#musicbrainz

Sciencentistguy: matroska support would be great, but currently nobody is working on it. See also https://github.com/quodlibet/mutagen/issues/354

There is also a Picard ticket

atj: I guess it is :)

But before I can publish the code I still need to clean up git history and implement one important feature that is missing.

will be interested to take a look

Granted ;)

is Spotify oAuth the missing feature?

I think this screenshot shows what is missing and what it is superior to a-tisket pretty well.

Yes Spotify is missing, but the feature I meant is caching

what's the stack for the app?

What's the use for caching? Shouldn't people import each thing just once?

Caching is useful for later reference in edit notes and for the post-submission screen (cover art, ISRCs, artist URL rels)

Ah, ok

The app is built with Deno, so 100% TypeScript

The web frontend is made with https://fresh.deno.dev which pre-renders TSX (preact components) and re-hydrates the dynamic parts on the client side.

nice

where will you host it?

I dunno, so far I've only been testing locally. The fresh app could theoretically run on https://deno.com/deploy but that limits access to the file system and makes caching of all the JSON responses impossible currently.

Some new features which can be seen in the above screenshot:

1) Automatic detection of the script of the titles

2) Guessing of the language using a tiny neural network

3) Handling of iTunes API responses which contain video tracks and/or alternative releases

4) Detection whether iTunes API results are streamable and/or purchasable

5) List of regions which will be ignored when determining worldwide availability (e.g. dependent islands)

6) For each release property a list of preferred providers can be specified (not yet exposed to the UI, higher artwork quality and duration precision are automatically preferred)

so, nothing much then ;)

Oh and as you can see I still have to work on the design.

One week ago this was still bare HTML and I've only added a bit of hand-written CSS so far to make tables, messages and input elements less boring.

kellnerd: i was looking at this the other day, seems quite nice https://shoelace.style/

atj: Hmm, while I certainly like the look and some of their ideas, I'm not a fan of UI frameworks which need JS in order to work at all.

I mean for dynamic components that's ok, but static content should not degrade (or even not be rendered at all) without JS.

kellnerd: wooo

Hey aerozol, I certainly hoped to attract your attention for feedback and help on the design and UX :)

kellnerd: the other one i was looking at was Bulma, which I don't think requires client side JS - https://bulma.io/

yes, CSS only

I definitely need to make the layout responsive and find a better place for the cover art and I still haven't integrated alternative property values from other providers, so there will surely be bigger UI changes.

CSS only = ❤️

in my limited experience using flexbox will make a lot of layout issues much easier to fix

yep, flexbox is already in use for the icon overlays of the input elements

Bulma uses flexbox and seems to have some nice utility classes

But everything below the lookup section isn't using any grid/box layout so far as I didn't want to define breakpoints myself.

I will check Bulma

I added a release, but just noticed that I had two tracks swapped. I can't swap them until I remove the discid I added. Could I please get a few yes votes on https://musicbrainz.org/edit/100453375 ?