So long Xorg, thanks for all the fish, hello XLibre?

“For a while we had hoped to get an actual Xorg release, but that's notgoing to happen anymore, so I'm finally going for a full fork,” Enrico stated. “This fork [XLibre] became necessary, because it’s the expressed wish of the current Xorg group's majority to abandon the project, let it rot forever, and block any substantial contributions, let alone new features. I'm leaving it to the reader to deduce which corporate interests are behind that, and instead just moving on.”
Click to expand...

See:
https://twitter.com/x/status/1930727192964514137
View: https://x.com/LundukeJournal/status/1930727192964514137
 
I have discussed with Enrico a little in the past, he is doing some really good stuff (he was last setting up full chroots for automated builds for OpenBSD which a while back I was helping to advise him on (the /var/sysmerge stuff is fiddly)).

There was a small speculative IRC channel with guys from the most well known WMs that we invited Enrico to but he seemed too busy to join. I do feel a little more collaboration between him, Oracle and OpenBSD could be beneficial but I suppose he doesn't want to get bogged down with any potential bureaucracy.

The Xorg board consists of people forwarding Wayland, not because of some conspiracy theory but purely due to historic quirks of how Wayland started out. There is too much legacy to bother moving them, so I honestly think a fork is the best approach. They are two very different projects, serving different purposes and neither can replace the other.

Xenocara/Xlibre are two great options for taking X forward and its not like the primary UNIX/Linux Xserver hasn't been forked a few times before. Perhaps a project refresh might also attract a few newbies back.
 
When I tried Wayland before (back to X11 then), I felt it too less defined for use-cases to replace X11. I was quite confused (and still being confused) with input method interfaces having (seemingly) incompatible multiple versions.

So my current assumption is that Wayland should be implemented as runtime services of UEFI that virtualize all "screen outputs" and "GPU services including CUDA, OpenCV,...)" and completely disallowing every OS'es running on it to directly access GPUs and use its runtime services.

This way, the confusing parts can be completely splitted out to OS layer and above.
 
I've done a bit more research into this and the character behind and lets just put it like this, have you set out one day expecting a nice walk in the sun and found yourself in the middle of a minefield...

Say no more, but I kinda regret starting this thread as it wasn't my intention to kick over a political hornets nest.
 
jardows said:
I just wish he'd picked a different name. The "libre" moniker is overused, and I've always found it a bit cheesy.
Click to expand...
Heh, actually agree. Something like Xopen would be good. Perhaps even go for some sleaze-words like Xmodern, etc ;)

RypPn said:
I've done a bit more research into this and the character behind and lets just put it like this, have you set out one day expecting a nice walk in the sun and found yourself in the middle of a minefield...
Click to expand...
If he does good work, it doesn't really matter. Some of the best and most useful inventions benefitting humanity were developed by people with complex personalities ;)

RypPn said:
Say no more, but I kinda regret starting this thread as it wasn't my intention to kick over a political hornets nest.
Click to expand...
I think most in this thread are quite supportive of the project. We are used to discussing *way* more political lobbying here.
I'm not sure how new you are to these forums? (If so this is a strong post to start with) When you are a regular, you will see what I mean. Its all good discussion though :)
 
RypPn said:
Well I've been here since 2017, but it depends on how you define new. I just don't post very often as I tend to just use whatever gets the job done tbh. I do like FreeBSD, but I'm not wedded to it.
Click to expand...
Ah fair enough, my apologies. Then you have probably seen at least your fair share of turmoil haha.

I have just had a snoop of how some people are reporting this (i.e Lunduke) and it is very sensationalist.
 
Yup, indeed, he seems to delight in poking the bear, I've no axe to grind with redhat or gnome, although I don't use either.

cyric I can take it, it's still better than no reaction.
 
It would be really nice to see some of those features integrated, like change namespaces. It still concerns me that my jails system all share the X11-unix socket and they can all "see" each others screens and keystrokes. Or at least - that X11 doesnt accomplish lower level segregation.
 
