https://kotlinlang.org logo
#feed
Title
# feed
b

Big Chungus

06/01/2022, 10:14 AM
We've passed another thousand kotlin libs at 9007 kotlin libraries on MavenCentral after last night's kamp.petuska.dev scan! An increase of 977 since 2022-03-17. Apologies for the spam - over and out! K K
🎉 15
💯 4
K 8
👍🏻 1
👍🏾 1
a

altavir

06/01/2022, 10:41 AM
Good dynamics. Do you have a plot somewhere?
b

Big Chungus

06/01/2022, 10:50 AM
No, but if you search for "kamp" in #feed you should surface enough data points to plot it yourself
I used to post such numbers every 500libs and now padded that to every 1000libs
Kamp 2.0 will have a statistics page to let you plot the data by time, targets, etc... But I have no ETA when that will be ready
a

altavir

06/01/2022, 10:52 AM
I know. But it seems like dynamics looks too good, for not showing it.
b

Big Chungus

06/01/2022, 10:54 AM
Totally agree, but I've been spread too thin recently to be able to focus on that 😀
a

altavir

06/01/2022, 11:57 AM
Yeah, I feel your pain.
s

seb

06/01/2022, 6:26 PM
We have 136k KMP artefacts in the Package Search index 🙂 https://package-search.jetbrains.com/
FYI some stats are currently broken, we're aware of that
b

Big Chungus

06/01/2022, 6:31 PM
Key word here is artefacts. Kamp groups them all up under a single logical "library" unit. I.e. if a lib publishes artefacts for 10 targets, it would still count as a single mpp lib
However your 136k is still way too big to be explained by just the target artefacts. How do you define an MPP lib in that count?
Also is your artefact a single published file (pom, jar, etc...) Or a maven module?
c

Colin White

06/01/2022, 8:30 PM
Looks like the results for
io.coil-kt
are out of date? It says the latest release is
2.0.0-alpha09
b

Big Chungus

06/01/2022, 8:33 PM
Yes, it's a known issue with the scanner. Basically if there's a package at the group level, it picks it up and does not scan deeper. https://github.com/mpetuska/kamp/issues/61
👍 1
So a good number of kmp libs might still be hidden because of that.
s

seb

06/01/2022, 8:58 PM
@Big Chungus yes agreed that "artefacts" is a vague definition. I think, but need to check, that it's the number of KMP versions of all KMP libs we have in the index. Some numbers are a bit dodgy in our stats as I mentioned earlier; this could be likely if it's the number of artefacts of KMP versions of KMP libs, which is a fairly useless metric 😂
Unfortunately I'm in the US so comms with the person who worked on that, who's in Taiwan, will be difficult for a couple weeks
I have asked for further scrutiny tho
b

Big Chungus

06/01/2022, 9:05 PM
Thanks, I'd really like to know more about your metrics since your numbers are vastly different, even though both approaches seem to be similar 😄 I also noticed that for example kotlinx-coroutines-core and kotlinx-coroutines-core-js are both listed on the package search, implying that they're counted individually.
s

seb

06/01/2022, 9:07 PM
I have to admit I'm not overly familiar with the logic here :) I'll try to find out
b

Big Chungus

06/01/2022, 9:09 PM
Thanks. Looking forward to more details!
👍 1
s

seb

06/01/2022, 9:31 PM
Looking at the code, seems like it's indeed counting each version of each artifact of each group where it detects it's KMP (based on Gradle Metadata, right now)
x

xxfast

06/02/2022, 12:17 AM
itsover1000.jpeg
😄 3
3 Views