> GNOME’s Wayland session is now stable and usable with Orca.
They did that via an out-of-band D-Bus protocol, rather than going full Wayland. I'm all for keeping D-Bus around for backwards-compatibility (lots of things use AT-SPI2, and I certainly don't want a repeat of the CORBA -> D-Bus migration), but Wayland's AT support should be first-class, not relegated to proprietary GNOME extensions.
> Assistive technologies or other accessibility clients currently connect to the compositor through a D-Bus protocol, defined in the Mutter repository linked above. By exposing this interface via D-Bus rather than Wayland, we make it easy to withhold this communication channel from sandboxed applications, which shouldn’t have this level of access.
Excellent excellent call out. This feels very wrong direction, this using a side-band for accessibility!!! I had no idea.
It's surprising to me that security would prompt the push to D-Bus. Wayland's design with the compsitor as the message bus center was built to center security in the architecture, to make the compositor the arbiter of data flows. I'm struggling to picture what the issue was here with sandboxed applications.
Compositors are incompatible with one another and have totally different levels of functionality. Most compositors don't do any proper kind of security and permissions and just limit data flows in the most primitive and restrictive way possible. Using something with existing security but the possibility of cross-communication like DBUS is just easier here. Although I expect a lot of confused deputy type of exploits there...
> Compositors are incompatible with one another and have totally different levels of functionality.
I see far more similar about compositors than different. They have freedom of implementation & do different things, but they offer to apps a very common set of protocols, and some of the newer edgier protocols have more varied adoption. But I see standards & growth & development forward as 'what Wayland is', not "incompatibility" (that's why Wayland proper is purely 100% protocols & not a bespoke impl) and "totally different" (???).
Finding formal concensus to make special stable has been hard. But there is a pleasantly large amount of overlap and interop, even if compositors end up implementing more than one flavor of say gamma control.
> Most compositors don't do any proper kind of security and permissions and just limit data flows in the most primitive and restrictive way possible.
When starting a composite yeah, you figure out who has input and send it there. You render the surface the apps give you. Dataflow is inherently quite unidirectional for a long time when beginning a desktop compositor.
But that doesn't feel true at all today. Most compositors support advanced screen sharing, virtual pointers, and other fairly advanced data-flow protocols, almost all built security in mind, requiring more advanced routing & permissioning, as is very visible with the screen share flows of most apps.
> Using something with existing security but the possibility of cross-communication like DBUS is just easier here.
Agreed, but you yourself in your very negative comment on Wayland in this topic seem quite opposed to second systems in broad basis (and pretty radically down on Wayland in general). I'd like to see the display system protocols able to better tackle talking about and managing what's on the screen, rather than building a second parallel system for saying and working with what's on the screen.
I also think there's a decently strong basis already there in the current protocols. There's a security context to start. https://wayland.app/protocols/security-context-v1 We see more advanced flows like XDG Foreign for apps to expose surfaces and mix them onto our own apps.
At-spi d-bus was already right there & done, and it felt easy to keep using it. It was mostly convenience I feel like, and the justification of security feels like a blunt hammer to deny exploration. https://gitlab.gnome.org/GNOME/at-spi2-core/
Maybe it is actually better to do these accessibility operations over D-bus than over a Wayland protocol. Just because something existed in Xorg does not make it the best solution.
I like that this article isn't just a bunch of GNOME hate. But I freely admit to being frustrated that this switch is being forced when they're still developing basic accessibility functionality for GNOME. Once I switch to Wayland, I have to find new ways of:
- Adding hotkeys
- Replacing all my utilities that involve screenshot + OCR
- Making sure Orca works to some extent (low expectations here, but I need some of it to work)
- Reversing the gamma ramp
- Screen magnification
I have mapped out some approaches to most of these, but I full anticipate one or more to be show-stoppers or items I have to attempt to implement from scratch. I'm looking at 100 hours of work, minimum, to make this switch, and am basically just left out in the cold to figure all these workflows out under more hostile circumstances. It's also a chicken and egg, since I need to bootstrap into the new environment without any of my old tools working.
It's great that GNOME is taking some of this seriously, but they're forcing a very difficult transition, and it's frustrating that they think accessibility is in a usable state, or that we're low prioritiy enough to not matter. I'vbe heard the word "edge case" used a lot, it really stinks to be in a category where your entire computer use and professional career are considered an edge case.
it only takes popular distributions making it default, then all sorts of things will start depending on it, and it will be difficult to not switch eventually. As somebody that has written a lot of xlib/xt/motif code years back, and that still has the full O'Reilly X series set on my bookshelves, I would prefer X11 to continue; but just like sysv->systemd it seems things are moving against that being the case.
