#metabrainz

Hi, all. Is there a recommended version of perl to run a slave mirror?

I guess this is kind of a loaded question. Doc's say at least (at least version 5.18.2) yet many modules wont install with newer versions... https://bpaste.net/raw/1cfd466281c2

ephemer0l: I'm running it fine in 5.22.2 FWIW

reosarevok: I tried debian 8 then moved to testing... all was full of fail. :'-(

cloud provider's ubuntu image wouldn't boot.

What are you trying to run it for?

headphones or something?

yeah

I had it running on another box, but a perl update broke it

Hmm. Then the VM might be a better option I think? It includes the search server, which you won't get if you just set up the server itself (unless you already have *that* running)

I'd sooner run it native. all the boxes are in datacenters

How did the update break it? If updating will break it we should probably look into it, so can you add a ticket at http://tickets.musicbrainz.org/ with as much info as you have? What version, what broke specifically, etc

coroco perl module was not compatible with 5.22 back in the day iIrc

I set it all up last week and I basically just followed install.md

or something that sounds like that module

what host OS?

stable perl is 5.24 on my main box

(I'm not a huge expert, for that you'll need some others, but)

npm and cpan drive me nuts. packages should be packaged by the distro.

Heh

I don't necessarily disagree, but wouldn't that mean you'd have to wait a year for updates on Debian?

um, dont use debian ;-)

Well, you mentioned you were on Debian, so :)

I'm in Lubuntu 16.10

With 5.22

I'm a Gentoo zealot

anti systemd, and all

If 5.24 doesn't work, then we should look into that - bitmap?

I'll see how much progress I make with setting it up on my main box. I'll bring up issues, here.

Ok!

Thanks for pinging me back

ruaok / Quesito: https://wiki.musicbrainz.org/Development/XML_We... says "Non-commercial use of this web service is free, sign up or contact us otherwise". Does that still apply or do we want everyone to register even if non-commercial?

(also, non-commercial links to the live data feed doc page, which is a bit weird when this is specifically not about the live data feed)

reosarevok: i think we'd like everyone to register no matter the use case. Either way we really need to ensure our message is clear and consistent across everything.

That's why I'm pointing that one out :) you should talk to ruaok or something and decide *what* to actually write there :)

But I thought it looked a bit odd

thank you for the catch. I think we may need to go over all the musicbrainz pages, etc to ensure this. Any catch will help--but yeah--ruaok and I are meeting on monday and we can discuss and make a plan. ;)

Yay, plans

:) more verbiage! yay!

(people, don't edit that much dammit! :D )

:P

I think it's important for adoption that a read-only API be usable without needing to register, though it's fine to encourage it.

Particularly if registration implies secrets; open source software should be able to ship with those read-only APIs working out of the box, without the user having to register or the distributor having to find a way to package secrets.

Leftmost:

[listenbrainz-server] paramsingh opened pull request #123: Import time to fix bug in get_listens. (master...import-time) https://git.io/vD7np

good thoughts. I tend to agree. I think it's a balance (although always must confirm with ruaok) of staying true to our open source roots and encouraging people to do the right thing if they are earning revenue uses our data.

Agreed. I tend to think that as soon as you're profiting from the work of others, you should be doing _something_ to give back.

(although honestly Leftmost I'm still learning a lot....) ;)

If you weren't still learning a lot, I'd think you were slacking. :-D

yes. but as a whole we are very transparent--and this is what differentiates us and we need not start down the path of secrets...

:P

I'll take any thoughts on this to help me learn!!!

another thought--if we have messaging on wiki pages that can be edited, how can we assure that our messaging stays true? (besides checking it constantly)

I just mean in terms of authentication secrets and the like. The use case I always keep in mind is an open source personal library app. Users should be able to fire it up and search for music (or books :-D) without configuration.

Well we've never had an issue before with the wiki that I can remember

I still would like to move documentation to github, but

MediaWiki allows fairly granular protection, which could be used for important messaging.

(Though I too want to move toward git.)

is the moving to git under discussion? (i'm assuming a lot has been discussed, that i have no idea about) :P

It was discussed, but ruaok doesn't like the idea much because it makes it harder for random people to fix errors

ah

Leftmost: where are we on BookBrainz? how is it going?

Leftmost: ignore previous q-I'll ask a more defined question some other time--that was super general.

Hi everyone,

I am Nupur,

A second year undergraduate studying Computer Science at Netaji Subhas Institute of Technology ,Delhi

I wanted to contribute to the project “New Machine Learning Infrastructure” under Gsoc this year.

I have acquired basic knowledge of the skill set required for the project ie "Python, C++, Machine learning, scikitlearn, postgres”.

I have been studying algorithms and data structures, programming in C++ since last 1 and half years, and have Coursera certification too. Further i have attended Machine Learning workshop and made a few basic projects implementing functionalities of the scikit and numpy libraries. I have also been working with django, flask and building chatbots at the messenger platform.