#bookbrainz-devel

So I know that testing -site is still broken, and that's because we lack a fake web service to hit. I'm looking at https://github.com/typicode/json-server as a way to fix that.

Uhm.

Tests in -site shouldn't rely on a live -ws, it should be testable on its own.

You should mock some WS responses instead.

I don't know any mocking libraries for JavaScript though, so I can't guide you in the right direction, but relying on a live ws is not best practice.

Leftmost: ^

Freso, that's what I'm doing.

json-server and its ilk are mock web services.

Alright then.

Looks like json-server doesn't have the flexibility we need, but the idea is to create a simple, static web service which is run by the testing suite as needed.

[bookbrainz-site] leftmostcat pushed 1 new commit to master: http://git.io/pSde

bookbrainz-site/master 97c6619 Seán de Búrca: Don't silently fail and spin if we can't get updates

bookbrainz-site/master 4df3da2 Seán de Búrca: gulp before running mocha

bookbrainz-site/master e175679 Seán de Búrca: Ensure public/js/ exists in git

Hey guys

hey

Hey, how goes?

Good, I've fixed alembic, so I can get a proper set of schema migrations saved

Sweet.

Are we all here?

Or, dammit, did I fail to account for DST?

DST has started?

Leftmost: yes you did, has it started for you?

It has.

I forgot about that when I put the meeting in my calendar, then forgot I had forgotten about it.

Ahh, will an hour's time be too late for you?

Nope.

I have nothing scheduled for this evening except DS9, beer and hacking.

Ah, I need to start watching TNG again :P

Has Spacebase DS9 gotten any updates?

Leo_Verto, that's DF-9, not DS9. :)

Leo_Verto: It's called that in Germany? :o

Ohhh, a game

*general confusion*

so what is DS9?

(I think, anyway)

Yes.

ah

Unless there's another DS9

:P

Apparently the best Star Trek series, though the only one I've never seen an episode of :(

I personally like TNG more, but that's partly a nostalgia thing, since the first two seasons are awful.

Yeah, I'm stuck on episode 4 or something

original series I've seen about 20, but I still haven't got into it, and don't think I will

The episodes are all a bit

... old

I feel a need to get through TOS, but it's kind of a struggle.

I played around with json-server last night, but I'm afraid we'll have to write our own fake web service because it assumes too much about the structure of output.

So something like an express server serving static content?

ohhh

Actually, I think we get a mock server with apiary.io, which I've started adding us to

LordSputnik, testing should be able to be independent of any of that.

Leftmost: http://docs.bookbrainz.apiary.io/

eg. http://private-0c443-bookbrainz.apiary-mock.com...

As Freso said last night, we shouldn't depend on another service being up to run tests.

We should have an internal service which we can feed test data to and get a static response. It shouldn't be difficult to write, just annoying.

And time consuming, when we get the apiary.io one for free

(if we document properly, that is)

Not that time-consuming, and as I said, we shouldn't rely on an outside service for testing.

The apiary.io one can be useful for making sure something you've just written is sane, but running `npm test' shouldn't touch any service we're not running ourselves.

Ah, I didn't know you meant local testing

Oh, yeah. Sorry if I was unclear about that.

For just testing code you're writing, I don't care what any of us does as long as it works. :) I just meant for automated tests.

On travis-ci?

Yeah, that sort of thing.

eurghh

Committed a fix, then forgot to push it before regenerating the schema migration scripts

now I've reset and lost it :P

Ack.

bookbrainz-schema/master 46a7b92 Ben Ockmore: Fixed alembic env.py script to deal with bookbrainz and musicbrainz schemas.

Y'all should fire that slacker.

Tsk.

I don't even know who has firing power, so it's not easy.

I think we'd have to hire him ifrst

Shh, Leo.

Don't you come here and be talking all no sense like to me.

there's a fallout mod for EU4

EU4?

Europa Universalis

Grand strategy

Right.

got it during a sale two weeks ago, downloading right now

Hurray!

Leo_Verto: here? :)

yup

We ready to get this row on the showed?

mhmm

So the first thing I had wanted to discuss was some set of criteria for a "go/no-go" when it comes to saying we'll start collecting data for real.

Basically, what we want to have place and what we don't need to have in place.

Ok

Well, we need to have registration and login, for a start :P

Things I think we should have in place: somewhat stable code structure and deps, though this is a vague requirement. Things I don't think we need: guaranteed-stable schema, guaranteed-stable WS API, an absolute guarantee that we won't be wiping data.

I still think you should aim for an April 1st release.

FWIW, ruaok agrees.

Freso, I agree.

Leftmost: \o/

LordSputnik, so that folds into the HTTPS thing, plus some testing of the current auth code with that.

I think we should guarantee preservation of data

No-one will want to contribute if their effort will be wasted

And I think we'll be in a position to do that on April 1st

Alright. If we want to do that, it puts somewhat a somewhat heavier testing burden on us. I'd say we should freeze before release in that case.

do we have issues for everything we need to get done by then?

Leftmost: we could start dumping the database, so we have backups if anything is broken

I'd rather not have a freeze or limitation on development at this point in the project

LordSputnik, I think that's a good idea, but we also want to make sure that we don't have any serious schema issues.

Good testing is definitely a +1 though

I don't think we do, but it's a good idea to shake it out and make sure.

Leo_Verto: I think so, yes

I don't think it'd need to be a long freeze. Even just a couple days, and not a hard freeze either.

I'm just thinking a soft freeze of a few days where we can fix bugs and hammer the crap out of it.

I'm probably going to hold off less integration until the server-side stuff is sorted out

The thing is, *when* do we freeze when we're continuously developing?

I'd understand if we had a formal release, and no other changes to bb.org until then, but we don't

LordSputnik, I'm not sure I follow the question.

Hmm, freeze the week before the grand opening?

that'd mean only 1 week left for features though

Leftmost: do you mean just for 1st April, or for future releases too?

I'm just thinking that we take the 30th and 31st, maybe the 29th, to do no major structural changes and just test and fix bugs to make sure our data guarantee can hold.

ah

Ok, then that seems like a good idea to me

Maybe someone call out for user testing in -devel (maillist and IRC both).

*?

I thought you meant somehow regularly freeze development like MB does with beta/hot/cold weeks

Freso: good plan

yeah, we also need to populate the mailing lists

Oh. No, I don't think that's necessary yet. Just something so that we can be sure there's nothing seriously wrong before we institute a preservation guarantee.

Ok

So, back to the milestone thing

We already have relationship support

And support for adding/removing entities

So long as the entity properties we have now won't go away, I don't think any schema issues will stop us from starting collecting data

To me, the most important entities to have in place would be Creator, Publication, Edition, and Publisher. That would cover most of the "basic" data you could give about any given book.

Work?

We could make do with less, though, and I don't see anything there being a blocker.

Work is definitely important, but I'd place it second tier.

Well, the only things stopping us having Edition and Publisher are forms and display on the site

Okay.

They're present in the ws (although there may need to be some small tweaks)

Ok, so, freeze at 00:01, 2015-03-30?

LordSputnik, do you have some time after to chat about the schema? Nothing important, I just want to continue our series on Making Sure Leftmost Knows What the Hell He's Doing.

UTC

Sounds good to me.