#metabrainz

weeksio, Not the view I prefer but it will work.

Now just to add the imports and some log statements.

Freso: looks like a hack on their end, it's not fully integrated yet, but new api accounts and keys can be generated now.

also in solr-5.3.0/server/logs i think

opatel99: Re: screencast, I'll /query you in a sec.

opatel99: Re: essentia, did you see the Gist with my own old notes about an acousticbrainz beets plugin?

In particular, http://www.last.fm/api/accounts only shows api accounts made prior to the migration, but http://www.last.fm/api/account/create allows creating new accounts, with a note saying that you need to save the credentials shown on the submit page since they're not available elsewhere :)

kepstin: 🙌

kepstin: And yeah, I'll poke OpenDNS about it too.

weeksio, I made the code changes.

opatel99: ^ can you check all the other *Brainz sites one more time? Just to make sure. :)

I am almost done fixing the imports. Are you able to recompile and deploy the code for me?

sure

That should be correct. Give it a try now.

You are building from mb-solrquerywriter in your /home, correct?

yes

MBXMLWriter.java:[60,67] cannot find symbol

What symbol? I probably made a typo. :)

MBXMLWRiter

r not R

:P

There should be more than that. Can you paste service the output?

It doesn't like something I added.

mb-solrquerywriter/error.txt

That is actually a better solution. :)

Freso: musicbrainz is only one not working. Will look at gist

weeksio, Try it now.

What command are you building with? mvn clean install or something else?

opatel99: Ah, looks like I never wrote this down. Re: "Auto-fetch extractor static binary, or just provide d/l link?

"

Anyway, my thought was to simulate what the replaygain plugin does, which also calls an "external" program.

CJ_: check the error file again...I overlaid it, but it seems to be the same

(Dependent on which format it's replaygain'ing.)

weeksio, Add -e and run it again, please.

oh yeah...clean

mvn package

Are you running it now? I can try.

opatel99: And the path to that being provided in the beets config file (possibly falling back to checking for the _extractor in $PATH).

But look at the replaygain plugin to begin with, as that's in a somewhat similar position in this regard.

mb-solrquerywriter$ mvn clean install -e > error.txt

weeksio, Okay. I sudoed to your user and will run it myself. Thanks. :)

kay...I'm outta here for now

weeksio, Thanks.

no, no....thank *you*

Oh yeah, meetings have started again

Woo! Meetings!

i've a ton of cd's that i burned and the burn software used accurateRip tags... MB doesn't like modifying these files... any info?

I was right. It was a typo staring me straight in the face. Doh!

anyone here know about picard and whether it supports tagging with non-numeric track numbers?

LordSputnik: https://github.com/stasszczesniak/bookbrainz-ws...

darwin: huh, i'm not sure. You might run into problems with mp3, since I think the id3v2 track number field is numeric format...

kepstin: right, but for the containers etc. that are supported.

i wonder how many players would break if you put something non-numeric into vorbis/flac/whatever

might not be worth it even where it is possible

(it's more correct.......)

and probably players don't support it because they haven't had to in the past, etc.

sure, but players use that field as a numeric value for sorting tracks on an album in order

if anything, it would make sense to put a presentation track number in a different people and e.g. foobar2000 users can make a custom display that shows the alternate field

in a different field*

LordSputnik: also http://tickets.musicbrainz.org/browse/BB-162?pa...

(that could work in mp3, too, since you'd put it in a TXXX)

well, I don't disagree that id3 is useless

because it is poorly specified

but

not that much to meet about though, I've just been trying to figure out a way to migrate data to the new BB schema, and guiding GCI

fwiw, storing non-integer values in the track number field would at a minimum break all gstreamer-based apps and rockbox (they'd parse the track number as 0 or missing)

and it would almost certainly break many other apps :/

so yeah, that's pretty much a no-go.

so yeah, best to store a 1-based position number in the track number field, and store the presentation/display track number in a separate new field.

otherwise you're just gonna annoy people.

Doesn't really make sense to store a non-integer in the track *number* field anyway

LordSputnik: well, sure, but we support that in musicbrainz for stuff like vinyl/cassette numbering with sides and whatnot

but musicbrainz also stores a separate (hidden) position field that's just an integer

LordSputnik: Update : https://github.com/stasszczesniak/bookbrainz-ws...

The previous commit was deleted.

stanislas: why add a space before \ ? Is that really what PEP8 says?

checking now

Also "== False" and "== True" are correct in this case - they are interpreted in a special way by SQLAlchemy

5 minutes, everybody.

Grab your tea and cookies now!

(you could add a comment for that so that future contributors don't come along and "correct" it!

hmm, tea and cookies? sounds good

Freso, I have cookies but I need you to provide me with some tea.

CJ_: I don't even have any myself. ;_;

LordSputnik: 1) It is not in PEP8, it was changed automatically, maybe my own restrictions are wider than PEP8

Freso, Well, at least I have something to report in a meeting. :)

LordSputnik: 2) I will suppress it to keep it out of checking

LordSputnik: (and of course revert the changes)

<BANG>

CJ_: Go right ahead. :)

(Who has the OAuth/Discourse topic?)

Okay, I'll go while we wait for CJ_.

I started looking at the solr setup. SOLR-2 specifically.

Oh, okay. :)

Freso, I was typing. :P

I added some logging code but the server is configured in such a weird way that I'm not sure how to get it deployed to test.

I will get with weeksio on that later. I also plan on working with him to iron out the configuration into something much more manageable.

Do we have any doc about the deployement ?

Not that I know of.

And if there is, I would wager that they need a decent bit of working to be correct.

We are moving to use Chef for all deployements now, i can help on this if i can get basic steps

Nice. I am only familiar with Puppet.

Maybe you two should talk more after the meeting? :)

CJ_: fin?

Freso, zas and I? I do not think we are at a point to make a chef build yet.

Yes, I believe that is all.

zas: Go ;)

Well, weak week, mostly partying ...

🙌

but i improved the write speed on tails, and discovered that our lsi mptsas cards are quite slow

enabling write caching + few ext4 options improved things

something like +40%

tails is our stats server, it needs to write a lot (all metrics from all servers)

i removed the encrypted swap on jiji

so we now have swap stats for it too

i also improved gateway cookbooks, ie. nginx conf is now tested before deployement

and as usual a bunch of syswiki updates

various cookboks fixes/improvements

and performance tuning

i passed quite a long time on tuning kern/tcp options on gateways

that's all i think

Alright. ruaok is on his way. I guess I'll go then.

I... mostly did GCI stuff, but not a while lot. GCI seems to have quietened quite a bit compared to the first two weeks, so I actually have time for other things now. So I've caught up with some other private projects, had family time, friend time, etc.

*a whole lot

fin, I think.

OH.

Since last meeting, ruaok has appointed me to be in charge of these meetings.

That might be a relevant note here too. :)

ruaok: Go. :)

\ø/

What is GCI?

CJ_: Google Code-in

Mostly the same as above for me, except I spent half of my Christmas "holidays" in bed with a cold because I have perfect timing

Ah, right.

anyways...

my main focus of the last week was tacos.

^ :p