#musicbrainz-devel

what, you imply that you have a LIFE??

tsk tsk

Nah. I just pretend.

I once had a life. :)

but you'll get this done then, yes?

I wonder if you can get a degree in cat herding.

"Hi, I'm a cat-herding major. With a focus on pedantic ADD cats. what is your major?"

I'll get the SQL done, yes.

awesome, thanks. :)

ruaok: There is already and SQL patch, if you want to start trying to roll it.

I will on monday.

ruaok: It just doesn't have any new and changed indexes.

k

Freso: needs adding to upgrade.sh, adding to CreateTables, FKs at least :)

(before it can be used for testing)

Got my SQL in review.

darn... apparently i should use gmake instead of make :P

ianmcorvidae: Uhm... mind writing that on the ticket? :D

Freso: uh, sure

ianmcorvidae: ^^

ruaok: pong

hi warp.

I did some tidying up of your tickets.

removed that one ticket for series.

ok, thanks

please take a look at your tickets and make sure all is well.

I would like to announce the final set of tickets on monday.

and start the first data conversions then as well.

i'm going to send out a summary email to everyone who has schema change tickets, just to make sure we're all on the same page.

does anyone now how to pull an FTP dir from ssh?

I'm going to start a fresh fullexport import now, so I can do some testing of MBS-1839 tomorrow.

tonyschuite: wget knows ftp, that should do?

warp: thanks. I appreciate that. I think everyone should do that on their branches.

no wget on freenas :(

i thought i could use the fetch command but it doesnt exist

ruaok: my schema change tickets both need a tiny bit more work which I will also finish tomorrow. I've had a bunch of feedback on them already, so hopefully they'll be ready then.

ok, great.

* Ian McEwen: MBS-5491: check for 'vs.' and 'feat.' in PossibleCollaborations report

* Ian McEwen: MBS-6012: when extracts get an empty string back from wikipedia, don't try to display them

(UI changes is a bit misleading for MBS-6023, because the vast majority of the changes shouldn't affect the interface at all)

* Ian McEwen: MBS-4364: update tests to ensure VA doesn't get arbitrary attributes added to it by merges

* Ian McEwen: MBS-4364: also don't merge IPIs

ok, got it... was a syntax thing... i can use fetch :)

er

huh

that test doesn't fail locally D:

heh, of course, we don't consistently have artist ID 1 be VA D:

ianmcorvidae: locally you don't have perl v5.old on ubuntu 10.weird.

warp: nah, it's the DB constraint, I didn't refresh my test DB

ok :)

Project musicbrainz-server_beta build #378: STILL FAILING in 4 min 31 sec: http://ci.musicbrainz.org/job/musicbrainz-serve...

Ian McEwen: MBS-5912: actually list cover art relationships report

warp, ocharles: I'm fixing that test and pushing to beta for ci, but in light of our conversations, shall I subsequently open a review for it?

ianmcorvidae: in general I'm fine with small fixes to make beta not fail. IMO that doesn't have to go through review.

hm, apparently bitbucket isn't closing PRs automatically

it is

having just pushed out five changes it didn't, guessing no :)

(it takes a few seconds for bitbucket to realize a pull request is closed if you just push it, but so far it has always closed them for me)

hm

must be taking a lot longer than usual then, I waited a few minutes

or something about that particular pull requests is confusing bitbucket.

Good god ISRCs are so DUMB

Can't exec "./admin/InitDb.pl": Permission denied at /usr/local/lib/perl5/site_perl/5.10.1/Carton/CLI.pm line 320.

“When the supply of registrant codes for use with the United Kingdom country code of “GB” has been exhausted, new registrant codes will be allocated for use with a new “UK” country code. This is expected to happen in early 2013.”

fock!

hawke_1: lol

the US has been using QM codes for a while

hawke_1: why is that so dumb?

tonyschuite: perhaps the script isn't +x ?

also, there's a reason they're trying to figure out a replacement for the ISRC system :P

reosarevok: 'we didn't plan ahead and make our code long enough, so we’ll go through some weird contortions to shoehorn something into the existing system'

tonyschuite: you can try putting perl in front of it, so something like carton exec -Ilib -- perl ./admin/InitDb.pl -- --createdb etc...

warp: what do you mean?

and national prefixes will disappear” but that’s not mentioned in that document of course

I wonder why they chose DG

(instead of a user-assigned one like QM)

nikki: “[CP and DG] are used by the International Telecommunications Union to identify certain technical installations (on Clipperton Island and Diego Garcia respectively) and no ambiguity can arise.”

nikki, because they are the voiced alternatives to /T/ and /K/!

Oh, bah, my explanation was cooler

yes, it is. :-)

hawke_1: it still seems odd to me

nikki: I think they just went “well, what can we use that won’t break too many peoples’ validation systems?”

since their validation specifies that everything in ISO 3166-1 is allowed…

…that means that DG is already allowed.

of course that doesn’t explain the inconsistency, but hey whatever

technically DG is only "exceptionally reserved" :P

nikki: yes, they say that too.

I didn’t paste the whole bulletin. :-p

and if they're worried about breaking stuff that doesn't believe QM to QZ exists, then they're a bit late

Yes, it’s completely stupid.

Now can I move on to the stupidity in their validation criteria?

don't underestimate how dodgy most software is

“An ISRC is always invalid if:… it is the same as the code assigned to a different recording It is a fundamental principle of the ISRC system that two different recordings must have different ISRCs. If an ISRC is found that has already been seen assigned to a different recording, then an error has occurred. The source of the error should be investigated.”

How do you know if a recording is different? Well, it has a different ISRC of course!

(well, or through some other method of determining wtf a recording is)

So yeah…ISRC is pretty much fucking useless.

well... quitting for now

thnx for the tips!

ciao

Project musicbrainz-server_beta build #379: STILL FAILING in 28 min: http://ci.musicbrainz.org/job/musicbrainz-serve...

Ian McEwen: MBS-4364: correct Edit::Artist::Edit test to not use the special ID 1 for testing

feh

no doubt we're using 1 elsewhere then

oh, nope

but still something I should have seen, heh

ianmcorvidae: shouldn't you use gids and not rowids anyway to ensure the test remains trustworthy over time?

reosarevok: possibly? I mean, it's for VA, which really is always ID 1

except in tests

well, except in that one test, now fixed :P

ianmcorvidae: oh, ok, it is a test *for* VA

I thought the problem was someone had given rowid 1 for a random test and it had broken now because of enforced no-gender no-type

ignore me then :)

yes, that is what happened

because that test used rowid 1, when rowid 1 is, in every other case, always going to be VA :)

NOT READY, build status is red_anime.

partitions are hard (and cool) :D

ianmcorvidae, does that mean you're fighting geordi? :p

reosarevok: no, it means I'm discussing how to improve the dated/undated rels merge patch with ocharles :)

Oh

With partitions?

reosarevok: to be clear, when I say partition I mean as in PARTITION BY, the SQL command :P

oh

:)

reosarevok: which you may not know exists, I admit :)

Yeah, that wasn't in my introductory course :D

:)

we may have a way to git rid of some of my perl hackery and turn it into SQL, plus combine it with other SQL to make the whole situation shorter

oh

import running. tomorrow will be more testing and fixing of schema change branches. now sleep, goodnight!

though I just thought of why I'd done it in perl in the first place, hm.

bitbucket rid of it!

svn rid of it!

Ok, I'll shut up now

bitbucket isn't a VCS, it's a hosting service :P

guessing you meant hg rid of it or darcs rid of it ;)

aw, poor bzr

time to perforce rid of this conversation, I think