IRC log of musicbrainz-devel on 2012-08-15

Timestamps are in UTC.

00:41:31 [nikki]
nikki has joined #musicbrainz-devel
00:43:08 [nikki_]
nikki_ has joined #musicbrainz-devel
01:03:40 [kepstin]
kepstin has joined #musicbrainz-devel
01:29:10 [ruaok]
ruaok has joined #musicbrainz-devel
03:33:01 [kepstin-laptop]
kepstin-laptop has joined #musicbrainz-devel
03:45:36 [DremoraLV]
DremoraLV has joined #musicbrainz-devel
03:55:01 [nikki]
demosdemon: I would probably do a mixture of either 2 or 3 with 4. i.e. edits close on the hour, so we might as well just set the expire time to then anyway since that's effectively how it works, but also change the text to say something different for the (hopefully rare) times when the expire time has passed but the edit hasn't yet closed
03:55:10 [nikki]
1 seems like a lot of work for very little gain
06:05:04 [mathieu_]
mathieu_ has joined #musicbrainz-devel
06:06:23 [mathieu_]
Hi all, when performing "carton install --deployment", I get "! Couldn't find module or a distribution ExtUtils::MakeMaker (6.59)". How do I fix it? It seems that only 6.55_02 is known.
06:06:53 [mathieu_]
I'm on debian sequeeze
06:30:57 [ruaok]
carton install ExtUtils::MakeMaker
06:31:08 [ruaok]
usually fixes it for me. it installs whichever it can
07:29:44 [warp]
hello!
07:30:48 [ijabz]
ijabz has joined #musicbrainz-devel
07:40:44 [ianmcorvidae]
... XML::LibXML->load_html apparently can't deal with an & in a @href because it tries to parse it as an entity?
07:41:36 [warp]
ianmcorvidae: I would expect you'd need to use & in xhtml, so that behaviour sounds correct.
07:41:44 [ianmcorvidae]
heh
07:41:55 [ianmcorvidae]
well, it means I can't use html_ok on statistics pages, I guess
07:42:32 [warp]
ianmcorvidae: related: http://tickets.musicbrainz.org/browse/MBS-4568 :)
07:42:42 [ianmcorvidae]
heh, yeah
07:43:11 [ianmcorvidae]
* ianmcorvidae sees if the links actually work when I put & in the query strings
07:43:28 [ianmcorvidae]
heh, they do
07:43:29 [ianmcorvidae]
cool
07:43:34 [warp]
I'd prefer we keep doing well-formed xml, even if the output is otherwise only valid html5 and not xhtml5.
07:43:39 [ianmcorvidae]
then I guess it found an at least semi-real bug
07:46:39 [ianmcorvidae]
just adding stupid get, check-html tests for Statistics controllers, heh
07:46:54 [warp]
:)
07:47:14 [warp]
* warp looks at MBS-5086
08:00:05 [ianmcorvidae]
* ianmcorvidae writes more failing tests
08:42:06 [reosarevok]
reosarevok has joined #musicbrainz-devel
08:55:05 [warp]
warp has joined #musicbrainz-devel
09:05:33 [nikki]
ocharles: http://tickets.musicbrainz.org/browse/MBS-3478 is missing a due date
09:12:50 [stefans]
stefans has joined #musicbrainz-devel
09:17:08 [ijabz]
ijabz has joined #musicbrainz-devel
09:17:40 [ianmcorvidae]
man, 45 open codereviews
09:17:54 [ianmcorvidae]
wonder how many we'll have by next monday when anything can get merged again :P
09:29:10 [ocharles]
nikki: thanks, added
09:29:28 [warp]
ocharles!
09:29:44 [warp]
ocharles: help!
09:31:33 [warp]
ocharles: we use Catalyst::Plugin::Session::Store::Memcached to store sessions. when it cannot save a session it "Catalyst::Exception->throw"s an exception.
09:32:08 [warp]
ocharles: which on my dev machine results in a blank page. any idea if I can catch that exception, do something, then continue handling the request?
09:32:14 [warp]
or will I have to replace the plugin?
09:48:06 [ocharles]
you mean Store::Memcached::Fast ?
09:49:25 [ocharles]
hmm, no
09:49:31 [ocharles]
* ocharles tries to find the source to that plugin
09:50:25 [warp]
Cache::Memcached::Fast is passed into Catalyst::Plugin::Session::Store::Memcached on my dev machine, but not on production I believe.
09:51:22 [ocharles]
I don't know how to avoid that exception i'm afraid
09:51:29 [ocharles]
I'm not sure what part of the request cycle it's in
09:51:34 [ocharles]
but if it was during dispatch, we should already catch it
09:51:49 [warp]
I get the exception after dispatch has finished.
09:52:56 [warp]
(which makes some sense. reading the session (start of request) is not fatal, but saving it (end of request) is.
09:53:02 [warp]
s/^.//
09:54:59 [ocharles]
right
09:55:09 [ocharles]
probably best to pop into #catalyst on irc.perl.org
09:57:56 [warp]
* warp reads some documentation first.
