This guy's fix of a 13y old xserver bug was reverted because the fork he made to apply the patch was deleted, so the merge request was reverted after it had been approved.

gitlab.freedesktop.org/xorg/xs…

@Nanook if you can locate the offended subjected and invite him to the discussion I believe he would be best able to explain your question

I thought it would be of interest here being a very critical community of IBM/RH products. Unlike many other projects we do like the Alan Coopersmith of X/Wayland has been one of the most helpful and polite persons ever encountered in FOSS-land.

We shouldn't forget what we have here is a for-profit corporation and we are outsiders, clients, customers. Behind closed doors, what happens in IBM stays in IBM.

randr: fix unconditional byte-swap in ProcRRGetProviderInfo() (!1977) · Merge requests · xorg / xserver · GitLab

The list of the associated provider's capabilities was always swapped unconditionally, while it should only be in case of client having opposite endianess. Reported-By: dasha_uwu Signed-off-by:...

^GitLab

This just tells me you are 100% systemd dependent, you know no other way. What I would suspect is that you wouldn't even take the chance of trying something else and moving it.

systemd open bugs and security issues are between 4 and 5 digit at all times through its history.

runit in the past 10 years has had one bug ?? its code wasn't kept up to latest glibc library and had to be touched up/patched, which was done 9 months ago. Now we are down to boring 0.

But if you really want to get into complex server service supervision and dependency management you need a little study of s6 and maybe 66 (see Obarun). This is state of the art.
skarnet.org has been up and running for years, it serves mail, git, web, mailing lists, who knows what else, and it publishes its record of down time instances (even power failures)

Void was one of the first distributions that adopted systemd, it has been running runit ever since getting out of it. Debian 7 (not that long ago, didn't have systemd, it was introduced in Debian 8). s6 and 66 have been supported by unofficial repositories

skarnet.org/software/systemd.h…

The term "modern" has been heavily abused by the "gang" it is an empty term providing zero rational arguments of why "modern" is better and in what way is it "modern".

sinit is proof that as far as init goes all you need is a few lines of code, with "modern" equipment the difference in performance between sinit and systemd is measured in nanoseconds.
Init is done and finished as commonly known stage 1, which also creates the escape route for the final stage (call it Z) of shutdown or its more complex alternative, reboot.
What goes on between stage 1 and stage Z is init irrelevant, should be and should have been.
The frozen empty argument that systemd floats as defense is that it is better than sysvinit or BSD scripts, because it can start all those things in stage 1 in parallel, one process doesn't have to wait for another. When you say that daemontools, perp, anope, runit, s6 do the same but in a better more efficient way they pretend that those systems copied systemd. If you turn their argument around on them and call them "modern" they will point out that they are older than systemd, daemontools are.
So you can never really hold a rational argument with the religious clergy of the linux-hydra. Their thought is as clear as any clergy discussing astrophysics.

The clue is that if you package OpenRC (a sysvinit improvement) runit, s6, and sinit, and combine them in one superpackage to coexist in one system (they can) the space of this package is less than half of systemd.
You can use them on BSD, on solaris, or any unix fork, while system is only useable in linux.

So funding for this evil kernel and the owners of this systemd X/wayland complexity is one, the goals are now clear, and the make MS even look innocent in tactics and development.

I gave you a link above to study if you are trully interested, but please don't come here with IBM fanboy tactics to tell us off.

MX is nothing without antiX, just a theme creator. The only difference between MX and debian is pid1, the rest is just like debian. So not a true distro, just a theme creator like Mint or Ubuntu or Manjaro.

Basically any of those fake distros claiming to be systemd-free using elogind are just fakes, advertising the technicality/legality that pid1 is NOT systemd, but everything else is! Gentoo and Void are the latest to turn rogue and adopt it, artix, MX, devuan were designed since day1 to utilize it.

Like emperor Trump says, those with wealth and power can enforce their irrational will against any rational system of thought. Rationalists are public enemies. God Bless AmeriKKKa!

So, yes, systemd is better, it is modern!

And I need to get back to work, because I don't get paid to debate or run upgrades.

I think you missed my pun about sysvinit and 20th century relics.

What init is and what services are "started" up is a very blurry picture due to systemd.

This is stage 1 in runit (+ bootlogd optional)
Anything else is stage2, they may seem hierarchical by traditional thought, but not many wait for the previous to finish . s6 goes a step beyond this and doesn't separate runit-init with runsvdir (service supervisor) it is all one process that supervises all services, and is able to log and provide a tty earlier than any other system.
Advantages handy for deep enterprise commercial grade systems, not for home/hobby users. For us runit is plenty.