> but worst case I'll run a Wayland compositor on my X11 server for that
This is actually surprisingly easy, too. My first experiments ran a wayland app in cage ( https://github.com/cage-kiosk/cage ), but in order to better handle multiple windows I switched to sway with a lightly tweaked config, and other than some input weirdness (I think it struggles with modifiers held while switching focus) it works pretty well.
I understand that Wayland doesn't support Xlib function calls. What are the options for those of us who rely on C code which which uses Xlib for 2d graphics and event handling? If I have to switch to e.g. GTK, I'd not only have to rewrite all the Xlib calls, but all the code which depends on this, as Xlib and GTK do graphics ad events quite differently.
Ariadne Conill is turning Xwayland into a full X11 replacement: https://github.com/kaniini/wayback. She doesn't want to maintain it long-term, so if you're interested in contributing, please do: the more people understand the codebase, the better.
There is a missing piece in Wayland tha I’m worried about—or rather, a key functionality that will need to be sorted out: Remote Desktop, but not in the sense of access to an existing display — rather, fully in-memory accelerated compositing for multi-user, concurrent desktop sessions to the same machine. FreeRDP has been doing some work on that, but we need a mature Xrdp replacement and I don’t think that is around just yet.
TLDR (non-AI): The author raises very valid points about accessibility under Wayland. The "And Now We Don’t Have a Choice" subtitle seems to address that only Gnome seems to have usable accessibility under Wayland right now but can easily be misunderstood as "we have no alternative to Wayland".
The author also mentions X11 sucked in many ways and that this transition also is a moment to rethink what proper accessibility should fulfill.
Wayland is a perfect example of the second system effect and the sunk cost fallacy. X11 had problems, but instead of fixing them, developers threw the baby out with the bathwater, designed something completely byzantine and weird. Time was invested, so no stopping now or ever. The design of wayland introduces more problems than it solves, only by plastering over with yet another extra protocol that you have to implement in addition to wayland you might regain some lost functionality. Or paper over a design bug that will never be fixed.
Wayland is a dead horse. Made edible by sufficient fermentation and mobile towards the finish line by declaring the rot spreading over the line to be a victory. That Gnome just now gains parts of the functionality it had in X11, years later, while still being actually behind X11, while being an incompatible mess that leaves behind all other desktop environments to the detriment of the Linux ecosystem as a whole isn't actually a victory. Its the victory of the slightly-less-ruined after a devastating war.
You can't fix a protocol that simply isn't designed for how modern graphics hardware works.
Both macOS and Windows have upgraded their display stacks over the decades, but it was seamless because unlike Linux, nearly all applications dynamically link the system library which they can upgrade. Linux is late to the party here because everyone wants to make their own toolkit.
X was designed for multiple remote terminals receiving drawing commands over a network, not locally hardware accelerated graphical interfaces and functions that rely on close coordination between the hardware and display server (e.g. hardware planes, vrr, hdr).
Fixing X would require a new protocol to the point that it isn't X anymore, aka Wayland. There are arguments that not having a reference display server has led to problems though.
Wayland also doesn't even remotely resemble anything that would be fit to talk to modern graphics hardware. DMA buffers and DRM (direct rendering manager, not the digital restrictions management) are an afterthought, a separate protocol that is not even all in stable yet. Vulkan usually doesn't work. Latency with Wayland got worse and won't get better because "frames have to be perfect". Tons of unnecessary blitting and latency be damned. All while actually tearing in Wayland is as bad as in X11 here.
Fixing X would require a protocol that is mostly X, but of course incompatible because you have to rip out some protocol bugs. But Wayland isn't X minus the bugs. Wayland started as a little bit of broken bitmap-pushing and a whole lot of hot air. And even with tons of extension and auxilliary protocol development, multiplied by tons of unnecessary reimplementations in tons of compositors, it isn't even where X11 was when Wayland started. Wayland fixed nothing yet, broke a lot, fragmented the community, brought pain and misery.
> Wayland also doesn't even remotely resemble anything that would be fit to talk to modern graphics hardware. DMA buffers and DRM (direct rendering manager, not the digital restrictions management) are an afterthought
You don't use Wayland to talk to graphics hardware, you use Wayland to communicate with the display server.
The Wayland protocol lets apps negotiate an area to write it output to and how it gets written there is completely up to the application, whether it involves the GPU or not, OpenGL, Vulkan etc.
This is in contrast to X where the app use X APIs to draw textures, which are then pulled by the compositor (copy, rip latency/performance), and then sent back to the X server to display.
