#metabrainz

Leftmost: canonical?

Canonical suggests to me that it's the official alias for the described entity.

OK, let's stick with primary

Spaced out that is_primary doesn't fit our boolean naming scheme. Feel free to change that.

LordSputnik pushed 2 commits to bookbrainz-sql: https://github.com/bookbrainz/bookbrainz-sql/co...

zas: When you say to have the accel attached to P, would <Open in &Player> suffice?

Leftmost: any idea why our editor IDs are so sparse?

The last 5 registered users were 131, 132, 133, 142, 143

Oh, or maybe not, ignore me :P

But we definitely have gaps

Not sure.

Oh, hey, I hit a bug in prod and found another testing case: I went to Publication, created a new edition from there, and it failed creating the data. The revision and entity were created but not the data.

LordSputnik, my idea for improving error handling back then was to have unified error displaying system, of course our architecture has changed a lot since then

(I suspect doing it from publisher will do the same.)

so I'm unsure of how relevant that is these days

Heh, very relevant.

Our error handling is bottom-notch right now.

Leftmost: eww, that's a WS error isn't it

Hopefully that'll get caught by the GCI tasks

Leftmost: are you looking into it? It's probably a site bug in incorporating the publication ID into the data sent to the handler

That's a WS error and a UI error. I suspect the publication BBID didn't get set properly in the interface and that the WS isn't doing insertion transactionally.

I can look into it, just haven't gotten to it yet.

OK

And yes, all of the revisions (CRUMB) should be transactional, I don't know why we haven't done that already :P

Leftmost: hmm, a WS error while making entity data should result in an Exception, which would prevent the transaction occurring

Effectively at the moment a transaction is begun at the start of each WS request, and committed when session.commit() is called

But that shouldn't ever be called if there's an exception

Can you look into the WS side and I'll look into the site side?

Hmm. Dropdown is not working for the Bookbrainz site. I have the site server , ws, and redis all running on the correct, default ports

opatel99, did you run gulp before starting the site code?

opatel99: yeah that's what I would suggest too

^

We have to make alias set ID on entity data NULLable, don't we? We have historical data that requires it.

I think it probably should be any way

Thanks. I thought gulp watch would also compile it, but I guess not...

Oh

opatel99: yeah, I tend to run "gulp; gulp watch"

That way, when gulp watch encounters an error, I just press "up" and hit enter to start things off again

Leftmost: We could make it so that every entity data has an alias set, even if the alias set has no aliases?

I did think of that, but I'm not sure how I feel about it. I'll set it as NULLable for now and think more on it.

This may be that I have not yet understood Bookbrainz yet, but why are books like Pride and Prejudice or Old Yeller entered? Does it still need to be done, or is there a reason?

not entered^

opatel99, just hasn't been done yet.

Can that be automated from another database?

We've had some discussion about pulling in data from the Library of Congress and/or Deutsche Nationalbibliothek.

We've also talked about using user scripts to prepopulate BookBrainz forms, and actually about making some GCI tasks for that. LordSputnik, did those ever get entered?

I would enjoy doing that. When a book title is entered, automatically search that via an official API like Lib of Con and populate known fields?

I think the idea was along the lines of going to the LoC site, finding the book you wanted, and inserting a button like "Add to BookBrainz". That's how I'd envisioned it, at least.

Oh. Greasemonkey script.

opatel99: yeah, exactly that :)

They will be BB tasks, but in a week or two

(probably)

Exactly which? Referential ambiguity!

Other GCI stuff ;)

leftmostcat pushed 4 commits to bookbrainz-sql: https://github.com/bookbrainz/bookbrainz-sql/co...

I would advise against trying to anticipate what tasks may or may not appear later on in the contest! ;)

Are all the Jade -> React tasks claimed?

Nope

There are two of those available, 1 beginner

So what if the bookbrainz collection times out on me? GET / - - ms - -

In the WS?

That is the sites error

No errors printed from the site code?

Failed to load resource: net::ERR_EMPTY_RESPONSE

opatel99: which URL is that?

localhost:9099

Leftmost: are you sure that edition bug creates a revision and entity?

Oh here we go

WS error: 504 GET /ws/message/inbox/ undefined Error: There was an error accessing the web service

opatel99: have you set up ./config/development.json ?

LordSputnik, https://bookbrainz.org/revision/830

LordSputnik: http://hastebin.com/xapasowemu.cpp

development.json

Leftmost: It doesn't give me any revision/entity locally...

Double the fun.

opatel99: Fixed now.

I was in copy mode :P

It seems to have the effect of pausing the running process on the server

Leftmost: I suggest we just fix the bit of broken code in the site for now, because I can't see for the life of my how a revision can possibly be made in this case :P

LordSputnik: Thank you

Even the SQLAlchemy docs say that in the event of a commit raising an exception, the data will be rolled back

opatel99: no problem :)

Leftmost: we could perhaps make edition_data.create return None if there's no publication set, and then abort the request handler if that happens?

Sounds reasonable.

LordSputnik pushed 1 commits to bookbrainz-ws: https://github.com/bookbrainz/bookbrainz-ws/com...

LordSputnik pushed 1 commits to bookbrainz-ws: https://github.com/bookbrainz/bookbrainz-ws/com...

leftmostcat pushed 1 commits to bookbrainz-sql: https://github.com/bookbrainz/bookbrainz-sql/co...

LordSputnik: I saw the github issue you ported to Jira today. Since the project is heavily based on React, would you consider it natural to perhaps change the select2 dependency to https://github.com/JedWatson/react-select

opatel99: we've looked into that already, but decided against it

I'd consider a proposal on it, but I've worked with it in the past and it's pretty inflexible.

We've got a good React.js wrapper for select2 already

Gotcha.

LordSputnik, were we still considering trying the other thing that I can't quite remember?

Leftmost: If we were I've also forgotten, but I'm happy with select2 now

Apart from 4.0.1, which is not good :P

There are a couple things that annoy me a little. Worth investigating them when we get there. :-P

Leftmost: keep a note :)

(although I kind of doubt we'll find a better alternative)

When did you try react-select?

LordSputnik pushed 1 commits to bookbrainz-sql: https://github.com/bookbrainz/bookbrainz-sql/co...

At my day job. We were working with keystone.js (the project react-select was written for) and had trouble even making it do what we wanted within keystone.

opatel99: about 6 months ago, I think

(well, I tried it out on BB locally)

Oh, you did? Huh. :)

Leftmost: yeah, just before I made the new select2 component :)

My excuse for not knowing that is that UI scares me. :-P

Have a good one.

Leftmost: before closing that jade task, if Annie finishes, check the files have copyright headers, thanks!

Will do.

Hi Freso

Hi reosarevok

LordSputnik: Working on sandbox again?

I suspect he's in bed, and it looks to be working.

What error are you getting?

Same as when he was in copy mode. Error: There was an error accessing the web service

Can you give the full error message? The HTTP code and URL are relevant.

On http://localhost:9099/publication/create

But the weird thing is that this time, there is no error in console...

You're sure you're logged in? It doesn't show you as having hit any login pages prior to submitting, and 401 is the HTTP unauthorized error.

Just logged out and back in. Working. Thanks

The navbar did however show that I was logged in

Had you refreshed the page?