@bhauman have you ever encountered this error while reloading files by any chance?

ioc_helpers.cljs:42 Uncaught Error: No protocol method ISwap.-swap! defined for type cljs.core/Atom: [object Object]
    at Object.cljs$core$missing_protocol [as missing_protocol] (core.cljs:272)
    at Function.cljs.core._swap_BANG_.cljs$core$IFn$_invoke$arity$3 (core.cljs:756)
    at Function.cljs.core.swap_BANG_.cljs$core$IFn$_invoke$arity$3 (core.cljs:4303)
    at cljs$core$swap_BANG_ (core.cljs:4291)
    at file_reloading.cljs:303
    at file_reloading.cljs:289
    at file_reloading.cljs:289
    at figwheel$client$file_reloading$state_machine__12780__auto____1 (file_reloading.cljs:289)
    at figwheel$client$file_reloading$state_machine__12780__auto__ (file_reloading.cljs:289)
    at cljs$core$async$impl$ioc_helpers$run_state_machine (ioc_helpers.cljs:35)

Looks like an undefined atom

@richiardiandrea yeah some how your state of your runtime got way out of wack. That error should never happen.

Are you live reloading figwheel itself?

@bhauman well yeah 😄

I guess this atom here is not defined https://github.com/bhauman/lein-figwheel/blob/b187d9a4535e96d49f24c01579c30d11715b0fe8/support/src/figwheel/client/file_reloading.cljs#L275

on-load-callbacks

no I think the atom is defined

ah

yep I can deref it in the console

its a problem where the prtocols are blown away

the swap protocol got blown away off of the Atom

woah that's weird

you can't reload cljs.core

oh

ok that might be the problem then

because I am reloading stuff after an exception and it looks like I am reloading everything

figwheel prevents you from reloading cljs.core

cljs.core included

good, I am happy I have asked!

https://github.com/bhauman/lein-figwheel/blob/911b54e38711f689f8b54dea497bc588a902a2be/sidecar/src/figwheel_sidecar/build_middleware/notifications.clj#L149

no scratch that

anyways thanks a lot! I think I am very close to a first stable alpha

well so one of the things figwheel does is it ensures that you only reload files that are in the source directories and that actually change

yeah I was wondering that, because I am running lein-figwheel in parallel on the same project to see what happens

you must be sending a signal to figwheel that cljs.core changed

and that can't be true

because cljs.core doesn't change

yeah, no I am going to filter that for sure

but filtering isn't going to help I don't think

sorry, I mean filtering the change set

yes I understand but the change set should only be the file that changed from edit and its dependents

it should never be cljs.core etc

and the first compile needs to be ignored

unfortunately for what I can see, in boot when an exception happens, no file is written

so the next time I compile it rewrites all of them

which means everything changed

well thats a PITA

maybe I am missing something (will ask in #boot)

for now I can just filter the messages I send to the client

so what you can to is md5 the files

and see what happens

boot does it by default

but keep track

but if the previous fileset is empty

and only send the changed ones

save the state

in an atom

yep that is also already done in boot-reload, probably the case of an ex was handled differently

hmmmm weird

there is no (reset! prev fileset) when an exception happens at the moment

I have another random question, it reloads fine now, but the on-reload function does not get updated properly (aka, if I have a print in it, and I updated it, I never see the new string printed)...can this be due to pragma stuff and figwheel-load?