/etc/rc/sysinit

'''
01-sysfs -> /usr/lib/rc/sv.d/sysfs
02-procfs -> /usr/lib/rc/sv.d/procfs
03-devfs -> /usr/lib/rc/sv.d/devfs
05-root -> /usr/lib/rc/sv.d/root
10-hostname -> /usr/lib/rc/sv.d/hostname
15-hwclock -> /usr/lib/rc/sv.d/hwclock
20-kmod-static-nodes -> /usr/lib/rc/sv.d/kmod-static-nodes
25-tmpfiles-dev -> /usr/lib/rc/sv.d/tmpfiles-dev
30-udev -> /usr/lib/rc/sv.d/udev
31-udev-trigger -> /usr/lib/rc/sv.d/udev-trigger
32-modules -> /usr/lib/rc/sv.d/modules
33-udev-settle -> /usr/lib/rc/sv.d/udev-settle
40-console-setup -> /usr/lib/rc/sv.d/console-setup
45-net-lo -> /usr/lib/rc/sv.d/net-lo
50-misc -> /usr/lib/rc/sv.d/misc
55-remount-root -> /usr/lib/rc/sv.d/remount-root
60-mount-all -> /usr/lib/rc/sv.d/mount-all
65-swap -> /usr/lib/rc/sv.d/swap
70-random-seed -> /usr/lib/rc/sv.d/random-seed
75-tmpfiles-setup -> /usr/lib/rc/sv.d/tmpfiles-setup
80-sysusers -> /usr/lib/rc/sv.d/sysusers
85-dmesg -> /usr/lib/rc/sv.d/dmesg
90-sysctl -> /usr/lib/rc/sv.d/sysctl
95-binfmt -> /usr/lib/rc/sv.d/binfmt
97-binfmt.d -> /usr/lib/rc/sv.d/binfmt.d
99-cleanup -> /usr/lib/rc/sv.d/cleanup

'''

I have to disagree about MX-Linux... it's not just Debian with some window dressing tied on. It includes a very nice GUI toolbox for things like making a bootable copy of an installed OS, a Live kernel updater, a USB formatter, etc. I know there are terminal commands for these jobs, but a desktop user like me doesn't care to fool around with the terminal and, I'm downright phobic using the terminal. I feel like I'd mash the wrong button and accidentally ignite the Earth's atmosphere and wipe out all life on the planet or something. Gimme the point-and-click stuff any day.

That said, however: I see the whole Linux ecosystem being corrupted by Big Tech money from Microshit, Red Hat, IBM, Google, all those guys. BSD is largely free of all that Big Money corruption and politicization. Which is why I've been experimenting with a desktop OS based on FreeBSD. For me it's as much about ethics as just about anything else.

Beware of the enemies of your enemies are not always your friends. Some of those people have spent such a long time in the corporate world they have incorporated the madness into their own thinking and practices. Go some days back and this guy was perfectly confortable "helping out" IBM improve its image as a FOSS contributor. Now things changed.

Alan Coopersmith of Oracle who approved this guys MR for a bug fix replies to the list on the whereabout of this long term contributor as if the matter is about him being thrown off the list for "spamming" with his email having too many recepients. So the email where this guy is reporting and venting steam for being unjustifiably thrown off is WELL moderated because of too many Cc: recepients.

There is nothing good or of value coming off authoritarian gangsters serving their corporate agendas (profit and more profit, domination, dictatorial ruling of markets and slave-work whose exploitation can be translated to $$$$).

What happened to consolekit when IBM/RH (and don't dare correct me that back then RH wasn't IBM, RH was IBM long before systemd was announced .. a well fed child of IBM consulting .. a trojan horse into FOSS/linux ) . What happened to udev when they decided to incorporated into the hydra of code?

Why are so many upstream projects pushed to incorporate systemd mechanisms into their code? To alienate BSD and other unix forks and leave linux as a corporate puppet to creat the new IBM/Oracel-PC-Windows full of trojans and backdoors as their "client" prescribed! Russian and Chinese corporate employees need not apply!

Is Gnu sticking its finger deep into the deep-state/corporate honey? I'd be surprised if they are not the host of the gangster party.

#linux is bound to become a corporate non-free product, and there is little we can do to resist, especially if we are consuming "modern" hardware having specific demands of later kernels later gfx firmware latest X/way
So take your i9-ultras and Ryzen-9s and dump them as toxic waste!

I’m downright phobic using the terminal. I feel like I’d mash the wrong button and accidentally ignite the Earth’s atmosphere and wipe out all life on the planet or something. Gimme the point-and-click stuff any day.

To be phobic of a terminal is one thing, to blindly accept a nuclear reactor under your bed to power you alarm clock is another.
Some friend with similar tendencies who kept being convinced of living systemd-free but slipping back into the needle (Arch), used the argument all he cared about was to have his usb-stick mount and show in his filemanager when he inserted it so he can click and drag files into the "stick", or read a cd/dvd once it was inserted in the optimal medium drive.
So I wrote a few lines of a script chmod it into 0755, told him to save it on his desktop as an icon (filemanager maximized without borders with a background image) and asked him to click on it when he entered the stick (so it wouldn't run and parse system directories all day just in case a stick was inserted). I told him within 3-4" the writable/readable partitions of the stick would appear on the list of media on the side of his filemanager, whether ntfs or ext they were there. I also asked him how many hard disk drives he had in his system so the script wouldn't have to account for such, it would just read /dev/sdc /dev/sdd ..

But instead you have logind, polkit, dbus, sysusers, .. running all day, huge amounts of code, just to be able to do this one thing, automount an external disk and make it appear in filemanager.

I do know that this script would have to become more complex for general/generic use to all systems but it can also be edited by non-coder capable users just as easy to make it work well in a particular system.

So there is much gray area between working on console, not even a graphic terminal, and having a full blown carnival of Gnome or Plasma to deal with as overhead.