A Linux user for over 10 years, I’m very annoyed by the latest developments in distributions, with Pipewire, but also particularly with Wayland. Wayland doesn’t work completely, and despite everything, KDE and Gnome are gradually forcing it on us, and you have to be cunning to fully restore X11… To summarize, I had finally stabilized my MAO installations with Ardour and distributions based on Debian and Ubuntu, but I have to start all over again. I think that in the future I will return to my installations based on Debian 12 because LTS support lasts until 2028. I will stop supporting the development of new distributions. What do you think of Wayland, and should we believe that it’s a real threat to the installation and development of Ardour?.. or to Linux enthusiasts in general. Must we forever return to the console?
(this is a automatic traduction of my french text
hello, utilisateur de linux depuis plus de 10 ans, je suis très agacé des dernières évolutions des distributions, avec pipewire, mais aussi particulièrement avec wayland. Wayland ne marche pas complètement et malgré tout, KDE et gnome progressivement l’impose à marche forcé et il faut ruser pour rétablir pleinement X11… Pour résumer j’avais enfin stabilisé mes installations MAO avec ardour et les distributions basées sur debian et ubuntu, mais il faut tout recommencer. Je pense qu’à l’avenir je vais revenir à mes installations basés sur debian 12 car la prise encharge lts dure jusqu’en 2028.Je vais arréter de soutenir le développement des nouvelles distributions. Que pensez vous de wayland et doit -on croire que c’est une vraie menace pour l’installation et le développement d’ardour ?.. ou pour les amateurs de linux en général. Doit -on revenir eternellement à la console? =
Thanks for the reply.
So do you have an explanation why Linux and environment developers are moving so quickly toward Wayland (which doesn’t seem very resilient). I didn’t understand Debian’s recent move with Debian 13 to make a technological leap forward, which seems out of step with system stability.
I think Paul Davies did a lot to popularize Linux with Jack Audio and Ardour, and I don’t understand why the Linux ecosystem only seems preoccupied with Steam, OBS, GIMP, and LibreOffice (I’m exaggerating and being deliberately provocative).
Do you, as “small” developers, have any contact with distribution developers on this subject?
Obviously Desktop Environments are a different case but at an application level (like Ardour) isn’t this precisely what ‘XWayland’ is supposed to do? Isn’t it’s job to ensure that the thousands of legacy X11 applications continue to work seamlessly with Wayland? If XWayland does it’s job then things like Ardour shouldn’t need to change much at all should they? My real question and confusion is what about the Plugins? Currently on Wayland systems XWayland is handling existing X11 Plugin tookits, what happens if the Plugin toolkits for Linux start using Wayland directly, will things like Ardour still have X11 Plugin support via Xwayland and Wayland on the system will support Wayland drawn Plugins? Or are we going to get into some bullshit like Ardour only supports X11 and Wayland compliant Plugins have to be used in Wayland-native DAW? If it’s the latter then that will probably end most casual interest from Plugin developers looking to support Linux…
What has changed since the last post I referred to above, is that Ardour has meanwhle completely internalized the toolkit and more importantly the window and event-system (gdk).
That currently has abstractions for X11 (Linux, BSD), quartz (macOS), hwnd (Windows), and it would be possible to add Wayland support, though there are currently no plans to do that.
Short of XWayland, another option is just run a complete X server (Xnest) on top of Wayland. The crucial part will indeed be plugins though. Until now there was a unique native Window handle on all platforms… but now Linux may diverge.
Well, maybe the eventual way out will be web-ui, and run everything in the browser
The Wayland project was started in 2008, and the 1.0 specification for the Wayland protocol was published in 2012. The majority of distributions switching to Wayland after 17 years of continuous development does not seem “so quickly” to me.
I believe that Wayland will eventually become mainstream, and find its way into LTS distributions. Whatever layer lives on top of this to enable all the legacy X11 apps is something I’m not expert enough to speculate. But there will be one.
For the foreseeable future, though, Ubuntu Studio 24.04.3 LTS is supported through April of 2027. It offers X11 and Wayland sign-ins, and I wouldn’t think for a nanosecond about trying Wayland. Ardour 8.12 and the LSP suite run very well under X11, on my aging Lenovo desktop with NVidia GPU. I miss a more recent version of Pipewire (currently at 1.0.5 as far as I can tell) but otherwise it does everything I need it to do for my very occasional music production. In the meantime it’s a very good daily driver that’s enabled me to mostly discontinue dependence on Microsoft Windows.
