shivam-kapila: i don't have any role in the UI. All of the UI is SomalRudra 's work :)
2020-07-24 20654, 2020
iliekcomputers
shivam-kapila: awesome, thanks!
2020-07-24 20656, 2020
_lucifer
and its great :)
2020-07-24 20609, 2020
SomalRudra
_lucifer: hey! thanks bud
2020-07-24 20612, 2020
shivam-kapila
I literally like that
2020-07-24 20626, 2020
shivam-kapila
I was also thinking of something card based like he did. But certainly the greyscale idea goes awesome
2020-07-24 20606, 2020
_lucifer
Android Studio lets you code and use the designer simultaneously so if the same person is going to create the UI makes more sense to code
2020-07-24 20622, 2020
shivam-kapila
I know
2020-07-24 20637, 2020
shivam-kapila
I have put my hands in android too
2020-07-24 20640, 2020
shivam-kapila
Kotlin
2020-07-24 20643, 2020
_lucifer
you used AS, nice!
2020-07-24 20657, 2020
_lucifer
I am in the Java camp though!
2020-07-24 20611, 2020
iliekcomputers
why not react native ;) ;)
2020-07-24 20613, 2020
iliekcomputers runs
2020-07-24 20618, 2020
shivam-kapila
I have changed a lot of fieds
2020-07-24 20628, 2020
shivam-kapila
iliekcomputers: you know why
2020-07-24 20629, 2020
_lucifer
JS!!
2020-07-24 20637, 2020
_lucifer
I hate and fear it at the same time
2020-07-24 20638, 2020
shivam-kapila
No I love JS
2020-07-24 20644, 2020
shivam-kapila
and Python
2020-07-24 20653, 2020
iliekcomputers
typescript is really nice
2020-07-24 20600, 2020
shivam-kapila
That
2020-07-24 20621, 2020
iliekcomputers
you can't really work with plain javascript in a reasonably sized codebase
2020-07-24 20651, 2020
shivam-kapila
But I like the thing how fast is JS evolving
2020-07-24 20659, 2020
shivam-kapila
_lucifer: I have been in Java club in school... #Carmelian XD
2020-07-24 20629, 2020
_lucifer
yeah, I almost forgot.
2020-07-24 20659, 2020
_lucifer
i like java especially since the new release cycle. they are catching up with kotlin fast
2020-07-24 20607, 2020
rdswift
ruaok: I just tried the recommendations again. The list looks good. In fact, I think every one of the recommendations is somewhere in my listening history. ;-)
2020-07-24 20619, 2020
shivam-kapila
_lucifer: havent used in years
2020-07-24 20613, 2020
_lucifer
ah, Ok.
2020-07-24 20658, 2020
_lucifer
I recently found a person who coded Computer Vision apis in Java skipping OpenCV altogether
2020-07-24 20612, 2020
ruaok
rdswift: I *think* that is good. I think the current window is a month -- if the track had been played in the last month it should not show up in the list.
2020-07-24 20654, 2020
ruaok
but it should also have new tracks on it you've not listened to before.
2020-07-24 20614, 2020
ruaok
on some level there ought to be a balance between new and old stuff
2020-07-24 20641, 2020
_lucifer
ruaok: iirc, a few months ago (pre gsoc) there were some plans of rewriting something lb related in rust or go. do you remember what was it?
2020-07-24 20651, 2020
iliekcomputers
_lucifer: yes
2020-07-24 20656, 2020
rdswift
Some of them have definitely been played in the last month. I wonder if they were linked to a different "recording" of the same track?
2020-07-24 20657, 2020
ruaok
we have a lot of tuning and tweaking to do. we're just now understanding what we've created.
2020-07-24 20611, 2020
_lucifer
iliekcomputers: can you remind me?
2020-07-24 20611, 2020
iliekcomputers
our db writer is in python
2020-07-24 20632, 2020
ruaok
rdswift: that is possible.
