#musicbrainz-devel

navap: I pruned 33 branches of mine, mostly branches which I was reasonable sure were merged into master or otherwise no longer neccesary.

Yay, I'm down to 221

ask ocharles to do the same

:)

ocharles: Go prune your old branches on git please

master "only" has 74 now.

er, musicbrainz-server.git I mean.

well, mb in spanish :P http://i18n.mbsandbox.org/ I guess we need to figure out how to not use the same cached version for different languages next

but first sleep

nikki: Nice work

heh, I didn't do much

You did /something/

I just downloaded the file and told the server to use it ;) most of the work there is reosarevok's I imagine :P

and ianmcorvidae's for setting up mbsandbox in the first place

night!

Night

It's a little surreal to see MB in a different language

warp: off to dinner here soon but I can set up an account and such sometime this evening; the process is pretty fast :)

aw, looks like I currently don't have an account on rika.

warp: should you be warp, kuno, what?

I should be "warp"

okay

just a minute, then :)

(I probably am heading off to dinner soon)

okay; it's ready

hm, there are locales missing.

this is ubuntu, correct?

it is, yes

lucid

this is one of the things I hate about ubuntu.

on debian it's just "apt-get install locales-all"

heh

well, point me at something to apt-get and I can make it happen :)

cat /usr/share/i18n/SUPPORTED > /var/lib/locales/supported.d/local && dpkg-reconfigure locales

that should fix it. I hope.

it's generating away

darn, eo_US.UTF-8 failed ;)

What's eo?

esperanto

:P

ah

hence my amusement

warp: Generation complete, it tells me :)

How would I change the time zone for me?

set the TZ environment variable

ianmcorvidae: great, thanks. indeed bash stopped complaining.

e.g. "export TZ=Antarctica/South_Pole"

(apparently it's 2:49pm on the 21st at the south pole right now)

huh.

ianmcorvidae: so, I should run mb in a screen session to keep it running?

if you want it to keep running, probably

ok, np.

ianmcorvidae: I assume this runs against a test database on rika?

it cannot find "musicbrainz_collate"

ianmcorvidae: https://frob.nl/tmp/output.txt

kepstin: thanks

Weird, I'm gettig the same musicbrainz_collate problem on xpc.xabbu.net, and that was working just fine an hour or so ago

I'm not that getting that issue on mbsandbox though

odd

looks like it works with master

(running mbs-211,part2 -- editable webservice proof-of-concept)

error: Untracked working tree file 'carton.lock' would be overwritten by merge.

huh?

sounds like someone accidentally checked in carton.lock :/

navap: carton.lock is automatically generated by Carton, but it is also part of our git repository.

kepstin: no, intentionally! that has the exact versions we want, so it should be checked in.

navap: you can safely remove it.

What happens when we do local installs?

local installs?

I installed XML::Feed with carton earlier today because I checkouted a really old branch that required that

I presume that's when the carton.lock was created

navap: with "carton install XML::Feed" ?

Yes

I was following the instructions in the ~/README

navap: that's fine, that shouldn't cause problems.

carton has an uninstall command, which is slow but should work in theory. and otherwise it's easy to start over.

but a few extra packages shouldn't have any effect on the server.

So it seems a recent commit to master is what's causing my collate issues. If I checkout a branch from a few days ago on xpc I don't get collate problems, I do when I checkout master though

But interestingly both master and the other branch work fine on rika

it may be related to where it is looking for extensions.

I know we changed where the cube extension gets installed (in the public schema instead of the musicbrainz schema)

which also means we changed the postgresql search path (which is used to find extensions, among other things)

so perhaps musicbrainz_collate is in public on rika, and older branches don't have public in the search path

and the other way around on your server.

ocharles made those changes, so he should be able to tell you the details.

uh, what the hell! master works on rika, fails on xpc. bootstrap fails on rika, works on xpc.

navap: exactly what I was describing :)

But, but, it was partially working 5 min ago!

haha

the easy fix (if you're patient) is to do delete the database and do a fresh import of fullexport while you're on the master branch.

That would be on xpc, the older server right?

yes

That's easy enough, I've only got the test db on there

on rika you should be able to run the bootstrap branch by just merging master into it (git merge --log --no-ff master)

heh, turns out you can't drop a db that has active sessions to it. That's a good thing :)

What am I missing here?

(master)navap [at ip-10-117-57-215]:~/musicbrainz-server$ ./admin/InitDb.pl -- --createdb --clean

Fri Jan 20 22:17:29 2012 : InitDb.pl starting

Ah, the extra --

I copied the line from the INSTALL which says carton exec ./admin/InitDb.pl -- --createdb --clean

I imagine carton needs the extra --?

yeah, that separates carton options from options that it passes through to the program it runs

goodnight!

warp: merge in master, as I recall -- nikki was having that trouble

warp: Thanks! Both rika and xpc are fixed for me

re: collate

ah, it says that in the log

cool :)

navap, kepstin: is there any reason you're doing pull requests, and not going through the official process of submitting patches?

I figured the best way to test out github's review system was to actually do a pull request. I have no problem submiting the patch to review board though.

i would still prefer that, yea. I was hoping to use the cover art stuff to test reviews

but I guess that doesn't have as many collaborators

we should probably test github reviews if after the discussion it even looks like people want that

ocharles: How do you suggest we fix the robots.txt rewriting? Is that something I could do?

navap: I was looking at that, but haven't quite got a solution yet

I'm thinking of having root/robots.txt disallow everything, and then having a set of features conf files, and one will rewrite /robots.txt into /robots-production.txt or something

Would changing line 42 to root/robots.txt.production work? https://github.com/navap/musicbrainz-server/blo...