Announcement

Collapse
No announcement yet.

X.Org Server To See More API/ABI Breakage

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

  • #11
    Originally posted by johnc View Post
    Hmm... yet another phoronix article with the obligatory negative slant on everything.

    I thought beer was supposed to cheer people up and add optimism?
    Two big pluses to your "negative slant" remark. He's been doing that even more often of late and I hope the only reason is to drive people to comment.
    As for beer (or any alcohol), they are cns depressants, technically

    Comment


    • #12
      Originally posted by johnc View Post
      Hmm... yet another phoronix article with the obligatory negative slant on everything.

      I thought beer was supposed to cheer people up and add optimism?

      Binge drinking is defined as episodic excessive drinking.

      Males who drink more than 35 units of alcohol per week report being physically hurt as a result of alcohol, and 15 percent report physically hurting others as a result of their drinking. Almost 16 percent of binge drinkers report being taken advantage of sexually, and 8 percent report taking advantage of another person sexually as a result of alcohol within a 1 year period. Heavy drinkers cause approximately 183,000 rapes and sexual assaults, 197,000 robberies, 661,000 aggravated assaults, and 1.7 million simple assaults each year. [19] Binge drinking has been associated with high odds of divorce, spousal abuse, and poor job performance.[20] Binge drinking can cause adverse effects on the body including effects on blood homeostasis and its circadian variation, cardiac rhythm, ischaemic heart disease, blood pressure, white blood cell activity, female reproductive hormone levels as well as adverse effects on the fetus. There is also evidence from animal studies that binge drinking causes brain damage.

      http://en.wikipedia.org/wiki/Binge_drinking

      Comment


      • #13
        Originally posted by nadro View Post
        X.Org API changes too fast in my opinion. Changes should be more conservative I think (like in a MS OS'es). In current situation NV/AMD should provide some open source module, which will be a bridge between X.Org and a driver API. It will be allow community to add a support for a new X.Org releases on own hand. Only really big changes in a X.Org, will require re-design of a "bridge" module by a NV/AMD.
        Intel have no problems with the allegedly fast changing API.
        Open source AMD and Nvidia drivers doesn't seem to have any problems with it either.

        Its just AMD and Nvdia who seem to have a problem with it with their proprietary drivers.
        Nvidia is usually relatively out with a new driver, but AMD is slow.

        We need to continue to evolve X.org to make it cutting-edge and competitive.
        We needed to add kernel mode setting, a in-kernel memory manager, XRandR etc.
        We need OpenGL 4.2, improved video acceleration, etc.
        We still need work on multi-GPU (aka SLI / CrossFire), and hybrid graphics (aka Optimus / PowerXpress).
        Maybe still need work on multi-monitor, I don't know.

        Comment


        • #14
          Originally posted by AnonymousCoward View Post
          There is also evidence from animal studies that binge drinking causes brain damage.
          ahaaaaaaa!

          Comment


          • #15
            @uid313
            Yep, I understand it, but we must remember that closed source drivers have long verification process (mainly in AMD) and every changes in X.Org cause, that users have to wait for a working driver a long time. Second problem is a drop support for a legacy hardware. "Bridge" module will solve these two problems, but I think that AMD/NV isn't interesting of similar solution for it's own drivers :/ Now every, even small change in a X.Org header files cause a problems for an users...

            Comment


            • #16
              Well, in fairness, most people don't run out and grab the new X.org server release as soon as they're available (which is part of why the entire article is based on FUD).

              By the time the new version gets picked up by distros, the proprietary drivers could be updated. Well, at least nvidia's will be.

              Comment


              • #17
                The guys at Canonical swarm the AMD headquarters a few weeks before each Ubuntu release and hold the devs at daggerpoint for an ABI-updated fglrx.

                Comment


                • #18
                  Originally posted by johnc View Post
                  By the time the new version gets picked up by distros, the proprietary drivers could be updated. Well, at least nvidia's will be.
                  I know from my own experiences, it is rare to see nvidia not be somewhat on time with new X.org versions ... I run Archlinux which is bleeding-edge and thus tends to have these kinds of updates show up quickly (before many other distros) and i can't even remember the last time i wasn't able to update X.org because of Nvidia ... I have had to downgrade X.org a few times for other reasons, like bugs - but not because of the nvidia driver... it's always been other things, like last year when X had issues with my tablet on an update...

                  mind you, i do tend to stick with the nvidia-beta-drivers, unless i upgrade and it is unstable, so i am sure that makes it even less likely an X.org update wouldn't be compatible.

                  Comment


                  • #19
                    Originally posted by nadro View Post
                    X.Org API changes too fast in my opinion. Changes should be more conservative I think (like in a MS OS'es). In current situation NV/AMD should provide some open source module, which will be a bridge between X.Org and a driver API. It will be allow community to add a support for a new X.Org releases on own hand. Only really big changes in a X.Org, will require re-design of a "bridge" module by a NV/AMD.
                    Seems like a lot of work for very little payoff. Distro's should be conservative, upstream should have a bit more leeway to improve their code. Also seems like a layer of unnecessary abstraction (A driver used to drive drivers.) as well as a duplication of effort. I believe this is already how nVidia drivers are internally, seeing as they share 95% of the code between the X and windows implementations. Also such a layer may reveal things about the internal API that nVidia doesn't wish to make public.

                    Comment


                    • #20
                      God, here we go again.

                      If there is no commits/progress/changes on X people whine that it's old/legacy/crufty X and wayland will have to come to rescue us and save the day.
                      And if there are changes to clean up interfaces or prepare for interesting new features the same idiots complain about "more API/ABI breakage".

                      Comment

                      Working...
                      X