2020-07-24 20654, 2020
iliekcomputers
the script that takes listens from rabbitmq and writes to the database
2020-07-24 20610, 2020
rdswift
Anyway, they are all tunes I like. đź‘Ť
2020-07-24 20612, 2020
ruaok
I could see a need for "when did user X listen to track Y last"
2020-07-24 20626, 2020
ruaok
:D
2020-07-24 20628, 2020
_lucifer
iliekcomputers: and it has performance issue?
2020-07-24 20651, 2020
shivam-kapila
_lucifer: 3rd now. Going into 4th. Wow CV on JAVA.
2020-07-24 20651, 2020
iliekcomputers
we looked into it and influx was the bottleneck
2020-07-24 20604, 2020
rdswift
Is there some significance to them all showing "51 years ago"?
2020-07-24 20609, 2020
ruaok
I still think python will be a bottleneck before too long
2020-07-24 20621, 2020
_lucifer
oh ok. so that is now relevant or not?
2020-07-24 20639, 2020
shivam-kapila is interested in the GO move
2020-07-24 20640, 2020
ruaok
rdswift: shoddy UI for now. (timestamp 0, Jan 1 1970)
2020-07-24 20600, 2020
ruaok
_lucifer: probably still needed.
2020-07-24 20606, 2020
rdswift
NP. A work in progress. I understand.
2020-07-24 20617, 2020
ruaok
very much so.
2020-07-24 20620, 2020
shivam-kapila
_lucifer: we may get that done post gsoc.
2020-07-24 20620, 2020
ruaok
but a fun one.
2020-07-24 20624, 2020
yvanzo
reosarevok: Good question, need to be tested probably. 'There was a problem cancelling edit {edit_number_link}.' would be nice, did not test though.
2020-07-24 20631, 2020
_lucifer
okay, i was asking since i recently toying with rust trying to learn it
2020-07-24 20638, 2020
rdswift
And awesome so far!
2020-07-24 20659, 2020
shivam-kapila
iliekcomputers: 2 minutes more please
2020-07-24 20617, 2020
ruaok
_lucifer: I think out team has greater lean towards go -- we want to do it for performance. not sure if rust is the perfect choice for that.
2020-07-24 20650, 2020
yvanzo
reosarevok: This is a known limitation of edit note syntax though, I would just pay no attention to it until we change the syntax.
2020-07-24 20656, 2020
_lucifer
yeah ruaok, rust is the choice if you have memory safety issues
2020-07-24 20608, 2020
_lucifer
go is better if you have performance issue
2020-07-24 20613, 2020
ruaok nods
2020-07-24 20658, 2020
shivam-kapila
I had a senior who reduced the execution speed of a server by 90% while moving from Python to Go
2020-07-24 20612, 2020
shivam-kapila
Plus Postgres
2020-07-24 20628, 2020
_lucifer
go routines are really great that way
2020-07-24 20646, 2020
iliekcomputers
i'm sure python will be a bottleneck. but i'm hoping we can do a few smaller tweaks first.
2020-07-24 20651, 2020
iliekcomputers
like json to protobuf
2020-07-24 20608, 2020
_lucifer
protobuf is nice
2020-07-24 20615, 2020
shivam-kapila
how many weeks work will that be
2020-07-24 20659, 2020
_lucifer
i'd interested in helping with the protobuf and go stuff since learning server programming is on my bucket lsit
2020-07-24 20617, 2020
shivam-kapila
Oh server programming is just awesome
2020-07-24 20651, 2020
ruaok
iliekcomputers: we should start by analyzing the code to see how to avoid parsing and serializing JSON.
2020-07-24 20625, 2020
ruaok
we do it far too often and that is very expensive -- protobuf solves part of that, but factoring out necessary ones would be even better.
