"You are not worth catering to."

Based on some recent news, and an interaction I had.

#linux #flatpak #art #opensource

Based on some recent news, and an interaction I had.

license is CC0, feel free to share around everywhere without attribution. inspired by unix surrealism

@nelson this is one hell of a cope if I’ve ever seen one ngl

@julia ?

@nelson flatpak is great

@julia they've recently declared that the next versions of flatpak will no longer support systems without systemd

@nelson Flatpak W ngl

People that use other init systems always have a stick up their ass about non-native packaging anyways, it won’t make a difference

@julia most of the non-systemd linuxes are usable exactly because of flatpak, specially lightweight systems, such as postmarketOS. chimera linux itself pushing towards this idea of the custom OS that's made usable due to flatpak providing an amazing system-agnostic platform.

flatpak truly has no need to depend on systemd as of right now, they even promote the idea of installing flatpak on non-systemd systems in their website, because their containers shouldn't (and currently don't) have a dependency on anything other than the kernel and userland tools that talk to the kernel such as bubblewrap, why go this route now?

also, wow, you've never even took the time to talk to me, and now you come here just to shit on what i'm trying to convey? get out

@julia @nelson prejudice much

@julia @nelson flatpak is great , but also isnt this kind of a weird hill to alienate contributors on ?

also , the reason people like native packaging is exactly because of stuff like this .. if flatpak allowed more community input, im sure most people would feel the same way?

@fiore @julia nah, some people ARE annoying about all of this, but that's just The Loud Minority(tm)

I still think that there's great worth in a truly universal package manager and flatpak was close to being perfect already, this was just a very weird move

@nelson @julia yes i agree , thats how i feel too!! sorry i mightve phrased it wrong

@fiore @julia nahh it was good, thank you for replying!

@fiore @nelson the point of flatpak is to decrease ecosystem fragmentation and provide stable runtimes for applications

supporting entire different multiple different OS stacks is contrary to that goal

@julia @fiore what flatpak does at a fundamental level is setup containers and os-tree, there is no need and there has been no need to depend on systemd as long as the protocols are properly implemented, such as the freedesktop xdg-portal and wayland stuff

what flatpak does amazing is that it was able to support and depend on a specific subset or intersection of several systems at once, it doesn't matter how weird your system is because i'm pretty sure flatpak can be packaged for it, and thus, everything else can work for it

what this is doing is to make it much harder for weirder systems to ever get flatpak and thus lose a lot of support from those with intent of supporting linux as a platform

@julia @fiore what could actually cause some real issues when it comes to "supporting entire different multiple different OS stacks" could probably be drivers or some weird kernel level stuff, but for the most part, most of the userland essentially disappears for an application that's built for flatpak as a platform

@julia @nelson imean i guess i havent rlly looked into why flatpak would even need to depend on systemd so i dont rlly have anything interesting to reply here

but if the point is cross distro compatibility, deciding to cater only to Some Distros , with a technology that is to be completely honest kinda falling apart on itself (lets face it , systemd is not in its golden days anymore and hasnt been for a while, while alternatives have been getting a lot nicer to use), kinda makes no sense to me ? but idk , i rlly should look more into it i think .

> systemd is not in its golden days anymore and hasnt been for a while

since when and how

@fiore @julia @nelson

@julia @fiore @nelson julia hi this is an L I thibj.

flatpak itself is the abstraction you as the software dev target to make your software work on the weird distros without needing to care about them in specific. this is counter to pretty much half of the entire selling point of flatpak (the other half is the sandboxing)

@pj in any case , i am a happy systemd user on my servers . what im trying to say is that the push towards alternatives is there , and is something many people advocate for, and work actively towards . binding yourself to outdated technology is a bad idea imo..

Ok, so I'm like with you until "outdated technology" because where in the hell is systemd an "outdated technology" (as a project and its author that always has been striving to provide best Linux Desktop experience)?

@fiore

Also to be clear, I'm all for having diversity in software stack and be as compatible and interoperable as possible, but I'm also just a dev that cannot support everything always forever and having less scope and more predictable platforms is extremely better for me

(I'm also an Alpine/pmOS/Chimera user that has been on musl and non-sd init/rc for years)

@fiore

@kopper @julia @nelson this is kinda like what if chromium decided to only work on windows . not rlly but . kinda

@fiore @julia @nelson eh, more like wine only deciding to support wsl

@pj outdated since alternatives have been consistently been able to provide better performance with simpler and more portable systems.

