#musicbrainz-devel

But I imagine that doesn't affect your editing too much

well, what happen to languages other than english is that they'll do the same thing as they did pre-ngs

ocharles: qunit.t and selenium.t both pass here.

kepstin-laptop: yeah, that being "randomly being turned into feat. or not"

Which could work for tracks, but is annoying for recordings

We'll see

warp: excellent

final test run here

excellent all is good here too

warp: lead the way!

alright

ocharles: ok, logged in.

ocharles: all traffic to pingu.

is the closure compiler down or something?

compile resources is failing with 500 :(

yikes.

ok, it's back again, good

astro is back up

perhaps we should switch to using a local compiler at some point.

ok, all traffic to astro.

man, pingu is SO slow atm

only 8 workers have started up

ocharles: right, next upgrade we do pingu first.

we tried that before

it didn't make much difference

ok, pingu back

drat.

ok, both should be getting traffic again.

ok, all done

I can't tag because Rob still is the only one with the signing key

(for mb_server, not in general)

nikki: hi

luks: when you have a distributed system, presumably you want to be able to trust the version

bitmap: is it intentional that dragging files back from the right panel makes clustering use the release name from the right panel rather than the actual tag release name?

it's not got a huge amount of value for musicbrainz though

nikki: hmm, no that doesn't sound right

I noticed because I have some files here with only track ids (don't ask why I removed the rest, probably in a fit of madness), so they load as [other tracks], then when I try dragging them back and clustering, they cluster as [other tracks]

okay, I'll look into that :)

ocharles: exactly, it doesn't provide much value, and it continuously causes problems

I'm not really fussed about the fact that the tag is a little late

if somebody has issues with trusting their tags, they can just check git.mb.org

which doesn't prevent man in the middle

and dns spoofing, or whatever

oh, we are *that* paranoid :P

no

but that's what they are there for

*someone* might be paranoid

the issue is getting the key through a side-channel, otherwise a man-in-the-middle could give you a different key to check against.

just because we aren't, doesn't mean we shouldn't do it

the public key*

I don't think signed tags have any purpose for anybody except Linus Torvalds

especially for schema change releases, having a tag early allows customers to upgrade early

which is IMHO better than upgrading one week later

we have never been a week late

well, I didn't expect to convince you, it just seems funny to me :)

soundunwound stuff shipped

which version?

separate section

hrgglcgllmfrs version

agreed by ruaok?

"SHIP-IT. Whatever-the-fuck-it-is. Just fucking ship it."

is what I was told

heh

oh well, onwards we can move now

I'm expecting some userscripts to appear :)

did he add classes or ids?

no...

but he said he was going to

so does that mean we shipped something he wasn't finished with?

I think the branch wasn't up to date

so what happens then?

I fix it when I have actually finished dinner :)

but it's not very urgent the feature works fine

oh but you have time to close tickets while I'm trying to :P

:D

ocharles: any idm/ambient recommendations? :)

Hmm! Spacetime Continuum, Jan Jalinek, Pixel, Oneohtrix Point Never, Carbon Based Lifeforms, Senking

Oh, "Retold" by Nest

awesome, thanks! :)

luks: do you remember what the differences between orig_metadata, user_metadata, saved_metadata, and metadata in file.py were? it looks like they're being used incorrectly in a few places

not really, sorry

okay

except for just "metadata" the names seem fairly obvious.

but my guess would be: orig = from the file, user = from the UI (can be changed), saved = saved to the file (not sure why though)

that's my guess as well

so in a few places I think they're being used wrong, I'll have to experiment

I'm not sure about saved, because it seems it should be same as orig after save

yeah

there's also server_metadata which doesn't appear to be used anywhere :)

bitmap doy

oops

bitmap do you write any new metadata fields with the changes you've made to Picard ?

If so, can u please add to https://spreadsheets.google.com/spreadsheet/ccc...

I don't think there's any new tags actually, but I'll make sure to add them to that file if I'm wrong

tkx

woah, next is wide.

yea, takes a bit of getting used to!

nikki's clustering issue works fine after removing saved_metadata altogether. I'll just do that since it only seems to be causing trouble, and should be identical to orig_metadata anyway…

ocharles: the "Link Type" edit searches all crash for me on test.musicbrainz.org

(tried "collaboration", "member of band" and "wikipedia")

warp: yes, the plperl stuf isn't installed

ah

your instructions are good though

should I install that on hobbes then? :)

oh, I remember the problem

hobbes has PG built from source

doh

I had other things to do than fiddle recompling it from source

yeah, indeed.

if you want to, you can though...

otherwise I'll do it tomorrow

in that case I can wait. I don't normally build postgres from source, that's more rob's turf :)

:)

ocharles: you're going to merge http://codereview.musicbrainz.org/r/1356/ , right? or were you talking about a different branch from pecastro earlier today?

argh

ocharles: that's not quite an answer to the question ;)

it answers it I think :P

it's not been merged

argh means "yes, but I forgot" ?

it does indeed

ok, np then.

ocharles: http://codereview.musicbrainz.org/r/1284 has been on code review for 2 months now. is that still just waiting for a shippit or does it need more work?

yes, it still needs more work

yay.

ocharles: no real hurry, but what's your status re: reviewing my rate-of-change graph stuff?

since I'm thinking about it

not really started a review on that

okay, cool

I wanted to look at reviews more today, but I got distracted staring at BBC news :P

huh, we require plperl now

si

well

only for next

well, then soon enough