10:10:27 [djce]
djce has joined #musicbrainz-devel
10:20:29 [nikki]
nikki has joined #musicbrainz-devel
10:42:50 [ianmcorvidae]
huh
10:42:54 [ianmcorvidae]
apparently tests just fail for me now
10:42:55 [ianmcorvidae]
great
10:43:16 [ianmcorvidae]
heh, and jenkins too
10:43:17 [ianmcorvidae]
for five days
10:45:46 [ianmcorvidae]
(this is from beta)
10:47:43 [ianmcorvidae]
ah, yeah, the check thing that breaks since a bunch of rows don't actually fit it
10:47:52 [ianmcorvidae]
* ianmcorvidae just checks out beta before that got merged then
10:52:01 [ocharles]
oh, beta is broken :/
10:55:39 [nikki]
it is?
10:55:48 [ianmcorvidae]
not the site itself, the test suite
10:55:52 [nikki]
oh right
10:55:56 [nikki]
I was gonna say
10:55:57 [ianmcorvidae]
because of a database change that hasn't been applied to the actual database
10:56:11 [ianmcorvidae]
(enforcing whitespace trimming on a DB level, specifically)
10:59:48 [ocharles]
other way round I think
10:59:58 [ocharles]
the constraints are there, but the tests violate them
11:00:52 [ianmcorvidae]
ah
11:01:05 [ianmcorvidae]
that might be it, I'm getting a complaint about something violating a constraint anyway
11:01:18 [ianmcorvidae]
I guess webservice.sql probably violates it is what's up
11:03:20 [ianmcorvidae]
yeah, we have "Chemical Brothers" in there at the very least
11:03:42 [ianmcorvidae]
I think that's the one I was seeing fail
11:08:33 [ocharles]
ah
11:20:19 [stefans]
stefans has joined #musicbrainz-devel
11:39:51 [djce]
djce has joined #musicbrainz-devel
12:05:17 [djce]
djce has joined #musicbrainz-devel
12:34:51 [djce]
djce has joined #musicbrainz-devel
13:06:21 [mfmeulenbelt]
mfmeulenbelt has joined #musicbrainz-devel
13:21:37 [plaintext]
plaintext has joined #musicbrainz-devel
13:40:29 [djce]
djce has joined #musicbrainz-devel
13:43:57 [thejav]
thejav has joined #musicbrainz-devel
13:45:24 [thejav]
can anyone here speculate on how many releases are entered before their releaseDate? is there a significant amount entered retroactively? trying to figure out if using the API to check for releases w/ a releaseDate of "today" is sufficient for keeping an eye on new albums
13:48:04 [Muz]
It's often frowned down upon for editorial reasons pertaining to the quality and accuracy of the data.
13:48:27 [Muz]
It does happen though, but I'd be willing to wager it's in the minority of cases.
13:48:40 [kepstin-work]
I'd say that yeah, most of the releases are added retroactively
13:49:16 [kepstin-work]
* kepstin-work occasionally adds albums years after release, but that's mostly on lesser-known artists.
13:51:26 [kepstin-work]
thejav: I suspect that even for popular artists, major new releases often won't be added for a day or two, unless well-published details are available ahead of time.
13:54:15 [thejav]
OK, interesting. I don't think the API exposes a way to search albums by when their data was created, does it?
13:54:32 [kepstin-work]
no, nothing like that.
13:54:51 [kepstin-work]
If you're planning to do large volums of this sort of checking, you really should look into using the replicated database
13:55:57 [thejav]
yep, I should. i've emailed ruaok about it
13:56:09 [kepstin-work]
(our current largest api consumer is an app called 'headphones' which does this sort of 'new releases' checking)
13:56:29 [thejav]
any idea the logic it uses?
13:56:54 [kepstin-work]
thejav: it periodically requests the list of releases for each artist a user cares about, for each user.
13:56:58 [kepstin-work]
so rather inefficient.
13:57:49 [thejav]
yeah i'm trying to avoid doing that. what i'm currently thinking about is checking for releases of the last X days, and notifying users who care about those artists that released something.
13:59:40 [kepstin-work]
if you just have a single server that aggregates the queries required for all of the users, that might not be too bad - the issue with headphones is that each user runs their own copy that does the queries separately.
14:01:41 [ocharles]
note that over the weekend I will be writing RSS feeds that publish the latest MBIDs
14:01:47 [ocharles]
that might be of interest
14:01:55 [kepstin-work]
ooh.
14:01:55 [reosarevok]
reosarevok has joined #musicbrainz-devel
14:02:08 [kepstin-work]
right, that's *this* weekend :)
14:03:25 [kepstin-work]
* kepstin-work could probably fix us up with a BigBlueButton conferencing server for the weekend if folks would like that.
14:05:29 [reosarevok]
We need a cycles liveset at the end of the weekend, ocharles :p
14:05:37 [reosarevok]
To close it with style :p
14:05:43 [thejav]
ocharles: that's amazing! latest as in the newest created?
14:05:57 [kepstin-work]
thejav: precisely.
14:06:38 [thejav]
kepstin-work: FWIW I do have one server doing this, NOT going to make my apps poll it like headphones. *BUT* the ocharles RSS thing would be even better.
14:08:04 [ocharles]
thejav: yes
14:08:07 [kepstin-work]
thejav: well, the reason that ocharles wanted to write this rss feed is exactly for use cases like yours :)
14:08:35 [thejav]
so brilliant, really glad to hear this!
14:09:30 [kepstin-work]
ocharles: I'm a bit curious about the design tho - how far back do you expect the feed to go? e.g. could I use it to resync picard if a user hasn't started it in a couple weeks?
14:10:18 [kepstin-work]
resync riker*
14:10:41 [kepstin-work]
man, my name choice for my vapourware player is occasionally confusing :)
14:13:05 [ocharles]
i don't know how far back, i'll figure that out on saturday :)
14:13:16 [warp]
three years!
