Announcement

Collapse
No announcement yet.

DRM Updates Land For Linux 4.15; Torvalds: "There's Something Odd About DRM People"

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #21
    Originally posted by FireBurn View Post
    say he wants a timeline for the todo list
    do timelines for the todo lists in kernel exist outside of your imagination?

    Comment


    • #22
      Originally posted by coder111 View Post
      How did that .h file get this far up the maintainer chain without anyone noticing it is unused?
      it is documentation. how exactly do you want to use kernel docs?

      Comment


      • #23
        Originally posted by cl333r View Post
        Header files causing issues? Another reason to write an OS in Rust.
        start right now, i'm waiting

        Comment


        • #24
          Originally posted by agd5f View Post
          Linus is just whining to for the sake of whining. It absolutely makes sense to use machine generated register headers. They eliminate human error and enable more people to get involved. If there is a feature someone wants to implement, all of the register information is there and correct. It also makes debugging a lot easier since you have ready access to all of the additional registers you might need to look at or ask someone to dump. Adding it manually on a per feature basis is error prone and makes it harder for someone working with the code to see the full picture and capabilities.
          IIUC Linus is not against auto-generated code, he'd simply prefer having the original 'source' file and the tools used to generate that code, along with source code for said tools.

          Comment


          • #25
            Originally posted by pal666 View Post
            These 4 posts in a row are today's reminder that I rarely make a point without being an asshole.
            Tip: Sometimes, running the asshole metaprogram is the most effective and/or humorous way to make a point. Constant overuse can cause people to not take you seriously and tune you out, even if you have a good point. You're welcome.

            Comment


            • #26
              Originally posted by cl333r View Post

              But Rust is already mature, Servo is just one example. Since you're a Rust dev you could tweak Servo and integrate it into Linux to make it as fast as Servo, no?
              "Mature" isn't a destination, it's a direction. The Linux kernel's requirements are further down the road than Servo's.

              Comment


              • #27
                Originally posted by srakitnican View Post
                I've understood cl333r as a troll post and laugh about it. I see it got some people
                Actually, you're right. Now he's talking about building a rendering engine into the kernel.

                Comment


                • #28
                  Originally posted by pal666 View Post
                  do timelines for the todo lists in kernel exist outside of your imagination?
                  If Linus wants one yes, if not it was just a theory

                  Comment


                  • #29
                    Originally posted by DanL View Post
                    I think you're over-dramatizing. This is barely a 3.0 on the Linus Richter scale. He's not going to go off about such a minor issue, especially when the dev in question quickly fixed the issue and apologized: https://www.spinics.net/lists/dri-devel/msg157463.html

                    The tinydrm thing notwithstanding, it seems that DRM maintainers have been doing a better job of respecting merge window, meeting deadlines, and not sending in code inappropriately complex for a linux-rc version.
                    You probably need to re-adjust that scale considering how Linus has admitted that his rants have gone to far and said that he's going to try to avoid saying things like how he thinks someone is so stupid it's amazing they haven't gotten themselves decapitated by a door. The profanity laden rant directed at that security guy who broke userspace and refused to lift a finger to fix it, instead insisting userspace devs should just try to avoid having their stuff broken by his mistake, is probably about the top of the scale now.

                    The whole point of the chain-of-trust model that Linux uses is that you trust maintainers like David to catch things like this and not pass them on to Linus. It's not about senior maintainers like David making mistakes like this, it's senior maintainers catching them and rectifying them rather than passing them on to Linus forcing him to rectify them instead. With this sort of development model the logical thing to do is shunt David down the ladder a step or two and have him replaced with someone who can catch the mistakes and bad code and not pass them on to Linus.

                    Comment


                    • #30
                      Originally posted by cl333r View Post

                      Since you're a Rust dev you could tweak Servo and integrate it into Linux to make it as fast as Servo, no?
                      I see that you are trolling (browser renderer and OS kernel have almost nothing in common in terms of code they could share), but still

                      Originally posted by cl333r View Post

                      But Rust is already mature
                      for anyone reading this forum who might actually be interested:

                      No, it isn't. There are quite a few things in the language that need to be fixed/improved. One of the biggest ones is the fact that, since const generics are not implemented yet, the language only properly supports arrays up to 32 elements in size. This is an ugly hardcoded temporary hack, which will be replaced by const generics eventually (when they are ready), to allow arrays of any size. I wouldn't call a language with crap like this "mature". Some other important parts of the language are still only in the nightly compiler: `impl Trait` being the most famous one. Rust is really not achieving its full potential/glory as a language yet, without these things.

                      Give it another year or so, for things like NLL, const generics, impl Trait, specialization, const fn, etc. (the list goes on) to be implemented and stable. Then give it another year or so for the library ecosystem to mature and make good use of the various new language features. Then we can talk about Rust being at its full glory.

                      Yes, Servo exists. Servo is Mozilla's pet project that they literally developed Rust for. Servo and the Rust compiler are the biggest testbeds for Rust as a language, to see how well it actually works for large codebases. They don't exactly count IMO. Both of them will also get much better once they can start using the new language features. Notably, the Rust compiler has already accumulated quite a lot of legacy crappy code based on outdated coding practices, due to the language changing so much over time. A lot of it would not be considered "idiomatic Rust" by today's standards. This suggests that the language is still changing too much (i.e is not "mature").

                      tl;dr: A language with this much feature churn and regular changes + ugly temporary hacks due to still-missing language features cannot be called "mature", despite some large software being written in it already being in production and despite it being backwards-compatible with Rust from 2 years ago (Rust 1.0).
                      Last edited by tajjada; 17 November 2017, 01:20 PM.

                      Comment

                      Working...
                      X