BawdyAnarchist said:
It would be really nice to see some of those features integrated, like change namespaces. It still concerns me that my jails system all share the X11-unix socket and they can all "see" each others screens and keystrokes. Or at least - that X11 doesnt accomplish lower level segregation.
Click to expand...
I tend to move the X11 UNIX socket into the Jail, keeping it isolated. Though this does mean that the Jail must be on the same filesystem. Though actually even if the UNIX socket is inaccessible in the FS, this doesn't prevent code from being able to access it after some technical fiddling.

But obviously (and luckily) each Jail doesn't know the MIT_MAGIC_COOKIE of the Xorg session unless you explicitly tell it, so things are completely secure in practice.

But I suppose this is going off-topic a little.
 
1749235896039.png
 
What I find interesting, at least to me, is how many times MIT-X Windows has been declared dead but then pops up again under another name.
If we never had it to start with, would we really have any graphical environment as portable?
Yes, Xerox PARC, Mac, NeXt etc, but how easily portable are they?

Something like X has a lot of historical inertia, but perhaps actually taking a chainsaw and cutting out the deadwood is needed (I think a lot of us have been involved with codebases that are 20 years old).
If the contents of the email Charlie_ linked is true, it really sounds like "we don't want to play with you anymore so we're taking your ball and kicking you out of the sandbox".

I think forks sometimes generate renewed interest in a specific project and move it forward by leaps and bounds.
 
mer said:
would we really have any graphical environment as portable?
Click to expand...
X11 is not (only) a graphical environment, wayland is not a substitute of it.

mer said:
Something like X has a lot of historical inertia
Click to expand...
Indeed.

That are two reasons for me to insist on X11. Why are people that like wayland so fanatic? What are the real reasons behind wayland this fanaticism? Why they need to kill X11 and impose something that does not substitute it?
 
kpedersen said:
I have discussed with Enrico a little in the past, he is doing some really good stuff (he was last setting up full chroots for automated builds for OpenBSD which a while back I was helping to advise him on (the /var/sysmerge stuff is fiddly)).
Click to expand...
I read some of the comments on his merge requests, and they made me somewhat skeptical of the project:
gitlab.freedesktop.org

xrandr doesn't work anymore on xorg-git (#1797) · Issues · xorg / xserver · GitLab

After the recent commits xrandr no longer works, also kde kscreen fails. I tried installing libxrandr and xrand from git, but xrandr remains blocked without any output.
gitlab.freedesktop.org gitlab.freedesktop.org

However, I will defer to your actual experience working with him.
 
Jose said:
I read some of the comments on his merge requests, and they made me somewhat skeptical of the project:

However, I will defer to your actual experience working with him.
Click to expand...
Hmm. He does come across as pretty volatile. I do think there are complexities on both sides though. Some more dialogue from your same linked thread (read down from here a little):

He does seem to play ball with some members. But you are right, Xorg is pretty big and if there isn't better collaborations with other developers, it will be difficult to be a one man show, no matter how much I respect his work.
 
Wayland is better, in my opinion, for better fps in games and no screen tearing. Is this new fork of X going to get a port to FreeBSD? It would be interesting to do some game testing with it. 😀
 
kpedersen said:
Heh, actually agree. Something like Xopen would be good.
Click to expand...
Already taken by Python: https://pypi.org/project/xopen/. Dont like it? The R programming language is another case of staking a claim to that: https://github.com/r-lib/xopen.

Anaconda likes it (https://anaconda.org), and according to Wikipedia, X/Open means something else entirely: https://en.wikipedia.org/wiki/X/Open.

I propose openX as a new name for the fork, a la OpenSSL -> LibreSSL renaming 😂

I agree, Libre is an overused moniker, but at least it's not taken. We have enough name collisions as is - like Tesseract, for example. Is it a game? Is it an OCR lib? who the devil knows? 😈
 
Someone should spearhead a foundation so that people can crowdsource their efforts and monetary sponsorship. This is their chance to break free from Red Hats hegemony.
 
You must log in or register to reply here.
Back
Top