#musicbrainz-devel

that was many many many strong cocktails ago.

haha

seems almost irrelevant. :)

<3 nola

:)

free cocktails too.

we just realized that the cocktails you have to pay for are weak.

kinda odd, but hey it was fun while it lastest.

hospitality and alcoholism, louisiana specialties

that said, I've spent a lot of time on hosting related issues.

working with ocharles on deadlock mitigation.

Was wondering what had happened :p

and doing a lot of catching up on minor tasks. I do feel a lot more caught up on that stuff.

ocharles: did you already speak to deadlock issues?

as in part of my review?

in general.

should we put it on the agenda?

ruaok has changed the topic to: Agenda: review, github (ruaok), importing IA data (reotab), deadlocks (ruaok)

I did lightly touch on it, but there's not a huge amount to say there yet, so feel free to talk about it

ok. lets cover it later then.

thats my biggest concern for the time being and I would like to see if we can put that behind us this week.

thats it for me!

should I take over the meeting warp or would you like to contiue?

+n

ruaok: I can continue.

ok. back to you then.

ruaok: i'd certainly like to get it squashed too, esp. as it seems to be getting progressively worse :/

anyone else for review?

reosarevok_, nikkimini?

ijabz?

no don't think so

Nothing too important from me, apart from the IA thing later

ok, let's continue then.

ruaok: github?

right-o.

ocharles pointed out that our github migration was not complete.

we need to announce it and decide exactly what should happen with git.mb org

but all the projects except picard (as intended) are now under the metabrainz account

next we need to have our server mirror all the github repos for backup purposes.

then I think we should just shut down git.mb org

ruaok: +1

i can look into that if you want

which means the mirrors are just for backup, and not otherwise publicly available, right?

what else needs to be done to complete this, so that ocharles can stop complaining (about this one thing)? :)

ocharles: plz

warp: yes.

github serves all of our needs, backup is the only thing I am concerned about.

in general, we should consider how to divest ourselves of all the services we run that are not core services.

ideallly we should only host MB, MB sandboxes and our own wiki. (for wikidocs)

everything else, should ideally go away.

need to make sure issues are turned off for github?

mailing lists, forums, etc. off to best of class services that are willing to do it for free for us.

reosarevok_ has changed the topic to: Agenda: deadlocks (ruaok), importing IA data (reotab)

alastairp: as we already have that repo, i'm pretty sure that's all done

ocharles: ah

good thing to check, for sure.

Speaking of github, ruaok, did you talk to VxJasonxV in the end?

ruaok: we're having quite a few issues with (our deployment of) code review as well.

reosarevok_: yes, I did.

ok

nikkimini: good point.

warp: would you consider pull requests to be complete enough to replace CR?

and this isn't something I want to get done tomorrow,but over time.

it's a little less feature complete, imo

alastairp: agreed.

unless you use social constructs instead

alastairp: I'm still unhappy with not being able to get emails from github for pull requests

pull requests really aren't sufficient IMO

(e.g., a shipit tag, rather than something built in)

nikkimini: right. that only happens if you're not involved in the discussion

alastairp: I encounter so many issues working with code review, I certainly would except github issues to be an improvement for me personally.

like, you get the first one

alastairp: er, github pull requests.

but not if you don't participate

maybe github will evolve to be as complete as CR.

if we can host has many things on github as possible, we can have a single login. :)

ruaok: I get the impression that they keep it simple for a reson

so you can bend it to your will easier

alastairp: maybe they'll stop being wise. :)

I don't think they'll ever get an explicit ship it flag, for instance

but, lets not get mired in the details now.

(we should probably encourage smaller patches anyway, those are easier to review. I'll be guilty as charged with my /ws/2 json work ... )

sure

alastairp: we have discussed pull requests before

and yea, doesn't quite do it

just as a general thought for us. less hosting == easier and cheaper.

ok, so. turning off issues on our projects, backups and an announcement.

apparently atlassian offers their 'ondemand' cloud-hosted jira for free to open-source projects

anything else to finish off this topic?

kepstin-work: that would be fantastic. :)

if they can import our exitisting DB, I'd be very happy.

ok, I'll do the blog post and the issue turn off checking.

that'd be great, does anyone volunteer to look migrating jira? :)

+at

ocharles: once you have a clue how to do backups, let me know and we'll get those in place.

...

heh, I'm starting myself a free trial now

haha

I'll see if there's a database import at least

let pick this up next week then.

ianmcorvidae: great, let us know

ruaok: sure thing

thanks everyone.

hmm

over to warp/reotab

ruaok: we haven't talked deadlocks yet, have we?

We only want to move it if it doesn't block single sign in, I guess?

no.

I thought I changed the topic.

crappy hotel wifi.

lets do it after IA

ok

reosarevok_: go!

Well

Fiiine

warp has changed the topic to: Agenda: importing IA data (reotab), deadlocks (ruaok)

So, the IA have a very big set of music

oh, sorry. didn't see that was intentional.

And they are trying to map it to MB

They've done a lot via acoustIDs apparently, but now they could use some help in mapping the rest and importing what isn't there

and in a similar vein, smaller labels are starting to offer us data.

We're talking thousands to tens of thousands of releases, though, so fully manual won't work

So, mostly: who has any ideas?

so increasingly we need to have an ingestr system setup where we can get make this data available to our community.

and working on these systems without the data first, isnt going to work.

They have metadata, which is supposed to be fairly good since it comes from a community of annoyingly pedantic fasci... I mean, users

:)

so I am hoping that when people give us data, that people will get excited to build tools to use the data.

lol.

reosarevok_: I would say that we should ask them for the data.

in a DB dump format if at all possible.

and then we should host it and make it available.

for other people to build such tools we need an editable webservice though, which needs NES.

once we have data, people will build shit.

s/needs NES/we've decided not to do until NES is done/, let's be clear

warp: we can build simple tools to do RE seeing in the meantime

ruaok: a bit kludgy, but true.

it *a* path. not *the* path. :)

:)

ruaok, last.fm did give us some data, but its still not been made available to the general community

so, the answer is "give all your dataz!"