#musicbrainz-devel

yeah

I decided I'm ignoring the line-length part of PEP8, because it just results in a bunch of gross hacks to make it actually fit

some of that is because my code probably could be better, but not much

... heh, oops, it's pep8 checking the *dependencies*, that's clearly wrong

Project geordi build #9: STILL FAILING in 5 min 20 sec: http://ci.musicbrainz.org/job/geordi/9/

* Ian McEwen: PEP8: views.py, except E501

* Ian McEwen: PEP8: search.py, except E501

* Ian McEwen: PEP8: __init__.py, except E501

* Ian McEwen: PEP8: mappings subdir, except E501

* Ian McEwen: PEP8: ignore E501, as it results in nasty hacks

Huh. last.fm doesn't associate a release MBID with tracks. Is that because they're using pre-NGS structure?

Project geordi build #10: STILL FAILING in 2.8 sec: http://ci.musicbrainz.org/job/geordi/10/

Ian McEwen: Don't test dependencies in virtualenv for pep8-conformity

Project geordi build #11: FIXED in 3.1 sec: http://ci.musicbrainz.org/job/geordi/11/

Ian McEwen: fix my idiotic exclusion of the actual *code* in the if block

apparently the pep8 test runs a lot faster when it confurms

conforms*

the line length thing is really good because it forces you to introduce both small functions and meaningful variables

rather than cramming an insane amount onto one line

* nikki: Add files for converting Discogs data dumps

* Ian McEwen: PEP8 fixups for discogs scripts.

heh

hey, there's even three or four lines that *didn't* touch :P

the empty ones? :D

nah, also those that aren't indented :P

it's no secret that I'm a tab lover.

you also had some spaces after [ and before ], and some extra spaces after operators :P

hm. would be nice if github would show me the differences without the indentation

ianmcorvidae, still can't log in to geordi.

VxJasonxV just told me about that earlier today :)

Leftmost: add beta. to the url it takes you to

Leftmost: will fix *that* bit now :)

Thanks.

hm, but that's all the whitespace changes gone :/ I knew about the tabs vs spaces thing, but not the others

Ian McEwen: Use beta.mb as the base for oauth, until it ships to production.

Leftmost: should be fixed

nikki: oh, hm

Out of curiosity, what exactly does matched by index mean?

nikki: https://github.com/metabrainz/geordi/commit/e05..., and 83 and 92 (two spaces after +), and 116/118 for the spaces inside the []

Leftmost: it means that the data as we got it already had an MBID in it, since the provider of the data did their own matching

ah, thanks

Ahh.

the first was just a typo, I wouldn't do that intentionally

and the has_key() -> in changes you saw on the whitespace one

yeah

Leftmost: the wcd index matching was acoustid-based

as in, they took the files (which they, of course, actually have) and acoustid'd them, and depending on some criteria that aren't exactly clear (but which seem to be "everything in the release has an acoustid that matches to a recording, and all those recordings are in the same release")

... they made release and release-group matches

the recording ones are purely acoustid-based

Ahh, okay. Thanks.

though I think they've now imported luks's process, which used durations and track names

(the results of his thing, I mean, I don't think they're actually doing it themselves :/)

Matching last.fm library/scrobble data against MB data is going to be difficult.

ianmcorvidae, wasn't that yesterday?

/irrelevant

anyone avail that can help troubleshoot an install error?

Project libdiscid » autotools build #21: FAILURE in 19 sec: http://ci.musicbrainz.org/job/libdiscid/./build...

* github: win32: autotools support for versioninfo

* github: win32: libdiscid.rc -> versioninfo.rc

Project libdiscid » cmake build #21: FAILURE in 23 sec: http://ci.musicbrainz.org/job/libdiscid/./build...

