok
        IIRC that was added because a surprising amount of people were like "oh, I'd totally use it if I knew it was there"
      • 2021-02-26 05701, 2021

      • reosarevok
        But dunno
      • 2021-02-26 05710, 2021

      • Lotheric
        maybe change the bottom "Use beta site" with something like "Use beta site (list of tickets that we are testing)" with an url
      • 2021-02-26 05717, 2021

      • reosarevok
        I mean, if people saw that tiny footer link there would be no problem in the first place :D
      • 2021-02-26 05730, 2021

      • reosarevok
        I dunno, do you think people would be annoyed if there was a banner?
      • 2021-02-26 05740, 2021

      • Lotheric
        let's find out, do it!
      • 2021-02-26 05742, 2021

      • Lotheric
        :)
      • 2021-02-26 05749, 2021

      • Lotheric
        I don't care, I already use beta all the time
      • 2021-02-26 05750, 2021

      • Lotheric
        hehe
      • 2021-02-26 05728, 2021

      • Lotheric
        if someone comes in here asking to speak with your supervisor then you'll know :D
      • 2021-02-26 05730, 2021

      • Lotheric
        haha
      • 2021-02-26 05726, 2021

      • shivam-kapila
        Mr_Monkey: Can you help me with the failing test here https://ci.metabrainz.org/view/Listenbrainz/job/l…
      • 2021-02-26 05741, 2021

      • legoktm
        reosarevok: IIRC the new summary endpoint was implemented for the mobile apps after they were dissatisfied with the prop=extracts output. Let me find where it's implemented so we can understand the differences
      • 2021-02-26 05749, 2021

      • bitmap
        reosarevok: I'm not opposed to trying it. my initial thought was every single update seems a bit much if there's no "never show this again" option
      • 2021-02-26 05716, 2021

      • reosarevok
        That's fair I guess, but once by release cycle doesn't sound too bad
      • 2021-02-26 05721, 2021

      • reosarevok
        That's once every two weeks
      • 2021-02-26 05733, 2021

      • reosarevok
        Most editors probably dismiss the "new edit notes" banner more often )
      • 2021-02-26 05754, 2021

      • bitmap
        re: MBS-9858 I'll have to check, I think I did change the rate of change code
      • 2021-02-26 05755, 2021

      • BrainzBot
        MBS-9858: Statistics/Time Graph/Rate of change: incorrect handling 31/30/28 day months? https://tickets.metabrainz.org/browse/MBS-9858
      • 2021-02-26 05714, 2021

      • shivam-kapila
        Mr_Monkey: nvm. Fixed it. The PR is ready now
      • 2021-02-26 05748, 2021

      • _lucifer
      • 2021-02-26 05758, 2021

      • _lucifer
        alastairp: yvanzo: this works ^
      • 2021-02-26 05751, 2021

      • alastairp
        looks fine by me!
      • 2021-02-26 05708, 2021

      • alastairp
        we're duplicating docker-server-configs/LB/config.%s.json, but I don't think there's much we can do about it
      • 2021-02-26 05709, 2021

      • _lucifer
        to my understanding, the use of the template is to avoid repeating the first line of the snippet. duplicating it for once seems fine to me
      • 2021-02-26 05721, 2021

      • alastairp
        yeah, if we end up duplicating it 1 more time we should take a step back and look at it again, but for I think it's ok
      • 2021-02-26 05712, 2021

      • _lucifer
        alastairp, should we proceed or leave the upgrade till the next time?
      • 2021-02-26 05751, 2021

      • legoktm
        reosarevok: https://www.mediawiki.org/wiki/Page_Content_Servi… explains how the endpoint works conceptually. I would say if you find the output from that API better, make the switch
      • 2021-02-26 05707, 2021

      • reosarevok
        legoktm: I guess there's no equivalent of the old intro option? (so, not just one paragraph but the whole intro section)
      • 2021-02-26 05721, 2021

      • reosarevok
        Tbf, maybe we should not display the whole intro section, but :)
      • 2021-02-26 05733, 2021

      • legoktm
        doesn't seem like it, because this is targeted to smaller things apparently
      • 2021-02-26 05743, 2021

      • legoktm
        do you have the ability to do DOM processing? https://en.wikipedia.org/api/rest_v1/page/html/Da… has much richer HTML markup so you could extract the first <section> tag to get the lead section..but then you'd also need to get the image too, hmmm
      • 2021-02-26 05722, 2021

      • legoktm bbl (will read scrollback)
      • 2021-02-26 05700, 2021

      • BrainzGit has quit
      • 2021-02-26 05719, 2021

      • BrainzGit joined the channel
      • 2021-02-26 05752, 2021

      • BrainzGit
        [musicbrainz-server] reosarevok opened pull request #1945 (master…MBS-9826): MBS-9826: Show days + hours when edit closing time is over a day away https://github.com/metabrainz/musicbrainz-server/…
      • 2021-02-26 05724, 2021

      • reosarevok
        legoktm: we don't show images anyway, not since THE COURT MESS (TM)
      • 2021-02-26 05706, 2021

      • MajorLurker joined the channel
      • 2021-02-26 05705, 2021

      • MajorLurker has quit
      • 2021-02-26 05701, 2021

      • llrcombs has quit
      • 2021-02-26 05720, 2021

      • reosarevok
        Yay, the waves of migrants are coming in
      • 2021-02-26 05735, 2021

      • reosarevok
        Migrant geese, that is :D
      • 2021-02-26 05746, 2021

      • reosarevok
        Time for birding soon!
      • 2021-02-26 05751, 2021

      • rcombs joined the channel
      • 2021-02-26 05711, 2021

      • ruaok
        <honk!>
      • 2021-02-26 05702, 2021

      • chaban
        reosarevok: looking for more low-hanging fruits (tickets)? Check out MEB-114 :)
      • 2021-02-26 05703, 2021

