kcomhnallso, bit of a newb question: IDEA is saying my 'configurations' declaration is 'unstable' and that it is marked with the @Incubating annotation. I'm on Gradle 8.10.2, but I can't seem to find where this is documented. any tips or insights? All I can find on this is the Kotlin dsl script handler reference page, which isn't entirely helpful. (i'm still looking around)
ParaIt's an IDEA inspection. Guava's @Beta triggers similar.
* dreamreal sees gradle and recognizes the actual formal response is and should be a shrug
kcomhnallPara: ah okay. I kind of figured so.
kcomhnalldreamreal: heh
dreamrealit's not that I don't care, it's that the rules around gradle are "the devs don't care and have said so"
dreamrealyou live with what they give you, they mean well and there you are
kcomhnalldreamreal: yeah. and as they say - everyone has to start somewhere.
dreamrealI use gradle and appreciate some of its strengths but I find I don't *like* it much.
dreamrealgiven a choice in my own projects, it's maven first, second, and third... and gradle comes in fourth. It's still way ahead of sbt, but that's only because I'd rather strip naked and roll around in a pool of rubbing alcohol that has bits of broken glass and thumbtacks in it than use sbt again.
dreamrealnot that I'm opinionated or anything.
kcomhnalllike most java devs - I was first introduced to maven. it was somewhat imtimidating I'll admit.
kcomhnallthen I realised I can still use a .pom file even with gradle and said..meh. I'll pass on that path.
dreamrealI think it's because people are introduced to "how" and not "why"
kcomhnallironically, if I have to use (or decide to use) Eclipse, I'll likely *want* to use maven. weird right?
dreamrealhah, why would you use maven if you had to use eclipse?
dreamrealis m2e THAT good over eclipse' gradle support?
kcomhnalldunno, I have a learning curve appetite or something.
kcomhnallm2e?
dreamrealthe maven eclipse plugin
deebois maven still xml only? day-to-day has been gradle projects for 10+ years so trying to debug some maven projects was a shock
deebolike going back to configuring spring in xml
dreamrealIt hasn't been LITERALLY XML-only since 2007, just nobody uses the non-xml since copy-paste is so effective
dreamrealconfiguring spring in xml... okay, honest question: why is XML hard for people?
dreamrealI mean, verbose, yes, agreed, conceded, it's very verbose, no argument even offered there
dreamrealbut... I don't understand the *difficulty* at all
dreamrealwhat am I missing?
deebodon't really have an answer for that, just got so used to not having to deal with it
deebono need to do much programmatic setup either in our case when using spring boot to be fair
dreamreal*nod* Understood - I just never can quite grasp the resentment for XML on "difficult" grounds.
dreamrealVerbosity, completely understood. It IS verbose. I don't use XML for data transfer if I can avoid it. But ... verbose is relevant because things have meaning, and in XML and SGML those meanings are well-defined and VERY clear if you understand a VERY few simple things... and most people don't seem to.
kcomhnallfor me, I just get confused. looking at a well-implemented xml file makes me dizzy. I can't keep up with all the tags. I hate html just as much though.
dreamrealI wonder why. :D
dreamrealI find well-written HTML or XML to be a joy. It's verbose, but a joy. HTML for teh web tends to mke me want to vomit.
dreamreal<div><scan> ... NO YOU JACKWAGONS OMG
dreamrealsorry, span, not scan, but same effect
kcomhnallthough, I start the file myself I'm not so bothered because i've already put in the time to understand it.
dragonmasterI'm an utter novice, but it seems perfectly reasonable to have <span>s within <div>s
dragonmasterIs that not analogous to statements within blocks?
ParaSpan element has always been meant for highlighting a section of text in some way.
ParaOf course nutsos force box model on for it, turn it into a piece in CSS grid and add all the other stuff until it's about as close to a normal, intended span as Starbucks coffee is to proper coffee.
ParaBut eh, maybe I should just create an <llm> web component and somehow monetize that.
* dreamreal giggles
dreamrealnevet now has a command that's run in clojure :D
p14nd4hello! can anyone educate me / point me to appropriate docs / search terms? an old project was using org.glassfish.jersey.bundles/jaxrs-ri, but jersey 4.0 was changed to "stop distributing jaxrs-ri bundle" without any explanation or documentation ( https://github.com/eclipse-ee4j/jersey/pull/5556 ). 1. any idea why that was discontinued? 2. am i now supposed to just explicitly include the dependencies
dreamrealI don't know what sci is here - this is done through a function call generated from clojure modules
dreamrealI have a markov chain trigram system in 50 lines of clojure, and dispatch via a shim in kotlin
DoofusCanadensisI recognize those words
dreamreal"14:05 <nevet> * channeling para: to it, and it takes quite a bit of effort to realize that for example the entire database functionality is a feature in itself. so yeah, been stuck with that"
dreamrealbasically, a markov chain that relies on three word chains so it definitely goes off the rails but takes a little bit of time to do it :D
DoofusCanadensisno, I said I recognized them
dreamrealmarkov chains don't have to be clojure but clojure's a good way to build 'em, plus it's fun to prove out polyglot design
Paradreamreal: sci, small clojure interpreter, does what it says on the tin, very capable, and sandboxed by default, unlike Clojure itself :)
dreamrealah, no. This is definitely put in the build
dreamrealit's still scripted so it gets loaded at runtime but eh
dreamrealYou didn't see how quickly it responded to the request to imitate you, but it's *really* fast
dreamrealand that includes te full trigram build process and the DB query
DoofusCanadensiseverybody wants to treat LLMs as "new" without acknowledging what came before
ParaIt's obviously getting confused from my incorrect usage of commas.
ParaAnd other random bits.
dreamrealPara: or by being a markov chain!
ParaI recognize that bit from not too long ago, though!
dreamrealmarkovs are hilarious. That markov chain thing was actually one of the things that led me to write nevet in the first place
ParaMarkov chains are more fun than LLMs.
DoofusCanadensisright?
dreamrealI wanted to get it to "be someone" to run sentiment analysis over the generated text, because my assertion was that they were generally toxic; a markov chain generator that used their body of content as a corpus would similarly be negative if I was right, but wouldn't rely on me pointing out SPECIFIC INSTANCES where I thought they were negative
dreamreallike, it'd remove ME from the selection process
ParaBe someone, or be someone's fool.
dreamrealin other news, their generated content *did* tend to be negative, as long as the generated length was long enough
ParaI did a fun agentic test today. It scans PR commit messages, checks that the messages match a few rules and then complains if not. Mostly just that the commit answers what was changed and why.
ParaBest part was that it complained about all five commits in the PR which added the agent itself.
dreamrealhaha
dreamrealnevet has pretty severe throttles over AI usage in most places
ParaWe had internally a bit of an agent spikeup on Friday and today.
ParaSomeone already blew 100 bucks accidentally in like two hours, so I'd say we're pretty much on track!
dreamreal:D
dreamrealmeanwhile the LLMs are like "whoops! Please, sir, may I have another?"