I think I've got a clue as to what is wrong with our DB server.
2010-07-21 20206, 2010
nikki_
oh?
2010-07-21 20209, 2010
ruaok
straight from the horses' mouth. :)
2010-07-21 20236, 2010
ruaok
yeah, it turns out the size of the DB is much larger than when a clean import is done.
2010-07-21 20254, 2010
ruaok
so, we need to dump our DB and reload it. to fix the biggest problems.
2010-07-21 20254, 2010
ianmcorvidae
interesting
2010-07-21 20215, 2010
ruaok writes notes to not forget anything
2010-07-21 20200, 2010
ruaok
and the ignite talks are happening and there is free beer.
2010-07-21 20208, 2010
navap
Has the master database ever been dumped/reimported before?
2010-07-21 20212, 2010
ruaok loves oscon
2010-07-21 20234, 2010
ruaok
yes, but not recently.
2010-07-21 20247, 2010
ruaok
not in 1062 days anyways.
2010-07-21 20234, 2010
ruaok
fortunately our tuning has bought us some time. but a dump/reimport will still take 1-2 hours. thats 1-2 hours of downtime. :-(
2010-07-21 20248, 2010
ruaok
exactly how to do that, I'm not sure yet.
2010-07-21 20253, 2010
ruaok
we may not have to.
2010-07-21 20259, 2010
ruaok
but I won't think about that right now.
2010-07-21 20212, 2010
ruaok
there is another dilemma I need to focus on first:
2010-07-21 20217, 2010
ruaok
free beer.
2010-07-21 20218, 2010
ruaok
:)
2010-07-21 20222, 2010
nikki
that's a dilemma?
2010-07-21 20223, 2010
navap
heh
2010-07-21 20244, 2010
ruaok
if there is beer left over, thats a dilemma. I don;t want to go there.
2010-07-21 20248, 2010
ruaok runs off
2010-07-21 20250, 2010
ruaok
ttfn!!
2010-07-21 20250, 2010
navap
Of course it is! Free == good; beer == bad ;)
2010-07-21 20255, 2010
ruaok
lol
2010-07-21 20253, 2010
nikki
bah... I just spent ages trying to remember how to select something from the database, then eventually remembered some code where I'd done the same thing... looked at it and found out I'd already made a function to do it >_<
2010-07-21 20209, 2010
nikki
now I don't remember why I was doing this
2010-07-21 20218, 2010
navap
haha
2010-07-21 20230, 2010
nikki joined the channel
2010-07-21 20253, 2010
nikki
whatever it was, osx didn't approve :P
2010-07-21 20224, 2010
alastair_ joined the channel
2010-07-21 20208, 2010
navap
warp: Looks like I was mistake about the Controller/User.pm merge, the conflict wasn't resolved the way I wanted.
2010-07-21 20252, 2010
navap reminds himself to rebase before asking to get his stuff merged in
2010-07-21 20223, 2010
Leftmost joined the channel
2010-07-21 20237, 2010
navap
ugh uservoice is making it harder to submit anonymously, which means they're making it harder for someone to submit more than 10 votes worth of ideas :(
2010-07-21 20241, 2010
dinog joined the channel
2010-07-21 20236, 2010
nikki
oh?
2010-07-21 20257, 2010
navap
Well it's not that bad, they're just making an email address a requirement and then matching that email to their user database to see if you have an account. So you can't use the same email if you want to post "anonymously".
2010-07-21 20213, 2010
navap
The good thing is that now they have a display name field for anonymous users.
2010-07-21 20235, 2010
nikki
ah
2010-07-21 20239, 2010
navap
I've thrown lots of random things up on UV anonymously because I'm out of votes (but only good ideas, if it's a bad idea it wasn't mine!)
2010-07-21 20205, 2010
nikki
haha
2010-07-21 20242, 2010
nikki
I don't even have a proper email address set for mine
2010-07-21 20214, 2010
ruaok joined the channel
2010-07-21 20220, 2010
warp
navap: I just learned from aCiD2 that rebasing is bad. you should merge master into your branch instead.
2010-07-21 20208, 2010
ianmcorvidae
merging and rebasing are bad in different ways :P depends on the project which should be done
2010-07-21 20225, 2010
ianmcorvidae
but there do seem to be more people on the 'use merge' side of the debate than the 'use rebase' side
2010-07-21 20228, 2010
warp
ianmcorvidae: I meant in the context of musicbrainz.
2010-07-21 20233, 2010
ianmcorvidae
yep :)
2010-07-21 20241, 2010
warp
most branches get applied to both next and master at different times. if a branch gets merged into next, then gets rebased onto master beforing merging in master,...
2010-07-21 20252, 2010
warp
the next and master branches will have diverged and start getting conflicts.
2010-07-21 20200, 2010
warp
I did not know this until about a week ago :)
2010-07-21 20212, 2010
ianmcorvidae
ooh
2010-07-21 20241, 2010
ruaok
can we dump next and simply replace it with a new copy of master?
2010-07-21 20256, 2010
warp
ruaok: the issue has been resolved already.
2010-07-21 20213, 2010
ruaok
what was the solution?
2010-07-21 20221, 2010
warp
ruaok: merge master into next and resolve any conflicts.
2010-07-21 20235, 2010
ruaok
ok
2010-07-21 20237, 2010
warp
most of the conflicts were my doing, because I was doing the rebasing.
