1. 7

  2. 4

    I totally agree with the author’s point. Whenever one brings up the issue with Weston and Wayland being too “thin” for any real stuff, these smartasses just tell you to write a Weston plugin. However, Weston is horrible to work with and literally a joke. I’m sure it’s still better than X11, however, it’s not worth the price.

    Here’s how it should go: There should be official foundation libraries on top of Wayland. I don’t mind if Wayland is thin, it should stay a thin protocol, however, I don’t want to write an input handling library every time I write a new compositor, and I also don’t want to rely on third-party-libraries to do the job for me.

    1. 1

      When you use X11, do you use only the stock X11 libraries, or are you using a toolkit? As far as I remember, to use a gamepad with X (and not just as a mouse), you will need 3rd party libraries too.

      I am also on the opinion that writing compositors is something the vast majority of people targeting Wayland are simply not interested in. Just like in the case of X11: there are many more apps than window managers & related tools.

      1. 2

        I use the stock X11 libraries exclusively, though I don’t do gaming on my computer so can’t really know much about it. The point is, if you write a game, you most likely use a toolkit like SDL or something, which does the job well for you. For normal GUI-applications, you don’t need gamepad support.

        My stance is: For a healthy ecosystem, it should at least be more straightforward to write a Wayland compositor than an X11 window manager. I am one of the developers of dwm, so maybe my view is a bit different than that of a normal user.

    2. 2

      I’m currently writing a window manager for Wayland using the pretty solid wlc library. I have no idea how many machines this will work on, but wlc has been the starting point for quite a few tiling WMs so given time it may in fact reach the level of portability that X has, with less of the cruft.

      I’m going to write this up once I get it working (just finished basic layout this morning). I can already run it within X but haven’t tried outside it yet. I’m hoping that it will (a) provide a good case study for the viability of Wayland now, and (b) be a WM I’m comfortable hacking deep inside of.

      1. 2

        With windows and MacOS X, you get the whole thing baked in too, and they succeeded. You even have a lot less options than with Wayland.

        And X being easy to develop for? God no. With toolkits, maybe. But you can use the same toolkits for Wayland. The raw X11 libraries are definitely not easy to work with - luckily, noone in their right mind does so, and uses Qt, GTK or similar toolkits. Most of those work with Wayland too, easing the porting effort considerably.

        Granted, when writing a window manager, Wayland may be harder (I don’t know, haven’t tried), but X is definitely not easy either. I have contributed to a number of X11 window managers, watched some of them evolve from concept to daily use - I would not call that development easy. So many weird stuff!

        The article, as far as I understand, boils down to Weston being a PoC, and crap. Well, possibly. There are a number of other compositors though, with the prominent desktops also working on their own. As a GNOME user, I’m fairly sure I’ll get an improved Wayland experience, compared to X11, once I switch over.

        The X11 story is very similar: either you use the Big Desktops with all their bloat and baggage, or hope that your tiny WM and additional tools can be massaged into working the way you want them, that they support multiple heads, and won’t crash when you plug in extra devices, or when the laptop comes back from sleep.