14:24:56 [plaintext]
when adding copyright headers for gsoc sourcefiles, should I add my name, or 'MusicBrainz'?
14:28:08 [kepstin-work]
or, for that matter, MetaBrainz :)
14:28:18 [warp]
plaintext: your name
14:29:16 [warp]
you cannot really transfer copyright without paperwork
14:41:07 [djce]
djce has joined #musicbrainz-devel
14:49:44 [plaintext]
thanks
15:09:26 [hawke_1]
hawke_1 has joined #musicbrainz-devel
15:37:44 [kepstin-work]
sweet, I can give us a free BigBlueButton instance with a branded landing page for the musicbrainz hackfest this weekend if we want to use it for video conferencing.
15:38:53 [reosarevok]
Is BBB where you work?
15:39:08 [kepstin-work]
I work at blindside networks, which is a bbb development house and service provider
15:41:43 [kepstin-work]
basically, there'd be a page like http://moz.bigbluebutton.org/moz/ created where everyone participating could join the conference for the weekend.
15:44:41 [demosdemon]
oh BBB is cool
15:46:29 [demosdemon]
My university is trying to build something similar, http://fidelio.cacs.louisiana.edu/fdd/ivc/ but the students working on it have all graduated and we can't find anyone interested to work on it
15:46:57 [kepstin-work]
yeah, bbb actually originally came out of a university project as well
15:47:19 [kepstin-work]
when it was working, the people developing it spun out a company to maintain and support it
15:48:39 [demosdemon]
Integration with moodle you say...
15:48:47 [demosdemon]
* demosdemon strokes beard inquisitively
15:48:56 [kepstin-work]
why yes, we do have integration with moodle.
15:51:00 [demosdemon]
We finally upgraded to a 2.0+ version of Moodle, government institutions are notoriously known for slow rollouts of new technology
15:51:29 [demosdemon]
Which is sad, considering our position in the computer science education rankings
15:52:21 [kepstin-work]
we have a moodle demo video at https://www.youtube.com/watch?v=ShB04IoNgk8 :)
15:52:41 [demosdemon]
Out of pride, I will not suggest you visit our website. It was built in the 90s and well, they're too cheap to hire someone to rebuild it
15:53:06 [demosdemon]
still uses ssi -_-
15:53:32 [kepstin-work]
yeah, we hired a designer, our sites are sexy ;) http://www.bigbluebutton.org/ http://www.blindsidenetworks.com/
15:53:55 [demosdemon]
I know, the software is pretty too
15:53:56 [warp]
oh wow, moodle
15:55:40 [demosdemon]
I took an online class over the summer, and they used Blackboard Collaborate… which I will emphatically say blows huge chunks
15:57:46 [reosarevok]
* reosarevok isn't sure whether that's a good or a bad thing
15:57:47 [demosdemon]
I'm definitely forwarding BBB to our Moodle department
15:57:52 [reosarevok]
(to "blow huge chunks")
15:57:58 [demosdemon]
think vomit
15:58:01 [demosdemon]
chunks
15:58:03 [reosarevok]
Oh, ok
16:01:34 [warp]
(I was a sysadmin at a school in 2004/2005, moodle was one of the things I adminned)
16:02:05 [demosdemon]
About 5 years ago, UL switched from Blackboard to Moodle
16:02:22 [demosdemon]
there were some transitional pains, but so far I think it worked out really well
16:03:01 [warp]
performance was a big challenge for us, we were running it on windows. and PHP on windows was terrible at the time.
16:03:56 [demosdemon]
ew, dear lord. I'm sorry
16:04:30 [kepstin-work]
http://musicbrainz.bigbluebutton.org/ - whee.
16:04:35 [kepstin-work]
(password is 'guest')
16:06:36 [demosdemon]
oh Chrome, why you gotta give me issues with flash constantly
16:07:01 [kepstin-work]
hmm, which operating system?
16:07:19 [demosdemon]
Mac
16:07:50 [demosdemon]
I need to disable the Adobe plugin, which forced itself upon me when I installed the CS suite
16:07:50 [kepstin-work]
we do test regularly on macs, but I'm not sure what browser we use - I think we test chrome
16:08:13 [kepstin-work]
does chrome have a built-in flash on os x like on windows and linux?
16:09:38 [demosdemon]
also, I've noticed a bug with Chrome. If you have click to enable plugings turned on, when you click to enable a flash container, flash doesn't receive keyboard events
16:10:00 [demosdemon]
I have to always allow, quit and restart chrome, and try again
16:10:09 [kepstin-work]
huh. that would be annoying.
16:10:29 [demosdemon]
defeats the whole purpose for that setting
16:50:59 [demosdemon]
oh good, my university is already considering BBB
16:57:20 [ruaok]
ruaok has joined #musicbrainz-devel
17:00:02 [ocharles]
ruaok: sent you a work planningy type email. i can hang around for 30 mins now if you want to discuss it, otherwise I'm going to go and enjoy some sun :)
17:00:04 [ocharles]
"sun"
17:08:49 [ruaok]
british sun you mean. light rain?
17:09:17 [ruaok]
when are you working for mb next? tomorrow, friday?
17:09:19 [ocharles]
it's mildly warm outside :)
17:09:20 [ocharles]
monday
17:09:24 [ruaok]
ok.
17:09:25 [ocharles]
i worked mon-today this week
17:09:44 [ruaok]
I have to finish out income tax return today and send out invoices.
