[QUESTION] Why Clojurescript and not JVM? about closh

[dundalek]

I don't want to give any promises, but it will hopefully be done in a month or so.

[neominik]

JVM is faster, but has a lot slower startup time. Waiting 10 seconds for your shell to start up every time would be a no-go for me.
Otherwise I would also like to see a JVM version.

[dundalek]

I am hoping that in the future it could be compiled to native code with Graal which should make the startup time be less than a second.

Anyway, the JVM port is in progress in the master branch. It is mostly done, there are just a few issues that need to be ironed out before a release.

[dundalek]

I come from js and node background so it is easier for me to hack on. Plus lumo is awesome and it does a lot of heavy-lifting. Basically when I saw lumo I got the idea I could just hack the REPL a bit and got the prototype working.

I think with the new clj tool and clj.main Clojure proper gets more hackable too. It would be a cool exercise to abstract the code and implement platform specific code for JVM support. I would like to have that eventually down the road.

[NightMachinery]

That'd be awesome! JVM Clojure is both faster, more stable, more secure, and has a lot more libraries! 👍🏻

…

On Thu, Mar 29, 2018 at 5:57 PM, Jakub Dundalek ***@***.***> wrote: I come from js and node background so it is easier for me to hack on. Plus lumo is awesome and it does a lot of heavy-lifting. Basically when I saw lumo I got the idea I could just hack the REPL a bit and got the prototype working. I think with the new clj tool and clj.main Clojure proper gets more hackable too. It would be a cool exercise to abstract the code and implement platform specific code for JVM support. I would like to have that eventually down the road. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#66 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/Aii--l9Lsly6aFD7oIya8nBTyvj_xb7Mks5tjOEzgaJpZM4S_IbJ> .

[zcaudate]

this is exciting. when will this be out?

[dundalek]

Just FYI here are remaining issues that need to be completed for the JVM integration so that it can be released for v0.3:

• Fix remaining failing unit tests regarding file redirection
• Fix cding to non-existing directory which ends in infinite loop
• Handle SIGINT properly
  • For example when a process runs (like sleep 10) and ctrl+c is pressed then shell quits.
  • Make sure it also interrupts properly when running in process thread, e.g. (Thread/sleep 10000) and pressing crtl+c
• Bonus: Port the autojump integration
• Fix glob expansion when switching directories

[zcaudate]

would it also be possible to have nrepl support?

[dundalek]

@zcaudate What would be the use case for having nrepl?

[zcaudate]

@dundalek: I'd be looking at it as a replacement for lein

[zcaudate]

it'd be essentially like having both the shell and the repl in one.

[dundalek]

@zcaudate I see, it should be possible to include nrepl middleware to start an nrepl server. Could you create a separate issue so we can track it there?

[jeroenvandijk]

I think this can be closed as there is a JVM version now..? 🎉 #87

I'm already playing with it => #92

[dundalek]

I agree. Unfortunately more blocking issues were discovered. There are separate issues opened, I label them with bug and jvm tags.