The fail case might have tried to truncate or fallibly allocate extra (and truncate or abort if allocation failed due to limits).

Any of these would be better than hard stop error.

It could have been turned into an unoptimized fail case and a warning, so the issue turns into degraded performance and not global outage.

oh *brother*

woof

So I think in horror games, losing should actually be reserved for the very end of the game, and instead build tension throughout. That is how a game can be made scary. If that is the purpose, of course...

Having recently dealt with the Stalker entity in Receiver 2 and having by heart PUMPING HARD due to fear, the tension is all what makes something scary.

Knowing how you lose or win in the game just releases you from that. Mystique lost, tension released.

And one of the things I've started noticing in indie horror games is that they often have some kind of monster, a threat, a jumpscare, loud noise, etc. in their losing sequence.
Those are the ways to provide a release from tension. Maybe that's why people feel like horror isn't scary anymore?

I did not intend to have a pun in there

My first guess was LLMs and the second/backup one was Cloudflare

Print an ASCII art of a middle finger?

Real customer feedback? Bullshit.

I've seen creative channels that I watched for years become shittier and more boring when they started using AI for scripts and some visuals.

And now we are here. The AI is the new shiny buzzword, and everything is just shitty because of it. Working features deprecated, replaced with AI.

Then there was a moment with image generation models, particularly when midjourney came out. That's when the shit hit the fan and the AI-generated images started getting popular. I still generated some, but it got boring real quick.

Oh, absolutely, except() is better. I like using it when I am certain something won't happen and explain the reason in the message. Turns the code comment into just code.

Link please?

Same for option, maybe have that info in the None variant itself.

So, a monad... I think.

My main pain point with rust's errors is tracing back the original source and the bubbling up. Otherwise it's nice that it's not just some special arcane thing.

#rustlang

From what I have seen in the rust docs, there either was a traceback at some point or an attempt to add it. It doesn't work because the Result is just an enum with an error variant, and that type is arbitrary

This could be fixed with a wrapped error type, where the wrapper handles the tracing.