these swagger docs are not really verbose, I need to check the docs
2021-10-06 27928, 2021
gcrkhehim[m]
ruaok: There are only a few devs who are working on native funkwhale clients and I don't think its a nice idea to put pressure on them. We can add API endpoints, but we need to do this the right way. But in this case I think everything is in place to get it somehow working
2021-10-06 27928, 2021
monkey
Yes, the combination search endpoint - get first returned track - get listen_url -play content seems like it would work well.
2021-10-06 27956, 2021
outsidecontext
this integration of FW in the LB player would be awesome indeed. I'd love it
2021-10-06 27935, 2021
tandy[m]
+ an ability to link your funkwhale instance to your listenbrainz profile would be useful, that way people could easily share their music libraries with listenbrainz
2021-10-06 27949, 2021
tandy[m]
s/+/plus/, s/an/the/
2021-10-06 27900, 2021
tandy[m]
s/+ an ability to link your funkwhale instance to your listenbrainz profile would be useful, that way people could easily share their music libraries with listenbrainz/plus the ability to link your funkwhale instance to your listenbrainz profile would be useful, that way people could easily share their music libraries with others on listenbrainz/
2021-10-06 27917, 2021
tandy[m]
so you arent just limited to the open.audio instance
2021-10-06 27928, 2021
gcrkhehim[m]
monkey: You cannot specify the fields to search, only do a string search
2021-10-06 27943, 2021
monkey
Thanks for checking that.
2021-10-06 27938, 2021
monkey
I'm guessing that might result in less precise results, judging from using the youtube search
2021-10-06 27936, 2021
monkey
On the other hand if MBID is available and returns a result that's ace
2021-10-06 27942, 2021
gcrkhehim[m]
You'll get artists, tracks, albums and tags in different arrays, so you can filter this way
2021-10-06 27943, 2021
gcrkhehim[m]
So you can emulate a album search by doing search?q=$name and then only look at album in the response
2021-10-06 27944, 2021
monkey
That might require doing two searches in a row, one with MBID and one with the song title + artist
2021-10-06 27901, 2021
monkey
Yep, I saw that, it's really convenient
2021-10-06 27913, 2021
monkey
And If I look up an album title and get a result in the album array, I can go fetch a song from that (BrainzPlayer on really manages to play single songs at the moment)
2021-10-06 27948, 2021
monkey
That's really convenient
2021-10-06 27940, 2021
yvanzo
reosarevok: About our previous discussion on entering a second date period: Wouldn’t #2288 make it impossible?
2021-10-06 27912, 2021
Etua joined the channel
2021-10-06 27930, 2021
reosarevok
No, that seems to just merge them into the same *block* but not into one relationship (it still shows it's a dupe)
2021-10-06 27905, 2021
reosarevok
Obviously I can't currently submit, because they're detected as dupes, but I could set two dates and they were shown differently and correctly in the editor
2021-10-06 27921, 2021
nawcom joined the channel
2021-10-06 27922, 2021
Etua has quit
2021-10-06 27957, 2021
Etua joined the channel
2021-10-06 27937, 2021
Etua has quit
2021-10-06 27905, 2021
reosarevok
bitmap: meh, I'm not sure what's the best way to do this in Perl either. I'm doing horrible things like this already: my ($CLEANUPS) = $file_content =~ m/const CLEANUPS: CleanupEntries = ({.*};)\s+function testAll/gsi;
2021-10-06 27926, 2021
reosarevok
(which will break as soon as someone changes the URLCleanup file)
2021-10-06 27934, 2021
reosarevok
Honestly, duplicating might be the simplest option
2021-10-06 27931, 2021
reosarevok
Maybe we can have a test that checks the regexes are the same or something, if we want to make sure we don't forget to duplicate changes
2021-10-06 27904, 2021
bitmap
ok. I'd prefer importing them into URLCleanup, but whatever you think is best there
2021-10-06 27914, 2021
zas
I'll switch to herb in order to test nginx params (related to listen drops we get atm)
2021-10-06 27934, 2021
reosarevok
yvanzo: what do you feel? Turns out parsing with either JS or Perl to save into its own file is quite annoying
2021-10-06 27933, 2021
reosarevok
So we have the easiest two options of either keeping the match and having a separate file with only regexes which we keep in sync by hand, or having the regexes on a separate file and importing it into URLCleanup, it seems
2021-10-06 27916, 2021
bitmap
in vscode I can press F12 to jump to the definition for something, & in vim you can do the same with the right LSP plugin, so I think it's not so bad to import them
[listenbrainz-server] 14MonkeyDo opened pull request #1643 (03master…monkey-brainzplayer-events): Use browser events for communication with BrainzPlayer https://github.com/metabrainz/listenbrainz-server…
2021-10-06 27902, 2021
monkey
This brings us a step closer to an embeddable BrainzPlayer ^
It will be very handy for embedding it in other pages
2021-10-06 27947, 2021
tandy[m]
Oh awesome i just saw that
2021-10-06 27938, 2021
monkey
I've thought about the standalone player project a bit more, and with that PR it seems to me we're not far off from being able to use it in other websites.