17:09:53 [ocharles]
ok, we can catch up on this another day
17:09:54 [ruaok]
after that I can focus on team organization.
17:09:59 [ocharles]
yea, that's cool
17:10:02 [ocharles]
the google doc is probably more important
17:10:04 [ruaok]
your email is great, btw. I can work with that.
17:10:17 [ocharles]
awesome
17:10:18 [ruaok]
and yes, the google doc is near the top of that list.
17:10:22 [ruaok]
one thing...
17:10:38 [ruaok]
if you ever get stuck in this situation and feel like spending a few hours on ingester, do that.
17:11:16 [ruaok]
brewster pinged me on this yesterday and it would really be nice to move forward on this. now that someone else cares about this.
17:11:26 [ruaok]
the IA now wants something of us, so lets give em a hand.
17:11:38 [ruaok]
I put ingester on our proposed hack list.
17:11:52 [ruaok]
(for the weekend)
17:12:44 [ocharles]
mmm, i'm still not set on whether ingestr makes sense though. i'd probably want to explore direct ingestion instead
17:13:03 [ocharles]
i'll probably avoid it for now until we can talk about it more
17:13:21 [ocharles]
there's work to do on jira to keep me busy for sure, it's just a case of what to pick that I wanted to address
17:13:22 [ocharles]
I think :)
17:13:45 [ocharles]
ok, i'm off now, cya!
17:14:00 [ruaok]
ping me when you get home for a minute.
17:14:09 [ruaok]
I'd like to talk a litte more in ingrestr
17:15:11 [reosarevok]
So, what's ingestr exactly?
17:15:34 [reosarevok]
(as in, how would it put the stuff in MB?)
17:15:42 [ruaok]
a tool that we can feed random XML files to and it will index them and let us search them.
17:15:59 [reosarevok]
Oh, so it's still user-imported afterwards
17:16:02 [ruaok]
then once people look at the data, we can have people establish mappings.
17:16:13 [ruaok]
this field maps to the release editor field X.
17:16:33 [ruaok]
once we have those mappings editors can import data from ingestr to the release editor.
17:16:58 [reosarevok]
So, it's like CD Stubs on steroids
17:17:08 [ruaok]
yeah.
17:17:19 [ruaok]
with much less structure than the cd stubs.
17:17:58 [ruaok]
my idea is to just get some data exposed.
17:17:59 [reosarevok]
huh, not really... I mean, if we are to be able to feed them to the RE, shouldn't it have *more* structure than the CD stubs?
17:18:06 [reosarevok]
(at least in step 2)
17:18:28 [reosarevok]
For people to be able to say what goes where, it needs to support all that "where"
17:18:38 [ruaok]
well, I mean that the data that comes in is essentially treated as if it has no schema.
17:18:43 [reosarevok]
Oh, ok, yeah
17:18:59 [ruaok]
and collaboratively we can establish the schema.
17:19:25 [reosarevok]
What I was imagining was data coming, us telling it "this is it", and it mapping to a "CD Stub", only much less artificially limited
17:19:30 [ruaok]
or we could have people review albums and mark [] ok to import [] needs fixes [] trash it
17:19:44 [ruaok]
and then we could run an import on all the ones that were good.
17:19:59 [reosarevok]
Join that to a better use and visibility of cdstubs, and that could actually work I guess
17:20:07 [reosarevok]
(or "cdstubs")
17:20:17 [ruaok]
yeah, its a hybrid of that.
17:20:22 [reosarevok]
Note that for the IA thing, we can't say "trash it"
17:20:32 [reosarevok]
Since they need to map it to something
17:20:33 [ruaok]
[] duplicate
17:20:46 [reosarevok]
If it's a duplicate, we need to feed them back "what" it is a dupe of
17:20:49 [ruaok]
which then establishes a proper mapping
17:20:51 [ruaok]
yep
17:21:05 [ruaok]
and that IA has lived up to their promises so far.
17:21:16 [ruaok]
so we really ought to work on this and help them get this data into MB.
17:21:21 [ruaok]
which is really in our interest.
17:21:29 [ruaok]
on oct 22 we're having a meeting at the archive.
17:21:34 [reosarevok]
Yeah, I read
17:21:46 [warp]
hello ruaok
17:21:50 [ruaok]
two mentors will be there. as will I. I will fly in ian.
17:22:07 [ruaok]
and I'll invite callerno6 and derwin
17:22:10 [ruaok]
hi warp
17:22:11 [reosarevok]
I think we really need to make cdstubs more visible in general, including the ingestr "cdstubwhatevers"
17:22:24 [ruaok]
reosarevok: yes.
17:22:38 [ruaok]
reosarevok: it would actually help ocharles and I if you could suggest steps for the former
17:22:47 [hawke_1]
* hawke_1 looks for CDstubs
17:22:53 [ruaok]
the latter might be be done on another site.
17:22:56 [reosarevok]
The only "easy" option I could think of is what I mentioned the other day, for each artist page where any cd stubs match the artist name, show "there are CD Stubs that might belong to this artist"
17:23:05 [ruaok]
not as another monolithic part of musicbrainz-server
17:23:18 [ruaok]
yep.
17:23:39 [ruaok]
if someone could draw up a wiki page/google doc with recommendations, I would be very happy.
17:23:42 [reosarevok]
(which *could* lead to people adding them to the wrong artist, but well, if they were to do that they'd have done it manually too)
17:24:12 [hawke_1]
reosarevok: Also if people had a dashboard they could see CD stubs for their artists, right?