@pj thats when something becomes “outdated” in software

But all those alternatives are worse in terms of providing functional features (as in they lack them). Like I've been there, done that (openrc/dinit) and there's only one project that can somewhat compete with current status quo.

@fiore

@pj init and daemon supervision is not and never has been something an application developer should worry about

No, I'm talking as a developer relating to other developer when e.g. Flatpak devs say they want to use X that is used by majority of users because it makes their job easier and more rigid.

@fiore

@pj uh ok . could you be more specific in which features you are talking about ?

@fiore logind, tmpfiles, sysusers, user services (which are quite a recent thing still in openrc), socket activation, more...

@pj as i said , i have not looked into the specifics of how flatpak would even need this. so i cannot say anything about it

And because of that I hate this whole discourse because people don't really look into the specifics and just pile on because of buzzwords like "systemd"

@fiore

@pj @fiore i still don't quite understand why would flatpak, NOW, depend on systemd even though they worked this hard to have no hard dependencies on pretty much anything except a couple of userland static utilities

Have you seen the talk that was referenced by the original post that spun this whole shitfest?

@nelson @fiore

@pj @fiore

> sysusers

https://git.pinkro.se/Rose/gardenhouse/sysuserd.git/

> tmpfiles

https://git.pinkro.se/Rose/gardenhouse/seedfiles.git/

> socket activation

super-servers have existed for a while, and i'm honestly dubious of the performance claims systemd gives to apply it for every single possible socket -- it feels like a "trick" to shave miliseconds out of boot time more than anything else

eventually i'll be doing performance tests, and if it turns out to actually matter, then something similar will be implemented in openrc

@navi @pj like why would anything even depend on whether socket activation is available this is making me lose my mind

You don't have to really but it's up to you to not use it and not the project that wants to use it

@fiore @navi

@pj @fiore yeah i still don't understand why would a package manager like flatpak require systemd

@nelson Nice draw

@nube thnkx :)

@pj @fiore @julia since they've began using AI to write code and file bug reports

@nelson

adding this to the potential tumblr sexyman list

@noahebalon evil systemd flatpak tumblr sexyman

@nelson

it’s already following few sexyman requirements by being pathetic ngl

@nelson this situation really sucks but your art goes hard as fuck as always keep up the good work

https://blog.sebastianwick.net/posts/so-peerpidfd-gets-more-useful/

@nelson @fiore

@pj @nelson the literal blog post you linked contradicts you

yeah, sorry I don't have all the references to everything that is happening in projects I losely follow, I would need like multiple hours to undertand what is hard to understand from the talk and references but it's not just single thing and people should stop complaining on internet when they don't understand why something is done

@fiore @nelson

@nelson flatpak is hot garbage. I used to be critical of l'eunuchs maintainers in favour of dev packaging their own software, but after looking at what flathub has to offer I changed my mind. Linux maintainers are among the most sane and competent crowd compared to that.

@newt i was a huge flatpak proponent because it makes universal packaging far more doable, i don't have to think about weird distro details because i can ship one binary that just works, in both MR. Ubuntu LTS, and that one distro with dinitctl, musl and freebsd coreutils

it allowed for OS freedom without compromises other than perhaps some extra disk usage and containers being set up temporarily

i guess i should'nt have trusted techbros with the future of "freedom".

(the blogpost is referring to systemd-appd btw.)

@fiore @nelson

this is getting printed and put on my wall its so cute look at him

@fiore big tumblr sexyman vibes. someone should draw him making out with the onceler or himself.

@pj @fiore ... so they're just going to have a hard dep on systemd for this?

@pj anyway, my true issue with all of this was never systemd specifically but just a pattern i've seen repeated in the FOSS community where there's a clear disconnect between users and developers, with this entire thing about AI and LLMs being integrated deeper into the stacks of software we build our "freedom" into

flatpak itself becoming less portable because some people decided it should be that way, then refusing to answer to anything, is what actually gets me very mad. "You are not worth catering to."

@nelson sorry, you really can't have your cake and eat it too. The problem with Flatpak is that what you ship isn't a single binary, it's effectively a docker container but slightly different.

The problem is that people end up vendoring all kinds of dependencies with their packages because there is simply no other (easy) way. For one, VSCode famously shipped its own git. Except it was an outdated and vulnerable version and nobody really cared about it for months.

