Fork me on GitHub
#off-topic
<
2019-07-30
>
ghadi01:07:05

Yup they were fast this year

theeternalpulse04:07:15

I don't know how many of you do js development and use npm, but I've noticed more packages advertising things like donation links, even "for hire" advertisements when installing dependencies. Some even fetch donation status from xhr calls during install-time. Something about it just seems...wrong. Is there a maven equivalent to this?

schmee07:07:25

oh, JVM summit 😮 and a new @ghadi talk, yay! 😄

borkdude09:07:04

gotta watch that talk

borkdude09:07:30

anyone wants to play with this data https://github.com/borkdude/clj-kondo/tree/dump-analysis/analysis#data before I release it tomorrow?

lilactown15:07:06

@borkdude does the ns graph work in CLJS?

borkdude15:07:49

@lilactown the implementation or the analyzed source code?

lilactown15:07:16

oh, sorry. I want to analyze CLJS code and get an ns graph

lilactown15:07:32

the impl I assume is in CLJ

borkdude15:07:41

if you can find a library which creates such a graph, you should be able to get it working. I already ported some things to planck here: https://github.com/borkdude/clj-kondo/tree/dump-analysis/analysis/script

borkdude15:07:57

I used loom which is only for CLJ I think

lilactown15:07:50

ah that wasn’t what I expected. so this can only analyze and produce an ns graph of a CLJ project?

borkdude15:07:19

no, you can analyze CLJ, CLJS and CLJC

lilactown15:07:40

😄 yay

borkdude15:07:09

but clj-kondo itself only runs on the JVM and natively

lilactown15:07:56

that’s fine. I’ve been having this issue where with a CLJS project that has multiple builds, it’s hard to know what ns will actually show up in a bundle after awhile

lilactown15:07:23

so I was planning on building an ns graph viewer into shadow-cljs (since it already has done that analysis), but this is even better!

lilactown15:07:37

it would help in the link above if you posted an example of the actual EDN format

lilactown15:07:04

(if that’s what you’re looking for feedback on)

borkdude15:07:38

Is the pprinted data at the top not clear?

lilactown15:07:16

is that the actual data? I didn’t think that was valid EDN

lilactown15:07:42

I’m dumb

lilactown15:07:15

I would prefer to see literal EDN that I could paste into a REPL as well but I get the gist now

borkdude16:07:58

ah ok. yeah, the reason I did it like this was to make it a bit more succinct

borkdude16:07:56

feedback in terms of what data is additionally needed to make things would also be welcome

borkdude22:07:34

^ data is changed to EDN now

lilactown22:07:55

are there certain fields that are guaranteed to exist vs. are optional?

lilactown22:07:28

might be nice to notate that in some way

borkdude22:07:33

some fields like :doc are only there if there is a docstring 🙂

borkdude22:07:55

I'll think about this

Alex Miller (Clojure team)22:07:10

no fields are guaranteed to exist

Alex Miller (Clojure team)22:07:27

the compiler has an option to elide all metadata from vars

borkdude22:07:52

this data is collected by static analysis, so not to worry about that