17:24:35 [reosarevok]
hawke_1: probably, yeah, or a "report" of some kind
17:24:46 [reosarevok]
I'm just thinking of the ways that should be easy enough to implement
17:24:52 [reosarevok]
(dashboards are a bigger project)
17:25:18 [reosarevok]
(but just displaying the CD Stub thing? that could even be a hack weekend thing I bet)
17:25:20 [hawke_1]
Also doesn’t help that you have to completely fill in details when you add a CD stub
17:25:39 [hawke_1]
Ah, the problem is that CD stubs aren’t associated to an artist
17:25:39 [reosarevok]
hawke_1: see my comment before about making cdstubs on steroids
17:25:59 [reosarevok]
hawke_1: sure, display them when the name matches
17:26:10 [reosarevok]
That still applies to a good amount of them I bet
17:26:28 [reosarevok]
And yeah, ruaok: is there any reason cdstubs have a limited amount of optional data?
17:26:38 [hawke_1]
Right, I just meant that it is more than a simple lookup. We could certainly do an educated guess.
17:26:48 [reosarevok]
I guess if someone wants to enter all the fields, we should let them...
17:27:07 [ruaok]
reosarevok: they were basically a knock off of freedb.
17:27:27 [ruaok]
I'm ok with doing away with CD stubs and replacing them with something better.
17:27:58 [ruaok]
but, ocharles thinks that with the new edit system we could have this data in the normal system and then change its visibility as we need.
17:28:04 [ruaok]
that may be a perfect solution,
17:28:17 [ruaok]
so we may not want to spend much time on it before then.
17:28:30 [ruaok]
but a simpler CD Stubs with a key-value store wouldn't be much work.
17:28:41 [ruaok]
in the space of a day we were able to get some interesting things done.
17:29:10 [reosarevok]
Well, my point was more "have a second 'system' that is the same as the first, do the auto-ingestion there, have a way to move things from it to normal-MB"
17:29:29 [reosarevok]
So, basically, "make CD stubs equal as releases"
17:29:39 [reosarevok]
Because CDStubs *are* that system already, pretty much...
17:30:03 [kepstin-work]
i suppose the main difference between this and cdstubs is that we have artist mbids?
17:30:33 [ruaok]
kepstin-work: yep
17:30:53 [reosarevok]
Except that we do not...
17:31:17 [reosarevok]
I mean, how do we have mbids with ingestr or the IA data?
17:31:21 [reosarevok]
Or am I missing something?
17:31:32 [ruaok]
oh, wait. misread kepstin. cdstubs do not have mbids.
17:31:44 [ruaok]
they don't get them until they get moved to mb proper.
17:32:05 [ruaok]
we could give them MBIDs when they walk in the door, sure.
17:32:13 [ruaok]
its not like we're going to run out of MBIDs. :)
17:32:34 [kepstin-work]
no, I'm talking about artist mbids - I mean, for the data you're getting from IA does it include the mbid of the artist of the release?
17:32:48 [ruaok]
in some cases, yes.
17:32:52 [reosarevok]
kepstin-work: at most it could be a guess
17:33:12 [reosarevok]
Unless they're mapping it manually, and they're not...
17:33:15 [ruaok]
well, they are acoustid'ing everything.
17:33:34 [ruaok]
so if you get a handful of matches you can make apretty good guess about matching an album to mb.
17:33:46 [reosarevok]
Yeah, sure, but the ones that they actually match with acoustID are the ones that they already map themselves
17:34:00 [reosarevok]
They do have cases of "this fits several releases" though
17:34:04 [ruaok]
but you can infer more matches from that.
17:34:09 [reosarevok]
Guess for those, they could have an artist id if it all fits
17:34:15 [reosarevok]
(or even an RG ID)
17:39:12 [ruaok]
MBChatLogger: off
17:39:12 [MBChatLogger]
is not logging
17:48:19 [MBChatLogger]
is logging
17:48:24 [demosdemon]
MBChatLogger disagrees
17:48:51 [reosarevok]
MBChatLogger, on
17:48:51 [MBChatLogger]
is already logging
17:48:56 [demosdemon]
MBChatLogger: log
17:48:56 [MBChatLogger]
is already logging
17:48:58 [reosarevok]
MBChatLogger, wtf
17:48:58 [MBChatLogger]
I'm logging. I don't understand 'wtf', reosarevok. Try /msg MBChatLogger help
17:49:04 [reosarevok]
MBChatLogger: off
17:49:04 [MBChatLogger]
is not logging
17:49:08 [MBChatLogger]
is logging
17:49:12 [reosarevok]
MBChatLogger: dieabit
17:49:12 [MBChatLogger]
I'm logging. I don't understand 'dieabit', reosarevok. Try /msg MBChatLogger help
17:49:14 [warp]
MBChatLogger: off
17:49:14 [MBChatLogger]
is not logging
17:49:22 [MBChatLogger]
is logging
17:49:33 [reosarevok]
ruaok, restart the thing :p
17:49:34 [warp]
MBChatLogger: DROP TABLE puid;
17:49:34 [MBChatLogger]
I'm logging. I don't understand 'DROP TABLE puid;', warp. Try /msg MBChatLogger help
17:49:35 [kepstin-work]
is it still logging #musicbrainz?