@newt vendoring dependencies is good in my opinion, keeping everything be a global, shared dependency, just makes stability worse instead of better, since you now have a bunch of different programs having to be in-sync to share one common API

yeah, cases like vscode will always happen, but i feel like that's the job of the mantainer, and much easier to manage than keeping everything stable at an API-level

sadly, we're not in a sane environment where that's even possible, linux sucks, and flatpak, to me, was the best option regarding what we had

... the other options were making mantainers package every app ever and break things along the way, providing "appimages" (ew) or being snap, which is just flatpak but worse(tm)

@julia @fiore @nelson supporting multiple different OS stacks is the entire point of flatpak. otherwise you might as well just use native packages for whatever one distro you support

@nelson also, being on hachyderm is a mental diagnosis. I've never seen someone worthy of having a conversation with on that instance.

@newt i am not sure how but it's like the biggest concentration of fossbros in the internet, it's kind of insane

@nelson vendoring dependencies has both benefits and drawbacks.

My guess here is, there just isn't one size that fits all. I would prefer most of my software installed via semi-traditional means (I use NixOS btw). With most open source packages, this can be rather easily automated. At the same time, some huge pieces of software like vidya from Steam is better isolated as much as possible, especially since it doesn't really need to interact with the rest of the OS that much.

@newt i'm just speaking from the perspective of someone who makes tools

most of the time i can just statically compile fucken everything and let people download binaries or just compile stuff themselves since it's all self-contained because i'm just based like that

but when it comes to making games and similar software, everything gets a bit more complicated because of mesa and other stuff which pretty much has to be dynamically linked

i'm of the school of thought that the best thing here would be to have some system on which we can have a unified platform for all of linux which doesn't dictate any hard rules on how the underlying system works, plus the actual overhead for containers is alright for this kind of thing, it's not like we're running a virtual machine

eh, this entire thing sucks, linux sucks, i'm moving to 9front and disappearing off the face of the earth /j

@nelson they aren't "fossbros", most of them are corporate employees being paid to work on Linux. The instance itself is (was?) run by a Github employee.

@newt LMFAOOOOOO

oh wow i am considering to just defederate altogether

@fiore @nelson @julia

yes that’s the point of flatpak

however… i don’t think flatpak should start to require systemd like it requires the linux kernel to function

i have been using flatpak for a long time without systemd and the bit about not requiring nothing more than a few common librairies to run is precisely the reason i chose flatpak

@stilic @fiore @julia yeah it's awesome that you can just port a specific piece of software to your weirdo little distro and as long as all pieces are in place, boom, you now have support for all the regular software you'd use on any other system

@nelson@wetdry.world @julia postmarketOS switched to systemd

@tragivictoria @julia that project isn't finished and they will continue supporting OpenRC.

@alexia @julia that's because existing installs couldn't be migrated. All new installs are using systemd

@tragivictoria @julia No, new installs only use Systemd by default for GNOME, Plasma and derivatives. Unless that changed within the last few weeks.

@julia @tragivictoria There's also some devices for which, reportedly, systemd wouldn't be possible for. Though this is something I've heard about I've not yet seen it noted on any wiki.

@kopper @fiore @julia @nelson@wetdry.world yeah and that abstraction doesn't need to support weird init systems nobody uses.

Seriously, everyone here talks as if flatpak was about to cut off support from half of it userbase. Non-systemd users are minority of minorities, all major distros moved to systemd decade ago. At this point keeping support for them is holding the vast majority back.

@tragivictoria @fiore @julia from what I'm aware this systems support will be done by adding a new component to systemd (systemd-appd). I would really be interested in knowing why that has to live in the systemd monorepo

@kopper @fiore @julia @tragivictoria

The talk is on youtube and only around 30 minutes long https://youtu.be/1AXBfsiaQNk?t=16221

You might need to do some googling on subsandboxing if you don't know what that is, I didn't and had to do so.

So after that googling it took me roughly 45 minutes to get an idea on why it was useful.

@fiore @julia @nelson Flatpak as a project in 2025 was basically completely stagnant because there were not enough contributors, not enough maintainers to review PRs etc. . So I'm uncertain on the community doing much here.

