Fork me on GitHub
#cursive
<
2022-11-11
>
onetom00:11:41

Is it just me, or others are getting these exceptions on startup in IntelliJ IDEA 2022.3 Beta (Ultimate Edition) Build #IU-223.7571.4, built on November 9, 2022 ? 1. Cursive (1.12.5-eap-2-2022.3) java.lang.Throwable: Activities registered via registerPostStartupActivity must be dumb-aware: cursive.build.BuildSystemUtil$$Lambda$1473/0x00000003015c8748@10af016c 2. IDEA core java.lang.IllegalStateException: java.lang.IllegalStateException: No HighlightingSession found for ClojureFile (class cursive.repl.ReplFragment) in ProgressIndicator 670549150: running=true; canceled=false in map: ClojureFile (class cursive.psi.impl.ClojureFileImpl) -&gt; HighlightingSessionImpl: myVisibleRange:(2697,5816); myPsiFile: ClojureFile; ClojureFile (class cursive.psi.impl.ClojureFileImpl) -&gt; HighlightingSessionImpl: myVisibleRange:(0,2813); myPsiFile: ClojureFile

cfleming08:11:28

It’s not just you - I have the first one fixed for the next build. I haven’t seen the second though, but I’ll check it out tomorrow.

cfleming02:11:33

The first one is fixed in the current EAP, out now. However I haven’t seen the second one, and no-one else has reported it through the tracker as far as I can see. Is that when you’re debugging? I can’t see how you’d receive it at startup.

cfleming02:11:58

Do you have a full stacktrace from the logs for that one?

onetom00:11:29

On 2022.2.2 I'm having issues with cursor focus being lost, especially, when pair programming over the native client of http://pop.com On 2022.2.3 Load File in REPL (`Cmd-Shift-L`) is loading unnecessary files into remote nREPLs On 2022.3.x I'm getting the mentioned exceptions constantly, which is very distracting, but we found other small issues too, as expected from EAP versions I'm feeling like when I have a small rock get stuck in my shoe and i can't get it out 🙂 It feels like IntelliJ is constantly getting faster and smoother, so I would really like to be on the latest version. I found the new UI pretty good too, after some tweaking, so that has a strong draw too 🙂

onetom01:11:57

ah, yeah, i can completely lose symbol resolution in 2022.3.x, within minutes after opening a project, despite it being fine initially

onetom01:11:11

so it's not really usable at the moment 😞

onetom01:11:08

I must be fatiguing to build a Clojure plugin on top of this hyper-object-oriented JetBrains editor platform 🙂 I just peeked into java.net.HttpCookie recently and I was baffled to see, that it's still relying on System.currentTimeMillis() for it's constructor and hasExpired() method and even some expiryDate2DeltaSeconds() private method, used from some assignors java.util.Map... My head is spinning from it and I'm still not sure, how could I mock the time in it. If it would be using java.time.Clock, I could just use a different one...

onetom01:11:49

I mean, even the JDK is so convoluted, so I can imagine that the JetBrains API being even crazier.

Ivar Refsdal13:11:44

I was able to (in practice) override System.currentTimeMillis using javassist/bytecode rewriting. Test verification: https://github.com/ivarref/hookd/blob/main/agentuser/test/agentuser/core_test.clj#L53 I've seen a few test libraries and byte-buddy do this as well... Welp. I spent way too much time on this 😕

🙏 1
onetom03:11:32

oh, that's reassuring, if clojure libs are doing such wizardry too!

onetom03:11:03

for java.time use-cases i gave a try to the tick.core/with-clock macro (from https://github.com/juxt/tick) and it works reasonably well, just needs some practice and getting familiar with the various clock options...

onetom03:11:48

i see you were also testing with java.net.HttpCookie 🙂 so much time is wasted on this issue world-wide probably. either because ppl can't believe it's not supported to fake time or they actually write tests, which use wall-clock time, constantly slowing down test feedback loops...

Felipe Nascimento13:11:58

hey guys, Can cursive auto complete node depedence?

cfleming19:11:16

No, sadly - I’m considering some options for that.

👍 1