17:49:43 [demosdemon]
It needs to be restarted
17:50:04 [MBChatLogger_1]
MBChatLogger_1 has joined #musicbrainz-devel
17:50:04 [MBChatLogger_1]
is logging
17:50:04 [kornbluth.freenode.net]
Users on #musicbrainz-devel: MBChatLogger_1 ruaok hawke_1 reosarevok thejav plaintext nikki warp mathieu_ DremoraLV nikki_ navap_ Jormangeud dimonov sezuan luks alastairp_ MBJenkins kepstin-work ianmcorvidae MiX-MaN bitmap lfranchi murdos pthulin Wizzcat Muz voiceinsideyou dukelet0 djinni`_ MBChatLogger demosdemon alastairp navap Mineo adrianw sampsyo flamingspinach njh ocharles adhawkins-away noobie DWSR night199uk frewsxcv
17:50:18 [demosdemon]
ohhai MBChatLogger_1
17:50:30 [reosarevok]
hah
17:50:42 [reosarevok]
It was working here when we said on, it just didn't say it
17:50:44 [kepstin-work]
http://chatlogs.musicbrainz.org/musicbrainz-devel/2012/2012-08/2012-08-15.html - it was logging all our attempts to turn it back on
17:50:53 [ruaok]
lol
17:51:35 [demosdemon]
oh silly bots
17:52:32 [MBChatLogger]
MBChatLogger has joined #musicbrainz-devel
17:52:32 [MBChatLogger]
is logging
17:52:32 [card.freenode.net]
Users on #musicbrainz-devel: MBChatLogger ruaok hawke_1 reosarevok thejav plaintext nikki warp mathieu_ DremoraLV nikki_ navap_ Jormangeud dimonov sezuan luks alastairp_ MBJenkins kepstin-work ianmcorvidae MiX-MaN bitmap lfranchi murdos pthulin Wizzcat Muz voiceinsideyou dukelet0 djinni`_ demosdemon alastairp navap Mineo adrianw sampsyo flamingspinach njh ocharles adhawkins-away noobie DWSR frewsxcv night199uk
17:53:47 [demosdemon]
oh crap
17:53:53 [reosarevok]
thanks, MBChatLogger
17:53:55 [reosarevok]
Very useful
17:54:03 [kepstin-work]
lol
17:54:06 [warp]
nice
17:54:09 [kepstin-work]
it caught up :)
17:54:31 [demosdemon]
oh
17:54:33 [demosdemon]
bye
17:55:57 [plaintext]
hi ruaok
17:56:51 [ruaok]
plaintext: hi!
17:56:53 [kepstin-work]
something's weird with the logging right now; it was continuing even while the MBChatLogger was not here.
17:56:59 [ruaok]
the logs import died today.
17:57:06 [plaintext]
oh, how come
17:57:34 [ruaok]
I sent you the email from the log daemon.
17:57:42 [ruaok]
some bug in the code form yesterday.
17:57:57 [ruaok]
see if you can find it. I can try and re-run it and isolate the line that caused the trouble
17:58:02 [ruaok]
then we can run the rest of the log.
17:58:09 [ruaok]
but that will have to wait 2-3 hours.
17:58:15 [kepstin-work]
http://chatlogs.musicbrainz.org/musicbrainz-devel/2012/2012-08/2012-08-15.html#T17-54-31-824026 - MBChatLogger logged some stuff while it wasn't connected?!
17:58:19 [plaintext]
on it
17:58:28 [ruaok]
I have pressing foundation stuff to take care of.
17:58:31 [ruaok]
k
17:58:41 [plaintext]
okay
18:00:04 [plaintext]
the weird thing is that the number of fields separated by spaces should be constant for all the lines
18:00:13 [plaintext]
I could add a try block and print the line where we get an error
18:00:30 [ruaok]
plz
18:00:37 [ruaok]
then I can re-run the sanitzer.
18:02:01 [ruaok]
feh. I added ingestr to the list of projects for the weekend, but failed it his save.
18:02:02 [ruaok]
lame
18:05:41 [demosdemon]
The only thing I can think of that explains why logger got those lines even though it was disconnected was if freenode's server still sent the lines to the socket before it terminated it and logger got it when it restarted
18:05:52 [demosdemon]
but, whatever, not a big deal… just boggles my mind
18:06:24 [plaintext]
ruaok: pushed
18:10:57 [plaintext]
by the way I was really desperate about the manifest error, so I tried the CR/LF thing, and turns out that was causing the error
18:13:47 [ruaok]
LOL. noes.
18:13:50 [ruaok]
that sucks. :(
18:13:53 [ruaok]
sorry I couldn't help.
18:14:15 [plaintext]
well, that's what I get for using windows :P
18:14:32 [plaintext]
but it works now, so I moved the script away, and also added pie charts
18:18:00 [ruaok]
plaintext: OT question: how much longer are you at uni?
18:18:09 [ruaok]
I'd love to have you back next year. :)
18:18:44 [plaintext]
wow :) I'd love to be back as well
18:18:52 [plaintext]
well, 1 semester, then masters
18:19:11 [the_metalgamer]
the_metalgamer has joined #musicbrainz-devel
18:20:34 [ruaok]
awesome then. :)
18:20:42 [ruaok]
don't make any serious plans for next summer then. :)
18:21:23 [plaintext]
:)
18:22:59 [reosarevok]
Or rather, make them :p
18:23:16 [reosarevok]
Decide what you'd like to work on
18:23:18 [reosarevok]
I guess
18:23:35 [ruaok]
reosarevok: +10
18:23:43 [ruaok]
next year you might have your own ideas, plaintext
18:23:50 [reosarevok]
You're allowed to :p
18:23:58 [ruaok]
yes, he is.
