hard to say details when I am not sure what I am doing. I wanted to have a backend node service, but I am unable to setup the lifecycle hooks to that too.

what do lifecycle hooks have to do with a backend node service?

this is what I mean, that I don't know what I am doing:)

just describe what you are trying to do, not how you are attempting to do it

:target :node-script with hooks, so that when I build it, I can start the built script

get rid of any shadow-cljs terminology

using hooks here just confuses me

sorry, this is more difficult than I expected:D i want to write a backend server and have the normal devtools available while writing it

you mean when you start the build you want to start the node script?

after the build

this is not supported, and also not desirable

the intent is that you have shadow-cljs built the :node-script

that's all right, at least I asked after trying it for a day anyway

you then, in a separate terminal, run node the/output.js

this gives you a clear place where the output goes

yeah, I don't know what to write the script in that starts pm2 or similar thing

pm2 then can monitor output.js and restart the node process

well, with hot-reload you shouldn't need to restart in the first place

if you actually need to restart you can ctrl+c the process and restart it

I mean you can, and you know the tools to use for it

i don't think i would have hot reload in any of these cases

you just won't find anything in shadow-cljs to do this for you

yeah, thanks, I just misunderstood the docs, thought it would work because the :target wasn't specified in the example and I wrongly assumed it's more general than what actually is. But I don't need shadow-cljs to do this auto-restart, just would be nice to not have to manually ctrl+c and restart 200 times a day

anyway, thanks again for the help, I will figure something out

I agree that restarting 200 times a day is not great. hot-reload/REPL give you the option to drastically reduce that. I usually start my clojure REPL once a week or so, same could be true for a node script in theory. All depends on your workflow.

I don't have a flow of work. It's sporadic hiccups. 😅

sweet, thanks!

or :devtools {:enabled false} to disable it entirely

super!

@U05224H0W I think I have an issue with vanilla CLJS advanced vs shadow-cljs advanced in the area of keywords, but I haven't gotten to the root of this. But I suspect it may be related to the keyword optimizations in shadow. Is there a way to disable that in advanced, to debug this issue?

In vanilla CLJS I don't have this problem, to be clear

what is the problem?

no, you cannot disable it

only thing you can set is :compiler-options {:shadow-keywords true} which optimizes it even more

but I never tried that with :esm, so no clue if that even works

let me try a repro

hmm, no that wasn't it. I thought that keywords optimized at build time weren't correctly matched against keywords created at runtime, but it works. the issue I'm having is with clojure.spec.alpha which dynamically processes some input at runtime, which works with vanilla but not with shadow... I'll debug further

ah, it was my bad vanilla CLJS has something like (set! *main-fn* ...) and then the arguments are the command line arguments and I relied on that, but when porting to target esm that no longer works (as expected, probably)

luckily it works now with manual process.argv processing 😅