#musicbrainz-devel

Hmm

You broke the link :)

Ok, should be OK now http://musicbrainz.org/doc/Products

ah, oh well spotted !

rika update: edit dump was corrupted, I think by way of wget weirdness, it's redownloading, I'll work on it through the rest of today

ianmcorvidae: Bummer

But I bet it doesn't take as long to apply as it does for me :)(

heh

it just takes forever to download

while you're here, any chance of putting tig on rika?

It's a console front end to git.

No package, but it was a simple compile on my box at home.

definitely not compiling anything right now, I need to shut down and pack my desktop

(moving on Sunday)

How about if I build it, and you just run the install (as root) from my home directory at some point?

adhawkins, you can compile and install locally, just ./configure --prefix=$HOME

Yeah I know, but figured it might be useful to others.

ah, well

until ian is able then, that's a solution

:)

Yes, true enough.

Ok, build in my home directory. Didn't realise just how easy it was, only 5 source files!

make sure you had $HOME/bin to your PATH

it's no on by default

*not

Yeah, I already have for something else on there.

I would also suggest install anything not shipped with ubuntu to home directories.

+ing

Not sure why the colours are so washed out...

some nasty antialiasing in that image, though perhaps those are just jpg artifects. do not use .jpg for screenshots please!

:)

adhawkins: what do you need in tig?

because that looks like something git log is capable of

It's just a useful history browser. It's interactive as well, you can scroll through the commits and stuff.

i know, but so is: git log --graph --pretty=format:'%Cred%h%Creset -%C(yellow)%d%Creset %s %Cgreen(%cr) %C(bold blue)<%an>%Creset' --abbrev-commit --date=relative --no-merges

In the screen shot, the bottom half is details of the selected commit.

right, you can add that into git log too, but it'd be for every commit

Yeah, it's useful to be able to scroll to a specific commit to see what changed.

The app is tiny, has come in handy occasionally.

sure, i'm not against it

not my machine after all :)

It's in my home directory now anyway. Thought it might be useful to others, but as you're all Git Gods I guess you don't need anything that makes things easy :p

i'm really not against it at all

Just pulling your leg...

I use magit myself

ok :)

magit

it's an emacs plugin

ocharles: is it me, or is admin/InitDB.pl just a very unreadable version of script/create_test_db.sh? :)

warp: haha

warp: I have thoughts on InitDB actually...

if you want to hear them now I can summarize

ocharles: I'm trying to create the database in script/create_test_db.sh. and was attempting to make the database selectable in InitDB

ocharles: I think I'm somewhat done, in the sense that things work. but the situation has only become messier, so I hesitate to submit this to review, I'm quite unhappy with it.

understandable

so yes, summarize if you have the time.

The first thing I don't like is GenerateSQL

I don't like that I have to manually keep Drop* up to date

I can find no mention of that in InitDb.pl

That's not used by InitDb.pl

but it generates the drop scripts

ah

I'd rather have a program that is given <some input>, and an operation like 'create tables', 'index', 'drop foreign keys'

As a developer, I now only have to maintain one file

InitDb would initially be a script that makes a set of calls to this

I like the concept of https://pyrseas.wordpress.com/ too, but for some reason the guy wants to add a web interface to that project

which starts moving it towards phpMyAdmin or something, which is not what I want

not sure if any of that is helpful/sensible

I haven't started any work here because I couldn't find enough motivation; what we have now works

hm

ocharles: describing the database schema in something a bit more machine readable than scripts which need to run against an SQL server sounds like a good idea in theory.

ocharles: I have no experience with such an approach.

ditto

but it can also still be comment annotated SQL

but I think yaml is going to be more flexible

it's nice because we could load a tree of dependencies and do a brief pre-copmile before even doing SQL

but that's also not so important, because either way something will break :)

but anything like that sounds like a fair amount of effort, and not important enough to work on now.

yep

but if we've got a big mess atm, then maybe it is worth looking at

ocharles: do you have idea what kind of stuff InitDb can do which create_test_db.sh doesn't?

well, take arguments for one, and kick off mbimport

another thought is maybe we consider adopting mbslave as our official database importer/creater/replication thingy

and remove all of this from mbserver

I have no idea what that entails.

the would make a lot of sense if you just need a replicated db, but don't some users actually need mbserver

ocharles: anyway, now I'm still no step closer on deciding what to do about this work I'm doing now :)

adhawkins: you can run a local webservice with just the replicated database.

a full mbserver with everything is really only for development I'd say.

warp: I can? How?

no u can't only gives u searcg

Just to clarify, my VM is configured as a slave.

not lookup

teatime , fish curry !

huh? why wouldn't the webservice work on a replicated database?

hello ruaok!

hi

having a problem with the schema update... anyone around can help?

yep

I have a brand new server and I've downloaded everything using git

if you're importing a clean data set, make sure to use HEAD, not the release tag.

so I have an empty DB because I can't import the latest bz2s of the database

the release tag has a bug that doesn't cleanly import.

yep got the head from the schema change

what is the commit you're working with

?

I've created the empty DB using carton exec ./admin/InitDb.pl -- --createdb --clean

dec346a ?

you want 18e31b9f02b1816

otherwise imports fail.

ocharles, warp: should I move the tag to this commit so import work?

the v-2012-05-15-schema-change tag that is)

ok how do I get that directly? sorry a little shaky on git

git fetch

git checkout master

git pull

ruaok: the tag shouldn't be moved to include anything not running on production. but I assume 18e31b9f02b1816 is on production? if so, no objections.

warp, maybe I'm misunderstanding but I thought mbslave just provided yo with a replicated database and no observer code

that should get you there.

therefore no code for the webservice

s/observer/mbserver

warp: its not on production.

that code is never executed in production, so pushing a hotfix is silly.

ruaok: yeah, I'm not worried about that particular commit

ruaok: but I think there have been more commits between the release and your 18e31b9f02b181614c398aa277bea720eb3950fd commit.