#musicbrainz-devel

/

      • navap_
        How would it be if nikki gave modbot a fake email and then warp or ocharles added an email confirmation date in the db?
      • 2012-09-18 26200, 2012

      • nikki would be fine with 'bot' overriding 'limited user'
      • 2012-09-18 26218, 2012

      • nikki
        the thing about limited users is mainly of interest to me to see which users are brand new and might need more help/advice than normal, a bot is something totally different
      • 2012-09-18 26237, 2012

      • navap_
        That's what I thought
      • 2012-09-18 26226, 2012

      • navap_
        Although we'd want other bots to have a working email address at all times
      • 2012-09-18 26202, 2012

      • navap_
        I used $EDITOR_MODBOT http://pastie.org/4747361
      • 2012-09-18 26202, 2012

      • nikki
        yeah. but that seems separate to me
      • 2012-09-18 26233, 2012

      • nikki
        like you could have a bot with 10,000 auto-edits and a confirmed email address but it would be a limited user because it has fewer than 10 non-auto-edits
      • 2012-09-18 26232, 2012

      • ocharles
        warp: I was mostly looking for the 'do we partition relationship lists' discussion
      • 2012-09-18 26217, 2012

      • ocharles
      • 2012-09-18 26246, 2012

      • nikki
        heh, that page doesn't display very well
      • 2012-09-18 26256, 2012

      • warp
        ocharles: ah, that took place on the mb-devel mailing list iirc.
      • 2012-09-18 26204, 2012

      • ocharles
        nikki: no?
      • 2012-09-18 26215, 2012

      • ocharles
        warp: ok, i'll have a dig
      • 2012-09-18 26243, 2012

      • nikki
        nope. the left side of the text is cut off by the menu thingy
      • 2012-09-18 26226, 2012

      • ocharles
        that's odd, because the main text has a left margin
      • 2012-09-18 26242, 2012

      • ocharles
        displays nicely in firefox and chrome for me
      • 2012-09-18 26211, 2012

      • nikki
        it's cut off in opera 10.10, opera next (probably 12.50 something) and safari for me
      • 2012-09-18 26229, 2012

      • nikki
        there is a margin, it's just not big enough
      • 2012-09-18 26247, 2012

      • nikki
        o_O
      • 2012-09-18 26259, 2012

      • nikki
        that swap track titles with artist credits button is totally not where I expected it to be
      • 2012-09-18 26244, 2012

      • MiX-MaN joined the channel
      • 2012-09-18 26239, 2012

      • warp
        ocharles: we have some modbot failures, I assume due to changes in the release.
      • 2012-09-18 26241, 2012

      • warp
        I'm guessing caused by the changes for MBS-4599
      • 2012-09-18 26241, 2012

      • mb-chat-logger
      • 2012-09-18 26204, 2012

      • ocharles
        warp: i see the problem, do you want to fix it?
      • 2012-09-18 26213, 2012

      • ocharles
        it seems that the release label being edited doesn't even exist now
      • 2012-09-18 26234, 2012

      • ocharles
        and indeed, the release no longer exists
      • 2012-09-18 26224, 2012

      • warp
        ocharles: no, I'm still working on my first schema change ticket. I just noticed the modbot mails.
      • 2012-09-18 26224, 2012

      • ocharles
        and i really hope that Deleted Editor stuff is correct
      • 2012-09-18 26250, 2012

      • ocharles
        http://musicbrainz.org/edit/18919168 apparently it is correct
      • 2012-09-18 26238, 2012

      • MBJenkins
        * ianmcorvidae: MBS-5301: scheme-independent links for gravatars
      • 2012-09-18 26239, 2012

      • MBJenkins
        * ollie: MBS-3902: Add database constraints that groups cannot have a gender set
      • 2012-09-18 26240, 2012

      • adhawkins
        Anyone comment on the latest server release?
      • 2012-09-18 26250, 2012

      • adhawkins
        Just updated my VM and it seems to be chewing processor when nothing is connected
      • 2012-09-18 26206, 2012

