At CS school 25 years ago one of my mate was developing a chat client and I kept on finding ways to mess with him, e.g. sending him text messages full of %20 or extra metadata, building a fake frozen version of his app etc.
Posts by Omar
#oopsie
Anthropic secretly installs spyware when you install Claude Desktop
www.thatprivacyguy.com/blog/anthrop...
Razor1911's 40th anniversary demo is a nice retrospective of early demoscene/cracktro/bbs aesthetic 😍 www.youtube.com/watch?v=2Anb... (cc: razor1911.bsky.social)
Hey! Today at @limbicgames.bsky.social we announced Corsair Cove. I hope you'll like it. It's about to be released soon! Don't forget to wishlist it!
store.steampowered.com/app/1368140/...
That bug alone made me extra confused for a good 1-2 hours, because GIF capture does a specific override leading to 60 FPS being displayed in most user-land places while it was actually running at 1000+ and I was confused that some of my debug marker wouldn't show.
Coincidentally I've also been bitten by a bug in test engine where suspending a test script (to manually interact with the UI) in the middle of a video capture would leave me with unthrottled framerate, much faster than monitor's, making some single frame markers not reliably visible.
The varying framerate was mostly obfuscating my understanding/repro of the bug but once I understood it I repro-ed in automation without caring much about that part. That said, test engine can simulate any DT.
Even writing the repro has been initially tricky, because it ties to pixel-perfect mouse positioning & varying framerate (making debugger use trickier). Using automation to repro reliably now. The 4-5 days worth of confusion have materialized into a pile of on-screen debug geometry and log entries.
I've now been 4-5 days deep into understanding and debugging a group of confusing & entangled issues.
I realized I haven't been in this situation for a long while: Most issues I tend to understand fast (even if I don't always have an immediate solution).
Always carries additional emotion when the widgets are coming from you! Glad it is still trustable and useful :)
OpenAI: We’re burning money like the Joker. A miracle needs to happen for us to turn a profit
Microsoft: Please please use our AI systems, we’re teetering on the edge here
Anthropic: I wonder what’ll kill us first: lawsuits, regulations or model collapse
Media and universities: AI is here to stay
(it says "Post not found" above so i can't see it 😭)
Ark VCS has finally reached 1.0! 🎉
youtu.be/XSnIw4siQ6U
get it from ark-vcs.com
Here are the slides for my "Live++: A Bag of Tricks" talk presented at Guerrilla Games last Friday.
Enjoy!
PPTX:
liveplusplus.tech/downloads/Gu...
PDF:
liveplusplus.tech/downloads/Gu...
#cpp
🌞 SOL CESTO – 1.0 OUT NOW ON STEAM 🌞
It took us nearly 3 years, but Sol Cesto is finally complete! We hope you’ll enjoy our weird roguelite and all the surprises we’ve added 👀
Please help us spread the word 🙌
For CJK glyphs since they tends to fit in same rectangular area it could be more efficient in a grid, but it's difficult to predict things and you want a generic system. And you can easily repack later, it's easier once you know your contents.
Kinda want to see a physics simulation here :)
Many possible algorithms for iterative additions. One can always repack & optimize occasionally (easier to pack perfectly if you know your contents ahead).
Fancy offline packing for game sprites can include rotated images too, it can look bizarre.
What I have frequently heard programmers say over the years:
(1996) "Lines of Code" is a terrible metric
(2006) "Lines of Code" is a terrible metric
(2016) "Lines of Code added" is a terrible metric
(2026) I have to use AI because it's the only way to meet my company's "Lines of Code added" quota
Well that project homepage doesn't particularly explain what property based testing can mean.
I'm not sure what's the official lingo for what i'm doing, certainly not unit testing. The system simulates user inputs within a real app/gui, so it's pretty holistic. They frequently fail for reasons others than the one I initially expected them to fail for, but always for a good and useful reason.
"redhibitory" doesn't seem like the right English word for what I meant, if even that's a word because it is currently underlined by my spell checker.
"Prohibitive" would be better suited.
I can never understate how automated tests - not unit tests - using my custom framework have improved my productivity. Tests are focused on tricky code, unobvious behaviors or when combinatorial explosion would make manual testing redhibitory.
(capture is in slowed down mode)
Thanks for your continued support!
Image credits/links:
- Pyrite64: N64 Engine/Editor github.com/HailToDodong...
- Amiberry: Amiga emulator amiberry.com
- Bento Blocks (good puzzles): store.steampowered.com/app/3311670/...
- Intel 8080 Taito/Midway arcade board emu: anthonydelciotto.com
Hi hi, Dear ImGui 1.92.7 is released!
50+ changes (e.g. tables reworked columns freezing with hidden columns, reorder from context menu, webgpu for emscripten 5 & wgvk native, nav keyboard/gamepad to open context menus, & many fixes/improvements)
github.com/ocornut/imgu...
⚡️🇺🇸🇮🇷US: "Dear Iranians, we have come to liberate you from the evil oppressive Regime which is trying to kill you." Two weeks later: US: "Dear Iranians, sorry, Plan A didn't work. Plan B is we bomb you into the Stone Age."
Best April Fool's was a working stb library release:
I have been working on data inheritance recently. The kind of stuff that you usually see with prefabs, but I believe that all the assets should support data inheritance.