FSMaxB @FSMaxB@mastodon.cloud

@alva mostly nil dereference I guess?

@alva It's sad that there's not more stores that sell consumables without packaging.

@alva that's fair!

@alva but what you were saying reminds me of the plastic bags in 7/11 stores in thailand. In the bags it says something about saving the environment because of recycling and them using thinner bags or such. But: Why do you even put EVERYTHING in those bags in the first place then? (And for heavy stuff, like big bottles, they sometimes even give you to bags)

@alva it is relevant for the environment and it shows you that even if you avoid eating animals, you still fall victim to how food is sold in stores.

It does not invalidate your point, just adds to it.

@alva but then again most vegetarian and vegan food is still wrapped in plastic ...

@alva "freilegen" 🤔

Just found some interesting code in jemalloc's implementation of a mutex:

if (ncpus == 1) {
goto label_spin_done;
}

https://github.com/jemalloc/jemalloc/blob/1df9dd35154ca460facbd74f779a13dcece78dac/src/mutex.c#L51-L53

Microsoft just open sourced their STL implementation! https://devblogs.microsoft.com/cppblog/open-sourcing-msvcs-stl/

@alva Compare and Swap in a loop would have the same result I guess. Although with quite a difference of what happens behind the covers.

@alva Sounds like "Living on my own" by Freddie Mercury.

But it's definitely strange that this happens in language. Essentially increasing the "no" by saying it twice instead of negating it like boolean algebra.

"Smooth GNOME experience in a wie variety of devices" https://www.youtube.com/watch?v=qAjPRr5SGoY&feature=youtu.be

@alva 🤔 seems you are correct. I definitely have code without including the New header anywhere but using placement new. But I think the header is probably included transitively via some other standard header.

Every day you learb something new.

@alva I'm 99% certain this is not the case.

@alva Yeah Rust is a tool that you can use to make things fast, but it's up to the programmer to use it that way.

But to be fair, if you solve the same problem both in idiomatic Rust and Python without particularly optimising anything, it is not unlikely that the Rust code runs faster and with less memory. (but also that writing it took longer)

@alva And writing a C++ API for a C++ Library is utterly impossible. There are so many choices to make and all of them are wrong.

@alva I'm struggling with similar thoughts. It's not that I want to grow it forever, but I definitely prefer it long, but this heat is seriously making me consider cutting it right off.

@alva I think some subsystems have test-suites (some of the file systems and graphics subsystems I think). There are also some independent building and fuzzing efforts.

But in the end we users are often the first line of testing (especially early adopters).

@alva First look at the jslib implementation of the crypto looks alright.

@alva This article is missing some crucial information: https://help.bitwarden.com/article/what-encryption-is-used/
Based on what they are writing the security could be either quite ok or utterly broken. I think I need to look at the code, I didn't find a whitepaper.
What is suboptimal though is the choice of key derivation function. PBKDF2 is not memory hard, so it can be bruteforced efficiently on FPGAs and ASICS. (Or GPUs)