- cross-posted to:
- [email protected]
- [email protected]
- cross-posted to:
- [email protected]
- [email protected]
Timothée Besset, a software engineer who works on the Steam client for Valve, took to Mastodon this week to reveal: “Valve is seeing an increasing number of bug reports for issues caused by Canonical’s repackaging of the Steam client through snap”.
“We are not involved with the snap repackaging. It has a lot of issues”, Besset adds, noting that “the best way to install Steam on Debian and derivative operating systems is to […] use the official .deb”.
Those who don’t want to use the official Deb package are instead asked to ‘consider the Flatpak version’ — though like Canonical’s Steam snap the Steam Flatpak is also unofficial, and no directly supported by Valve.
You must log in or register to comment.
Canonical’s SteamSnap is Causing Headachesfor ValveThanos snapped the uptime
-
Ubuntu wants to own snap, with their own proprietary store etc which runs against alternatives like Flatpak and goes against the FOS ethos
-
Snap is slower and worse than Flatpak (the most popular alternative) in most ways, with very few pros that will likely be caught-up-to soon too
-
Let me simplify that:
Canonical’s SteamSnap is Causing Headachesfor ValveUbuntu used to get a lot of undeserved hate but lately the hate feels deserved. Ubuntu has been the face of the usable desktop Linux for a long time and they just keep tripping over themselves every time they try to move forward.
Their intentions are usually good. A lot of things they propose usually end up being adopted by the community at large (just not their implementation). They seem to just yank everyone’s chain a little too hard in the direction we’re eventually going to go and we all resent them for that.
Off the top of my head, there was Upstart (init system), there was unity (desktop), and now snaps (containerized packaging). All of these were good ideas but implemented poorly and with a general lack of support from the community. In almost each case in the past what’s happened is that once they run out of developers who champion the tech, they eventually get onboard with whatever Debian and Rhel are doing once they were caught up and settled.
Valve’s lack of interest in maintaining the snap makes sense. The development on the Ubuntu platform is very opinionated in a way where the developers of the software (valve) really want nothing to do with Canonicals snaps.
On another note: my favorite thing about the Ubuntu server was LXD + ZFS integration. Both have been snapified. It was incredibly useful and stable. Stephane Graber has forked the project now into INCUS. It looks very promising.
The ZFS stuff was exciting! Are they still incorporating zfs in current releases though?
I think they are! I’m still trying to do more with ZFS everyday.
I do think the idea behind snap isn’t all about pushing the Linux platform as such forward, but to specifically gain a market advantage to Ubuntu.
Why else is finding documentation for changing the default store so difficult? And I don’t think you can even have multiple “repositories” there–quite unlike all other Linux packaging systems out there. (Corrections welcome!)
Wasn’t there MIR as well?
Yeah, I think as the replacement for x before Wayland?
LXD got better with the AGPL license, so Canonical did the right thing there.
(I know they added a CLA but it’s still way better than the permissive license they had before)
This might be an unpopular opinion but I really don’t get this trend of wanting to containerized just about everything, it feels like a FOTM rather than doing something that makes sense.
I mean, containers are fantastic tools and can help solve compatibility problems and make things more secure, especially on servers, but putting everything into containers on the desktop doesn’t make any sense to me.
One of the big advantages Linux always had over Windows is shared components, so packages are much smaller and updating the whole system is way faster, if every single application comes with its own stuff (like it does on Windows) you lose that advantage.
Ubuntu’s obsession with snaps is one of the reasons I stopped using it years ago, I don’t want containers forced upon me, I want to be free to decide if/when to use them (I prefer flatpack and appimage).
Debian derivatives that don’t “reinvent the wheel” is the way to go for me, I’ve been using Linux MX on my gaming desktop and LMDE on laptop for years and I couldn’t be happier, no problem whatsoever with Steam either.
I agree with a lot of your points but I do think containers a great solution.
I’ve been a really big fan of Universal Blue lately. It presents a strong argument for containerizing everything. Your core is immutable and atomic which makes upgrades seamless. User land lives in a container and just gets layered back on top afterwards.
