Is it expected that adding cljs.spec.alpha (and using it) to increase the (optimized) code by roughly 60kb? shadow-cljs reports this:

cljs/core.cljs  174.24 KB
cljs/spec/alpha.cljs  35.84 KB
cljs/spec/gen/alpha.cljs  27.33 KB

Seems I’m hitting https://dev.clojure.org/jira/browse/CLJS-1701 — please vote if this is important to you 🙂

@orestis we're aware of that one, but there's not an obvious way to tackle it - the main reason there's very little activity

issues like often require some brain storming & cleverness - anyone should feel welcome to investigate a plan for it

(note I would say we're not really concerned about the size of the spec stuff, if you include you need a lot of it)

only the part about not shaking out unused specs

multimethods have a similar problem but I'm skeptical that's solveable due to dynamism

honestly I think the same may apply to unused specs

(keyword ...) etc.

but we can leave that one open if someone gets a really great idea that hasn't been considered before

Yeah, I guess anything with a global state would have this issue. I’m not dying but just was surprised this happened. Thanks for the feedback!

@orestis it's not global state that's the issue

namespaces etc. are global state - but references are used statically

you don't invoke fns as strings

but specs & multimethods use keywords, dynamic lookups

this is probably an argument BTW to keep specs separate from code, which is my preference

then loading specs is optional

and can be controlled

Yeah, I can do that. I planned to have some validation done at runtime anyway, so the issue will become moot, it just caught me off guard.

I’m using CLJS in anger recently, and the experience is very very nice. Many thanks for that 🙂

cool!!! 🙂

How are people doing translation of front-end stuff? I’d like to have something as easy to use as transifex, but it seems I’m on the hook for generating a gettext .pot file, or xliff file, then somehow integrate it back in my code. I’ve seen tempura as a potential library, anything else?

I have a library that is intended to be used with Node.js. It needs to read some static files at runtime that I would like to package with the library

is there a a reasonable way to distribute static files that can be read at runtime? analogous to io/resource?

essentially I want to be able to read a file from the classpath in Node.js

Hello again, so I got fighweel-main mostly working except that in a node app if I run bad code which causes an error from the repl it gets printed in the node process to stdout. Is there a way to opt out of that behavior? I don’t think lein-figwheel behaves that way.

I keep seeing these kinds of log messages [goog.net.WebSocket] Opening the WebSocket on

that feel when you get advanced optimizations to work