This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2024-06-10
Channels
- # announcements (1)
- # babashka-sci-dev (16)
- # beginners (5)
- # calva (3)
- # cider (42)
- # clojure (103)
- # clojure-europe (79)
- # clojure-nl (2)
- # clojure-norway (17)
- # clojure-sweden (2)
- # clojure-uk (8)
- # clojurescript (7)
- # community-development (7)
- # core-async (1)
- # core-typed (18)
- # cursive (9)
- # data-science (1)
- # dev-tooling (14)
- # events (1)
- # fulcro (2)
- # hugsql (2)
- # hyperfiddle (1)
- # leiningen (7)
- # malli (11)
- # re-frame (14)
- # releases (9)
- # ring (18)
- # shadow-cljs (55)
FYI - I've created #C076WKL025D for discussions specific to CIDER's development, so we won't spam you here with topics that are not of general interested to everyone. (e.g. CIDER's release planning, etc)
I added a bookmark to cider-nrepl docs to this channel. (Since they are very hard to find using Google or similar means 😃 )
In other news - we didn't get any issue reports for nREPL 1.2-beta2 so far, so I'll likely cut a new release later this week. As a reminder - it brings back the ability to interrupt code evaluation on Java 20+ and improves a lot the bencode performance.
I haven’t gotten around to implement this in Calva. But it all looks great to me and I don’t see any particular troubles with implementing support.
Yeah, it's really just a matter of loading the JVMTI agent, so it should be fairly simple for jack-in, and for the other use-cases it's basically an entry in the docs that users need to load it themselves.
E.g. something like https://docs.cider.mx/cider/basics/up_and_running.html#enabling-nrepl-jvmti-agent
> it brings back the ability to interrupt code evaluation on Java 20+ How did you implement this without Thread/stop being available?
We considered another option as well (see https://github.com/nrepl/nrepl/issues/296#issuecomment-2098870752), but the JVMTI option seemed better overall.
I try to support pretty much anything gracefully. In practice things that most people are using (e.g. nREPL) have some minimum value, but I try to make sure that Cursive handles it all.
Interesting. nREPL targets Clojure 1.7, but if we get to a point where no tool supports it we might switch to targeting some newer Clojure. That's why I asked.
Sure, again it’s easier due to my policy of mostly doing things by static analysis and minimal eval-only use of the REPL. So it’s easier to maintain support for older versions. I actually don’t test older versions so it’s possible some 1.7 stuff might have crept in, but no-one has complained about it either.