[–] 1 point 3 months ago

Buddy, there were trivial rust crates that both predated async rust, and are almost as old as Zig, that mostly if not fully provided what you're describing and more.

One non-notable example I used a very long time ago is pipeliner, the current implementation of which simply uses crossbeam channels behind the scenes.

That crate provides an extension trait to Iterator, and its .map() is unordered (there is an ordered option too), which means .map().next() returns on first completion, out-of-the-box, working with any iterator.


I feel like it's not rare for matklad's blogs to be weird like that. Like that time he compared synchronization methods, and concluded that rust's std Mutex was bad because another implementation was faster under high contention, not realizing that the other implementation used a back-off strategy in its locking which you could easily do with std Mutex::try_lock().

Who knows, maybe Andrew will call him a noob too one day πŸ™‚.

  • source
  • Banging interview. The interviewer is fine, he just asks some fairly simple questions, but Andrew has so many good things to say it works. It's so refreshing to see an actual human being in tech for once. goated with the sauce no less

  • source
  • [–] 4 points 11 months ago

    I suspect this will happen more and more as large corporations try to shoehorn AI into literally every corner. Atlassian is equally guilty of this.

    It's like Dead Internet Theory, but for code repos.

  • source
  • [–] 2 points 11 months ago

    I’ve not had the time to do a proper sit-down with Zig, despite all the cool things I’ve heard about it; but at least it seems that the team’s got their heads in the right place. Hell yeah.

  • source