2020-07-24 20634, 2020
iliekcomputers
ruaok: agreed
2020-07-24 20600, 2020
iliekcomputers
overall, i'd say doing both of those things with testing and roll out would be a month's work of solid work
2020-07-24 20608, 2020
MajorLurker joined the channel
2020-07-24 20625, 2020
alastairp
> 9:37 PM <shivam-kapila> I had a senior who reduced the execution speed of a server by 90% while moving from Python to Go
2020-07-24 20635, 2020
shivam-kapila
I would be interested in it after Gsoc
2020-07-24 20640, 2020
ruaok
yeah, for someone who could really focus on that.
2020-07-24 20641, 2020
alastairp
careful here - this isn't always python->go, it could just be using the database better
2020-07-24 20644, 2020
ruaok
which is not I.
2020-07-24 20652, 2020
iliekcomputers
ruaok: yep.
2020-07-24 20607, 2020
shivam-kapila
alastairp: He specifically mentioned about python being the bottleneck
2020-07-24 20611, 2020
alastairp
a colleague of mine saved a client from paying $9000/month on a database server to a $50/month server just by adding some database indexes
2020-07-24 20617, 2020
alastairp
I don't believe that for a moment
2020-07-24 20617, 2020
iliekcomputers
i'll write up a priorities doc after gsoc, and we'll see how much time volunteers can give to things.
2020-07-24 20624, 2020
ruaok
alastairp: yea, I fixed that part already. there is a code fragment around that batches listens and commits every 3 seconds.
2020-07-24 20639, 2020
ruaok
that tripled pipeline insertion speed.
2020-07-24 20658, 2020
ruaok
but ingestion has python eating tons of CPU... mostly json shit.
2020-07-24 20622, 2020
ruaok
no its not python itself. no.
2020-07-24 20602, 2020
ruaok
unecessary parsing and serializing and improved insertion mechanics are the biggest issues.
2020-07-24 20626, 2020
alastairp
yeah, right. if you're doing that lots of times then it's definitely going to be a problem
2020-07-24 20601, 2020
ruaok
but I got timescale to insert 5000 listens every 3 seconds and it wasn't sweating.
2020-07-24 20612, 2020
ruaok
influx was sweating just watching timescale
2020-07-24 20625, 2020
MajorLurker has quit
2020-07-24 20625, 2020
alastairp
so instead of hitting the db every time something comes in, we batch them up?
2020-07-24 20628, 2020
alastairp
that sounds neat
2020-07-24 20632, 2020
shivam-kapila
> i'll write up a priorities doc after gsoc, and we'll see how much time volunteers can give to things.
2020-07-24 20632, 2020
shivam-kapila
I would be interested to continue. Plus I expect online classes so will probably have more time
2020-07-24 20651, 2020
ruaok
not in production. that code is floating around one of my machines in BCN.
2020-07-24 20655, 2020
ruaok
not urgent now.
2020-07-24 20655, 2020
shivam-kapila
iliekcomputers: redone and pushed. rebase was a PITA
2020-07-24 20627, 2020
_lucifer
shivam-kapila: they always are
2020-07-24 20628, 2020
ruaok
I've had bad luck with rebasing. it feels like groundhog day.
2020-07-24 20647, 2020
shivam-kapila
Oh I thought its only me
2020-07-24 20621, 2020
ruaok
which is why I made timescale-rebase and timescale-rebased-again and timescale-final.
2020-07-24 20647, 2020
shivam-kapila
I would say perfect choice
2020-07-24 20658, 2020
ruaok
for each branch I merged the last timescale branch into a clean branch off master, resolved all the conflicts in one go and was done.
2020-07-24 20628, 2020
shivam-kapila
Timescale part has been really good for me. I had been a NoSQL guy, but now I love PG
2020-07-24 20659, 2020
ruaok
the switch from mysql to PG was one of the best decisions ever.
2020-07-24 20609, 2020
ruaok
even though people have given me a lot of shit for it.
2020-07-24 20630, 2020
ruaok
but mysql people love failures under high load and other stupid shit.
2020-07-24 20642, 2020
iliekcomputers
shivam-kapila: the travis build didn't trigger for some reason