2021-10-06 27901, 2021
tandy[m]
im excited!
2021-10-06 27903, 2021
monkey
I still have to figure out how to deal with some duplicated code; BrainzPlayer can live in a separate package, but we currently also need access to some API service in the codebase. Separating the two might be a good self-contained project if you're interested in putting your hands in it ;p
2021-10-06 27955, 2021
tandy[m]
monkey: yes! if you could give me some details i can get stuck in
2021-10-06 27923, 2021
monkey
Awesome ! let me write that up
2021-10-06 27907, 2021
ruaok
lucifer: alastairp : this report https://sentry.metabrainz.org/organizations/metab… is being caused by a container that I didn't update yesterday. but which one? is there any way to find out which container this issue came from?
2021-10-06 27926, 2021
alastairp
I see the tag `environment: test`, which seems to indicate to me that it's coming from something that's loading the test consul config
2021-10-06 27945, 2021
ruaok
ah, a good tip.
2021-10-06 27949, 2021
alastairp
good point though, we should be able to do something like set a tag which is the docker container id, which would make this easier to track down
2021-10-06 27959, 2021
ruaok
monkey: can you please merge master into your branch and re-deploy to test?
2021-10-06 27906, 2021
monkey
Sure thing !
2021-10-06 27912, 2021
ruaok nods at monkey and alastairp
2021-10-06 27945, 2021
monkey
Stupidly I looked up to see if I missed your arrival at the office and you were physically nodding at us…
2021-10-06 27912, 2021
ruaok
nope, no office for me today. I'll be in tomorrow.
2021-10-06 27917, 2021
monkey
tandy[m]: So as I was saying one of the steps to have a standalone BrainzPlayer is to remove the existing links with the rest of the ListenBrainz codebase.
We currently do 3 things with the LB API in BrainzPlayer: refresh spotify token, refresh youtube token and submit listens to LB
2021-10-06 27956, 2021
monkey
The first two methods will need to be implemented outside of BP and passed to it as props.
2021-10-06 27956, 2021
monkey
In our case they are implemented already in APIService but need to be passed in as new optional BP props. We will be able to create a BP instance with `<BrainzPlayer refreshSpotifyToken={APIService.refreshSpotifyToken} …`
2021-10-06 27956, 2021
monkey
and inside BP create `<SpotifyPlayer refreshSpotifyToken={this.props. refreshSpotifyToken}`
2021-10-06 27917, 2021
monkey
The third method to submit listens should be implemented directly in BP I think, which means copying and adapting this bit of code :
ruaok: i see you got the answer already. to add to it, if the tags above the trace contain a url, it'll usually tell the container, test.lb -> web-test, beta.lb -> web-beta so on.
the topmost component creates a context and sets the theme, the individual sub components can just access it like context.tabColor so on.
2021-10-06 27921, 2021
lucifer
monkey probably knows more and may have thoughts on how to do this.
2021-10-06 27906, 2021
monkey
Haven't tried myself, but I did see it used as an example fro globalcontext, probably a good way to do it
2021-10-06 27940, 2021
lucifer
afaik a page reload will be needed for theme change. because the entire website is not in react. there's entire pages in just plain jinja and the navbar and footer are also in jinja.
2021-10-06 27918, 2021
lucifer
at that point it might be just easier to send a different css file to the frontend depending on the theme.
