#musicbrainz-devel

Leftmost: if you want to do more than add a check now and then we should whitelist your ip so you can run the VM at home and test out your changes there.

also you'll need access to the private nagios repositories on github

I have access to the repo already.

oh good :)

nikki: A Picard site?

What are we brainstorming on/about?

gsoc ideas

oh well, I added it :P http://wiki.musicbrainz.org/Development/Summer_...

"Actually do something with MusicBottle" >_>

Freso: isn't that one already on the list (one should mention that this is started as "MusicBottle" though

I don't know. I haven't looked at the list.

Can anyone tell me which of (number, position) is the number (integer) as per TOC and wich one is an arbitrary string like "A3" (for vinyl) in WS/2 (xml)

position is the integer

thx

(or should be)

I was kind of confused since having "A3" sounds weird for number and position, but I didn't find an example

as you said, position is the integer. Example: http://musicbrainz.org/ws/2/release/edd0a17e-3b...

JonnyJD_: Yes we used to only have the track number and that was called position which made sense, but then when the user defined stuff like A3 was added it had to be given a name and it was given 'number'

which I didn't like at the time

I think there was a bit of discussion about it

The other way around would have been nice, but that doesn't work with the surrounding names

it's the number in the same way that catalogue numbers are catalogue *numbers* despite often having letters in them :P

Would it make sense when I try to find a place to put this information in http://wiki.musicbrainz.org/Development/XML_Web...

whereas position is simply the position in the list of tracks

nikki: true dat. But I am used to cat#..

nikki: that is actually a good way to remember it though (going with cat# ~ track#)

maybe trackno would have been clearer

ijabz: why would that be clearer? That is just short for track_number.

I also have no idea what would be a better name and we shouldn't change the API anyways

because 'number' doesn't make it clear we are talking about 'track number' it could mean something else like 'number of edits'

obviously we can't change the api,. I wasn't proposing that

Hm, I can't find a good place in http://wiki.musicbrainz.org/Development/XML_Web... to put that information.

A human readable spec of the returned xml data is missing anyways.

that whole page needs redoing /MBS-5307

But that is neither human-readable (in the actual sense), nor is there room for descriptions

thx for the link

ianmcorvidae: is it possible to host a node.js application (WS/2 documentation) on MB servers? (MBS-5307)

heh, so I'm not the only one who doesn't like our docs

nikki: there is also a similar ticket for pymbngs, since that is basically a pass-through of WS/2: https://github.com/alastair/python-musicbrainz-...

And I personally think, python-musicbrainz-ngs is only usable, when WS/2 gets better documentation.

That is in contrast to python-musicbrainz2, where everything is wrapped and a good documentation of the wrapper was available.

I ended up making https://wiki.musicbrainz.org/User:Nikki/ws/2 (well, starting to, I didn't finish before I got sick of jumping through hoops to get the ws to return the info I wanted) because I could never make any sense of the actual ws/2 docs :/

You might want to link that wiki page in MBS-5307

nikki, https://wiki.musicbrainz.org/User:Nikki/ws/2 is just I wanted and similar to how I do the search page couldn't you merge that in with the actual page to make a new page

'ola

the issue is the way the current doc has been written rather than the software being used (wiki etc)

buenos dias

ocharles: re:rate limiter being applied to home network as one computer, if user is unfairly being hit here I suppose its not unreasonable for them to use a VPN to get a different ip per computer

yep, they could do that

or they could not run taggers at the same time in several computers and wait a little bit :p

(I mean, not that I'm against them using VPNs, it just seems like too much effort for little gain)

reosarevok: this affects people on larger networks

Im finding a user might have purchased SongKong and Jaikoz, using SongKong for mass matching a set of songs, and Jaikoz for fine editing of a set already processed

e.g., a network with different people who want to tag their collection

there is a legitimate problem

ocharles: hmm, but do those usually still expose only one IP?

why shouldn't they be able to use both

ocharles: I mean, home networks, sure, but...

reosarevok: depends how good the network admin is :)

ijabz: if they're the same user, why should they get double the amount of hammering only because of using two programs? :)

(if they're several users under one IP I agree though)

I mean, it's not like we could differentiate them anyway, but still :)

So far I like Bitbucket, but it bothers me a little that I can't see (or can't figure out how to see) comments on previous diffs.

Leftmost: you click the comment icon

next to the file in the file list

i also think they show up in the "Activity" tab

Aha.

Because they have bought two licensese (and for each license paid musicbrainz for use of the web service)

Didn't catch that, thanks.

ijabz: oh, true. That sounds like each license should come with uses of the unlimited paid ws thingy then? (wonder how the development of that is going...)

yeah , well athst I what thought but unfortunately (for me) access to that is going to be an additional cost on top of the %10 already paid

i think the unlimited paid ws thingy is vapourware

Well, since the license is a one-off, it probably shouldn't come with unlimited uses. But I'd think it'd make sense to have enough uses for tagging a reasonably large collection once

(included at no extra cost)

ijabz: I think the ratelimiting per IP should only apply to unauthenticated users. authenticated users should get limited at the account level, and possibly have more relaxed rate limits so they get a benefit from being part of the community in some fashion.

heh, I know work *has* been made in it

So I assume it will be released at some point, but dunno when :)

warp, sounds like a good idea

ijabz: vapourware? why?

maybe I haven't got the defn of vapourware right, but you worked on this a a long time ago and nothing ever came of it

Ive kind of given up on it ever happening

vapourware is constantly hyped but nothing will ever be delivered

that's not the case here - there's just more important stuff on the roadmap, like schema changes

and we are massively resource constrained

ocharles: isn't some or all of the 3scale related code in secret repositories?

Schema changes are important at the moment but f this work had been finished a year ago it would have provided Musicbrainz with some much needed extra income

I would say its more important tha most of the work done over the last year

you should talk to rob about it then

maybe that extra interest would help him understand where to schedule it

Im sure he is aware of that

well, you're a customer - if it's something you want you should apply more pressure :)

If I remember correctly, ianmcorvidae has most of it done but it was waiting for some specific stuff

But I might be mixing projects :)

Hmm, ocharles, just had a thought and maybe you know the answer. What would happen if I tried to merge a CD release with a discID into a vinyl release?

Would we end up with a dropped discID, a vinyl with a discID, or would it fail the merge?

reosarevok: probably a disc id on a vinyl release, but I can't remember

:(

I mean, not that it should happen while I am checking all destructive edits, but still

reosarevok: I don't have much free time at the moment :(

ocharles: wow, that is a fast shippit.

:)

Here comes the first 500 of the day!

502

502...

on beta?

Prod

not on there yet

it was merged in beta on monday

Oh, ok

Seems to be back now...

Or maybe not

Yeah, back but slow

Nah, 502 again

warp, ocharles: no idea what causes these I guess?

504...

:(

Seems it works again

Do all the reviewers get an email when I update a pull request on bitbucket?

navap: probably.

Hi all, I'm trying to run a Musicbrainz mirror using nginx as a proxy to the FCGI process and it seems to work fine for the most part but I'm seeing lots of "[info] 10346#0: *344153 recv() failed (104: Connection reset by peer) while sending to client" in the error logs. Any ideas what causes this?

(I assume it's the MB CGI failing, but I can't see any errors from Musicbrainz)

warp: do we have a branch with a real cpanfile in?

ocharles: I have one, but it's not finished.

ocharles: github warpr/cpanfile

ok, I also have one that is finished :)