This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2023-03-21
Channels
- # announcements (26)
- # babashka (115)
- # babashka-sci-dev (5)
- # beginners (48)
- # calva (69)
- # cider (4)
- # clj-commons (11)
- # clj-kondo (1)
- # cljfx (29)
- # clojure (109)
- # clojure-art (1)
- # clojure-czech (1)
- # clojure-europe (33)
- # clojure-nl (1)
- # clojure-nlp (3)
- # clojure-norway (7)
- # clojure-uk (1)
- # clojurescript (63)
- # clr (1)
- # data-science (41)
- # datalevin (1)
- # datomic (11)
- # emacs (58)
- # etaoin (11)
- # figwheel-main (1)
- # fulcro (5)
- # google-cloud (12)
- # helix (2)
- # honeysql (21)
- # hyperfiddle (22)
- # joyride (53)
- # malli (52)
- # off-topic (27)
- # portal (4)
- # re-frame (19)
- # releases (3)
- # ring-swagger (5)
- # xtdb (30)
chatgpt gave me the following gem
(doseq [neighbor unvisited-neighbors]
(conj stack neighbor))
From that and other code my impression is that it tries to translate code from an imperative language? It would be sort of a nice achievement, leaving aside that it doesn't work :)> From that and other code my impression is that it tries to translate code from an imperative language?
Yes I think so, too. It also sometimes makes locals and tries to update them in recur forms, same thing.
I wonder if you can prompt engineer it with consider everything in clojure is immutable
or something
I’ve seen it so similar stuff. Generally if you remind it Clojure data structures aren’t mutable it will try again and get it right.
I want additional money to work with JS, because of health issues. When your package dependency tool refuses to install a module that is inside the lockfile, and inside the package.json
, even after you basically nuked the whole node_modules
, no errors got reported, you already removed all your cache and such, and now you're thinking "why do I do this with myself? Why didn't I try something simpler like being a quantum scientist, or something less dangerous like lion tamer?"
Sometimes you also need to remove package-lock.json
. Or update npm
. Or sacrifice a lamb or something.
Sacrificing a lamb? I've been sacrificing goats. Did it change to lambs in a node update?
I'm just getting exposed to this: I've had to step in at work and rebuild the frontend CI pipeline (switching from Bamboo to BitBucket) and I've had to learn far more than I ever wanted to know about npm
and yarn
etc...😱
"Mamas Don't Let Your Babies Grow Up to Be Programmers" https://www.youtube.com/watch?v=i85ob2DackI
So, in my case - it was a package that had some requirement to be at a specific version of node (even though it was running just fine on a previous version). The issue is that it was failing to install, but yarn is parallel so the error was not on the last line, and it was being obscured by the much more aggressive "warning" that other packages were emitting 😞
I often redirect npm / webpack etc. outputs into a text file. So I can read and search it with my editor. This isn't a joke.
Stuff just breaks so often in this ecosystem and you get showered in warnings left and right no matter what you do.
Yeah- I actually kind of like JS relative to a lot of other languages. If you're disciplined about not using the dodgy parts, it can be treated kind of like a Scheme with M-expressions, and modern implementations are surprisingly fast given the nature of the language. But big parts of the ecosystem look like what happens when you take "Worse is Better" to its logical conclusion.
I’ve been trying to write a test for a webcomponent, and I stranded at the setting up the test framework step. The docs say it should just work (no), if you use typescript it should work (I don’t), or otherwise if you use this plugin it should work (nope). I really like webcomponents, I really like googles Lit library, storybook is great, but configuring new tools is a nightmare.
Y'all making me feel better about the "un-framework" angle I took when pitching my web framework. It seemed like a silly approach since it has more interesting virtues, but when I wrote https://tilton.medium.com/simplejx-aweb-un-framework-e9b59c12dcff (about the fundamentally same JS version) I kept yapping about it just being plain HTML. Everyone meant well building our ecosystems. The static nature of HTML along and the primitive nature of early JS, along with both being the only game in town, drove us to innovate atop the beasts, almost always meaning some kind of pre-processing/bundling. But now here we are on the Seventh Circle of Easier. We need a moral here about good intentions and road paving. Twenty years on JS is vastly improved and browsers and HTML are fine targets, but inertia. Who uses them as is?
I really encourage looking into webcomponents. It’s the web natives answer to the question: what if we allowed developers to extend html? In a sense also unframeworked . Lit is googles little helper library that makes working with that browser api so much more bearable
IIRC webcomponents don't work with hot code reloading. Meaning, you'll have to reload the whole page to see proper changes. But it's been quite some time since I read about them so maybe something has changed since then.
Hot code reloading in the JS world isn't so hot anyways. One of the popular build tools du jour just explodes with out of memory errors fairly regularly. Apparently since almost 2 years!
Hot reloading is awesome, but should that tail wag the development dog? I have the same question about TDD. Regression testing is great, but should we reject a strong approach to development that does not lend itself to TDD?
Also note that they are working on making it possible with lit. It's not an insurmountable problem.
btw, I am still documenting/polishing this for official release, but methinks this covers the goals of Web Components, atop vanilla HTML: https://kennytilton.github.io/web-mx-quickstart/#/html-composition. Check me on #CKCBP3QF9 if you need help getting started.
Don't.... tell me about web components... Atom bet on that front, and decided to make everything a web component. Turns out, like everything on the JS world, that it was deprecated and removed on recent Chrome versions, meaning that the whole editor is broken if they didn't add a polyfill to re-add the API.... which they did, and we on Pulsar found that the polyfill is not 100% compatible with what the API was doing, again, breaking some plug-ins
Finally, the new API landed, and it's basically completely incompatible with the old one. Not "incompatible, but we can craft a cool function that will convert most of the old code into the new format" - literally, "you can't convert the old format on the new one, ever"
Yeah, when I realized Web/MX was, in effect, delivering web components I went to check up on the official version. It did not seem to be a happening deal. Thx for the confirmation.
Well there was a v0 spec that only shipped in Chrome. I don’t know the details, but if that was what Atom was using, yeah, that got removed. It was never implemented in other browsers than Chrome tho, unlike v1, which is now (and has for a while been) fully implemented in the major three engines.
Not entirely related, but funnily (and sadly) enough I stumble upon browser bugs and compatibility issues more often than upon bugs in e.g. React. Especially when it comes to the bloody web audio. When I encountered a particular pesky issue, asked some experts, and they told me they ended up just writing their own audio stack in wasm because it's easier and more portable than dealing with web audio.