#musicbrainz

oh god, is that still actually going on? also does the amazon link *still* fetch cover art?

I actually did not link a amazon link to a release because the coverart was unsatisfactionary, and I was afraid it would over ride the proper one?

what's wrong with the word "unsatisfactionary" :/

stypid spell check doesn't give me any alternative

*u

It’s clearly unsatisfactory.

I like "unsatisfactionary".

It sounds like a youth movement.

While "unsatisfactory" sounds like the place where ikea tables are made.

unsatisfactorily

that. doesn't sound right to me. and it being clear isn't true when i couldn't remember it nor. so it being "clear" is only for people who aren't me. wich is everyone, but it's stil la bt grating ot be told that. actyally

ha CallerNo6

also thanks guys :D

was CatQuest looking for "unsatisfying"?

dissatisfying ;)

I don't know, unsatisfactionary is a perfectly cromulant word

oh. "dis". I like it. More negativity.

dis is a for of light fog. in .no

form*

ohh, nice, looks like 'date' is not a mandatory field

CatQuest: I read the "clearly" as a reference to your spellchecker, FWIW, not to your ability to spell.

SothoTalker_: Few fields are.

is there a list of fields that are strictly neccessariy? :-)

SothoTalker_: Fields that have their field name in bold, generally, IIRC.

hmm.. well my spellchecker is dumber than I am :/

E.g., the only mandatory fields for artist are "Name" and "Sort Name", which are the only two bolded field names: https://musicbrainz.org/artist/create

for releases only artist and title, ok :)

so i have to check for all the others.

hah! nikki says the root of the word comes from french "satisfactoire", so she guesses because it wasn't created from english morphemes, just imported, and that's why it's oddly not "unsatisfactionary"

SothoTalker_, you can also look at createtables.sql for fields with "not null".

my py is already ugly enough ;)

preach!

hmm maybe i need ot eat something so my brain works better

do you know of a release that has more than 2 labels?

that question could mean more than one thing.

lots of mb releases have more than one release label set, because more than one logo appears on the release.

some have more than one label in the sense that they were a joint venture between two unrelated labels, with two catalog numbers.

(relatively rare)

ugly

i am still thinking about this release: http://musicbrainz.org/release/ccd965f1-6292-43...

sorry for that :D

SothoTalker_, yeah, that's the former case. Many editors feel that "release label" means "list of imprint/logo thingies visible on the packaging"

maybe we should have "special this is not the lable-label" fields/type of that into it? ybe will that shut the "it's not *The* lable peopel up? i have no idea

If you want to filter them, you can try discarding any release label that doesn't have a catalog number paired with it.

I'm just very, very, very happy that the stuff i edit is in other circles than the people wo work loads with labels :/

CatQuest, IIUC the 'release label' field will eventually migrate to a relationship, so yeah, there might be a different rel for parent companies or whatever.

I always worry I'll come across someone who complains about some label+catno i added/made and say so and such is wrong and whatever and i wouldn't know how to argue back becaseu i kno only what i see on the release :/

CallerNo6: there are relationships to add other labels. :<

SothoTalker_: Yep! Relationships for all the things!

why the long face?

nobody uses them ;)

or is that a hipster mustache? I'm not good at emoji.

the rel for things thing i mean not the peopel not using it

the rel editor is awesome, the only thing missing is typing aliases to the ReleaseE/entityadding and things would be s mcuh better

tying not typing

ok, now with https for everything :p

<CatQuest> stypid spell check doesn't give me any alternative – <chirlu> It’s clearly unsatisfactory.

Is the reference clearer now?

stil vagly implies i should have *clearly* bee nable to see that of my own accord :)

SothoTalker_:

s=848&conditions.0.args=349

"at least 500 edits" disagree that "nobody uses [release-label relationships]".

.. i just thought about what if we had an automatic url shortener for complex search-urls like some sites (like twitter) has?

🙌

Unnecessary. I have a branch to make the regular URLs much shorter.

chirlu: it's really, really hard to read intent outa text, especially with irc, and especially with emoticons :/ sorry

well i didn't know about that.

also, there are edit searche, and jira ticket filter searches

I need to finish it some day, but it is working.

i hope you do not mind me 'abusing' the ws api while i write my script :S

having an mbs.co/?gfgfdg isn't a bad idea

chirlu: Looking forward to see the PR one day in the not too distant future. :)

it's probably going to make jesus irritated since he does weird ulr hackings with it :P

SothoTalker_: As long as you don't send more than the 1 req/second, you're not abusing.

more like 1 per minute

and if you *do* tne it'll give you an error anyway?

Ticket is MBS-5562 (and it’s by jesus2099).

MBS-5562: The edit search is generating very long URL https://tickets.metabrainz.org/browse/MBS-5562

hah, that's nothing :) don't worry

SothoTalker_: That's far, far, far from being abusive.

again I'm nto serious, I was joking.

sorry

is there a way to just get the number of results?

SothoTalker_: For… ?

so that i can then calculate how many requests i need to get all the data ;)

but then, i can also do that when i get the first 100 entries.

the xml gives a count

I wonder why that's not included in json. maybe the assumption was that it's easy enough to get len(releases)?

(once you've done all your requests)

the json gives also a count

does it?

yes

but this just means i have to treat the first request specially

oh, pfft, I was looking at a lookup, not a browse request. yes, there it is. "release-count".

SothoTalker_: Subsequent requests don't include the count?

they do

Just update the count with each request?

(Theoretically, the count may also change from between two requests.)

i can fetch a maximum of 100 entries per requests

so if the count is > 100 i need to make a second request with an offset to get the remaining data

Yep.

i only need the count for this ^

But what if there are 399 entities, and by the time you're doing your 3rd fetch, two more have been added, so there's now 401.

then i have to treat the last request specially, too :D

for want of an entity, the kingdom was lost

SothoTalker_: Or just get the count from each request and evaluate then whether you need to fetch more.

hmm.. meep :p

Then you don't need to treat any request differently.

wouldn't the logic be something like `if release-count==100 make another request`?

CallerNo6: More like...

oh, wait, that's wrong. not release count, but number or releases returned in the last request.

offset=0; get offset–(offset+100) releases; if count from request == total number of releases fetched: done; else: offset=(total number of releases fetched) and repeat

Freso: yeah, the person behind bs1770gain (and r128gain before it) is known to have had a bunch of personal conflicts with users and other devs too

"2016-11-10 To all you great people out there in the US: Thank you for electing Donald Trump!"

if you need an alternative replaygain tool, feel free to use https://github.com/kepstin/regainer - requires recent ffmpeg command-line tool, written in python3 with mutagen, works on vorbis, opus, mp3, mp4, flac

sampsyo: ^

That's cool! This wraps a built-in feature on ffmpeg, it seems?

yes, I use the built-in ebur128 scanner in ffmpeg, then just use mutagen to write tags

Nice. That might be just what we need. Thanks!

(kepstin: I became aware because sampsyo pointed it out in #beets.)

it's not that well documented; if you need help just poke me here.

I shall!

I suppose the '--help' option should work, and if you want to implement something in beets itself, the code should be a decent reference

note that I'm parsing the text output of ffmpeg's log messages for the gain values - this isn't /guaranteed/ to be stable, but there's no other way to do it atm.

(without actually using ffmpeg as a library, at least)

That makes sense; thanks!

We do a lot of parsing ffmpeg output, so that wouldn't be anything new for us. :)