Shared components work brilliantly in a fantasy world where nothing uses new features of a library or depends on bug fixes in new versions of a library, and no library ever has releases with regressions or updates that change the API. That’s not the case, though, so often there’ll exist no single version of a dependency that makes all the software on your machine actually compile and be minimally buggy. If you’re lucky, downstream packagers will make different packages for different versions of things they know cause this kind of problem so they can be installed side by side, or maintain a collection of patches to create a version that makes everything work even though no actual release would, but sometimes they do things like remove version range checks from CMake so things build, but don’t even end up running.
Shared containers work beautifully for a lot of things, though, many programs aren’t all that sensitive either. Making snaps for the tricky ones makes sense. Having snaps for all of them is ridiculous.
I can count the software requiring repo-pins on one hand on my desktop. For those, snaps make sense, replacing the need for any pins. Snaps are less confusing than pins. IMO.
It reminds me of Python programming, with requirements pinned to version ranges. Some dev-teams forget, and their apps won’t work out of the box. Sometimes, software still works ten years later, if they only use the most common arguments and commands from the packages.
Snaps <==> Virtualenv.
Kinda saw this coming sooner or later.
I remember asking in one of their articles if they had planned to reign over (or partner up) the project over to Valve once it was ready and said they had no plans.
That’s the problem with doing everything yourself.
You also have to maintain everything, yourself.
Fuck snaps 🖕
Yeah. Hey wait? Fuck you!
I’m sorry but Linux Council has already decided your fate.
Are you threatening me master jedi?
I am the Linux Council!
Oh yeah? Then I’ll make my own Linux Council! With blackjack and hookers!
I thought it was the GNU wizards circle that decides these things.
Are you telling me I have been going to the wrong meetings?!
I swear this Linux fragmentation will be the death of it.
🫰Fuck 🫰Snaps 🫰
Good. Snap is an abomination.
I know the “Arch BTW” meme exists for a reason, but one of the reasons I haven’t been able to drag myself away from Arch-based distros in recent years is that it allows me to always have current versions of my software while also just not having to care about all this appimage/flatpak/snap brouhaha.
I guess it’s somewhat of a “pick your poison” kind of situation, but I find dealing with the typical complaints about Arch based distros to be both less of a problem than detractors would have you believe, and less of a headache than having to pick one of three competing alternative packaging approaches, or worse, to use a mix of them all. Standing on the sidelines of the topic it seems like a small number of people really like that these options exist, and I’m happy for those people. But mostly I’m grateful that I don’t have to care about this kind of thing.
Edited to add: Seeing how this thread has developed in the past 5 hours convinces me anew that “on the sidelines” is where I want to stay on this topic. 😁
Yup, the AUR is a godsend. I barely touch the other methods and only use AppImages when I’m too lazy.
Appimages when you’re lazy? The fact you have to chmod them is annoying compared to an AUR helper
I mean Dolphin gives a prompt for the same thing when I launch it.
I’ve always found the most time consuming thing about arch is having to spend half your life telling everyone you use it.
I haven’t used it in years, so hard to remember now.
Nah, it’s repeating the installation process until you finally get enough stuff working to have internet, and then you can bootstrap installing every other bit of software that you need. Thank goodness for rolling release - I can’t imagine having to go through that again.
You can install Arch directly from a UEFI shell over the Internet: https://archlinux.org/releng/netboot/
If your BIOS has a UEFI shell that supports DHCP, HTTP and IPv4 PXE you can load the ipxe-arch.efi over HTTP and start installing.Does UEFI shell have wget?
Depends on the version. All of them (the newer ones with networking) have TFTP. Some even have HTTPS. I think HP Servers even have HTTPS-Boot with client TLS certificates.
None of it works with Wifi though. iPXE has wifi support for some devices but you obviously can’t start it over the Internet. You need to flash a ROM you don’t need or use a USB drive to load it. Then you can boot Linux from the Internet. (That also works if you don’t have a UEFI Shell in BIOS). https://netboot.xyz can also boot other OSes than Arch.The important question is if the UEFI shell can run Doom.
