Thank you for taking the time to write and publish this. It may seem simple but I couldn't find a similar article back in the day. I've been using Netlify ever since (found it easier to begin at the time) but using GitHub pages seems to be just as simple.

@tkjoine Would 'static' here accommodate a self hosted ClojureScript site? No server, database, server side requirements, but fully 'active'

Good question. TMK it should, but maybe I am missing something.

Maybe this is another example of Serverless computing 😁

Hey @U0XTHMGMP, what does your current folder structure look like? and what does your static_website.cljs look like? I just switched over to Java 8 and it’s still working for me.

What is the -d option anyway? I do not see any documentation for it. https://clojure.org/reference/deps_and_cli When I take it off things seem to work fine.

Interesting. -d is a ClojureScript compiler option: https://clojurescript.org/reference/compiler-options#output-dir allowing us to set the output dir of our compiled ClojureScript. You can check out more options at the above link, or via command line:

clj -m cljs.main -h

Having said the above, I am able to reproduce the error your specified. It happens for me if I run the following command

clj -m cljs.main -d "demo-clojurescript-gh-pages/out"  -c demo.static-website -r

Interesting, if the directory demo-clojurescript-gh-pages is present there is no problem.

Yup. I wrote a footnote that hopefully explains this in greater detail because I figured it might not be clear https://betweentwoparens.com/deoploy-clojurescript-to-github-pages#fn-project-name

Let me know if you have any suggestions for how to make this clearer as I feel it is something that will come up more often than not. Suggestions and feedback are very welcome 🙂

It might be good to add a { :paths ["src"] ... in the deps.edn. I think that is the default. Also from the instructions is sounds like the clj command should be run from the src/demo directory.

Awesome. That makes sense and would explain what happened to you. I will update. Thank you!

I think the actual issue was that the deps.edn file had not yet been processed successfully. The unknown -d option having prevented a successful first run. I think what did the trick was running clj -m cljs.main -h which effectively retrieved clojurescript. I then created the "missing" directory and the next run succeeded.

Oh wow. That’s interesting. I cannot speak to your case, but I can say that I deleted all local versions of ClojureScript and noticed that even if I ran the clj -m cljs.main -d ... command from the correct place, it would not download ClojureScript. I then deleted each directory generated by clj (because I still had the previously generated dirs from successful runs) out, demo-clojurescript-gh-pages and .cpcache and identified that if a .cpcache directory is present, ClojureScript will never (?) be downloaded (even if it is nowhere on the local). This includes if you run clj -m cljs.main -h and also if :paths is specified in deps.edn. In your case though, I would have to know where the commands were run. Either way, I def’ learned about some interesting behaviour 🙂

Me too. Thanks.👍

I have also been looking at the --output-to option for the production build.

Yeah, that’s a good one too. I will post a more “best practices aligned” version of the demo in a few days to show how I might actually configure this if I was trying to maintain a static website and not just a demo