but I'm a bit confused as to what actually failed for me, because we run init_db in the tests. but I saw an error running init_db locally on my machine
Hi Indian peeps, we have moved our meeting from Saturday to tomorrow at CP in Delhi
ansh: riksucks , it that interests you
monkey
lucifer: I added a commit to your multi-stage-build-branch, supporting relative paths in webpack config so I can use it outside of docker. Small change so I don't expect any difference from what the PR changes.
Everything seems to be working great on my side for the whole PR. Wanna give it a spin?
alastairp
monkey: that's great!
kgz joined the channel
lucifer
monkey: changes lgtm.
monkey
Added a comment about paths as well, I think my brain isn't working well enough for this :p
lucifer
i wondered the same and had tried copying at same path but then realized it would copy over node modules as well. easiest way was to keep copying to a separate place.
mayhem, alastairp, reosarevok: i ran the script for daily jams manually for users who hadn't got the daily jam today to test the fixes. i checked manually and see dails jams in your and jukevox's accounts
hehe, still need to do more timing testing. it ran fast enough on wolf and i manually checked the sql queries and this produces less queries.
aerozol
Glad people are digging into the signup UX. I've hit my hours for this week, but itching to get those LB mockups started
lucifer
however, best to test on an isolated vm so that the numbers are comparable.
aerozol
akshaaatt: I can't see the 3d cover :(
akshaaatt
Yeah aerozol, I broke my code while converting the library to typescript
I was hoping no one would notice…😂
aerozol
Haha that's all good I was just having a sneaky peek at your message to monkey anyway
Re. signup woes I think when we have centralized oauth we will have to have a way of redirecting users back to where they can't from anyway. Bit cheeky to just assume everyone wants to go back to mb?
I liked @elomatreb[m]s suggestion of having them input the email but not needing to confirm it to start using LB. I don't know if we would ever need them to confirm it? Maybe if we start having issues? I wonder what other modern sites do
rdswift
alastairp, reosarevok, et al: I know I'm a bit late to the user account structure discussion (I'm blaming that on my timezone), but I've put together a brief discussion of the way I've set it up for the sports pools sites that I manage.
Each of the pools was originally on a separate site, with a separate database. Last year, I combined the databases and sites so that someone participating in more than one pool only had to have a single account for logging in and managing their information for all of the pools.
I'm not for a moment suggesting that you need to use this structure, but perhaps it might prompt some ideas when the time comes to combine the various *Brainz accounts.
once we confirm the regression is fixed and we should probably move to py3 first and then look further into improving rest of stuff. newer sqlalchemy versions have `raisedload` feature which will give an exception with stack trace when we access a lazily loaded property. which is probably what we want.
monkey
The interactive part is what sells it. It'll be fun to have on LB.
A friend suggested making it into a sort of banner at the top of a user's profile (sort of like they have in facebook etc.), but unfortunately the charting library doesn't take advantage of the extra horizontal space, so not great for that. Perhaps we'll find a good home for it
alastairp
lucifer: yes, right. Lots of small steps to make, as with all projects
monkey glares at the small steps
lucifer
indeed
alastairp
but I agree, get it to same level as it currently is, and py3 would be great
monkey
"but I want my cake NOW!"
alastairp
I agree that individual custom queries per entity may be better
lucifer
there's also the rmq library issues, moving to kombu will probably stablise sir.
alastairp
neat. I've tried to stay away from that side of things
