spellew: the tests use a redis docker container and need to get the docker image. Tests should run properly once the image is downloaded and containers are launched. Could you paste the output the next time you try to run the tests?
"01:35:08 <zarcade_droid> It's just a bit sad that that link isn't shown on the sidebar" - note that the relationship type and whether a link is shown in the sidebar are generally not related ("official homepage" excepted). Sub-Fandom pages could be added to the sidebar while still using "fan page" or "biography" links.
MBS-9512: Changing recording name to empty string should not be allowed
yvanzo
yes
reosarevok
Ok :)
(was wondering since it said "mitigate" at first)
yvanzo
ya but bitmap found a definitive fix after all :)
Darkloke has quit
akhilesh
Mr_Monkey: did you tried endpoint implemented until. Only relationship of work lookup is left. I am reading about caching which is our next task. I understand the storing and getting data from cache but still confused with invalidation of cache? Have you any idea about cache invalidation for existing setup?
Mr_Monkey
akhilesh: Haven't tried the endpoints since yesterday, I was waiting for work to be done to review more thoroughly.
As for cache, I'm not the best person to ask. My knowledge of it is limited. zas has recently been testing setting up cache at the gateway level, so I think it would be better to keep cache implementation for a later date and move forward on the endpoints.
akhilesh
Sounds good to me for now!
Mr_Monkey
We might end up not needing to set up cache for any endpoint
You will need a cursory understanding of redis as the cache layer for credentials, but api/app.js is already set up for that as it is now.
Hi bukwurm ! You wanted to talk shop?
reosarevok
yvanzo: thanks for the hints re: the constants for select! :)
Should be good now
akhilesh
Mr_Monkey: Ok, but it is good to me if you will check implemented endpoint and see the output and give your valuable suggestions please. It will help me to speed up the project.
bukwurm
Mr_Monkey: Hey
Mr_Monkey
OK akhilesh. I'll do a full review tomorrow
bukwurm
Are you available right now?
Mr_Monkey
Hi bukwurm
Yup
bukwurm
Great
Mr_Monkey
So what can I help you with for the import project?
bukwurm
I was thinking of moving in the direction of final import process.
First of all, do you have anything specific direction for the project?
Mr_Monkey
Not at this point, no
bukwurm
Alright
Mr_Monkey
But the direction you were going previously seemed right to me
bukwurm
Ok
Mr_Monkey
I think there are some questions regarding the UI I would propose to make sketches for
bukwurm
Ok
First, I'd like to talk about the importer
The changes since then that have been made are:
1. Moving the project under a new repository - bookbrainz-utils
2. Adding a `cleaner` sub-project (apparantly in typescript) which doesn't appear to have much functionality as of now
So there are some fundamental changes I am thinking of making
Back then I had just started out with nodejs , so a number of issues had crept in
First change I'd like to have is a separate micro-service for consumer (one that pushes the data into db)
Mr_Monkey
How is it currently run?
bukwurm
Currently, it's coupled with a lot of things
Presently it's just a script, which can fail silently
It can be very easily set up as a separate service/worker
Nothing much needs to be changed
It can be even set up as a cron
Mr_Monkey
Yup
bukwurm
We can even run it at will though right now
But I imagine a situation where 3rd party individuals can use the API to import data from any place
The server pushes the request and data to queue, and the workers listening can feed that into the db
But it's not really really needed right now
Mr_Monkey
It's a potential future goal, yes
bukwurm
The next thing is - we should ideally have an error queue which we can monitor
In case an import fails, it could be possibly due to some bug. Instead of logging it - we can ideally push it to the queue.
Mr_Monkey
OK, sounds sane
Easy to monitor
bukwurm
If errors exceed while monitoring and in case we find there's an issue - we can just start some worker to refeed the message back to the original queue
after fixing it
Else we can simply flush the queue
Again, not a critical change but good to have
Next thing - and a major one
The other end of importer
Is the auto import we were running
By reading and parsing the dumps from OpenLibrary
Mr_Monkey
Right
bukwurm
When I was writing the code I had little knowledge of event loop and all
Cyna
bitmap: I still dont understand how do i get $form
bukwurm
So I wrote a fairly complicated code for multiprocessing and what now
Cyna
I've put the statement you told me, in the required place.
bukwurm
I don't think it's effective in a resource constraint project such as bookbrainz
Mr_Monkey
I was looking through it, yeah, it seemed pretty complex
bitmap
Cyna: push the code again and I'll check :)
bukwurm
It's clumsy, and as there were no threads then in js
Cyna
Sure
bukwurm
It's not async
I think we can get rid of the module
And replace it with a much cleaner worker threads now stable
Mr_Monkey
OK, that sounds like an important job, and a good idea