This situation only somewhat changed in 2026. Apparently this particular move is informed by systemd-appd allowing for subsandboxing (which should remove some issues with flatpak'd browsers and steam) as well as better permission management.

@Isofruit @fiore @julia yeah it's some cool stuff, hopefully they get it working without systemd because it'd be a bummer to lose universality with flatpak since it's pretty much awesome as it is right now

@nelson @fiore @julia I'm not sure they can?
I mean, you'd need:

- Something that associates running processes/apps with an id to group them together (e.g. identify all processes related to firefox)
- Something that can store permissions based on that id
- Have every systemd.service that would've been called from e.g. portals check those permissions when permission question comes up
- Maybe more

Sounds like work. Though I'm not deep in systemd to make that statement with any validity.

@Isofruit @fiore @julia

a good first step would be to make it optional, temporarily, so that implementations can catch up to it, because this is legit some cool stuff, again, the issue is losing universality

@nelson For one it's not even released yet.

For another, those distros could just not update their flatpak package for the time being until they have systemd-shims ready, I assume?

It's not like flatpak was an insanely active package before and if distros were not using systemd I'd assume they had a decent amount of systemd-shims ready already for other purposes..

@Isofruit oh it's not that simple, these people have spent a good while cloning and abstracting specific systemd details, like implementing alternative dbus servers and logind and all that magic stuff

these distros rely a lot on flatpak because a lot of things just don't work "natively" in them, specifically distros like alpine or musl void, since their libc isn't even the same, and flatpak provides an amazing, userland-agnostic platform that just works, no matter if you're on fedora or in some weirdo distro only 3 people in the planet use

appd might take a while to reimplement, and hanging onto old flatpak versions is somewhat unsafe. thankfully, i've heard that this thing will actually just be a complement, where you will just have two versions of flatpak at once, from which they will slowly but surely upgrade from one to the other

@kopper @fiore @julia @nelson so are we saying flatpack is like java (or at least, the selling point of java)?? runs on 8 billion devices, target an abstraction (JVM) which itself runs everywhere?? (light hearted, but also kinda making a genuine connection in my mind)

@goes2hard @fiore @julia @kopper its more like a zip file full of everything an app could ever need, or like an artificial operating system where your app can exist without being aware of the wider operating system, either to isolate it, or to give it room to exist comfortably

@kopper @julia @fiore @goes2hard @nelson

everyone in this thread actually go fucking outside and touch grass istfg

@kopper @goes2hard @fiore @goog @julia ok

@goog @fiore @julia @kopper @nelson never. grass killed my grandma

@fiore @julia @goes2hard @goog @kopper grass turned my dog into string cheese.

@kopper @julia @fiore @goes2hard @nelson

i can tell you didn’t grow up with enough guidance in your life

@fiore @goes2hard @julia @kopper @goog man did you show up just to be aggressive on a thread that doesn’t even involve you or anything you care about

@nelson Oh btw it gets funnier! The original guy from the flatpak thread is a vibecoder!

@nelson god i love unix surrealism

@tajoma same.

@nelson sorry to hear you got so much backlash over it. I think it started an important debate, and I'm glad you made it for awareness.

@anthropy thank you.

"the Flatpak fandom is dying.
repost if you're a true flatty."

#unix_surrealism

https://wetdry.world/@nelson/116625964273423922

@julia @fiore @nelson >supporting entire different multiple different OS stacks is contrary to that goal

if that's the case, then why doesn't flatpak exclusively support immutable distros? bad card

@fiore @julia @tragivictoria like I dont think this is about the init system part of systemd

@kopper @fiore @julia systemd is not only init system. It's a suite of tools

@tragivictoria @fiore @julia what makes you think I am not aware on that considering my use of "the systemd monorepo" and "the init system part of systemd"

@nelson yeah, me too. I work on Linux professionally (embedded stuff mostly).

Static linking on Linux is dead. Glibc doesn't support it, it used dlopen() left and right. Same with many libraries like Qt. You just have no other choice, should you use this software.

Linux has always sucked. Before systemd, there were other drama queen software packages. The only thing that's new here is that now corpos like Microsoft and IBM run the entire show.

@newt musl my beloved and glibc my behated forever

even the fucking x11 protocol implementations have hard dependencies on dlopen so you can't even statically compile them naturally

linux sucks, and has always sucked, and will probably always sucks, but everything around me seems to suck harder lmfao

@kopper @fiore @julia what was the point of that question then? Like if you know that, then why ask

@fiore @julia @nelson@wetdry.world what contributors?