      • adhawkins
        Doh
      • 2012-09-18 26207, 2012

      • adhawkins
        Ignore me...
      • 2012-09-18 26233, 2012

      • adhawkins
        It's processing a replication packet.
      • 2012-09-18 26259, 2012

      • adhawkins
        Is this one particularly big?
      • 2012-09-18 26214, 2012

      • reoafk joined the channel
      • 2012-09-18 26257, 2012

      • adhawkins
        Hmm, for some reason it's still processing them since 6th September...
      • 2012-09-18 26225, 2012

      • ocharles
      • 2012-09-18 26252, 2012

      • ocharles
        adhawkins: which process exactly is chewing processor?
      • 2012-09-18 26224, 2012

      • adhawkins
        I think the reason I'm seeing it is that it's processing replication packets *very* slowly
      • 2012-09-18 26240, 2012

      • adhawkins
        1068 23761 94% 1 30
      • 2012-09-18 26217, 2012

      • adhawkins
        I think I usually get a couple of hundred in Stmt/sec
      • 2012-09-18 26245, 2012

      • adhawkins
        Postgres seems to be the main CPU user at the moment.
      • 2012-09-18 26224, 2012

      • adhawkins
        Not sure why it got so far behind on the replication.
      • 2012-09-18 26253, 2012

      • warp
        ocharles: looks fine.
      • 2012-09-18 26202, 2012

      • ocharles
        wow, adhawkins you should check the postgresql log then
      • 2012-09-18 26205, 2012

      • ocharles
        cause that certainly aint right
      • 2012-09-18 26236, 2012

      • adhawkins
        forgot to load "LWP::UserAgent"?) at
      • 2012-09-18 26246, 2012

      • adhawkins
        There's a load of those in the old slave.log, that's why it's behind I think
      • 2012-09-18 26248, 2012

      • ocharles
        i thought we already fixed that
      • 2012-09-18 26258, 2012

      • ocharles
        you're at the latest tag?
      • 2012-09-18 26204, 2012

      • adhawkins
        Just updated today.
      • 2012-09-18 26216, 2012

      • adhawkins
        Yeah, it was sitting retrying one packet (2012-09-03 14:08:09.53593+00) over and over
      • 2012-09-18 26229, 2012

      • ocharles
        and git status looks good?
      • 2012-09-18 26233, 2012

      • adhawkins
        Finally moved past it a little earlier today.
      • 2012-09-18 26235, 2012

      • ocharles
        oh
      • 2012-09-18 26238, 2012

      • ocharles
        wait
      • 2012-09-18 26248, 2012

      • ocharles
        if you were running with bad code, you'll have to kill replication and get a fresh interpreter
      • 2012-09-18 26251, 2012

      • ocharles
        that might be the problem
      • 2012-09-18 26215, 2012

      • adhawkins
        Well, it's now moving through the packets again.
      • 2012-09-18 26227, 2012

      • kurtjx_ joined the channel
      • 2012-09-18 26250, 2012

      • adhawkins
        Only 11 more days worth to process...
      • 2012-09-18 26238, 2012

      • the_metalgamer joined the channel
      • 2012-09-18 26257, 2012

      • ijabz_ joined the channel
      • 2012-09-18 26217, 2012

      • hawke_ joined the channel
      • 2012-09-18 26221, 2012

      • plaintext joined the channel
      • 2012-09-18 26237, 2012

      • hawke joined the channel
      • 2012-09-18 26244, 2012

      • reoafk joined the channel
      • 2012-09-18 26255, 2012

      • CallerNo6 joined the channel
      • 2012-09-18 26235, 2012

      • MBJenkins
        Yippie, build fixed!
      • 2012-09-18 26236, 2012

      • MBJenkins
        ollie: MBS-4599 hotfix: When editing a release label that no longer exists, fail
      • 2012-09-18 26218, 2012

      • ocharles
        ijabz: http://ocharles.org.uk/v3.html not sure you've seen this yet, but it's probably of interest to you
      • 2012-09-18 26245, 2012

      • ijabz
        ocharles why not create as a google doc so it can be commented on
      • 2012-09-18 26257, 2012

