Announcement

Collapse
No announcement yet.

Mir 1.0 Expected To Come Early In Ubuntu 17.10 Cycle

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

  • #21
    Originally posted by bregma View Post
    It's entirely feasible. Is there a use case? Do you know of any applications written directly for the Wayland client API as opposed to using a toolkit like GTK+, Qt, libSDL, etc.?
    Toolkits can never cover all possible use-cases. When the toolkit doesn't provide the features necessary, you need to talk to X11 or Wayland. Lots of applications need to do this.

    Looking at my own distro, openSUSE Tumbleweed, here are some the applications that directly require libwayland: gammaray, ibus, mpv, and vlc-beta, as well as parts of the gstreamer library. That is despite the fact that openSUSE only has limited Wayland support right now. You can expect that most programs that currently depend on X11 directly will also need to depend on Wayland directly. There are hundreds of such applications in openSUSE right now.

    Comment


    • #22
      Originally posted by bregma View Post

      It's entirely feasible. Is there a use case? Do you know of any applications written directly for the Wayland client API as opposed to using a toolkit like GTK+, Qt, libSDL, etc.?
      At least in the Gtk2 days, Firefox reached to X11 behind the Gtk abstraction. I haven't actually checked if everything is now mediated by Gtk3 without Firefox knowing anything about X11 or Wayland, but I had assumed that complex apps that of have toolkits on their own despite using Gtk (Firefox, Chromium, LibreOffice) would do some interaction directly with the window system protocol. Is that no longer the case?

      Comment


      • #23
        Originally posted by Sethox View Post
        In all honesty though, competition is good... if Mir ever get to compete (get released in time). I mean the ideal scenario is that Mir could improve Wayland (on certain aspects) and vice versa. Just being positive here.
        I think that is already evident in the progression of Wayland development. Before Wayland developers had competition to rage against, their development progress was fairly stagnant. Even with competition, we are talking about an already accumulated 7 to 9 years of active development to get it to the beta quality state that it is in now. In terms of what Wayland and Mir are able to offer to the end user, it looks like, at this point, it may be better security? There are theoretical performance wins, but it may take another 5 to 7 years to see them. And now that I am trolling, I think Windows 10 has a better shot of increasing Linux adaption than the effect Wayland and Mir will have in the next 5 years.

        Comment


        • #24
          Originally posted by bpetty View Post
          I think that is already evident in the progression of Wayland development. Before Wayland developers had competition to rage against, their development progress was fairly stagnant.
          This is a common claim by Mir supporters, but I have seen absolutely nothing to back this up. openhub doesn't show any increase in development after the Mir announcement, in fact the rate of patches dropped slightly around that time.

          Originally posted by bpetty View Post
          Even with competition, we are talking about an already accumulated 7 to 9 years of active development to get it to the beta quality state that it is in now.
          Wayland isn't "beta quality", it has been stable for years now. It is the compositors for the particular DEs that have taken a few years to get into shape, and they could only start seriously after the Wayland API was declared stable in 2012.

          Originally posted by bpetty View Post
          In terms of what Wayland and Mir are able to offer to the end user, it looks like, at this point, it may be better security?
          And less tearing and flicker, better handling and match-up of windows and window contents, better multi-monitor handling, better multi-video-card handling, better handling of video and games, simpler and more straightforward API leading to less buggy rendering, easier support for newer extensions such as improved input handling, and a single frozen application doesn't freeze your entire desktop.

          Originally posted by bpetty View Post
          There are theoretical performance wins, but it may take another 5 to 7 years to see them.
          And you base this conclusion on..what, exactly?

          Comment


          • #25
            Originally posted by TheBlackCat View Post
            This is a common claim by Mir supporters, but I have seen absolutely nothing to back this up. openhub doesn't show any increase in development after the Mir announcement, in fact the rate of patches dropped slightly around that time.

