#musicbrainz

aCiD2: first thing to do: add `PerlInitHandler Apache::StatINC` to your apache config

restarting apache on every change is extremly annoying

:)

I put that in the vhost, yea?

yes

hmm, didn't know that one

i have a few cds with no tags at all

is it actually possible to tag them?

Sure, as long as you know what they are (or someone else knows and has submitted PUIDs for them :P)

i don't know what they are

i dunno. can't work the application out >_<

if it uses um, hash things, why do matches have a rating? o.o

i have the right album on the screen now. i can't get it to save >_<

the button's greyed out

i found the cd on the website, it says the album has puids

but if i drop the files there none match :/

Miles: are you using picard, picardqt or the classic tagger?

picard

ok

after you have done a lookup from picard once, will have opened a musicbrainz search in your browser.

as long as you don't close your browser, you should have green tagger icons on the site on every album. so look up the album you think it is, then click the green tagger button, which should send that information back to picard.

yeah

then i drag the cd from clustered albums

onto that

right?

yes. if it's clustered

otherwise you can drag the directory directory from the browser pane.

directly

ill try from there

nah still no matches

i guess it's just odd.

good luck. you might want to try picardqt btw, it's still alpha, but most users here seem to already prefer it over the old pciard.

s/pciard/picard/

ah

well i run kde anyway

so yeah, i might

luks: doesn't picard use python-musicbrainz2 ?

no, it doesn't

hi guys :)

luks: What does it use?

QHttp + QXml

hi brian

so, that requires qt4?

Makes sense since it uses QT.....

warp: yes

hey luks, this may sound strange, but can you think of anything that would cause a single given release to just never load right in QT?

found one, tried from IE6, FF, regular server, uk server, etc

BrianFreud: look at comments on the ticket :)

luks: ok, so for my commandline mucking about i better use pymb2 then? :)

yes, for anything

picard uses pymb2 too

just wondering why picardqt doesn't.

but i rewrote to avoid threads

ah, right

er, *used

Couldn't you have hacked pymb2 to be threadless or something?

no, because urllib doesn't support async sockets and the xml parser python uses can't do incremental parsing

I see

python is slow enough on it's own, threads in python are even slower because you can have only one thread running python code at the same time

so async sockets are really the fastest solution

ok, found the bug ticket....

So a merged label was blocking the release from loading properly from the server?

fun :P

no, the release points to a label which doesn't exist

gotcha

which was caused by a merge, but I don't understand how exactly did it happen :/

ok, so maybe it was a indicator of something more than the "minor" issue I thought it was :P

yep, it's pretty serious issue

glad I reported it :)

almost figured it too minor to warrent a ticket

ffs, textmate keeps freezing :(

yalaforge is the guy doing pymb2 ?

ok :)

I did a bit on it too, but not much (gonna try and squash bugs as they appear though)

ah nice. from the bits of conversation i picked up earlier, i guess yalaforge just started on the label stuff?

sort of, I contributed most of the label support

oh!

how usable is that? :)

he's done some unit tests, and maybe some other stuff since my code was commited, but it's still quite new (only seems to support getById, but that may be a limitation of the XML Web Service)

Usable for my needs, depends what you want to do :)

i mostly just want to get at the catalog # when i have a release mbid

You can do that

great

thanks :)

hehe

I'm not sure if matt's docs are up to date, so you're best bet is to generate local docs from the source code

oh, another thing.

when i did the parsing myself in my evil php code, i stored local copies of all the xml i downloaded, as a sort of cache. I need something like that again for my tool, any idea what the best approach for that is with pymb2 ?

hmm, so instead of accessing the online data, it would access a file?

oh, so you wrote that cue generator warp?

BrianFreud: erm.. no?

ah, thought that was the evil php code you meant :P

cue generator does sound evil though..

all of the interesting bits for which you would want a cue file have already been lost if you try to reconstruct it from other data.

warp, http://bugs.musicbrainz.org/browser/python-musi... if you set netloc and path to be your filesystem, and changed http to file, that might work

But I'm not sure if urllib can read file://

<3 pyparsing

aCiD2: hm, ok

warp, or reimplement IWebService... there may be a better way, but i'm not sure

aCiD2: ok,.. i'll ask again after i have a better feel of things :)

:)

aCiD2: does pymb2 enforce the 1 second between requests thing, or do i have to do that myself?

I don't think it enforces that

ok :)

warp: http://rafb.net/p/t0Ur0A81.html - this is what picardqt used before i ported it to qhttp

it's basically a wrapper around pymb2's web service

but I'm not sure if you want just a cache, or store the xml files for archiving purposes

well, I've been playing - not too successfully - with trying to write an xml output plugin

would be useful for local database purposes

XML Output for PicardQT?

yes

luks: i want to cache everything i lookup for a few weeks/months, until i look it up again.

have no python and very little idea how to write it though, so it's going... slow :P

:)

luks: the idea is to have a commandline go over the entire library and update tags based on the existing mbids.

s/commandline/commandline tool/

I

warp, Picard can do that though?

... I've been thinking about that kind of tool as well, but I have the database handy, so I don't have to do those web service tihngies.

aCiD2: picard is not a commandline tool

true :)

i have some other ideas too which i'd like to hack into it.

hehe

luks: the code from your pastebin is copyright you and GPL? MIT/BSD?

originally GPL, but I hereby relicense it to WTFPL :)

thanks :)

wtfpl?

google it ;)