Of course it can. What did you expect?
I haven’t done a vanilla arch install for years either, because if that sort of thing is fun for some folks great, but it was only fun once or twice for me. I think a lot of the vanilla arch faithful end up scripting it for fresh installs.
But, FWIW there’s always endeavouros, manjaro, and I’m sure other Arch derivatives I’ve forgotten about. I just did an endeavouros install on new hardware I was given last weekend, and it’s certainly no harder to install than Ubuntu.
I think if we could drag users (at least gamers) away from these Debian/Ubuntu based distros we could have developers just shipping packages that wouldn’t need to be compatible with some ancient LTS library release, and maybe we wouldn’t need appimage/flatpak/snap at all anymore (or at least only in rare cases).
From my perspective as someone who is both getting into gaming on Linux and also not much of a power user, Arch would have to make the installation and maintenance process a lot simpler to attract more people, and I’m not sure that’s something they actually want to do.
Looking at the official Arch installation guide, the average gamer may be overwhelmed by the process here, especially if they’re not comfortable with the terminal. Something like Linux Mint, on the other hand, has a built-in GUI installer with reasonable partitioning defaults, and it comes packaged with stuff like an app manger and update manager, something that will feel much more familiar to someone coming from windows.
If you want, or are interested in looking at an easier to manage Arch install I would suggest CachyOS, EndeavorOS, or Garuda Linux.
Debian catching strays over Ubuntu’s ubullshit.
I think Debian has a place in the Desktop market, it’s just not gamers or anyone wanting anything new (unless they of course go the flatpak route). Not a perfect analogy, but it’s kinda like gaming on Windows 7 these days because it “just works” for you. Sure you can, but you’re not getting the best of anything that way and all the underlying libraries are outdated and some things just aren’t going to work at all.
Well, the only thing holding me from switching to Debian is the absence of up to date KDE packages…
What about in unstable or testing? I moved to Arch from Debian because I wanted faster releases and it just made sense to move to rolling instead of testing Debian install.
Debian-trixie is pretty up to date and flatpak fills the holes perfectly.
100% all this. Canonical has been pushing snaps for awhile, and I wonder if the 12 year LTS for Ubuntu is part of that strategy - want something newer? It’s in the snap store. snap is terrible, worse than flakpak and appimage - but just as you say, as an arch user I don’t have to care. Whatever I want is probably in the AUR if not the main repos. Rolling distros, done right (arch), are an amazing experience.
I was like you using arch packages for everything until ferdium was hit by a terrible bug that broke its zoom function and wont be fixed in months since it is an issue with electron. Now I use the appimage version of it to downgrade to an older version.
And then there are the rust programs that you can only find as aur git packages which means installing a bunch of dependencies and all the crap that cargo puts into my home dir just to build one binary, for that I just instead take the appimage version or sometimes the binary if they release it and place it in ~/.local/bin.
Another good thing is that the appimages get compressed and take less disk space, for example the libreoffice package in arch is 600 MiB while the appimage is 300 MiB.
And the great thing is that I can just drop my homedir into any distro and as long as I make sure that fuse is installed everything will work out of the box.
It sounds like NixOS would solve all your problems. And it makes coffee too!
Still figuring out Nix for my daily driver (too many fun customizations moving away from gnome), but love it so far and built my last self-hosted box with it. So easy to redeploy after the initial learning curve. Boot from ISO, partition with disko, upload SSH key to decrypt sops, install.
Oh and since practically everything can be overridden, its easy to get bleeding edge using existing packages!
Nix is objectively a bad language. Glancing past the syntax (which is atrocious and the parser is completely unhelpful) it’s just a bad idea to mix lazy evaluation and dynamic typing, the only way to get understandable error messages with nix is to put asserts all over the place, hoping one of them gets triggered by the evaluator.