* github: win32: libdiscid.rc -> versioninfo.rc

good to have Jenkins in the chat, because I did not receive a mail about this (yet)

but it also looks like jenkins doesn't start with a clean working copy

ocharles: next time you're around, how does the EMAIL_BUGS feature thing work? It's supposedly sending things to root, which I think supposedly sends them to us, but I certainly haven't been getting any emails :P so

In the log it says a mail was sent to my address. The address is correct, but I did not receive it yet (and never received mails from jenkins)

hm

I don't know about jenkins emailing :/

Project libdiscid » autotools build #22: FIXED in 14 sec: http://ci.musicbrainz.org/job/libdiscid/./build...

* github: autotools: no conditional LT_INIT

* github: linux: always leave room for \0 in ISRC buffer

* github: autotools: no conditional LT_INIT

I think I can't fix that build without ocharles, because it needs a build file reset. ^

I have access to all the servers and jenkins stuff, what exactly do you need?

changing from "cmake ." to "rm -r CMakeFiles/ CmakeCache.txt; cmake ."

in the build config

I kind of was thinking jenkins builds in a fresh environment every time, but this doesn't seem to be the case

ah

added, building now

Project libdiscid » cmake build #23: STILL FAILING in 12 sec: http://ci.musicbrainz.org/job/libdiscid/./build...

heh

CMakeCache.txt

but you can change it to rm -rf too

that makes more sense, I suppose

and yeah, I'll do that

Project libdiscid » cmake build #24: STILL FAILING in 13 sec: http://ci.musicbrainz.org/job/libdiscid/./build...

woot

it sure doesn't exist. I renamed it.

that looks like a real failure this time :)

looks like it's trying to send emails to 'github' and failing, also, so I'll turn off the "Send separate emails to individuals who broke the build" feature

gut the individual in this case is given as "Johannes Dewender <github [at JonnyJD.net]>"

the only place where jenkins could take "github" from is the mail address..

yeah, but it's missing the @JonnyJD.net and doesn't seem to have a way to know that

it'll send to the configured address for you specifically still

why would it be missing the @JonnyJD.net? It's part of the commit and in a standard way.

looks like a bug in the plugin to me

probably

disabling is still the fix :P

ah, it doesn't autocreate an email address at *all* for things, it seems

so it's just sending it to nobody :/

Yippie, build fixed!

github: cmake: finish versioninfo rename

hm, ocharles put my brainz address in the config explicitely, but that didn't reach me either.

yeah, that bit should still have been sending

but I think the broken email may have made the whole email-sending bit fail

so we'll see

yeah. Anyways. Would be great if that works. I like to know when my build fails and I might not always be in the chat.

yeah, understood :)

nobody is probably the user that runs the jenkins build. And mail might not be configured on that machine.

Now that is fixed, I should go to bed.

I guess you just have to break it again so you can check it works :P

Well, breaking stuff for jenkins seems to be a daily activity anyways

haha

so we don't have to force it ;-)

It should be the same issue for mb-server

I don't think mbserver sends emails, though, since all the main devs idle in here nonstop :P

Usually my fault anyhow.

speaking of, I've discovered collection subs are making subscribed entity pages 502!

but it's not actually your fault :P

It's sort of my fault, it's just that the query is so slow.

Make your collection smaller! :-P

I'm not looking at *mine*, lol

mine works fine

Make other people's collections smaller!

or I could add a status column to edit_release so it doesn't have to touch the edit table :P

True.

omfg. reo and nikki both submitted invoices. hours early. minutes apart. O_O

:P

you'll get mine in two hours, unless I decide I'm not going to get any more work done and ship it out a bit early :P

:-)

the paypal account almost has enough money too.

well, reo mentioned it, so I was like "oh yeah!" and thought I'd surprise you :P

you did. :)

I need to add another $200 to the account.

nikki: would you like everything but $200 now, or the whole bit later?

like monday or so.

eh, the whole bit later is fine

:-)

ok

just delays the order of japanese music a couple days ;)

ruaok, it was mentioned to me that the FreeDB gateway could use adding to nagios. Would it be okay for Ian to give me the necessary access to do that?

yes. that sounds like a great project for you to look after.

ianmcorvidae: given that I use a credit card for that, it doesn't even do that :P