#musicbrainz

[16:22] nikkiI really really really want to paste the tracklist first

+1000

thomasvs: rock bands usually use selftitled, unless they specifically claim untitled

A fair amount of electronic singles, for example, are untitled...

(but there we use Track Title / Track Title unless tracks are *also* untitled :) )

It's kinda tricky. What does the artist use on their site or whatever?

nikki: it seems to me that the primary purpose of discID is to find the correct release metadata for tagging. So it’s just a key (with some additional utility), and forcing tracklists to match it is excessive.

FWIW, there’s no reason we could not have a different duration calculation applied for VCD medium discIDs…

I know that some people, rob in particular, want to just get rid of disc ids

hawke is terrific

since we already have lookups that use track times

WHAT ?!

I only wish I had written my card index numbers into my tags when I started this whole exercise. That would provide exaclty the kind of unique information I need in my folder names :-(

OMGWTF !

My only objection to that is that disc IDs are a solid source of track duration info, at least for CDs.

so the idea of adding more features for disc ids seems funny

and we are short enough on that already

it’s the only thing I use to tag my CDs before listening

(on computer of course)

(Ironic I know given that I want to loosen the coupling between tracklists and discIDs)

nikki: dropping TOC is terrible idea

hawke_: well, to some extent. you can already get different disc ids from the same disc by putting it in a different computer sometimes :P

nikki: Sure, but not *drastically* different

aw man, I wish ther was asimplified picard webapp thing, so i could spendtime in lecutres re-tagging my usb thingy full of music

at school there is weird proxies so picard don't work

CatCat, Use an ssh tunnel ;-)

Munger: a) i have no idea how to do this b) it is likly that the overly-restricted school machineds won't allow this

I'd love ot grock ssh tunels thoguh, i could accsess my home machine from anywhere?

CatCat: in what way does picard not work?

picard doesn't do magic - it's just standard http requests that return xml.

yeah, hence my question

CatCat, Proxies generally don't touch port 22, which is the ssh port number. You set up a tunnel on your home machine, and route all traffic for certain port numbers on your school machine to your hoe machine, which then passes stuff on for you.

well it's not that picard doesn't work, it's that the school proxies are so complicated that I have no clue how to make it work, you need ot link to some weirdass file or what not

probably an automatic proxy script…

indeed

next summit, i will harass one(or more?) of you to explain with examples how to do this using my and another persons machine

CatCat, http://www.revsys.com/writings/quicktips/ssh-tu... - I used to access IRC from my old company by tunneling through my home machine, hence defeating their firewall

but thank you

i will read this

A better one - http://www.rzg.mpg.de/networkservices/ssh-tunne...

hn

i am more basic

ooh the second one is interesting, thnaks

i'll still need to harass one of you

but hey, the more i learn the more i can be of use to musicbrainz

lmao, imagine catcat as a developer one day :P

(everybody run away from bad code)

\o\ o /o/

Munger: If I was tunneling to my home, my home IP would get end being ban and the result would be worse :)

Why I don't switch ISPs -- the small, local operation, even though more expensive, has helped me troubleshoot various connectivity issues, even to a bittorrent tracker.

jesus2099, All the company sees is that you are communicating with an external machine through port 22, which is encrypted. They have no way of knowing what traffic is going through the tunnel. So you pick an unused port number in Picard for musicbrainz.org. Set up your laptop to forward all traffic to that port to your home IP:port 22. At the home end of the tunnel you catch all traffic and send it to the correct machine/port

Even if the company was doing traffic analysis to try to determine what kind of traffic was going over the encrypted connection, it should appear no different from someone using HTTPS. At least, I would think so given you would be talking to a web server.

well, https is somewhat detectable since it uses standard ssl/tls setup and usually is on a specific port

the connection setup for ssh looks different

Doesn't even have to be port 22. Anything that the proxy/firewall has open that you aren't using for anything else will do

but yeah, just run ssh on the https port, and most scanners will just pass it through without checking

kepstin-work, Because the whole thing is wrapped up in ssh, they cannpt read any of the real headers

Munger: there's a certain amount of stuff sent unencrypted protocol-wise at the start of an https connection that is sufficient to detect that it is probably https.

