#musicbrainz-devel

full-import looks to be going ok

Project musicbrainz-server_master build #473: SUCCESS in 7 min 6 sec: http://ci.musicbrainz.org/job/musicbrainz-serve...

* Ian McEwen: MBS-6690: pass along stash/opts to serialize_entity when inc is passed

JonnyJD: You could release libdiscid 0.6.0 during the summit and we could have a release party. :p

lol

I guess the MBS release isn't the monday right after the summit this year

that's probably for the best

though it *is* right after rob's wedding thing, so I guess it's good ollie tends to do the releases :P

I could also release directly on the thursday before. Still a reason for a release party ;-)

JonnyJD: Or do a 0.6.1 release; change log: "no changes, just released for the party!"

Or just make ianmcorvidae release some Geordi thing

That thing goes through versions like they're candy anyway

you'll have to make me actually finish something first :P

and I mean, geordi doesn't have an actual version system, so :P

celebrate whatever is the latest commit hash to musicbrainz-server

hey, it says "version X of the mapping"

works for me :p

oh, heh

well, it does have a whole bunch of different versions of various things

hotfix released

Seems to work fine :)

time to revisit lolo now

504

lolo just in?:p

was in

http://www.salon.com/2013/08/22/why_using_the_r... <--

dropping the weight significantly looks ok

freedb.mb vacuuming after me finally getting to updating it for that pesky schema change we did three months ago

(slow to get to things? me? never)

ffs last.fm stop transliterating my japanese tracks

I should really get around to writing something I can use instead of last.fm at some point :/

as long as it doesn't interfere with doing mb stuff, do it! :D

chrisdone has already written a personal scrobble server

I saw something kuno was using to do something like scrobbling to identi.ca

which talks the last.fm api

I hate the last.fm API is the problem

if you don't care about the social aspect

right

I mean, now that we have track MBIDs again it's back to being roughly tenable, but

then you'll have write plugins, no?

half the plugins are broken anyway

imo

which is possibly our fault since we killed track IDs and then brought them back

but

there's a ticket for that :)

piggybacking on an existing thing that's more social-network-oriented isn't a bad idea

to a degree

it seems like pump.io has better ways of dealing with events that aren't outright "posting a status" type social events

but with adjusted priorities maybe more "cutting edge" work can be done alongside mb-server

which is why I'm curious what warp's using

(ref: sidebar on https://identi.ca/warp )

anyway, ocharles, chrisdone's scrobble-server thing does seem like it could be a reasonable start if we figure out where our track IDs go in tags and I can coerce mpdscribble to use that insetad of a recording ID

ok, no [error]s in lolo's log for the last 30 minutes or so and asterix is completely out

ianmcorvidae: hacks

think we're good there!

ianmcorvidae: but pump.io is pretty nice to develop clients for once you get past the OAuth stuff.

ocharles: sweet :D

kuno: well, can't possibly be as bad as last.fm's quasi-oauth that nobody ever uses (since the legacy protocol is better), at least :P

ianmcorvidae: http://gitorious.org/warpr/relatedmusic/blobs/p... and http://gitorious.org/warpr/ppp

nice, hardcoded AAC exception code

(yeah, yeah, hacks, you warned me)

heh, I have no idea how I would use an alternative thing to scrobble to

what player do you use anyway?

cog

and a version of the last.fm scrobbler which attempts to convince me to upgrade it every time I start it

well, it looks like at the very least the chrisdone thing ollie was mentioning can figure it out when you just proxy

though the official lastfm scrobbler possibly uses their weird quasi-oauth second-gen API that nobody else bothers with

most open-source players have made it possible to switch the domain that the scrobble api connects to; but you can make any player use an alternate by simply making the last.fm api domain go to a different box by editing your hosts file.

whoo, freedb.mb now post-schema-change

now to catch it up on three months of replication packets

ianmcorvidae: that's probably an issue with music player daemon, perhaps it is now fixed.

kuno: which one, the transliteration?

kuno: that has to do with the fact MPD scrobbles the recording ID and last.fm has the wrong name stored

the transliteration is probably because you have the 'fix my tags' option enabled on last.fm

if you turn that off, it should unfix your tags

ianmcorvidae: not being able to get an MBID out of .m4a files.

oh, do they have an actual option now

kuno: oh, yeah

ianmcorvidae: huh, it's two separate options now. There's "translate artist names" and "auto-correct spelling" and you want to disable both :)

yeah; only the auto-correct one was on

we'll see if it works after this scrobble

my last.fm history has some horrible misencoded messes thanks to an accident with my ipod scrobbler :(

interpreted utf-8 as latin1 one time when I was submitting a backlog

heh, nope, didn't fix it

it's the MBID thing, I suspect

hmm, the mbid stuff should only be turned on if auto-correct is enabled

maybe give it an hour or two to propagate the setting ;)

yeah, possibly

heh, 808451 rows in dbmirror_pending

so, no decision on how to store track mbids in files yet? :)

not that I've seen

right, so MUSICBRAINZ_ELEPHANTID it is, right? ;)

'course I was always fine with "the current tag is broken anyway, so let's call track IDs track IDs and recording IDs recording IDs", but I guess others aren't :P

since track ids have a superset of the info of recording ids :)

yeah, exactly

I mean, there's an argument for "people who consume MBIDs should resolve their type themselves" but that's maybe not something we can impose :P

would be helpful if we had a webservice that let you resolve the type of an arbitrary mbid :)

(certainly not without a non-internal webservice for that)

ocharles: assuming you're still up, I'm getting lots of emails failing to send to logster [at asterix]

ocharles: I assume this is related to something you're doing, so in case you aren't aware

ianmcorvidae: yea, I'm upgrading asterix

k, cool :)

well, forward me a logster related one

i'm only seeing nagios

okay, sent your way

Project full-import build #96: FIXED in 4 hr 30 min: http://ci.musicbrainz.org/job/full-import/96/

* Oliver Charles: Merge remote-tracking branch 'ian/mbs-6690'

Do we still have "modbot lag" of >5 minutes? I thought that was cleared up some time ago?

I haven't heard any reports in quite some time

the packets usually come out at like 2 after nowadays, anyway, so

https://musicbrainz.org/edit/23349418 has 0:4 Y:N

Hm.

edits with no votes don't close immediately any more

And has had 3 No votes since 2013-08-22 02:00 CEST.

it needs to be 72 hours from the first vote for it to be closed

Ah. Alright.

it'll close at 19:00UTC today, presumably

so, three hours :)

Yeah, it's designed to let people respond to no votes

Yeah.

Not that they always do :)

not that they're likely to do so, in this case, or so it seems, but :P