      • ocharles
        because I can't stand the google document editor
      • 2012-09-18 26216, 2012

      • ocharles
        and comment threads get really hard to manage there
      • 2012-09-18 26225, 2012

      • ocharles
        there is a github repository for this, and i'd prefer people create issues
      • 2012-09-18 26234, 2012

      • ocharles
        like we did for the caa
      • 2012-09-18 26201, 2012

      • ijabz
        don't we have a chance to discuss it at all then, its essentially a done deal
      • 2012-09-18 26220, 2012

      • ocharles
        ?
      • 2012-09-18 26234, 2012

      • ijabz
        jira isn't suitable for discussion, only more back and white this is wrong type things
      • 2012-09-18 26236, 2012

      • ocharles
        of course you can discuss it, that's why it's a collaborative repository
      • 2012-09-18 26247, 2012

      • ocharles
        i'm not using jira
      • 2012-09-18 26234, 2012

      • ijabz
        oh you mean using github, I remember that being kind of horrible
      • 2012-09-18 26244, 2012

      • ocharles
        i thought it worked well for the caa spec
      • 2012-09-18 26216, 2012

      • MiX-MaN joined the channel
      • 2012-09-18 26256, 2012

      • ijabz
        Its not very accessible, I can't even remember how you doit now
      • 2012-09-18 26221, 2012

      • ocharles
        just go to 'https://github.com/metabrainz/web-service-v3-design/issues' and click 'New issue'
      • 2012-09-18 26238, 2012

      • ijabz
        hmm, well I'll wait a few days for you to flesh it out
      • 2012-09-18 26217, 2012

      • ijabz
        But I still think some general discussion would have been good, and I can't see how using Github offers any benefits over using JIRA
      • 2012-09-18 26210, 2012

      • ocharles
        no real benefit, it's just very lightweight and easy to setup
      • 2012-09-18 26227, 2012

      • ocharles
        and i figured actually doing something rather than bikeshedding infrastructure was a better use of my time :)
      • 2012-09-18 26240, 2012

      • ocharles
        i'm only diving in writing this up as that's what I was told to do yesterday
      • 2012-09-18 26223, 2012

      • ianmcorvidae
        the only thing that bothers me is the ratelimiting is a lot more complex
      • 2012-09-18 26232, 2012

      • ianmcorvidae
        (at present, of course)
      • 2012-09-18 26242, 2012

      • Prophet5 joined the channel
      • 2012-09-18 26204, 2012

      • ianmcorvidae
        by being adaptive we're requiring clients to keep track of a lot more than "do X requests per Y time"
      • 2012-09-18 26210, 2012

      • ocharles
        good morning ianmcorvidae
      • 2012-09-18 26216, 2012

      • ianmcorvidae
        morning :)
      • 2012-09-18 26225, 2012

      • ocharles
        yes, that is true. i haven't considered that yet
      • 2012-09-18 26253, 2012

      • ocharles
        perhaps a timestamp to the next 'reallocation' of rate limit... stuff... should be in there too
      • 2012-09-18 26259, 2012

      • ocharles
        so they know how long to wait if they hit the limit
      • 2012-09-18 26218, 2012

      • ianmcorvidae
        it's still a lot more work for clients
      • 2012-09-18 26239, 2012

      • ianmcorvidae
        I'd really prefer we have something that's not more complex than our current system, but is just tuned better
      • 2012-09-18 26254, 2012

      • nikki wonders if /ws/3 can return more convenient results when looking up cds
      • 2012-09-18 26211, 2012

      • ianmcorvidae
        taking into account things like cache hit rates when we set our policy
      • 2012-09-18 26226, 2012

      • ocharles
        nikki: go on
      • 2012-09-18 26242, 2012

      • ocharles
        ianmcorvidae: if you feel like working on that area, do feel free
      • 2012-09-18 26243, 2012

      • nikki
        ocharles: well, I would do something that returns cd lookup results grouped by tracklist with a list of releases the tracklist is on (if that sounds like pre-ngs, that's because it is :P mb grew out of a cd lookup service, so it makes sense that it used to be very good at it :P)
      • 2012-09-18 26251, 2012