            Wayland isn't "beta quality", it has been stable for years now. It is the compositors for the particular DEs that have taken a few years to get into shape, and they could only start seriously after the Wayland API was declared stable in 2012.
            When I, and most others too lazy to be technically correct, talk about Wayland, it includes the compositors. It is a shorthand, because a specification means absolutely nothing. That said, I would bet that if you looked at the community at large, specifically the various compositors, you will see an uptick. That doesn't mean it was Mir driven, however. As far a beta quality... for the desktop I would say it is, I am not saying you can't release a beta quality product. In fact, if they could get it to a point where people could run it stably, you'd get free quality testing.

            Originally posted by TheBlackCat View Post
            And less tearing and flicker, better handling and match-up of windows and window contents, better multi-monitor handling, better multi-video-card handling, better handling of video and games, simpler and more straightforward API leading to less buggy rendering, easier support for newer extensions such as improved input handling, and a single frozen application doesn't freeze your entire desktop.

            And you base this conclusion on..what, exactly?
            It is based on an educated guess. How long do you think it will take before the above list of features is realized on the desktop? Canonical always says one more year... has been for the last 4. When someone finally releases a X replacement worth using, I can guarantee you most things on that wish list wont be fulfilled out of the box. There are too many moving parts for it to all come together.

            Comment


            • #26
              Originally posted by bpetty View Post
              That said, I would bet that if you looked at the community at large, specifically the various compositors, you will see an uptick. That doesn't mean it was Mir driven, however.
              There was no uptick in kwin, mutter, weston, or enlightenment after the Mir announcement. The only change is a decrease in Wayland development, which probably has more to do with the protocol being stable than Mir.

              Originally posted by bpetty View Post
              As far a beta quality... for the desktop I would say it is, I am not saying you can't release a beta quality product. In fact, if they could get it to a point where people could run it stably, you'd get free quality testing.
              Whatever arbitrary definition of "Wayland" you use, if Mir has had any effect at all it has been to slow down Wayland development, not help it. This is backed by the number of contributions to the code base after Mir was announced, by the fact that Wayland development would have had more contributions if the resources Canonical were currently dedicating to Mir were dedicated to Wayland, and by the fact that developers supporting Mir have had to divide their work between the two projects. Canonical's contributions to things like libinput would have been needed under Wayland as well, and there is no indication that any features of Mir have been useful to the design of Wayland. So at best Mir has had no impact at all, but much more likely is that Mir has had a negative impact.

              That isn't surprising, since Mir isn't really a competitor for Wayland except in terms of manpower. None of the distros or DEs targeting Wayland can even us Mir, Mir development lags years behind Wayland development, and Mir offers no new or interesting ideas, so it doesn't really provide any motivation for anyone involved in Wayland to change their behavior.

              Originally posted by bpetty View Post
              It is based on an educated guess. How long do you think it will take before the above list of features is realized on the desktop?
              For Wayland, most of them you get for free, or are requirements for basic usage (multi-monitor support). The only one we may have to wait for good support for is multiple video cards, but it would be easy to provide early on if anyone wanted (but that is more up to driver developers than anyone else). Wayland has a bunch of other advantages that will take longer to realize, but I intentionally didn't list those because that wasn't what you asked for. I only listed the advantages you get for free or are necessary.

              Originally posted by bpetty View Post
              Canonical always says one more year... has been for the last 4.
              Not surprising considering that the Mir developers have a remarkable lack of expertise in making display managers. That has nothing to do with Wayland, both core Wayland developers and window manager developers have been pretty realistic about the difficulties and timeframes involved.

              Originally posted by bpetty View Post
              When someone finally releases a X replacement worth using, I can guarantee you most things on that wish list wont be fulfilled out of the box. There are too many moving parts for it to all come together.
              Which ones, and why? As I said, I didn't list all advantages of Wayland, only those advantages that you get for free or are necessary. So what specific reasons do you have that those specific examples I picked are either unnecessary or don't happen for free?

              Comment


              • #27
                Originally posted by garegin View Post
                We actually do understand. Canocial was scared that Wayland wasn't going anywhere so they started to create Mir, then Wayland 0.85 came out and they decided to save face and keep course even though there obviously was no reason.
                If Canonical was actually scared about Wayland's development speed, they could've allocated people to work on it. They never did.

                Comment

                Working...
                X