#musicbrainz-devel

/

      • reosarevok
        So, I have tickets!
      • :D
      • Quite convenient and only slightly over my budget
      • ruaok will be happy to know I am flying Lufthansa after all :)
      • MBJenkins
        Project libcoverart build #6: SUCCESS in 19 sec: http://ci.musicbrainz.org/job/libcoverart/6/
      • dinog joined the channel
      • ruaok joined the channel
      • bitmap
        I'm getting nothing but 504 gateway timeouts on beta
      • ianmcorvidae
        that would seem like a problem
      • I got a 502, but same idea :P
      • huh
      • beta somehow got reverted to where it was the beginning of this week?
      • ocharles, warp: either of you awake?
      • er, or someone checked out master
      • meh
      • bitmap: beta should be back, I don't know what happened (ignore my last 5 lines or so :P)
      • bitmap
        yeah, works now for me, thanks :)
      • ianmcorvidae
        the servers got killed somehow
      • may have something to do with the big spikes in load average on hobbes though, hm
      • warp
        ianmcorvidae: yes.
      • (awake)
      • ianmcorvidae
        beta went down for some reason -- because the fabric script merges in master that's happened, so it now has the hotfix from last week
      • I haven't determined any real reason why
      • warp
        merging master in beta should be safe.
      • ianmcorvidae
        it brought in my fix for shutting up modbot :)
      • so jenkins is running
      • but that shouldn't be an issue
      • MBJenkins
        * ianmcorvidae: MBS-5369: make sure that EditReleaseLabel edits to remove labels still work
      • * ianmcorvidae: MBS-5369: Add a test to ensure removing a label name works correctly
      • warp
        SUCCESS \o/
      • ianmcorvidae
        yup
      • unfortunately we have to revert the release editor or that'd basically be a success for master too
      • er relationship editor
      • heh
      • warp
        both the MBS-4855 ticket and your question to mb-i18n are a bit vague
      • mb-chat-logger
      • ianmcorvidae
        MBS-4822 you mean?
      • mb-chat-logger
      • warp
        yes
      • ianmcorvidae
        so at present we have two strings, e.g. "Artist" and "Artist:"
      • and like wise label, release group, etc.
      • there's a lot of pairs of strings that vary only in terms of punctuation
      • colon and period being the big offenders
      • so the question is would it be okay to keep only "Artist" by adding "{item}:" for adding the colon
      • warp
        for the specific case mentioned in the ticket, it seems line 215 is just missing a period.
      • ianmcorvidae
        ah, yes
      • for that one that's true
      • warp
        (which would then make these two equal)
      • ianmcorvidae
        however it's the more general question of "is it okay to simply add a colon to an otherwise already-translated string"
      • i.e. "will you translate 'Artist' differently from 'Artist:' differently from 'Artist.' other than the trailing character"
      • warp
        I expect it would not be OK, but I don't speak any languages which use colon differently from english.
      • ianmcorvidae
        and additionally "will you translate 'Artist:' differently from 'Release:' (for example)"
      • (looking *only* at the trailing punctuation, in that case)
      • yeah, I'm not sure
      • hence the email :)
      • warp
        e.g. for japanese subtitles foo: bar is foo ~bar~, so for all I know they don't use colon for this case either :)
      • ianmcorvidae
        ah, but not using colon is fine, so long as they use whatever-they-use the same places we use a colon
      • because they'd then just translate, say, '{title}: {subtitle}' as '{title} ~{subtitle}~'
      • warp
        you mean have just the colon as a translateable string?
      • ianmcorvidae
        not just the colon, it would include the placement of the thing-to-interpolate
      • but pretty much, yes
      • warp nods.
      • the proposed one was for "{item}:" as a translatable string
      • adding a macro 'add_colon' or such
      • warp
        right, that seems to be lacking context
      • but that's all the opinion I have on this. I'm only fluent in english/dutch/frisian, and they all use colon the same way.
      • ianmcorvidae
        yeah
      • mostly I suspect v6lur or someone will come along and throw a wrench in things
      • warp
        :)
      • ianmcorvidae
        with our current set of translators it's usually Estonian that makes things hard :P
      • warp
        haha
      • ianmcorvidae
        (because of the case system, generally)
      • (the romance languages have genders but people are used to working around those -- less so with case languages)
      • can't even imagine how much brokenness we'd get with something that has weirder categories :)
      • warp stops reading wikipedia to figure out what "case language" means and continues on the schema change ticket.
      • heh
      • german also has cases, but the system's a lot less complicated than estonian's
      • warp
        ah, then a german example would probably help me understand it better.