18:24:04 [plaintext]
wow
18:24:16 [ruaok]
ok, accountants, post office, bank and then office.
18:24:18 [reosarevok]
I mean, it kinda sounded a bit like "we will now let you actually think" :p
18:24:23 [ruaok]
back soonish.
18:24:29 [reosarevok]
"enjoy"! :p
18:24:32 [plaintext]
hehe
18:24:34 [reosarevok]
(don't kill them)
18:24:35 [ruaok]
heh
18:41:41 [Jormangeud^]
Jormangeud^ has joined #musicbrainz-devel
18:45:16 [dimonov_]
dimonov_ has joined #musicbrainz-devel
19:07:23 [plaintext]
on codereview how do I know the hash of the last commit the diff covers?
19:07:34 [plaintext]
this is my old review request: http://codereview.musicbrainz.org/r/2100/
19:07:48 [plaintext]
I should have updated it much earlier :l
19:13:10 [warp]
plaintext: you should always submit a diff against whatever you branched off of
19:13:49 [plaintext]
oh so always the first commit?
19:13:56 [warp]
plaintext: presumably you branched from master. so do a git diff master...HEAD > myreview.patch to submit to code review.
19:14:09 [plaintext]
thanks!
19:14:10 [warp]
it doesn't matter if it's a new review or if you're updating an existing one.
19:16:31 [ruaok]
ruaok has joined #musicbrainz-devel
19:19:54 [ruaok]
* ruaok starts pushing the log to splunk again
19:19:56 [voiceinsideyou]
voiceinsideyou has joined #musicbrainz-devel
19:24:24 [plaintext]
warp: that gives me an empty file for some reason
19:25:52 [demosdemon]
plaintext: are you committing to master locally?
19:26:37 [plaintext]
uh oh, I'm still very new to git to know this
19:26:42 [demosdemon]
no
19:26:43 [demosdemon]
it's fine
19:26:45 [plaintext]
* plaintext should really read into how this works more
19:26:55 [plaintext]
I just use "git commit -m '...'"
19:28:09 [demosdemon]
find the hash for the first commit not part of this patch
19:28:22 [demosdemon]
and use git diff hash…HEAD instead
19:29:01 [plaintext]
that's what I tried first, but I might have tried the wrong hash because I had an error
19:29:07 [plaintext]
let me try again
19:30:09 [demosdemon]
It might be easier to use a gui until you learn some of the command line utilities, which one idk cause I've never used git on windows
19:30:23 [demosdemon]
github has one, I think
19:30:28 [demosdemon]
yeah, http://windows.github.com/
19:30:40 [plaintext]
thanks demosdemon
19:34:33 [ruaok]
plaintext: I moved the salts to a salts.py.default file.
19:34:46 [ruaok]
for reasons we already covered on the other stuff.
19:35:22 [plaintext]
okay
19:36:36 [plaintext]
I had no luck finding the correct hash :l maybe something else is wrong. I'll try the gui
19:37:12 [ruaok]
pushed.
19:37:17 [ruaok]
finding the right hash??
19:38:27 [ruaok]
ok, running the log again.
19:38:40 [ruaok]
lets see how long it takes to find the bad line
19:39:19 [plaintext]
ruaok: I had trouble updating the diff on codereview
19:39:33 [ruaok]
oh codereview again. :(
19:40:34 [plaintext]
yeah :(
19:47:21 [ruaok]
plaintext: this line failed
19:47:34 [ruaok]
1344952045.622 x.x.x.x "GET /user/vesterrusso614 HTTP/1.0" 404 11021 z=- up=10.1.1.23:80 ms=0.059 ums=0.059 ol=- h=musicbrainz.org
19:47:59 [plaintext]
aww
19:48:07 [plaintext]
2 spaces between url and HTTP/1.0
19:48:43 [ruaok]
:-)
19:49:18 [plaintext]
I'm not sure what format nginx logs are supposed to have
19:49:42 [plaintext]
didn't know there can be 2 spaces there :)
19:51:53 [ruaok]
lets just assume that extra spaces can happen.
19:52:11 [plaintext]
I'm wondering what would be the fastest method here
19:52:50 [ruaok]
http://docs.python.org/library/stdtypes.html
19:52:55 [reosarevok]
Do the extra spaces mean something?
19:52:55 [ruaok]
read str.split
19:53:00 [ruaok]
second paragraph
19:53:08 [reosarevok]
If not you can just auto-turn them into one
19:53:49 [ruaok]
sep=None on the split would do it.
19:55:04 [plaintext]
great
19:55:18 [plaintext]
I was already thinking about some regex and "\s+" :l
19:55:34 [ruaok]
yeah, that would an extra pass over the data.
19:56:40 [ruaok]
oh, crap. it doesn't work the way I thought.
19:57:39 [plaintext]
why?
19:58:02 [ruaok]
<== retarded
19:58:28 [ruaok]
>>> d.split()
19:58:29 [ruaok]
['1344952045.622', '74.121.193.90', '"GET', '/user/vesterrusso614', 'HTTP/1.0"', '404', '11021', 'z=-', 'up=10.1.1.23:80', 'ms=0.059', 'ums=0.059', 'ol=-', 'h=musicbrainz.org']
19:58:36 [ruaok]
that should be fine.