      • BrainzBot
        MEB-114: Contact page still says registration is required for talking in IRC channels https://tickets.metabrainz.org/browse/MEB-114
      • 2021-02-26 05757, 2021

      • reosarevok
        I guess I do have the meb repo locally too
      • 2021-02-26 05703, 2021

      • reosarevok
        So might as well
      • 2021-02-26 05737, 2021

      • ruaok
        cuz cloing a repo is so much effort?
      • 2021-02-26 05722, 2021

      • reosarevok
        No, but still more effort than doing more MBS triaging :p
      • 2021-02-26 05752, 2021

      • reosarevok
        Anyway, finally done sending emails for the autoeditor election
      • 2021-02-26 05704, 2021

      • reosarevok
        rdswift: for some reason this time it would only send 75 at a time
      • 2021-02-26 05750, 2021

      • rdswift
        That's odd. I'm guessing that it's some limit imposed by gmail.
      • 2021-02-26 05705, 2021

      • rdswift
        Might need to leave more of a delay between sending each message.
      • 2021-02-26 05748, 2021

      • reosarevok
        Maybe. Trying again even a few seconds later worked
      • 2021-02-26 05752, 2021

      • reosarevok
        Ok, maybe a min
      • 2021-02-26 05702, 2021

      • reosarevok
        So it could also just send 75, wait, etc
      • 2021-02-26 05713, 2021

      • reosarevok
        But of course, that might change, so maybe a bigger delay is smarter
      • 2021-02-26 05747, 2021

      • rdswift
        Sounds like some rate limiting for sure.
      • 2021-02-26 05700, 2021

      • rdswift
        I've not run into that because the most I've ever sent in one batch was about 50.
      • 2021-02-26 05709, 2021

      • BrainzGit has quit
      • 2021-02-26 05719, 2021

      • reosarevok
        Yeah, we have 260 or so autoeditors
      • 2021-02-26 05735, 2021

      • reosarevok
        Of course, a bunch of those emails just get rejected
      • 2021-02-26 05755, 2021

      • reosarevok
        And there's one who just replied with "I can't figure out how to unsubscribe from this" :)
      • 2021-02-26 05711, 2021

      • reosarevok
        To which the answer is, I guess, "stop being an auto-editor"
      • 2021-02-26 05710, 2021

      • rdswift
        Argh! Hopefully the rejectyed were just bad addresses and not the sending process. Unsubscribe... That makes me chuckle.
      • 2021-02-26 05740, 2021

      • reosarevok
        Yeah, it's just people who probably have had their address unchanged since they got elected in 2005 or whatever
      • 2021-02-26 05701, 2021

      • reosarevok
        For all we know, some might not even be alive, much less use the same email
      • 2021-02-26 05723, 2021

      • reosarevok
        And I mean, it's fair, we don't have any "de-auto-editor" procedure
      • 2021-02-26 05733, 2021

      • reosarevok
        So you might have stopped using MB for 10 years, and still get a mail every time
      • 2021-02-26 05733, 2021

      • BrainzGit joined the channel
      • 2021-02-26 05759, 2021

      • rdswift
        Perhaps an unsubscribe option might be a good idea (the next time you update the database). If people are bothered by half a dozen messages a year, then they probably aren't all that interested in elections and such anyway.
      • 2021-02-26 05700, 2021

      • reosarevok
        Well, in that case it should probably be "stop being an autoeditor", because that's like one of the few things you put up with in exchange for magic powers :D
      • 2021-02-26 05712, 2021

      • reosarevok
        (we don't have a "step down" process rn, other than "mail us")
      • 2021-02-26 05759, 2021

      • yvanzo
        !m _lucifer
      • 2021-02-26 05759, 2021

      • BrainzBot
        You're doing good work, _lucifer!
      • 2021-02-26 05724, 2021

      • rdswift
        That still wouldn't help with the lost (or dead). I mean, it's not like it's a great imposition or anything.
      • 2021-02-26 05718, 2021

      • yvanzo
        reosarevok: beta release could be appended after the footer “Use beta site”
      • 2021-02-26 05704, 2021

      • BrainzGit has quit
      • 2021-02-26 05724, 2021

      • BrainzGit joined the channel
      • 2021-02-26 05702, 2021

      • reosarevok
        yvanzo: it could, but as I mentioned to Lotheric, I suspect people just don't see the footer in the first place
      • 2021-02-26 05748, 2021

      • yvanzo
        well, at least two of us do see it :)
      • 2021-02-26 05732, 2021

      • reosarevok
        Sure :) I guess the main idea is "how do we make using beta more popular"
      • 2021-02-26 05707, 2021

      • BrainzGit has left the channel
      • 2021-02-26 05725, 2021

      • reosarevok
        What's with BrainzGit today
      • 2021-02-26 05730, 2021

      • yvanzo
        we do want advanced editors to use, not necessarily everyone
      • 2021-02-26 05751, 2021

      • reosarevok
        I mean, if something is broken it doesn't matter much who finds it
      • 2021-02-26 05702, 2021

      • reosarevok
        But I guess it might be confusing for randos if something is too broken
      • 2021-02-26 05712, 2021

      • BrainzGit joined the channel
      • 2021-02-26 05722, 2021

      • yvanzo
        of course it does matter, or we would just push everything in prod without using beta step :D
      • 2021-02-26 05712, 2021

      • BrainzGit has left the channel
      • 2021-02-26 05740, 2021

      • BrainzGit joined the channel