      • ianmcorvidae
        but for example "{age} ago" doesn't work in german because the preposition you have to use to translate 'ago' changes the thing into a different case than what you'd use for an age alone
      • i.e. 'am 7 Jahren' rather than '7 Jahre'
      • er, not am, vor
      • and the case ending isn't always just 'add n', so you can't just make it 'vor {age}n', which would of course be a gross hack anyway :P
      • I could easily see estonian translating "Artist" and "Artist:" differently, because the context provided by the ':' might change the case
      • warp
        Apparantly dutch still has some remnants of that case stuff. (from reading https://nl.wikipedia.org/wiki/Naamval )
      • (that still doesn't help me really grok the concept :)
      • Leftmost joined the channel
      • Leftmost joined the channel
      • ijabz joined the channel
      • ruaok joined the channel
      • ocharles: ayt? do we have a standard way to do an upsert/merge?
      • (I've probably asked for that before :)
      • jesus2099 joined the channel
      • jesus2099
        Did something change on when we receive email notifications of edit notes ?
      • I don’t think I received notification for « I tried to fix it. » in unvotable edit http://musicbrainz.org/edit/19163596
      • maybe I did receive but forgot but I don’t think so
      • Although I did receive on abstained edit http://musicbrainz.org/edit/19111101
      • Previously if you talked in an edit you would receive notifications…
      • ijabz joined the channel
      • ijabz_ joined the channel
      • ijabz
        warp, ping
      • stefans joined the channel
      • warp
        ijabz: ack
      • reosarevok joined the channel
      • ocharles
        good morning
      • warp: no standard way
      • and yes, you've asked before :)
      • warp
        ocharles: ok
      • ocharles
        warp: the 'standard' postgres way is to update, check the row count returned, insert if it's not what you expected, and do all this in a loop; you break if the insert completed without exceptions, and you loop if there was a primary key violation
      • warp
        what I have now is:
      • INSERT INTO cover_art_archive.release_group_cover_art (release_group, release)
      • (SELECT ? AS release_group, ? AS release WHERE NOT EXISTS
      • (SELECT 1 FROM cover_art_archive.release_group_cover_art
      • WHERE release_group = ?));",
      • ocharles
        that's what an approach
      • warp
        (which I haven't tested yet)
      • ocharles
        you can simplify it with a function in postgresql and use try/catch
      • but you can't do that outside a postgresql function without causing the exception to be logged, which i would like to avoid :)
      • warp
        ah
      • I'll stick to this, assuming it works. you can offer alternatives on the code review if you come up with something better :)
      • ocharles
        ocharles has changed the topic to: Arctic week! | http://musicbrainz.org/#devel | Agenda: replication packets (ocharles), schema change progress, nes3 (ocharles), github (ocharles), translations (ian), MBS-2377 DR (ocharles), MBS-3059 DR (ocharles), echoprint "partnership" (reotab)
      • ianmcorvidae
        ianmcorvidae has changed the topic to: Arctic week! | http://musicbrainz.org/#devel | Agenda: replication packets (ocharles), schema change progress, nes3 (ocharles), github (ocharles), MBS-2377 DR (ocharles), MBS-3059 DR (ocharles), echoprint "partnership" (reotab), translations (ian)
      • ianmcorvidae puts myself at the end of the line, everything else seems more important
      • reosarevok
        Not mine :p
      • (and heh, sounds like another looong meeting)
      • warp
        ocharles: I guess we should release today.
      • reosarevok
        Remember to revert the relationship editor first, if it's not done yet
      • ocharles
        i know
      • i'm doing it now
      • warp
        ok :)
      • ocharles
        well that was to reosarevok, but I guess it applies to your comment too :)
      • warp
        haha
      • >_<
      • ianmcorvidae hopes jenkins for master is fixed?
      • ocharles ponders how to revert the relationship editor
      • ocharles
        ianmcorvidae: not that i know
      • nikki
        is that why beta is so slow then?
      • DremoraLV joined the channel
      • ocharles
        possibly
      • but i don't think so right now, cause jenkins isn't doing anything
      • ocharles questions how well this revert is going to work
      • doing it on the day of the release is fun
      • warp
        :)
      • ocharles
        All tests successful.
      • ok, the site will be perfect. yay
      • ianmcorvidae: what exactly am I testing on the edit search?
      • ianmcorvidae
        ocharles: do a search, get results, then try to add a new property and hit search again
      • ocharles
        ok
      • ianmcorvidae
        before my fix that wouldn't work, but it had to do with the new jquery version, so I want to make sure my fix doesn't need reverting too.
      • warp
        hm, perhaps new jquery should stay merged in?