Live data from Hacker News

Long live Xorg, I mean Xlibre

dedoimedo.com

61–70 of 212 posts

Re: Long live Xorg, I mean Xlibre

#61
post #52

I don't get the Xorg vs Wayland fight. Feels like it mostly coming from people who don't contribute to either of them. If you like Xorg, use Xorg. If you like Wayland, use Wayland. If you're not happy about an issue, contribute to it.

The fight is because:

1. freedesktop based on xorg is outdated and needs to be modernized to keep up with competing operating systems.

2. The Wayland movement is and has been run extremely poorly.

Both xorg and wayland are bad so it's easy for either side to point that out.

Re: Long live Xorg, I mean Xlibre

#62
post #43

Well, it's my understanding that Xorg still cannot do per-monitor fractional scaling these days, have they fixed it? That was the major selling point of Wayland for me, as an occasional linux desktop user. Retina MacBook Pro was released in 2012, about 13 years ago. Personally, I don't think Xorg is in a position to sneer at its competitor for being "beta in quality" after "15 years into making."

This is why I jumped to Wayland: laptop monitor and desktop monitor with different resolutions/scaling. Xorg choked, Wayland worked. That said, moving to Wayland broke a ton of other stuff in my otherwise peaceful workflow. I really do not like Wayland, but I need to be productive on my machine.

People have sworn up and down to me that the fractional scaling stuff is possible in X but I read probably 10 guides on how to do it and it never worked. Bummer.

Re: Long live Xorg, I mean Xlibre

#63

Earlier quoted context omitted.

Honestly the belief of the developers don’t matter at the end of the day. The quality of code is the only thing who matters. And I don’t imply the quality of code is good or not, I have no idea.

From authors commits to x11 and discussions about that, looks like code was not great. source comment: https://news.ycombinator.com/item?id=44200000

founders beliefs are garbage

Well it's the code that matters. Not what their beliefs are.

code is also garbage

Well it's the thought that's important.

Re: Long live Xorg, I mean Xlibre

#64

[flagged]

[flagged]

You'd expect that the changes would've been reverted sooner if that was all there was to it, no? How come they're suddenly a problem?

that was my first thought too. if these commits were a problem they should not have been accepted in the first place.

was there no review process in place? and if there was no review, isn't that a sign that the project is dead? and if the project is dead what's with the sudden activity?

none of this makes sense.

Re: Long live Xorg, I mean Xlibre

#65

Having this specific guy becoming the lead/creator/maintainer of Xlibre, strengthens my argument about the kind of people who are against the Xorg->Wayland replacement.

[flagged]

I've been here for 13 years and I've never seen so many flagged throw-away accounts as I have in the series of threads discussing this fork. I was blissfully unaware of this niche corner of the conspiracy hive mind.

The README, their anti-vax rants, the "Make X great again!", the way they write... I'm getting Terry A. Davis (RIP) déjà vu.

Re: Long live Xorg, I mean Xlibre

#66

Earlier quoted context omitted.

[flagged]

If a new startup comes up, and someone points out "they've taken funding from Philip Morris / Altria", that is completely on topic, and will affect how some people evaluate the company. This is comparable.

Calling someone a "reactionary nutjob" isn't really informative. You can put someome's biases on display without insults, just as other commenters have done.

Re: Long live Xorg, I mean Xlibre

#67

Having this specific guy becoming the lead/creator/maintainer of Xlibre, strengthens my argument about the kind of people who are against the Xorg->Wayland replacement.

I couldn't be not against Wayland in those dark times when some distributions would make it as the default, at which time the flagship Wayland DE was using XWayland for its own panels. Which made a shitshow of a blurry mess on a HiDPI screen.

I couldn't be not against Wayland when in half of the applications I used clipboard wasn't working properly.

I couldn't be not against Wayland when screen sharing would either use X11 or not work at all.

I couldn't be not against Wayland when most of the complexity shifted from X server to compositors and toolkits, and it both diluted responsibility for the gnarliest bugs (go guess if it's the compositor or the toolkit! Added fun when it's the boundary of both!) and made it possible to write GNOME-only, or KDE-only, or wlroots-only software that won't work on another compositor because it needs a private implementation of an obscure protocol.

