Does boot-reload :ids take #{"main"} <-- like this for e.g. main.cljs.edn ?

that doesn't work as expected here

ah i think maybe the current snapshot is != master

So boot-cljs master, when it's building multiple .cljs.edn, the resulting .js files for each build all refer to main.out/path/to/assets

rather than their respective <foo>.out/path/etc

...that's because i'm specifying that in one (cljs ...) task that builds everything...

but how could i do anything different there?

ok so in my case, i can get by without using :compiler-options {:asset-path ... }, and everything just works (yay)

Is there an example repo anywhere around of, say, boot reload, cljs, less?

and cljs.test

@tel: maybe check out https://github.com/martinklepsch/tenzing

thanks!

another one to look at would be https://github.com/Deraen/saapas

no prob!

@tel: i'm updating an older example now as well

@tel: https://github.com/pandeiro/jamal has less and cljs.test

pandeiro: sweet!

@pandeiro: sounds awesome indeed

ohhhhhhhhhh

I see how people are doing cljs tests

that’s clever

just having their own .edn target

@pandeiro: very nice work

!m pandeiro

@tel: that's one way, the easiest imo... the other route is to get fancy and define a task that generates that .cljs.edn file

haha

I guess I should say “I see how @pandeiro is doing cljs testing"

@micha: thanks, i finally figured out using with-pre-wrap and with-post-wrap to do what i wanted -- i may be able to publish that work soon

tbh

but the approach in jamal is dead simple and works pretty well if your development is browser-based

it’s perfect

thanks 😄

Anyone using cljs-repl and vim fireplace?

I’m running into errors getting set up

even (start-repl) is failing with a Piggieback error

rather a weasel one

java.lang.IllegalArgumentException: No value supplied for key: weasel.repl.websocket.WebsocketEnv@3d295938

are .cljs.cache files a clojurescript thing or boot-cljs thing?

Cljs thing

Hitting the stale tests thing again, even with .cljs.edn files specifying :init-fns

Why would (sift :include #{#"(^index\.html$|^.+\.js$|^styles.css$)"}) NOT eliminate the various *.out/ dirs created in cljs compilation? (And are those dirs even supposed to be created with :advanced?)

if you want to remove things you can use the :invert option to sift to exclude things, like this

(sift :invert true :include #{#"\.out/"})

@micha: ok right... i can't just :include a regex and filter based on that?

and also we need to remove those dirs in the task when source maps are not enabled and optimizations are on

no, you can

yeah so that's why i'm puzzled my regex isn't working to remove those...

and even with source maps, the *.out/s shouldn't be needed right?

source maps needs the output dir

there should just be a single main.js.map or something no?

no, it uses the original cljs namespace files

which are in the out dir

right yeah ok, makes sense

does the sift behavior i'm seeing sound right to you?

seems like a bug

i'm not a regex expert

so maybe test in the repl if the regex is doing what you think it's doing

yeah very weird behavior

should i be able to use two different sift tasks in the same pipeline?

sure

does the regex work the way you expect when you test it with just a string?

yeah

i mean now i'm getting weirdness even testing stuff like #{#"index"} -- which is causing the entire target directory to disappear

(within a final sift task i mean)

do you have symlinks in there?

none at all

i mean it should be safe but you never knwo

yeah no this is very haywire, i think i'm gonna need to put the crackpipe down even

that’s a good start

don't be hasty

haha

we need to rewrite the sift thing using reducers

micha swears by crack, i’m more into coffee

some people swear by sleep too...

sleep is a crutch

yeah totally

crack is more like a prosthetic limb

or maybe a phantom limb

hey something i meant to ask you guys a while back

was the sound just very bad on the video recording of the boot presentation?

or was that the crack?

the sound was not excellent, i remember some last minute shenanigan about there not being enough lapel mics

ok, yeah i was able to hear parts of it but it was a struggle... i was kind of hoping maybe someone had fixed it somehow with post-processing

we should do a rifftrack

i don't know what that is but it sounds awesome

a comedic dubb of a b-movie you can find on the internet

oh yeah for sure

have people done that before on tech presentations?

not that i know of

another pioneering moment; this project is truly on hallowed ground