That is complete BS. The application doesn't talk to the graphics hardware alone and then just copy a bitmap into a Wayland buffer. You don't just magically talk to the GPU. There is this little problem called 'security', 'multiprocessing' and 'multiuser' in between.
DRM is literally the 3D acceleration driver framework for Linux. It has been around for decades and is same set of drivers that are used for any sort of accelerated graphics in X.
If you want to get rid of DRM you have to start over and rewrite all graphical drivers for Linux from scratch.
DMA is part of the basic architecture of modern computers. It is how you can do things like have fast USB devices or network devices because it allows device hardware to by-pass the CPU and write things directly to memory.
> DRM is literally the 3D acceleration driver framework for Linux. It has been around for decades and is same set of drivers that are used for any sort of accelerated graphics in X.
Yes, exactly. And Wayland ignored it for years, from the start, and only later slowly adopted it as an extension.
> DMA is part of the basic architecture of modern computers.
Yes, I know.
DMA is even older and not limited to Linux or Graphics. Back in the old VESA, SGI and Windows 3.0 times, DMA was a cool new feature (but actually old even then). When Wayland was conceived, it was boring and old. Yet Wayland didn't originally include DMA buffers, just later added it as an extension when it became obvious that they had just gotten rid of a 40 year old feature that was really really necessary for modern graphics...
The protocol is trivially extensible, and have had plenty of extensions that has significant altered how it behaves. Fixing X in a step wise manner wod have been perfectly doable. The problem is that the Wayland proponents didn't just want to fix the things there was agreement was broken, but also wanted to actively break things that would cause an uproar.
For literal decades they proved how malleable the protocol is, yes. Then they decided to ignore how malleable the protocol is instead of even trying to e.g. start by deprecating obsolete functionality and disabling it in future versions.
When XRender was introduced, for example, was the perfect time to deprecate server-side font-rendering. It's trivial to shim on the client side if anyone cared about the legacy functionality, and it takes a trivial amount of code to switch to using XRender instead for it instead (been there, done that, written a font renderer).
There's been plenty of opportunities to gut the legacy parts of the protocol that way, and reduce the complexity.
Grouping the main set of extensions and declaring that if the server reports a certain version number or above they
This didn't happen because redhat/ibm/microsoft took control of the project and decided they wanted to kill it. They have said this directly on social media that this is their explicit goal so it's not as if this is a conspiracy of some sort, it's just their publicly stated goal. They then deliberately stopped merging changes and left it to rot until the contributors got frustrated enough to fork it.
Now a massive gamergate-esque coordinated character assassination campaign against the creator of xlibre is ongoing even from publications that have many previous articles praising the same contributor at length. Corporate control of open source is very dangerous to the ecosystem as they also fund the "journalists" writing these articles who are willing to defend every decision they make.
Do you mean dbus as the other protocol? Because that's an intentional choice to separate out the protocol for drawing windows and responding to events and "desktop stuff" that's not related to being a window on the screen. They're not plastering over anything, that stuff is intentionally out of scope for Wayland.
The advantage is that anything can use the desktop stuff (cli tools) just by talking to dbus instead of having to be a wayland client despite having no windows.
Tons of stuff aren't in wayland that are actually graphics and window related. Wayland just forgot about it, then later on people realized that their design was far too primitive and simplistic.
DMA buffers, color management and pixel formats, scaling and DPI, image and video capture, tons of keyboard and mouse related things, all initially forgotten, now incompatible extras that are inconsistently implemented, frequently broken and forever in beta/staging/unsupported hell...
If you mean what Wayland literally calls "staging" it means that compositors should adopt it now, it doesn't mean pre-release. Once multiple compositors implement the protocol it can become stable.
I'm sorely tempted to take the backend of a Wayland compositor and write a new X server on top of it, with a focus on deprecating everything not actually used by modern X apps that don't have alternatives, which would be a lot.
I bet you could benefit from quite a bit of the Wayland compositor work on modernising the lower levels, and end up with something much simpler than current Xorg without ditching much compatibility.
I think XWayland is basically an X server that supports Wayland. It is based on Xorg though. But if you really think it has a bunch of features nobody uses, you could try to delete those simplify that way.
I don't want an X server that supports Wayland. I want a leaner, simpler X server without Wayland.
EDIT: To clarify, the reason I mentioned Wayland compositors is that it'd be an opportunity to pick a low-level rendering backend that has been written from scratch without the baggage of Xorg.
The "good parts" of X that modern apps actually use are comparatively simple compared to the low level bits - the protocol is trivial-ish, and you can get 90% there by implementing a small-ish subset of the protocol.
XWayland can be run in 'rootless' and 'rootful' modes.
Rootless is what it is typically ran as and allows you to integrate X11 apps into your Wayland desktop environment.
