Fork me on GitHub
#babashka-sci-dev
<
2022-03-22
>
borkdude13:03:15

Unfortunately I had to revert the declarative pods commits for now because this causes image bloat

cap10morgan14:03:30

ok, are you looking into that then? I'd need some guidance on where to even start there if you want me to.

borkdude14:03:39

yes, of course. just need to find the time, probably next weekend

borkdude14:03:01

if you want to take a stab at this yourself, just start from the state before the commit and slowly add things, then recompile and see if compilation gets slower and the binary gets a lot bigger

borkdude13:03:55

We'll have to inspect closely where this comes from. Usually this happens because of a runtime resolve or so.

cap10morgan18:03:26

yep, I think I found it and that was it. and it's just a bug. there was a reference to a local symbol resolve but when I extracted the referencing code out into its own fn, I didn't notice that and it just silently started referencing clojure.core/resolve

cap10morgan18:03:32

hopefully will be all fixed shortly

cap10morgan18:03:00

fixed pods PR is up

cap10morgan20:03:41

and all of the tests in the main PR are finally green

borkdude20:03:50

Excellent. I'm going to take a look tomorrow. Pretty tired now after a presentation at London Clojurians :)

👍 1
borkdude13:03:56

I'm sorry about that, but I'm happy I catched that. I thought it was related to trying out the m1 / graalvm 22 but luckily it wasn't that

borkdude13:03:37

Since I now have an m1 build of babashka (try the m1 branch + https://github.com/babashka/babashka/wiki/Compiling-for-M1) I see that it cannot load intel pods. Perhaps we should fall back on those, if they are able to run in macos