I couldn't be not against Wayland when your window manager breaking would bring your whole session down, complete with the breakage happening a lot.

I hear that there are problems with input protocols like "your toolkit has to implement several versions of the same thing and they all are half-arsed anyway", but living in the happy land of the Unicode's Basic Multilingual Plane, I know too little of this problem to have a say about it. But I'm somehow not surprised.

There is also this thing about Unix-like systems not being limited to Linux, and the story of Wayland elsewhere is way worse.

Most of the gnarliest points are now addressed, and Plasma got super stable around 5.26 onwards, however, I'm not really happy that we now have a triumvirate of Mutter, KWin, and wlroots, and that the bar for entry for new compositors is quite high. The fact that X11 allowed a proliferation of window managers is its advantage, not the other way around.

I'm wondering if someone gets to write a compositor that will be "engine, not policy" for Wayland, on top of which different desktop environments could be built. Like wlroots, but some steps further. It will probably have a year of glory, and then Wayland will be declared obsolete, crufty, insecure, legacy, and in dire need of being replaced with something lean and simple, and maybe written by an LLM. Good thing I'm going to be too old or too dead to give a shit by the time it happens.

Re: Long live Xorg, I mean Xlibre

#68

I get it about Wayland. Most of X is legacy cruft and even isolating that into XWayland or whatever it's called for backwards compatibility is better than having it front and center. But yes, there are use cases it doesn't cover. Example. My elderly mom uses Linux laptops that I've rigged to (1) always have an SSH connection open to my server machine, with reverse tunnels, and (2) run x0vncserver. Modern security peo…

KDE and Gnome each have their own Wayland compatible RDP servers. It's annoying it's not as convenient as older X vnc servers, but it's not impossible.

For wlroots systems there's wayvnc.

Re: Long live Xorg, I mean Xlibre

#69
post #65

Earlier quoted context omitted.

[flagged]

I've been here for 13 years and I've never seen so many flagged throw-away accounts as I have in the series of threads discussing this fork. I was blissfully unaware of this niche corner of the conspiracy hive mind. The README, their anti-vax rants, the "Make X great again!", the way they write... I'm getting Terry A. Davis (RIP) déjà vu.

In the Laundry Files by Charles Stross, doing too much computations in your head may give you a form of brain damage by inviting extra-dimensional feeders to feast on your central nervous system. I feel like while the details may differ, as far as the big picture is concerned, he is not wrong.

Re: Long live Xorg, I mean Xlibre

#70
Wayland is architecturally garbage. Not for technical reasons, but social ones. Although social isn't the right word to describe a technology that encourages toolkit/DM/etc fragmentation which in turn breaks core functionality in the wider application ecosystem. Some of this isn't the fault of wayland directly, but it makes it worse. AKA by design there isn't a standard way to iterate windows and detect buttons/lists/various control types/etc which massively complicates if not outright breaks screen readers, while at the same time continues to fragment simple things like theming (and copy/paste is even worse than it was 10 years ago) if an application isn't using the blessed GUI toolkit. And in say fedora, the gnome folks seem to be designing to a device type that doesn't exist, its terrible out of the box with multiple monitors/etc while at the same time being terrible on touchscreen devices. I don't run it that much, but was showing my daughter who is perfectly competent in windows and macos (she carries both around all day!), how to navigate gnome, and jut shocked how absolutely none of it is intuitive to someone who has 'expert' level knowledge of all the common OS/phone UI's. Sure it looks slick, because its not cluttered with 'garbage' like you know, maximize buttons, scroll bars, and other things one might click on with a mouse. All the while doing cool things like reflowing the terminal text whenever the scrollbar appears/disappears. Its this complete lack of supervision/understanding. Sure on a phone touchscreen scroll bars might get in the way, but on a large screen laptop with two 30"+ 4K monitors?

And frankly, as someone who works closely with some of these distro's, I think there is a silent majority who have the same opinion but aren't willing to pay the political tax in their ecosystem for standing up and pointing out the emperor is naked for fear of sounding like a Luddite and being sidelined.

Post reply on HN