• 0 Posts
  • 80 Comments
Joined 1 year ago
cake
Cake day: November 23rd, 2023

help-circle











  • Ted Ts’o was way out of line in that conference and was clearly channeling his inner ca. 2001 Torvalds.

    I think Rust is a better path forward for a majority of the kernel/driver code maintained currently, but it is definitely going to take time for it to gain a foothold. I also think there is some condescension on both sides that is completely unjustified and needs to stop.

    The hardline C devs that don’t want to learn Rust need to accept that at some point they will have to either adapt or pass the torch, and that no amount of whining or bitching in public forums is going to change that.

    The Rust devs that are getting upset because people are “attacking” their favorite language need to accept that there will be substantial and impassioned resistance to making broad language changes to a set of projects that have existed for decades. It would be an uphill battle for any language to try to supersede C in the kernel; this is not a condemnation or attack on Rust or its zealots, it’s a matter of momentum and greybeard stubbornness.



  • In fairness, “I don’t want to maintain bindings for a language I never intend to use” is a perfectly reasonable position.

    The typical answer here is for the language evangelist to implement and maintain the bindings, and accept the responsibility of keeping them in sync with the upstream (or understand that they will be broken for however long it takes for another community member to update them).


  • This is not always true. Some tablets are extended release and if you break them apart the timing is thrown off. You get a higher dose initially, and the dose doesn’t last the intended period.

    A family friend learned this the hard way when they were breaking a seizure preventive tablet in half to make it easier to swallow; they’d often have a recurrent seizure about an hour or two before their next dose time.






  • Maximum length is the biggest red flag to me and was the catalyst for me making the effort to switch to unique passwords per-account years ago. There’s just so, so many shitty homerolled security systems out there… and data breaches seem to be a perennial problem these days.

    There’s just no excuse for limiting the length if you’re doing security correctly (other than perhaps a large upper limit just to protect against someone DOSing the backend with a bunch of 100MB strings; 512 characters seems reasonable).

    By setting an upper limit, you’re basically saying one or more of these things:

    • We store your password in plaintext
    • We store a hash but our hashing function has an unnecessarily arbitrarily limited input size
    • The person/team implementing the backend has no idea what they’re doing and/or just copy pasted login code from stack overflow
    • We tried to get away with minimal password requirements but some middle manager wouldn’t rubber stamp it without arbitrary_list_of_bs