      • nikki
        now though you tend to just get a long list of releases which all look the same (since they were split from a single pre-ngs release) and I don't think there's any obvious way of knowing which ones are the same tracklist
      • 2012-09-18 26234, 2012

      • dsteiner_ joined the channel
      • 2012-09-18 26207, 2012

      • Leftmost joined the channel
      • 2012-09-18 26242, 2012

      • dsteiner joined the channel
      • 2012-09-18 26257, 2012

      • ocharles
        damnit, we STILL have an artist credit column on works, and it'll miss this schema change
      • 2012-09-18 26253, 2012

      • nikki
        make a ticket, put it in the next schema change fix version? :P
      • 2012-09-18 26259, 2012

      • ocharles
        that's what I'm doing
      • 2012-09-18 26249, 2012

      • ocharles
        there we go
      • 2012-09-18 26210, 2012

      • ruaok joined the channel
      • 2012-09-18 26218, 2012

      • ocharles
        ok, v3.html now has read access to all core entities
      • 2012-09-18 26226, 2012

      • ocharles
        yea right. let me know what I've missed :)
      • 2012-09-18 26242, 2012

      • ocharles
        ratings are one that I need to figure out still
      • 2012-09-18 26257, 2012

      • ocharles
        ruaok: https://github.com/metabrainz/web-service-v3-desi… / http://ocharles.org.uk/v3.html. That should keep you instapaper'ing for a while :)
      • 2012-09-18 26207, 2012

      • ocharles
        i assume this is what you were looking for from yesterdays meeting
      • 2012-09-18 26208, 2012

      • kurtjx_ joined the channel
      • 2012-09-18 26259, 2012

      • ruaok
        oy. a full spec, eh?
      • 2012-09-18 26207, 2012

      • ruaok
        I was just hoping for a design sketch. :)
      • 2012-09-18 26213, 2012

      • ruaok
        I'll dig in later.
      • 2012-09-18 26224, 2012

      • ocharles
        i don't know where a sketch stops and spec begins
      • 2012-09-18 26224, 2012

      • ocharles
        so i just got on with it
      • 2012-09-18 26212, 2012

      • ruaok
        well, the problem is that you just told us exactly what we're going to get.
      • 2012-09-18 26221, 2012

      • ruaok
        there isn't a way for us to comment on the high level picture.
      • 2012-09-18 26226, 2012

      • ruaok
        which is the stage we're at.
      • 2012-09-18 26222, 2012

      • ruaok
        well, you should probably stop working on that now.
      • 2012-09-18 26237, 2012

      • ruaok
        I can take that and distill the document I was looking for from it.
      • 2012-09-18 26241, 2012

      • ruaok
        then we can circulate that.
      • 2012-09-18 26214, 2012

      • ocharles
        how can i explain a high level picture?
      • 2012-09-18 26220, 2012

      • ocharles
        the high level picture is presumably the side bar
      • 2012-09-18 26257, 2012

      • ocharles
        i guess i haven't been quite clear on what you wanted
      • 2012-09-18 26245, 2012

      • ruaok
        next time I'm in london we should have a crash course in software engineering.
      • 2012-09-18 26201, 2012

      • ruaok
        so we can discuss exactly what a design doc is vs a spec vs requirements.
      • 2012-09-18 26245, 2012

      • ruaok
        but on a high level you should, in words or diagrams, talk about: granularity, the access model, rate limiting, transport layer, etc.
      • 2012-09-18 26214, 2012

      • ruaok
        and maybe give a little detail for one level of data access (e.g. here is how you fetch data about a release)
      • 2012-09-18 26213, 2012

      • ijabz joined the channel
      • 2012-09-18 26234, 2012

      • voiceinsideyou1 joined the channel
      • 2012-09-18 26213, 2012

      • stefans joined the channel
      • 2012-09-18 26235, 2012

      • voiceinsideyou joined the channel
      • 2012-09-18 26228, 2012

      • ocharles
        ruaok: you have local modifications to RemoveSpamAccounts.pl - can you either discard them, or commit them to master?