Rootful would allow you to run X11 desktop with X11 Window manager and the whole ten yards.
I don't know how well rootful mode is working because it is rarely used. I imagine it would take some work to make it fully functional. But it something that exists and I would expect that the Xorg/Wayland devs would like to see it fully fleshed out.
But that essentially gets you most of the way there.
> designed something completely byzantine and weird. Time was invested, so no stopping now or ever. [...] is a dead horse. Made edible by sufficient fermentation and mobile towards the finish line by declaring the rot spreading over the line to be a victory.
This applies perfectly well as a criticism of X11, you know.
Yes, but X11 has the excuse of being very old. Wayland was supposed to be a clean redesign, learning from the mistakes of the past. They learned from some mistakes, but they also completely ignored all of the good ideas, functionality and necessary additions that X11 and other systems like MacOS and Windows had.
They designed a system that was as simple as they could make it, trying to push the complexity that wasn't just bitmap-blitting onto others. And they isolated Wayland from all possibilities of making proper, compatible, common extensions. They tried to be the opposite of X11, because they recognized "push absolutely everything into the X11 server" as a mistake. Nobody ever used the XPrint extension. But they went too far, cut away all the useful stuff, all the necessary extensions, even cut away the "server" and just went with a library plus protocol. Now everyone has to reimplement their own server (compositor in wayland lingo), producing tons of busywork, splitting the community, ensuring incompatibility and pain. And everyone has to keep up to date on all the tons of necessary extra protocols that all the compositors have to implement all over again.
Yes, the same words do apply. But for different reasons.
More and more people are using GPT for editing. I admit to sometimes using AI to help me too. To be fair, I'm not sure what to think of this particular slick style quite yet.
Bad news for all my programming guides that use em dashes properly... ;D Luckily I have a Git history on them that predates LLMs, for whatever that's worth.
Ah, a lot of people do use LLMs for editing. And when I read this I can just hear GPT-4.x's style and tone of voice in my head.
None of the following are smoking guns, but together... well, either it's GPT or it's https://xkcd.com/810/ .
* Bullet point lists
* emdash
* extensive use of bold
* sentence fragments
* "Here's the good news:"
* juxtaposition over emdash: "This transition is happening — but we’re not being ignored anymore."
* Bullet points that simply MUST have a conclusion " The entire workflow? Gone." , " not just for me, but for every user who deserves to choose how they compute.", "And they shouldn’t have to." , "But we have to start over."
* "I hope it’s done right — not half-baked, not bolted on."
* "We lost an ecosystem."
* "That has to change. / And it starts with every compositor agreeing on what “accessible” actually means. "
etc....
Maybe it's
A) A human who is a very skilled writer with a particular style
B) GPT4.x
but my best guess is
C) Both: Human did the rough draft, then had GPT4 edit it into shape.
OFF but I liked most of these points ~3 years ago. Bullet points for "what you will learn" and "what you've learnt" for example.
Now articles structured this way make me sad. I am not sure if it is acquired distaste, just tiring, or recent articles are indeed worse; either way I am not very happy about them.
It'd be odd indeed if we'd banned something we hadn't heard of.
After looking into it, I think what happened is that a few submissions were flagged by users, for the cromulent reason that the articles weren't very substantive.
Edit: There also seems to be an aura of flamedrama around it. I didn't look closely, so I may be wrong, but if that's true, it's another reason why HN users would (cromulently) flag it.
Can you elaborate more on discussions about XLibre being banned on here? Are accounts that mention it getting banned? Is there just mass downvoting by partisans with some kind of agenda?
I get that X11 has security issues, but NVIDIA drivers and Wayland still seem to have no support now and no support planned for the future, so Wayland is a non-negotiable non-option for many (most?) Linux desktop users, including myself.
That said, with the custodian of X11 refusing to merge 1000+ patches including various bug fixes and security fixes, I'm excited for the prospect of XLibre - this is exactly what Open Source, as an idea, was invented to facilitate - user choice.
I have an RTX 4060 Ti with latest drivers installed, KDE on Debian. On the login screen, I can select X11 or Wayland. Selecting X11 lets me log in fine, selecting Wayland and logging in results in a black screen and the only thing I can do to get any video out at all is switching to another TTY.
I have an RTX 4060 with latest drivers, KDE, and Arch. Wayland works perfectly for me. Maybe Debian has some outdated packages that haven't caught up yet?
I've been totally breaking Linux installs trying to get Nvidia to work for 15 years now, and that's on X11. On the other hand I recently did the first OS upgrade that I've ever done successfully without breaking Nvidia and that was running Wayland.
