Hi. I want to write a server with ClojureScript, running on Node. Basically, I want to be able to use Express or Hapi, other JS libraries such as ones for i18n, but do so using ClojureScript and Hiccups with a view to starting with server-side rendering, and over time doing more client side stuff (so integrating reagent etc later). Here is a nice example of doing something like this using shadow-cljs: https://github.com/lambdaisland/npmdemo/tree/master/shadow-cljs I wanted to check in here on the following: - There are not a lot of examples of a setup like this to be found googling - is anyone here running a similar server-side stack with ClojureScript in production? - Any gotchas with shadow-cljs, which I have not used before? Any other thoughts and shared experience would be helpful. I could also do this on the JVM, but I like the idea of sharing frontend/backend code in future iterations. Also, I need to use gettext, and there don't seem to be any Clojure libraries for that (I want to use .po files etc.), and I need things like unicode URIs which in a very quick check, don't work as expected in Compojure.

@vikeri came across an old question of yours asking about RSA in goog.crypt. Did you find anything? I'd like to use RSA in react native and I couldn't find anything in goog.crypt either...

@pwalsh most clojure devs use clojure on the server-side but there are a couple people using CLJS on the server-side (using shadow-cljs). happy to help if you have specific questions.

@thheller at this stage the main thing is, how risky is it 🙂. Shadow CLJS looks very slick; I just don't know what I don't know. FWIW this is a pretty simple app - it gets data from a GraphQL endpoint and renders an ecommerce site, basically. Most of the business logic is on the API server which I wrote in Python.

no risk at all I'd say. equal or less risky than the other options. 😉

probably best to ask some people in #shadow-cljs. I'm a bit biased.

@pwalsh I’ve been using ClojureScript for a node.js server (via shadow-cljs) and it’s been a great experience

@mhuebert oh great to hear. any open code that I can peek at? Are you integrating with express, hapi, next.js or any other such framework?

Is there metadata about a function value that one can access in clojurescript? I'm putting the function that caused the exception in my error messages, and right now it prints the source of the function (which is awesome! much more than I hoped for) BUT I would also like to just get the name. Is that accessible in any form?

I just want the cljs$core$first part!

@idiomancy the name accessible via the .-name property on the fn but it will only be useful in development builds. :advanced optimized builds will strip that name or shorten it. and no there is no way to get other metadata about the function

ahh, interesting that makes sense. Same will go for the source, I imagine

source is always available, it just will make a lot less sense in :advanced 😉

lol yeah. Thanks, theller!

@darwin Thanks for Chromex! I'm using chromex-sample to figure out my own WebExtension. Just wondering how deeply you dug into the content_script init issue under :optimizations :none (referenced here: https://github.com/binaryage/chromex-sample/blob/master/project.clj#L57). It looks to me like the assets are being loaded from the page domain, rather than the extension domain. I guess under :optimizations :advanced the issue is sidestepped since everything's in one .js file. Has anyone else used CLJS for a WebExtension content_script? Where can I learn more about how asset loading works under :optimizations :none?

Yeah, so, it's doable but nontrivial @robert.mather.rmm. I've got an extension that injects a figwheel repl into any arbitrary webpage. I'll try to publish it tonight.

The only question I'm wondering about is whether or not to monetize it or not. Or to have an optional upgrade/donation button...

Anyone have opinions on which way to go with that?

I was thinking a 3 dollar monthly subscription, where 1 dollar goes to figwheel, 1 dollar goes to Clojurists Together, and 1 dollar goes to my project (and whoever is contributing to it)

That sounds like a fair price. Would it be for commercial uses only or everyone?

Not sure... But going that way, I could probably just, what, have a string somewhere saying "non-commercial version, click here to upgrade." And then just trust the user to upgrade if they're commercial?

I'm not really interested in splitting out feature sets

I think that’s how REBL does it right? It’s just stated in both the docs and its license that REBL is free for open source projects but you gotta have a datatomic or active paetron subscription to use it commercially.

Honestly, I don't expect to get much beer out of a dollar a month subscription. For me, it's more about helping to normalize charge-models around open source software. So I'm generally fine with a "trusted upgrade" or donation model, or not charging anything at all. It's more-so an experiment for the principle of it.

I believe so too

That makes sense. What I like about the trusted upgrade approach is that it gives people the option to try it out on a personal, OSS side-project, get excited about it, then purchase it for work.

Yeah

Okay, I'll try to take that route first

But if all the monetization bits prove to complicated, I'll probably just release a non-monetized version tonight, just to get it out the door, for people to play with - with a "trusted upgrade" versions soon to follow.

That sounds like a good plan to me.

BTW @jayzawrotny, I'll be using your template as inspiration for docs... The solution comes in two components: 1) the extension you install, and 2) a template that you run with lein new or clj -A:new which wraps figwheel's default template, with the extra configuration bits necessary to let the extension connect back to it. So your template docs are going to come in super handy as I build out that template, thanks!