kepstin-work, Yes, but it is all hidden inside the SSH connection - which is by definition secure and encrypted

Munger: yes, but ssh doesn't look like https - so a sufficiently advanced proxy might go "huh, this isn't an https connection! no going through!"

on the other hand, most don't check that.

kepstin-work, The point is, companies seldom block SSh as it's essential for remote maintenance

sure, companies don't... but other places do.

companies that primarily work with windows servers might :)

Munger: port 22 is blocked

it's really annoying, one place I used to work at had the default git port blocked

Run an sshd on a higher port then

is it? I guess anyway ok I just don’t go where I’m not supposed to

or port 80, if you can :)

you had to use the old slow git http method there, which was annoying

(ssh worked though, so I used that when I could)

Even the Great Firewall of China didn't block port 22 last time I was there (a year ago). that might have changed tho.

a company is smaller than country it blocks more stuff everything unproductive of course

it’s normals

reosarevok: what do you think about the “Spanish Classics” part of this: http://beta.musicbrainz.org/release/e8e5b699-bc... — is it part of the title, or should it be something else?

Don't use port 80 - that will screw up all your normal web browsing. Port 992 (secure imap) might be a good choice

hawke_: that's a series

hawke_: it’s a series name… in lack of series support I sometimes add it in disambiguation comment

Munger: you can really just pick any port above 1024 or whatever…

Well

A "collection"

(as in, it's not ordered etc :p)

Yep

Munger: port 80 for your remote sshd would be fine and not mess up anything, and might punch through firewalls that would leave port 80 traffic it doesn't understand alone

hawke_: I'd just leave it off the title though, and well, in other cases I'd suggest a wiki page

But this is Naxos

And Naxos has series sites

So... can wait until we get some series support :)

(man, I want series support :( )

+1 to that

brian_b, Also true, as you aren't forwarding port 80 to the tunnel. I tend not to do this as I have my own web server running on port 80, and it *would* mess that up

Munger: true.

there's also ways to tunnel TCP/IP over DNS, if you're truly hosed but you can do is DNS lookups

that one's fun, because it can sometimes get through paywalls and give you free internet :)

(newer services tend to intercept dns, tho)

free, really slow, internet

Hmm. Deaf Havana is sadly lacking on MB. Looks like I have some more work to do this weekend

hawke_: better than expensive, relatively slow internet!

reosarevok: look at what?

aw man tired

natta!

warp: the series support thing we talked about at the summit

natta CatCat

bye bye !

warp: I'm not sure if we talked of a planned date for it or not :)

reosarevok: is there an actionable MBS ticket?

warp: well, we had a fairly clear idea of how to do it on the summit IIRC

So if there's not, should be possible to change that :)

CatCat: still there ? you’re using a great IIRC and such → convertes to what it means… which one was it already ?

IIRC AFWTF etc. this stuff converter to full sentence

Strange to see a Naxos release with the same recordings as a BMG release.

hawke_: strange as in wrong? ;)

reosarevok: Strange as in right, but it’s the first time I’ve seen them shared between Naxos and a major label.

right? really?

heh, yes, that's weird

Unless they're Naxos Historical out of copyright stuff...

definitely not out of copyright

(I run sshd on the cvs pserver port, which seems to be generally unblocked)

I run sshd on 443 at home.

hawke_: huh, so Naxos is the one *reissuing* it?

Interesting

The Naxos release is http://beta.musicbrainz.org/release/f1f5ca67-5d...

ianweller: both?

warp: both.

ianweller: thats... impressive, how? :)

warp: haproxy shenanigans

warp: (there is a 2 second delay to connect over ssh: if haproxy doesn't detect anything that https uses for 2 seconds it assumes you want ssh)

hawke_: I saw, that's why I asked

I kinda expected a major to have found a great Naxos recording and tried to get money from licensing it

ianweller: whoa.

But the other way around, huh

Yep.

ianweller: cool.

yeah, there are ways of doing ssh and http on the same port, but my problem with that is sitting in front of some stupid proxy in some wifi hotspot

443 is more likely to work than 80

ianweller, warp, Krystof: I personally recommend 53, that can sometimes get around even more stuff

ianmcorvidae: oh, that's a good idea as well.

of course one can run sshd on all of these ports at the same time :)