Nvidia is just really really bad on Linux in general, so it's always a coin toss if you'll be able to boot your system after messing with their drivers, regardless of display server.
Nvidia under Linux has had a long and hard history.
For most purposes, including gaming, it is best to avoid Nvidia hardware. Using Intel for laptops and AMD for dedicated GPUs is kinda the best general approach if you are planning on using Linux.
Of course if you have a need for CUDA then Nvidia is the only game in town, but that is a different issue then Wayland support.
For a while Nvidia was fighting the Xorg/Wayland devs over GBM vs EGLStreams which has delayed Wayland support. This has to do with the API extensions that allowed Wayland to manage application output buffers.
Gnome was the only Wayland environment to try to support EGLStreams for Nvidia, but it really didn't do them any good.
A while ago Nvidia eventually switched over to GBM and EGLStreams is dead, which helped out a lot of people running non-Gnome Wayland desktops. But there are lots of problems with Nvidia drivers besides that right now.
The reality is that Nvidia doesn't care about consumer Linux desktop. Their primary focus is on Enterprise users in terms of people needing graphically accelerated desktops.
So right now if you are running Linux on your personal workstations/desktops/laptops you are essentially beta testers for whenever Enterprise Linux distros make the switch to Wayland.
> The reality is that Nvidia doesn't care about consumer Linux desktop. Their primary focus is on Enterprise users in terms of people needing graphically accelerated desktops.
What does this actually mean in terms of technology? What is Nvidia providing that works for RHEL but doesn't work for Fedora, or whatever?
> Can you elaborate more on discussions about XLibre being banned on here? Are accounts that mention it getting banned? Is there just mass downvoting by partisans with some kind of agenda?
YC moderators are hiding the articles from the front page.
x11libre got 2 active discussions here according to
A highly controversial guy making controversial patches to a somewhat heated topic around these parts is going to get flagged by users too, how are you so certain it's the admins doing the hiding? Even when not flagged the score to comment ratios will drag those off the front page quickly.
Both threads devolved into politics, and that's likely the reasons why they got flagged. There's no evidence these were hidden by moderators, as opposed to because regular users flagged threads.
Having seen some of your comments on one of those threads, there's a decent chance your comments contributed to getting those posts flagged.
The [flagged] indicator on a submission usually indicates user flagging. Moderators and algorithms just quietly downweight submissions without any visible indicator. So this isn't an HN moderator position, the question to resolve is why users would flag it.
In this case, I'd have flagged them too if I saw them. The "long live" post is an aggressive tirade that reflects poorly on the author and led to a poor-quality discussion. The second is a link to a git commit history, which is weird in its own right and provides no explanation, and the context provided in the comments shows that a generally dislikable figure with extreme political views is now leading a fork of X11 that has yet to prove itself viable. So I'd probably have flagged that one too as pointless drama until proven otherwise.
It's a fork of x.org that's trying to keep actively developing it more than the x.org developers are interested in doing (because they are working on wayland and see x.org as basically obsolete). That in itself is a reasonable goal but given the person responsible for it I would not be too optimistic about it.
XLibre is a fork of Xorg X11's codebase started by a developer who got kicked out of the Xorg project because he was making lots of changes that broke everything and had a hard time getting along with the other devs.
Another dev blindly applied his MRs assuming he had tested stuff before submitting the request and they had to go back and revert a bunch of stuff.
Broke nvidia compatibility, broke xrandr extension, and a bunch of other stuff.
This is false, he we the sole person submitting large amounts of code for about a year and was widely praised, until he forked and red hat went nuclear on his account for opposing their corporate goals, he was not kicked out for bad changes or anything of the sort. Trying to revise history on this is malicious.
The people like OA are spreading misinformation because the developer stated everyone was welcome to contribute and would oppose excessive politics. Keep in mind red hat is currently getting sued 3x over for blatantly racist policies which they used to ban other contributors and forced on their managers, they are evil people.
They did that via an out-of-band D-Bus protocol, rather than going full Wayland. I'm all for keeping D-Bus around for backwards-compatibility (lots of things use AT-SPI2, and I certainly don't want a repeat of the CORBA -> D-Bus migration), but Wayland's AT support should be first-class, not relegated to proprietary GNOME extensions.
Per Matt Campbell's article https://blogs.gnome.org/a11y/2024/06/18/update-on-newton-the..., this decision has something to with security:
> Assistive technologies or other accessibility clients currently connect to the compositor through a D-Bus protocol, defined in the Mutter repository linked above. By exposing this interface via D-Bus rather than Wayland, we make it easy to withhold this communication channel from sandboxed applications, which shouldn’t have this level of access.