#musicbrainz-devel

ruaok: which servers aren't in use now?

lemme look

keyop is a box we never got around to installing. and its the weakest of our boxes now.

but, that give me an idea. too.

jem and misty are being decommissioned they are too much trouble.

taz is in my office and may go into service or may get sold not sure yet.

and zim is dead in my office as well. RIP

ah

everyone else is at DWNI or Level42 in service.

ok, bedtime. board meeting in some 8 hours...

this week's gonna be a doozie.

nn!

night!

hello

hi kurtjx!

:D

hehe

looking at OGS

if that's what it's called

the schema before NGS

I don't think it has a name

I guess I just call it the current schema

we just refer to it as mason

i want to parse the dump file as text rather than loading it in a db

or at least that's what the server is called

hehe mason

kurtjx: no, you don't :)

haha

:)

but trying to figure out what each column is

i've got most

kurtjx: as I said, you really don't want to do that

k

just load it into the db, and then write your own exporter from that

you recommend just loading the damn thing

which columns can't you get?

k

a bit longer, but you will save yourself a big headache

darg

um a bit git-tarded over here

i did

git checkout RELEASE_20090524-BRANCH

but it seems to still look exactly like development master

hm, I would just get that from subversion.

warp: gotta link for that?

kurtjx: yes.

nice one thnx

also, the subversion tree has diverted from git

warp: halp

warp: yay

warp: how do I select "id-merge.merge_strategy" in jquery?

nikki: hi

ocharles: $('#id-merge\\.merge_strategy')

alastairp: should http://wiki.musicbrainz.org/Music_Hackday_Boston just redirect to http://wiki.musicbrainz.org/Music_Hack_Day or something?

I tried $('#id-merge.merge_strategy'), $('#id-merge\.merge_strategy') and $('#id-merge\\.merge_strategy')

hrm

nikki: yes

yeah, that third one should work.

warp: does " over ' make a difference?

it shouldn't afafik

afaik

nikki: you can probably get rid of the boston one. I don't think anything points to it

doh, was missing the hash sign

:)

warp: thanks!

you are welcome

warp: also, how do you feel about adding overflow-y: auto; or whatever it is to make the scrollbar always be present?

with some js I have here, toggling the select causes the layout to reflow when the scrollbar appears, which is a bit annoying

ocharles: hrm, wouldn't be happy about it. but I'd have to see the problem to judge if it is a worthwhile tradeoff (or if there are other possible approaches)

reflows are also annoying :)

ok, i thought you might not like it :)

warp: it's meant to just return a hash ref of the same structure as the wizard->value

iirc

it isn't seed track artists properly when they're not provided by the client. but the code is right there "my $track_ac = $track->{artist_credit} || $params->{artist_credit};"

so I'm about to start figuring out where this goes wrong :)

s/seed/seeding/

moin moin ruaok

moin!

coffee time. then board meeting time.

:)

back in 15

hi outsidecontext

hi ruaok

Oliver Charles got review request #1111!

one more award for me!

+1111 :-)

how do I ssh to hobbes?

I lost my shell history, so I forgot the port on zaphod

didn't you save the VPN instructions that djce gave you?

somewhere yea, but I never needed to use them for hobbes

was this an alternate SSH port that djce setup?

yea

on zaphod

warp: more halp!

recruitment agency fail.

I can't see what port djce has that setup on .

time to mount the old hard drive (oer missus) and see if I can find my shell history

I didn't save the email where djce told you to get access.

I figured you'd do it. ;)

mail djce and ask him, please.

aha

oh ruaok, we need to setup a new postgresql cluster on hobbes too for rc2

why?

I want the serverlog for the web server and the ci separate

so I can do log analysis on the main db

k.

did migration on rump finish?

or... I guess we'll have new server hardware soon so that will do the same thing

ruaok: finalising indexes now

wait. would we need to run two instances of PG then?

or can one instance run two clusters?

2 instances, I believe

seeing as you pass a data directory to pg_ctl

feh. what a pain.

like i said, with the hardware it won't be

can we just do it on new hardware then?

because hobbes can host the ci db, and the new server will host the main db

right?

sure.

we should pretty much forget about hobbes once we have the new hardware installed.

(until we need hobbes for the next server iteration)

the first parts will arrive today.

right

awesome

the bulk should arrive tomorrow.

though there are no signs of the memory yet. which is not so good.

no signs as in no email confirmation?

no shipping info yet.