'(Robert Smith)

> be NASA programmer > excited to write Lisp > code lookin like hieroglyphs > boss walks by > "wtf is that, alien code?" > no, just Lisp boss > boss laughs, Lisp now forbidden > mfw no cosmic parens > mission to Mars now in JavaScript > planet doomed > feelsbadman.jpg

@bygregorr user reported copy-paste doesn't work, that required patching in wl-copy other distros seem to not be able to make proper OpenGL/MESA (?) calls due to different webkit versions (?)

@stylewarning What broke first, the copy-paste or the webkit dependency pulling in a version your distro didn't have?

Last time I complained about packaging apps on Linux, people said it was all a well solved problem. Guess I have skill issues because: - Wayland and X have different ways to copy-paste - libwebkit2gtk-4.0 minor version bumps breaks API compatibility Frustrating...

Packaging apps for Linux in a way that is even somewhat portable across distros is incredibly frustrating. No wonder every distro effectively builds all their apps from scratch within their own package manager bubble.

If a coffee shop starts advertising pictures like these, it's guaranteed to have undergone the coffee-equivalent of enshittification.

@run_guns @LukasHozda the meta-WITH operator is UIOP:NEST fare.livejournal.com/189741.html cc @Ngnghm

@LukasHozda Many WITH- macros, all cluttering... Time for a meta WITH- macro you lass it a list of what yo WITH- and automatically generates allthe WITH-

Hiisi’s CL routes often signal validation conditions and let one wrapper translate them to HTTP. It keeps the happy path readable without losing typed errors. Unfortunately there is a risk of having a stack of (with-* ...) macros, but I keep my routes small hiisi dot app

@scheminglunatic are they really former then 😬

@stylewarning i had to stop using x:xs in haskell because i then remember my former romantic interests and get so sad i have to log off the computer.

The reason why richly-typed languages have shorter variables is because the information is in the types not the names. But for me, I think it's interesting because you actually get 3 or 4 letters not just 1 -- variable name as 1 lalamo token, not one letter.

@scheminglunatic I agree with you, and another source of information can be that which is convention. 'c' is character (C), 'xs' is rest of a list (Haskell), etc.

@stylewarning Another thing is like when I think about where information about a "thingy" ends up - it ends up somewhere. - in my mind, if it doesn't end up in the type it's in the name and if it's not in the name or the type it's in a comment right next to it. that kind of pattern

@scheminglunatic s/in addition to/somewhat driven by/ The average OCaml code doesn't read lexically similar to Haskell (to me), and Common Lisp definitely doesn't read like Python, despite similar typing stature.

@scheminglunatic I don't agree with this. I really think it's culture and convention, in addition to namespace ergonomics.

@snmsts I don't read sbcl-devel very often. What sort of things?

reader.tymoon.eu/article/444 ずっとshinmeraのsbcl-develでの発言がひっかかっている。

Cool new project from 🇯🇵: dotcl is Common Lisp on .NET by @snmsts.

@geistgrail @croloris Moreover I generally would not consider AI slop-shippers to be bad actors in the traditional sense (wanting to cause malice), but rather trying to gain notoriety with low effort. A lot of AI PRs are actualky aimed to fix and improve.

@geistgrail @croloris Disagree with (2). In my experience on GH so far, AI sloppers aren't reading any rules. They just blindly fire and forget.

Contributor Poker and Zig's AI Ban kristoff.it/blog/contribut…

@_Felipe Just a normal amount (?). If I were responsible I would do a careful deep dive to figure it out. Instead I just want to program. (:

@stylewarning An antivirus running on the Windows machine? Project contains many many tiny files?

lol @ this distribution of timings for building an identical program on different platforms and processors. macos arm64 - 6 min win11 arm64 - 20 min

@croloris Your own post sort of concedes that this isn't a problem intrinsically w/LLMs, and more of a problem of misuse and lazy engagement. I predict Zig will backpedal on (or worse, be entirely hypocritical with) their position in due time, and I'm not really an LLM-optimist.

@croloris My own 2¢: Regulating the tech used to produce a PR, I think, is a lost cause. Either a person is sincere and reasonable (≈ worth your maintainer time), or not. This calls for better moderation (à la "vouch" or similar) and more rigid participation/excommunication policies.

Closed/deleted the bounty. It was a bad idea on my part 100%. "I will pay you to interact with someone else's repository on Coalton's behalf" as a public offer was a mistake.

I put a bounty on a Coalton issue (add syntax highlighting support in another non-Coalton project) and in less than 5 hours some unknown people started pushing AI-generated PRs to this other project. Makes Coalton look bad. Never thought I'd need to actually moderate on GitHub.

@ludwigABAP lmfao 💀

cant make this shit up

And before anybody calls me out, yes, I am a paying customer, and yes, they really do charge *me* by the minute for their infra not knowing its head from its ass.

I'm just cherry-picking one response here but it applies all: Spare us the bleeding heart. It's literally been a meme that GitHub is in shambles. Three nines (89.99%) jokes, etc. Focus your energy internally on finding leadership that gives a shit, and not writing LinkedIn slop.