It is not clear to me what the purpose of Xlibre is. I understand the author has ideas about how to improve the code, but as a user, are there any practical differences between the two? What is gained by using Xlibre in place of the corresponding Xorg packages?
As of today probably no noticeable difference even though they have begun making changes and there are a few Distros (Artix Linux is one I’m aware of) that have ISO’s you can try with XLibre instead of Xorg so it is working in principle. I posted in another thread that there is also an unofficial Debian Repo with packages.
As far as I understand the general idea is to revert some of the supposed intentionally harmful commits that were made to upstream Xorg to allegedly hasten it’s demise and from there continue to work on it and improve it without Wayland prejudice and continue it as a proper X11 alternative to Wayland and replacement for the former Xorg. Whether they succeed or not will remain to be seen but I personally think it has a fighting chance if they remain focused on the work and not political posturing. Having read the reportedly offensive comments on the XLibre website in their entirety I found it a bit perplexing how it talks negatively about DEI and then on the other hand offers an open welcome to any and all competent folks who want to help, I think a lot of people read the first part and not the second… but truthfully it would have been better just to say nothing and simple judge any contributors by the quality of their code and leave it at that…
Wayback is also quite an interesting project. It should be able to run a complete X11 desktop inside Wayland. It is in the early stage of development, though.
Maybe a misconception about GIMP. The GIMP devs are a very very small team, and are pretty much on the same boat as Ardour (aside from the plugin UI problem) on views about GTK version incompatibilities, removal of useful widgets, and so on. They didn’t make the switch to GTK3 until very recently and have zero intention to switch to GTK4 because of further incompatibilities and unsuitability for complex applications.
Do you guys have plans to implement XLibre support sometime in the future, or would you treat it the same as X11? I know XLibre is the future of X11, as that’s actively maintained, though in a seemingly alpha or beta state as of right now.
The thing with all those x11 successor or compatibility layers is, that they (should) aim to provide compatibility. This means they should not require any support from an application side. It should just work by responding to the same calls as the ones to x11.
I’ve been working on some Wayland native plug-in (GUI) ports, mainly as a proof of concept, also as a means to test plug-in support in PreSonus Studio One (which requires Wayland, and has its own custom VST3 extension to provide Wayland native plug-in GUIs on Linux). I guess it - can - work, but, the issues mentioned previously in: https://www.kicad.org/blog/2025/06/KiCad-and-Wayland-Support/
Are basically just a subset of the kind of bizarre (and sometime hilarious) things that I’ve encountered along the way. And a lot of these issues seem to be by design, not by mistake, which is all the more worrying.
Can’t wait for the GNU coreutils re-written in rust which are rumoured to be slated for the next Ubuntu release. I guess that’ll be a fun ride - I just hope next time Wayland (or an overzealous compositor) unexpectedly drops me to the command line, that there’s still a command line there
@mike-acmtaudio Hi! I actually name dropped you in this thread because (a) your longtime development of excellent Plugins in various formats and (b) I had read your prior posts here about the err, uhm readiness of Wayland for Plugin development:
I ended up leaving the discussion because I was being misunderstood as ‘anti-progress’ and also for suggesting that Plugin developers should favour LTS and Stable platforms and that a lot of Linux folks aren’t necessarily using Gnome or KDE. My real point was that if XWayland does it’s job and both X11 and Wayland compliant DAWs could host X11 Plugin toolkits then Plugin developers should continue to make Plugins based on X11 for the widest compatibility rather than using Wayland directly. Since you’re one of the few Plugin developers I know and have had a long correspondence with and since you are actually DOING Wayland compliant Plugins which nobody else there seems to be doing I think their final drafted document could use some informed input like yours… Many people there seem to think just because their particular Desktop Environment works OK with Wayland that Plugins should just follow suit…
I personally believe if on top of the ridiculous GLIBC thing in Linux needing both X11 and Wayland Plugin versions is the new state of things we can probably say goodbye to some of the commercial Vendors we already have and we will cease to see as much new porting interest to Linux. I’m more than happy to be educated or proven wrong on any of these points…
Yes, I hope that’s not the case - I guess some of it depends on the support of frameworks (JUCE etc) and as it stands at the moment, there isn’t even an established / official standard for native Wayland GUI support in any of the existing mainstream plug-in formats (as far as I can tell).
I’m not anti-progress either, and much of this would be good - if it worked, but sadly, at the moment there are some deficiencies - that directly degrade the user experience - which simply can’t be fixed at the application (or in this case plug-in) level.
