package.el is not the best package manager is the world, unfortunately

other than pinning some packages to specific versions there’s little one can do

obviously if someone wants only stable package releases the simplest solution is to just disable melpa and use only melpa stable

@bozhidar: RE: Using melpa stable, does it work alright with prelude? I've thought about switching over to stable completely but I'm afraid it could mess other things up.

Right now, I have stable added to the list of package sources and just have cider and clojure-mode pinned.

nope, some packages prelude uses are only on MELPA

or haven’t had releases recently

Good to know. Thanks.

hi - a few of us are chatting that if you try to a big compile, or namespace refresh just after a CIDER REPL launches, then you sporadic compilation errors

I've had this for a while but blamed my own setup, but I think it's a CIDER issue

is it a known issue?

@jonpither: I think I’ve seen this. I have no idea if it’s a CIDER issue or not

You need to wait for the initial ns scan to finish before you do anything

which isn't great

definitely experience this as well

it would be nice to know that a) it’s still busy and then b) it’s done

via output in the repl buffer. i could totally live with that

file a ticket

with some repro steps and will figure out a solution

rock paper scissors, jonpither? -grin-

@jonpither @jonas I've encountered something similar - do either of you use refactor-nrepl? I think that warms some kind of analysis cache using tools.analyzer right after you connect

I do, and my colleagues do, so wouldn't surprise me

@bozhidar: @robert-stuttaford I can raise a GitHub, issue, but right now it would be vague (sometimes when I start a fresh REPL and try to compile, I get compilation errors). I think to repro you need a suitable large project

I'll raise it tomorrow

Try putting (setq cljr-eagerly-build-asts-on-startup nil) in your config and see if it still occurs

@cichli: will do

(had that in mine for a while now, but can't remember why...)

& I also haven't had the compilation issue in a while

Ah I remember - the reason I changed it was that it was causing weird issues when REPLing in to an already-running Clojure program with a system started - after connecting, lots of records would be instances of "old" versions of protocols, because something was re-evaluating everything. I tracked it down to the refactor-nrepl cache

I just created an issue by accident in CIDER

gah

should really go to bed