#musicbrainz

what a coinky - dink

added that hypnosis release to place a sample (Lemon Jelly)

and guess what...the very next track samples Nordine :-)

nordine is awesome :)

he's sampled on, or does quite a few new voice overs for hip-hop these days.

in 1990 i saw him an opening monologue at a grateful dead concert, at the request of the jazz musicians present -- it was fantastic

Now grooving to: The Kinks ~ Big Black Smoke (from Face to Face)

the cake is a lie!

yllona: i need to talk to you

sure thing... one sec

ruaok: Ping back. :) I'm not entirely sure how upgrade.sh can create the connection details and hence avoid the ``Can't call method "database" on an undefined value at ./admin/InitDb.pl line 42.'', plus, it's not going to help people running from a clean checkout of the branch?

14:37 < jonas> www.ifpi.com vs www.ifpi.org

Hahah

yeh, good one ^^

another 7 years before they have the chance of stealing ifpi.org, though :p

x...

ww

aaaaah, finally apple announced the release date of mac osx.

FauxFaux: did you update cgi-binDBDefs.pm to include the details for the RAWDATA database?

cgi-bin/DBDefs.pm

hm, do replicated DBs need RAWDATA?

probably not

i don't believe it is replicated anyway

inhouseuk: There's no default in DBDefs.pm.default. :/

starting here: http://bugs.musicbrainz.org/browser/mb_server/b...

:s

Okay, I suck. Thanks. :)

no problem :)

I swear I checked svn st about a million times. Hell, I swear I checked it on a clean checkout (at about 6am this morning). >.<

And the database dump is too old, sulk.

hi ru

ok someone help me!

anyone know enough japanese to ay 2this is my cat Burton" :/

inhouseuk might, if he’s around...

no, thats russian curtacy of his fife..

wife*

brianfreud: check it http://blogs.sun.com/plamere/entry/web_performa...

We got an F. :-(

I'm going to see if I can get O'Reilly to donate a couple of performance books.

CatCat: thought it was inhouseuk who just got back from a while in Japan... Who am I thinking of?

if some magically showed up at your house, would you read them and play with MB more to make it load faster?

oh paul redesigned his blog.. everyone seems to do that lately

ruaok: that’s something I’ve been paying some attention to, since tracking down that other thing...

exactly my thought.

35 to 50 http reqs per page, with page sizes ranging from 75 to 500k (a few are 5+ mb...)

brianfreud \o/

performance for mb needs rewarp

brianfreud: nope, never been to Japan

I'll see if I can get my paws on those books.

its because edit pages have had lots of littel patches, little flickerings

to add that and that and that over time

hmmm, maybe it was Kerensky97? - I know someone just got back from Japan...

partially it’s all the js that’s weighing it down

but also, all the little image bits kill it

heh, one thing to do is to combine all the sponsor graphics into one.

that would cut out 4-5 requests.

ruaok: i dont think the graphics really are doig nit

the gradiants being an image with text over it, instead of a css gradient...

each gradient bar = http req

oohh

you can do a lot by reducing the number of requests.. i.e. putting together several js / css files in one respectively

CatCat: the overall # reqs is an issue

brain is having ideas already

also i have an idea

lol, this is stuff acid2 and I talked about a while back, for his redesign

when you go to page liek "edit artist" or "edit track" you dont *NEED* the trackparser

but the 1st one we looked at getting rid of, re http reqs, was the gradients

cat: agreed, but I think there’s a even simpler solution

that was jsut an idea

mo: I think the js will already be in the cache then though

ru, something I was thinking about... invisible frame on every page, load the js in the hidden frame, then just *never close that frame*

it needs seperate guesscase suits fo different types of edits

Shepard: nope

heh

js only gets passed or loaded once that way, instead of over and over on each page

thats an insterestign ide?

idea

it’s hacky...

js on the page is awsoem ,its the loadign and loadign and reloading ad nausium that makes the site slow (for me)

just a random 3 am idea I had the other day, not sure how feasible it is, nor how usable it can be

yes its hacky

brianfreud: the browser caches the js, so no need for fancy tricks like that.

ruaok: not caching here...

your setup is AFU. :-)

under normal conditions its cached.

brianfreud: tell your browser to use its HEAD ;)

with all the various load tracking tools on (firebug, etc), or externally to the browser (firebug, etc all off, using packet sniffer)

either way, I see the same reqs made, and the same loading time for the js, even if the js code itself is cahced

well, we need to improve a number of things. including turning on gzip encoding.

IE and FF both still seem to req the data, even if it’s cached - if it’s cached they just don’t seem to bother to actually use the data :P

yeah, gzip would help

html-lint and rhino on the production js would help too

I gotta see what the CPU utilization would be... see if our front end web server can handle it.

lines like the icon one I mentioned to you, too - stuff that’s doing nothing but adding bits

“Milestone: Site Performance Optimization Update” :)

I think the XHTML2 milestone will bring a lot of optimisation already

I think so too

any ideas where acid2 stands with it? haven’t seen him much lately

he got sucked up by uni.

mostly popping in and out on im

figured that might happen

maybe we can nail him down over the xmas holidays.

well, with this server patch up and done, perhaps, between you, , luks, murdos, acid2, me, etc, we can do a “done by the end of acid2’s holiday break!” push at xmas :D

serious props btw ruaok... I thought I’d seen the update, and tested it, but I’m still finding fun little cool stuff - like the annotation icons - that make me smile at how great this new update is

well, we need someone to "own" the branch and make it theirs to move it forward. interested?

Let me talk to acid2 first? I think he kind of thinks of it as his already, don’t want to step on his toes...

but I'm glad you'r happy.

brianfreud: good call.

but I am keen on seeing that branch continue. and *now* is the time to do it.

agreed

if we hack on a bunch of other stuff in the meantime, we'll outdate the branch

well, I’d be more interested in “lets do a partial xhtml 2 conversion” and then trim the fat, streamline, then finish the xhtml conversion

rather then “try to do the entire conversion all at once, then try to streamline it”

the latter just seems like it’ll take much much longer, and be more troublesome re:testing

sounds about right.

for once I can ask this and be on topic...

Only reason acid2’s even really been worrying about supporting IE5 is the oldtagger- otherwise it’s pretty much a dead browser (finally! :) :) :) )

but oldtagger, and thus IE5 xhtml/css support, I figure also dies when TRMs die

yep.

so, with the “we don’t really care” that you got from the one crew, is the only thing keeping TRMs alive the wait for PicardQT b1?

last acid2 and I talked, IE6 css/xhtml support was causing headaches, but IE5 support was *the* nightmare in the redesign