2010-07-21 20250, 2010
ruaok
ah.
2010-07-21 20216, 2010
warp
navap: anyway, please just send me a patch with whatever needs fixing about that conflict.
2010-07-21 20242, 2010
alastair__ joined the channel
2010-07-21 20217, 2010
navap
aw I liked rebasing since it didn't create a new commit, it felt "cleaner". But I get what you mean about master vs next.
As to the patch: I moved the profile action (is that the right word?) out to it's own controller which is why I'm removing it from the main User controller.
2010-07-21 20210, 2010
warp
I'll apply it in a few minutes. I've been fixing some of the tests in t/*.t, but some uncommitted changes remain in my working copy.
2010-07-21 20258, 2010
warp
I need to double check if those changes are sound, then commit them. otherwise I'll probably forget hem as I haven't created a jira ticket for broken tests. maybe I should :)
2010-07-21 20217, 2010
navap
hmm my stuff rarely gets merged into next, it's only when the review is finished and the branch is being merged into master does it makes its way into next.
2010-07-21 20248, 2010
navap
So in my case wouldn't it be okay for me to continue rebasing since next isn't getting my branch?
2010-07-21 20206, 2010
warp
yes, rebasing is ok as long as the branch hasn't been released in _some_ form.
2010-07-21 20216, 2010
navap
Yeah
2010-07-21 20238, 2010
warp
navap: but if e.g. I cherry-pick something out of your branches because I need those changes NOW, things already break because rebasing changes all the commits in that branch.
2010-07-21 20250, 2010
navap nods
2010-07-21 20258, 2010
warp
so, try to avoid it :)
2010-07-21 20254, 2010
navap
I've been treating my github account as if it weren't public because it's a *much* nicer UI to look over what I've done than gitweb :(
2010-07-21 20215, 2010
navap
But that's obviously not the right thing to do.
2010-07-21 20240, 2010
ianmcorvidae
local install of gitolite or something?
2010-07-21 20241, 2010
ianmcorvidae
hm
2010-07-21 20253, 2010
warp
navap: using github is perfectly fine, it doesn't cause me any trouble it all.
2010-07-21 20217, 2010
warp
navap: but I do have your github as a remote, so I can pull out any of your branches whenever you ask me to look at something... or for stuff that is on codereview.
2010-07-21 20250, 2010
warp
navap: so in that sense, it is public, and there is a chance that I would like to grab stuff from your branches.
2010-07-21 20217, 2010
ruaok joined the channel
2010-07-21 20258, 2010
warp
and I try to publish most of what I do as soon as possible, because it's just a convenient backup.
2010-07-21 20209, 2010
navap
What I meant was that I've been rebasing/force pushing all the time because I've been treating github as private, but that causes problems if someone else ever pulls any of my branches and does something further on them without me knowing.
2010-07-21 20241, 2010
warp
now that I work from home, doing off-site backups is a bit more hassle than it used to be, so I tend to only do those about every few weeks.
2010-07-21 20259, 2010
warp
navap: yeah, I understand.
2010-07-21 20231, 2010
warp
navap: but from now, consider your github to be more public. especially when you've asked someone to look at a branch :)
2010-07-21 20251, 2010
navap
Yeah will do
2010-07-21 20249, 2010
warp grabs profile.patch
2010-07-21 20209, 2010
warp
'git am' doesn't like it :(
2010-07-21 20238, 2010
navap
oh whoops, I just did a git diff > profile.patch, I should be doing somethingwith format-path right?
2010-07-21 20241, 2010
warp
navap: can you use either 'git format-patch', or tell me which of your branches I should merge to get those changes?
warp thinks the rest of the day is a day for the beatles.
2010-07-21 20249, 2010
navap
The Beatles \ð/
2010-07-21 20211, 2010
warp
:)
2010-07-21 20215, 2010
navap
warp: Do I need anything special to run prove -l t/*t? I tried it once and just got a lot of errors (it looked like the tests weren't being run).
2010-07-21 20253, 2010
navap
Ah right, I don't have a musicbrainz_test schema set.
2010-07-21 20257, 2010
warp
script/create_test_db.sh
2010-07-21 20204, 2010
warp
but you probably found that by now.
2010-07-21 20238, 2010
warp
sadly, it takes a considerably amount of time to run all tests.
2010-07-21 20250, 2010
navap
So I have to create a new database, can I use whatever values I want for the database names as long as the schema is set to musicbrainz_test?
2010-07-21 20229, 2010
navap
On a side note, musicbrainz_db is the actual database name, so what's the significance of READWRITE? Is that just an easier way to reference the database in the MB code?
2010-07-21 20204, 2010
navap
Right, because you could name the database something different so the key stays constant.
2010-07-21 20248, 2010
warp
my databases are 'musicbrainz' and 'musicbrainz_raw'. the code refers to them as READWRITE and RAWDATA.
2010-07-21 20205, 2010
warp
navap: anyway, you don't need a new database.
2010-07-21 20217, 2010
warp
navap: script/create_test_db.sh creates a new schema in your existing database.
2010-07-21 20231, 2010
navap
I get ERROR: schema "musicbrainz_test" does not exist
2010-07-21 20253, 2010
warp
hm, odd. I don't remember having to create the schema manually.
