Xorg will be the default in 18.04 LTS

The architecture of GNOME Shell and Mutter is such that a GNOME Shell crash will end your whole session, killing running applications and returning you to the login screen.

Yes it can do so. :face_with_raised_eyebrow:

This temporary bug with Gufw which will not work on wayland will now be solved without using;

xhost +si:localuser:root

which also is a security risk. I agree with Will, out of the box users need stability and, may I add , some type of visibly working security interface, Gufw, which has been a staple for noobs and out of the boxers for some time.

1 Like

Doesn’t the entire Xorg Window System runs as root though, Dale?

Please consider this stack overflow question https://stackoverflow.com/questions/45465016/how-do-i-get-the-active-window-on-gnome-wayland (and the linked bug in the GNOME bug tracker, https://bugzilla.gnome.org/show_bug.cgi?id=788511) to be a blocker.

I realize that libraries like WNCK may need some kind of privilege escalation or user prompt to work on Wayland which more correctly and thoroughly isolates clients, but eliminating all inter-client communication poses tremendous automation and accessibility problems. Ubuntu already lags badly behind other platforms (COM on Windows, ScriptingBridge on macOS) in terms of end-user automation, and eliminating even the ability to programmatically move windows around is another major step backward.

No.

I have read a lot information about xorg’s security problems and it was one of the reasons why a lot of the community wanted to spur ahead with the more security hardened wayland but , I think for one, that a lot of the security problems with X were somewhat overly dramatized. I have been in the development testing stack for close to 8 years and I know there are serious bugs but I have never seen an intrusion on a development cycle or LTS using X with the exception of javascripts flips.

But the point is , that running xhost as root defeats the purpose of having wayland.

I mean, for the interim, why can’t we enhance (xorg) what we have for now. Perhaps the workaround is a simple one . That will give wayland devs time to come up for air during 18.10.

edit:

xorg not running as root on my machines.

1 Like

I think that without Xorg, a live iso cannot be booted. I think gparted won’t work with Wayland for example.

As an early wayland/weston adopter we tested it vigorously . It had promise but it didn’t work largely in the field. They have come a long way. Without an out-of-the-box graphical firewall for regular customers and new converts it renders wayland a critical security risk.

https://bugs.launchpad.net/ubuntu/+source/gui-ufw/+bug/1713238

Great decision, Wayland needs more work and more time, LTS needs battle tested and proven.

1 Like

Xorg sessions also have the benefit of better performance and/or more workarounds are available for some issues. For example, I don’t think these Wayland stuttering bugs exist in Xorg sessions:

Also related to smoothness, if you encounter problems with touchpad performance in Xorg you have the option of installing xserver-xorg-input-synaptics and then tweaking things with synclient. Libinput (the only option in Wayland) provides no such workaround.

We are working with upstreams to resolve these issues and more, so that Wayland becomes a more attractive option in future.

Ubuntu 18.04 will be Xorg default?
Not bad. But is there will be an option for wayland?

I heard that fractional scaling support is based on wayland, So Ubuntu 18.04 won’t have fractional scaling feature?

So sad for my WQHD laptop.

Yes, Wayland will be an option you can still choose on the login screen. Same as Xorg is right now.

1 Like

Where is the code of this shell port? Why you think will be preferable swap directly to Gtk4 instead of continues using St? It’s not a lot of work and ambitions to be done at once? It’s not preferable just fixed the wrong gnome-shell architecture and then introduced Gtk4 as an addition? The Gnome-Shell low latency it’s really the problem. Gtk4 it’s just probably better than Gtk3, but Gtk3 or St are not really the shell problem. St just is an incomplete library without future, because St is just a pretty close subset of the abandoned mx library (https://github.com/clutter-project/mx). Do you really mean then that you want a Clutter replace by the Gtk4 toolkit as the shell GUI in all cases, because St is just a little wrapper of Clutter to add support for CSS. I think the good idea will be remove St in favour of MX and just add support for Gtk4, but not as a replace of anything… Mx need to be a shared and independent toolkit library that internally is supported by an independent and shared clutter library. Currently Clutter is not independent, as you can not be able to create a fully window just with Clutter and this is wrong, because not one will make code for a library that depend of other library (gtk), to create a window. Also St is a not shared and is an incomplete library. So, a fully and Shared library need to replace St without add dependencies to Gtk. That will improved all at one.

Nobody is talking about replacing St with GTK-4 in this topic. At least not directly. You can safely ignore any mention of some future Gnome Shell 4.0. Ubuntu will continue to ship what exists – Gnome 3.x.

1 Like

@vanvugt Yes, just i see the gnome-shell 4 proposition and i agree that something need to be done there. I response to a very specific comments related with that and not with the thread.

For this thread in question, is more than clear to me, that this is the right decision. Wayland is like a baby, it will need to grow up first, to then be taken seriously. Specifically, what is more annoying to me is the impossibility to get the the list of windows and the active window on wayland (https://bugzilla.gnome.org/show_bug.cgi?id=788511). This missing stuff mean a lot of applications without a way to work on wayland, not matter if this can be justify with more security, is a big sacrifice for security.

That’s more or less the problem. This baby’s not going to grow up enough to be really useful. People, who had created working apps won’t waste their time recreating those apps to work with Wayland. Why should they?

1 Like

Yes the idea is also of replacing St with GTK4, which indeed will allow to get better results than the in-via-of-deprecation St toolkit.

But that might came also at a later point, as it might imply breaking all the extensions we have so far.

1 Like

I think you have a radical position. Why people should create an app that work on linux? For the same reason, they should create an app that will work on Wayland. But yes, Wayland need to prove first that it will be really better than X11 and this is not what currently happens… For the another hand (and probably the most important), Wayland should close the space between Linux and Android apps (Wayland is desingned to be a cross desktop display manager). For the big marked share of Android, this is currently a good thing for Linux and also for Android as will probably received updates and patches that come from Linux in general. For example: Samsung have an initiative to start Linux in his Samsung Dex as a replace of Android. Wayland will help a lot to this and that could also be good for Linux in general. Finally, I need to trust that everything that tries to standardize software’s it’s a good initiative and it deserves then our work to support it, because this is a door opened from where could come in the future others developers to contribute with our ecosystem. Tell me dreamer if you want, because that is what Wayland is… A dream for the future.

1 Like

This is exactly the point of why my comment.

It is Xorg plus a WM to get things going in any DE. Wayland (sort of) is supposed to do 2 jobs as one. But, Wayland for Gnome won’t work for KDE, or any other DE, whereas Xorg would work with any WM and DE, and with only a WM.

Today, I’d install Gnome and KDE (or Xfce or Mate) in the same partition and log out of one DE and log in to the other. But, would I be able to do that with Gnome on Wayland and KDE Plasma on Wayland?