cospaia

Clojure can to some extent help you get a full/true REPL into Java and any JVM language, but only to some extent. You'll have to use Clojure to get the full observability and program runtime-mutability.

Wow. I'm the creator of Calva and I can tell you that I do not have an opinion on how to use it. I would never say something like that, and now I wonder what I could have said that was interpreted this way...

By default Calva Paredit overrides three or so keyboard shortcuts. This is to help the users to not accidentally delete brackets, which in turn keeps the support questions down about broken Clojure code. I have tried to make it super easy to disable these overrides. There's a command for it. There's a button in the toolbar for it, and there's a setting for it.

There's also a setting for removing all Paredit bindings, even those that are not overriding any default ones. (This setting is for people who use Paredit, but want to provide their own shortcuts entirely.)

As a Clojure editor tool smith it pains me that this is the case. Especially for ClojureScript where I spend most of my time. I really want to fix this.

For JVM interop, I think Cursive (IntelliJ Clojure plugin) is smart enough to help with autocompleting Java libraries.

Thanks! Even though I am not super familiar with Ports & Adapters, this article made Polylith click in a way that no other article or video I've so far encountered has succeeded with. Composable mini-cores. Brilliant!

I am a member of the Interactive Programming cult, being dogmatic around one thing in particular: I demand that programming is joyful, else it is not for me.

Clojure changed my life. It made me realize how fun it is to code and switch back to being a programmer. This makes me a much happier person and thus a better husband and father. Never looking back. The Clojure Way just fits me perfectly.

Thanks for sharing this! It is a quite wonderful way to learn. Reminds me about Racket, which was my tool for following along the Land of LISP book.

I made a video about some of the benefits of using ClojureScript for developing VS Code Extensions. It's a short (9 minutes) demo of me hacking on an extension of mine, trying to relay what it is we Clojure peeps mean when we speak of Interactive Programming (aka REPL Driven Development). I find it super powerful when hacking on extensions, since restarting the Dev Extension Host can be pretty disruptive. And having to halt the program in the debugger to inspect it often does does not fit when searching for the cause of a problem.

This article on LISP interactivity inspired me: http://blog.fogus.me/2022/11/10/the-one-about-lisp-interacti...

Thanks! Please consider filing your suggestion as an idea on the Joyride discussions section: https://github.com/BetterThanTomorrow/joyride/discussions

Ah, so it's the editor war we are fighting here? =)

=)

Writing that made me challenge it. What if it is possible? https://github.com/BetterThanTomorrow/joyride/discussions/14

Which fine steak would that be?

It is definitely true that Joyride can't make VS Code as fully extensible as Emacs is. VS Code is not designed this way. And also some extensibility from an extension need to come from an extension manifest when an extension loads, so Joyride can't declare new commands for instance. I'm guessing that is the simplest thing in Emacs.

Having, draw a rendom number out of a hat, 50% of Emacs extensibility at your REPL is fantastic, and worth pursuing. Joyride can do everything a VS Code extension can do, except the things i mentioned above, and anyone who has used some VS Code extensions knows that that is a lot.

However, the self documenting part described above is not out of reach. Most of the code building up the VS Code API is open source. Same goes for many of the extensions. Looking up and navigating to any of that code is possible. Perhaps a lot of work would need to be put in, but it is possible.

Tbh, the last nights have been less about sleep and more about hacking on Joyride. =)

But usually I do get sleep. The question we should ask is if borkdude sleeps? He's doing about 100X as much as I do. Maybe there are hundreds of borkdudes.

As the co-creator here I'd like to add that Joyride is taking the very first baby steps here. We hope a lot of people want to try it and provide feedback to inform our next steps.

NB: The video has a target audience of people who know Clojure and its strengths. Especially how the REPL is used. I hope HN will see beyond the parens. =) It should make some sense anyway. Especially if you have some Emacs scripting experience, because that is where most of the inspiration to Joyride comes from.

Thanks Hacker News for helping me solve that mystery with performance in my Java program. Now I only have the problem left to solve. (And some of you might learn a thing or two about deployment of programs where performance is of utter importance.)

Thanks! It does and it does not. Here's 100 rounds where it does: https://github.com/PEZ/ghost-chase-condition/tree/master/tes...

But in other cases I have it seems to not help that much. Not sure how to minimize those, but will try figure it out.

I had this issue using JDK17 and 18 as well, so at least by default things are not better. =)

I wonder if there is a way that does not involve using `numactl`. It needs the container to run in privileged mode and that is not how the containers are run in the drag-racing benchmark...

I can't bind cpu and memory different.

  $ numactl --hardware
    available: 1 nodes (0)
  node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
  node 0 size: 15395 MB
  node 0 free: 2542 MB
  node distances:
  node   0 
    0:  10 


  $ numactl --cpunodebind=0 --membind=1 java PrimeSieve
  libnuma: Warning: node argument 1 is out of range

I love the detail with which you describe the experiments. Let's see what I get running this on my Linux box!

It was indeed this.

UPDATING again with that the solution is found, thanks to user CraigJPerry. The blog post is updated with some details: https://blog.agical.se/en/posts/java-bitset-performance-myst...

What is meant by that the runs are passed through a different core?

That's some pretty impressive results. Not seen the sieve perform on that level before.

No, I haven't seen this particular thing happening without Docker. It seems to be a NUMA issue, btw, see https://news.ycombinator.com/item?id=30231945

This seems to be the issue. Running it 10 times gives me quite consistent performance near the top.

    0 - Passes: 5876, count: 78498, Valid: true
    1 - Passes: 5870, count: 78498, Valid: true
    2 - Passes: 5944, count: 78498, Valid: true
    3 - Passes: 5903, count: 78498, Valid: true
    4 - Passes: 6051, count: 78498, Valid: true
    5 - Passes: 5874, count: 78498, Valid: true
    6 - Passes: 5902, count: 78498, Valid: true
    7 - Passes: 6041, count: 78498, Valid: true
    8 - Passes: 5906, count: 78498, Valid: true
    9 - Passes: 5889, count: 78498, Valid: true

Will run it 100 times now to lower the risk that this is just a coincidence.

Updating this to thank you all for this amazing help!

I've been mostly afk, and haven't been able to act on all the clues yet, but I have conducted some of the experiments and tried to summarize them in a reproducible way here: https://github.com/PEZ/ghost-chase-condition/tree/master/tes...

Thanks. I'll work a bit with unpacking all this. =)

Meanwhile. I just published an experiment where I have switched out the BitSet for an array of booleans. Just to demonstrate my observations that then the performance is perfectly stable. https://github.com/PEZ/ghost-chase-condition/tree/master/tes... Does that affect your analysis in any way?

Thanks! I'll check JMH out now. Some week ago, while I thought the problem was todo with something in Clojure - https://twitter.com/pappapez/status/1486339518902984707?s=20... - I ran async-profiler via clj-async-profiler. It turned out to not be granular enough. OTOH, the problem had different clues then and I should try it again now with the latest Java-only repro.

I added this experiment, and the results (spoiler, it isn't getCurrentTimeMillis) here: https://github.com/PEZ/ghost-chase-condition/tree/master/tes...