@tragivictoria @julia wdym?

@fiore @julia you're saying this move alienates contributors. What contributors?

@tragivictoria the ones not on systemd???

@fiore they're not big on flatpak anyway

Anti-systemd people are also the ones who often hate flatpaks (and wayland while we're at it)

@pj @fiore @julia @nelson as a user, i do feel like the project has ballooned a bit too much, taking on features that feel like they don’t need to be tied to an init system and/or service manager (such as systemd-ukify)– this means these features depend on libsystemd when they could work independently of systemd, which makes me feel a bit too “locked in” to the project (bus factor and all that)
that, while still having long standing issues (extremely vague log messages such as “A stop job is running …”, like, could you tell me at least what is holding up shutdown..?) has made me loose some faith in the project

@newt @nelson

> For one, VSCode famously shipped its own git. Except it was an outdated and vulnerable version and nobody really cared about it for months.

I'm not arguing in defense of VSCode for doing that because it's their problem but I imagine that's because a lot of distros, like Debian for example, doesn't comes with git pre-installed, so a lot of ""devs"" that doesn't know linux and install everything via snap or flatpak gets would complain that git doesn't work and instead of blaming them for being stupid would blame on vscode

@kumicota @nelson yeah, it doesn't matter why. Dependencies are a huge problem in Flatpak. Turns out, some software just cannot be shipped as a self-contained package without creating more problems.

@tragivictoria @fiore @julia that was not the question I was asking, but it doesn't matter because I've talked to a few people and this all seems a bit too overblown based on poor communication from some flatpak people, so there's no real reason to continue this conversation

@tragivictoria @kopper i literally dont understand why youre being si hostile to the discussion , both me and kopper arent “systemd bad” reactionary assholes , we are bringing up valud concerns that you are just dismissing as “youre the minority, you dont matter”.

@pj > logind, user services

there are alternative implementations that despite not being complete yet, are already in use by lots of people and have already proven that a standalone tool is better for this job (turnstile, for example)

> sysusers

what even is that

> tmpfiles

why is this even in systemd

@nelson this is the hachyderm founder https://hachyderm.io/@nova

I guess, she isn't around any more and other people run the server now.

@tragivictoria ??? what kinda strawman is this??

@nelson

I also am in the process of researching / tinkering to removing SystemD from my Debian systems.

This reinforces my desire to remove flatpaks altogether from my machines.

I'll keep it to either Debian packages or appimages.

@tragivictoria @kopper @fiore @julia

> nobody uses

by how many people got mad at the concept, by how many people use gentoo or alpine or chimera or void or (ew) artix/devuan or such on their machines

it's not nobody, i'd be surprised if it's not at least a few hundreds of thousands of people minimum -- though a user count is unobtainable given the nature of open source

and it's not like not hard-requiring systemd is holding anything *back* -- all the primitives necessary are in the kernel, there's no requirement for real systemd except "we want to use it bc it's """everywhere""""

@fiore @tragivictoria "users of service managers other than mine are all chuddy suckless reactionaries" mother of all copes

@kopper @fiore @julia yeah it's mostly about cgroups.

@kopper @fiore @julia which, AFAIK, can also be implemented by other init systems?

@fiore @kopper @tragivictoria

In my experience so much as criticizing systemd minorly can be enough to get accusations of being a schending Nazi who hates Linux and wants to destroy the desktop and whatever other gibberish people decide to throw at you.

@tragivictoria @kopper @fiore @julia I find it deeply ironic to talk about how there is no need to support "weird" system configurations "nobody" uses when we're talking about Linux, a "weird" system that is neglegted by most people because "nobody" uses it

@LivingCooki @fiore @julia @kopper Welp, it's the same thing, is this not? I'm not being a dick to app devs who don't support linux. It's called knowing own place. They don't need to support linux and that's fine really, especially considering our market share is miniscule.

@LivingCooki @fiore @julia @kopper people forget linux support is a bonus, not a requirement

@nelson I guess you got baited

@razze ja i even made cool ass art about it

can't even be mad at myself for that tbh

i just found myself an inmeasurable new hate for white floss linkedin guys on fedi

@julia this is the "linux users always have a stick up their ass about people not building for linux anyways, it wont make a difference" opinion from not even 10 years ago