19:59:00 [plaintext]
oh, so split(None) equals split()
20:00:09 [ruaok]
ok, running again.
20:01:12 [plaintext]
should I change the code and push it?
20:01:28 [plaintext]
if so, should I remove the try block we added for testing?
20:02:29 [ruaok]
I'll clean up and commit once we're past this issue.
20:02:34 [ruaok]
carry on with what you were doing.
20:03:26 [plaintext]
okay
20:12:23 [ruaok]
ran without trouble this time. :)
20:12:57 [plaintext]
great!
20:16:47 [ruaok]
pushed.
20:19:56 [ruaok]
time to find noms.
20:19:57 [ruaok]
brb
20:39:57 [noobie]
noobie has joined #musicbrainz-devel
20:40:08 [Leftmost]
Leftmost has joined #musicbrainz-devel
20:40:08 [Leftmost]
Leftmost has joined #musicbrainz-devel
21:53:57 [ruaok]
plaintext: log imported.
21:54:03 [ruaok]
it was only 6.5gb today.
21:54:06 [ruaok]
well done. thanks!
21:54:11 [plaintext]
wow :)
21:57:53 [reosarevok]
:)
21:58:02 [reosarevok]
(also, heh, "only" :D)
21:58:23 [reosarevok]
That's more than the space I have left in this disc
21:58:56 [reosarevok]
Is it all text?
21:59:10 [reosarevok]
* reosarevok 's mind fails to imagine just how much text fits there :)
21:59:27 [ruaok]
all text.
21:59:39 [ruaok]
and its one line per every request we send.
22:00:44 [ianmcorvidae]
heh
22:00:53 [ianmcorvidae]
that's after all the trimming/quasi-compressing we did yesterday?
22:02:32 [plaintext]
yes
22:13:11 [ruaok]
ok, off to really finish the taxes now.
22:13:13 [ruaok]
ttfn
22:44:40 [ruaok]
ruaok has joined #musicbrainz-devel
22:45:21 [reosarevok]
reosarevok has changed the topic to: Next agenda: Reviews, opening releases in robots (reotab), 3841, 2650, 4282, 3204, 1935 (ocharles)
22:46:07 [ianmcorvidae]
release-group might be a better target, reosarevok, but we can discuss it at the meeting I guess :)
22:46:50 [reosarevok]
ianmcorvidae: not too sure. would that still allow MB to appear when searching for a barcode for example?
22:47:02 [ianmcorvidae]
hm
22:47:13 [ianmcorvidae]
I guess not if they don't appear on the release-group page
22:47:41 [ianmcorvidae]
they do appear there
22:47:43 [reosarevok]
They do
22:47:51 [ianmcorvidae]
so dunno
22:47:58 [ianmcorvidae]
presumably they should be added to the meta tag
22:48:02 [reosarevok]
Yeah
22:48:05 [ianmcorvidae]
if they aren't there already
22:48:09 [reosarevok]
For /release, certainly
22:48:15 [reosarevok]
Also cat#
22:48:20 [ianmcorvidae]
heh, I don't see a meta tag for release-group at all
22:48:40 [ianmcorvidae]
nor release
22:49:16 [ianmcorvidae]
I guess kepstin's patch only added it to artist
22:49:29 [ianmcorvidae]
yup
22:49:31 [ianmcorvidae]
heh
22:49:32 [reosarevok]
Probably
22:53:56 [plaintext]
plaintext has joined #musicbrainz-devel
23:00:13 [murdos]
for reference, I agree that release-group might be a better target too
23:12:32 [ianmcorvidae]
ocharles, warp: do we have somewhere we could generate/display something like http://www.mbsandbox.org/~ianmcorvidae/cover_db/coverage.html other than on my sandbox? I think it might be a useful thing to have available for refactoring, but less so if it's only for me and/or only updated manually
23:26:55 [ruaok]
what is it not apprpriate to run that in mbsandbox?
23:27:27 [ruaok]
with nginx we could easily do something like dev.musicbrainz.org internally redirecting to a specific sandbox
23:27:28 [ianmcorvidae]
ruaok: it's not that it's inappropriate, it's that it's manually-updated and not very visible
23:27:46 [ruaok]
cronjob and add dev.mv org to the menu?
23:28:10 [ianmcorvidae]
possibly; it seems like it should be integrated with Jenkins
23:28:25 [ianmcorvidae]
(it's created by running tests, just by running tests with some extra incantations to run Devel::Cover as well)
23:28:38 [ruaok]
ah, hobbbes then.
23:29:35 [ianmcorvidae]
mostly I guess I'm wondering if others see this as a useful thing to make available
23:29:50 [ruaok]
I do.
23:31:04 [reosarevok]
So, is this a test for our lack of tests?
23:31:29 [ianmcorvidae]
something like that
23:31:31 [reosarevok]
How meta
23:31:37 [ianmcorvidae]
it gives us a good view into where we need to *add* tests
23:31:51 [reosarevok]
Yeah, it does sound useful
23:31:57 [reosarevok]
For those people who can navigate that table :)
23:32:20 [ianmcorvidae]
fixing some of the places in this I've already tackled, I've already fixed a number of not-yet-reported bugs
23:32:42 [kepstin]
kepstin has joined #musicbrainz-devel
23:32:46 [ianmcorvidae]
not necessarily important ones, e.g. an ISE on statistics pages when there aren't statistics (will never ever apply to production)
23:32:48 [ianmcorvidae]
but still