That said I still love the system to bits and fixing the language actually wouldn’t be that hard: Add static typing, with inference most stuff should just run as-is. Exception are the two or three uses of the y-combinator (yes, seriously, that’s how you recurse in nix) deep in the bowels of nixpkgs, e.g. during stdenv bootstrap. Just make
fixa primitive, or maybe a plain fold I don’t think nix even needs to be Turing-complete. Reason it’s not being done is that there’s enough other stuff to work on and, well, Stockholm syndrome. Good news: Nix doesn’t support constructing import paths dynamically, that would have been a nightmare to sort out in a static setting.I don’t mind most of the language having FP experience, but I agree the lack of static typing just sucks. I’m using the repl a lot to try and track down why things aren’t aligning quite right when trying different techniques to keep things DRY and organized. Documentation is also a headache as a newb with the multiple ways of structuring a repo while trying to grok all the implicits and how it all gets merged.
Is NixOS xdg base dir compliant? that is one of the reasons I don’t use flatpak.
Yes, all the environment variables are set automatically and programs respect them.
As someone who knows nothing about Arch, what do you do if your app exists only as a .deb file? Can you install it?
I know you already got another answer, but - it’s very rare that any software for Linux exists that is both 1) not present in the official Arch repos, and also 2) not packaged by a user for the AUR.
Probably 99% of what a typical user (I know we all define that differently) will want doesn’t even require AUR access - it will be in the official Arch repositories and will be up to date to within a few weeks of release.
There are some potentially substantial downsides to the AUR (it’s the Arch USER Repository - so these are not official arch packages) but IME the real world problems are minimal. I would suggest anyone who is new to the Arch way of distributing software should hit up the relevant page on the Arch wiki and make up their own mind before using the AUR - but it’s about being aware of what you are doing more than it is a real warning, if that makes sense. I suspect few Arch or Arch-based users don’t have at least a smidgen of AUR packages on their system. (Edit: That page is very thorough. I think it makes installing from the AUR sound much harder than it needs to be. For most people the command is just “yay -S packagename.” There are gui options that handle all packages including AUR, and yay is not the only cli option, either.)
Interestingly, there are some AUR packages that work by pulling down the deb and deconstructing it for installation on your system - AFAIK it can be that, RPM, a true “compile from source” situation, or I’m guessing some AUR packages are deconstructing snaps\flatpaks\appimages during the install. Whatever the origin of the files, they include a pkgbuild file that tells your system how to either compile or deconstruct and install the software.
I know I ran into this years ago. I think it was some collection manager app for a trading card game that someone had on GitHub and only had .deb releases. Eventually you will want to install something niche.
I have found no such instances. Software which is only officially packaged as deb will usually be unofficially repackaged on the AUR regardless.
I distribute an app I made for Linux, macOS and Windows. The Linux version I only have available as a .deb. Released recently and has about 200 users so far, but definitely exists. No Arch user contacted me yet.
When folks start wanting it, someone will package it for the AUR, or if it becomes even somewhat mainstream it will end up in the main repos eventually.
Possibly, though I wonder how updates would work then. Currently I have a Debian repository that contains a single package, and installing the .deb from my website also installs the repository so you get updates as with any other package.
If someone repackages it on AUR, I guess they will also need to update it every time I update the .deb, so it’s always behind? Of course it would be better if I provided a first party package for AUR, but I can spend time on that when there is actual interest. Most of my users are on Windows anyway.
Usually what you describe is what they do. Someone “owns” the AUR package (and it’s not quite literally any random user IIRC - you have to be accepted as an AUR maintainer I think) and they then take on the responsibility to repackage it whenever the author (you) releases a new version. There is also a mechanism for users to flag the AUR package as out of date in case that maintainer misses a release, and if they abandon it (or even if folks just don’t like how they package it) someone else can package it, assuming someone else wants to.
Sometimes the AUR maintainer is the dev themselves. I can’t think of a good example currently, but I know I’ve seen it before.
I don’t know the process for how things end up in the official repos, but I would guess it’s similar to however any other distros identify software they want to officially package.
They may not have to. For example, Plex on nixos just unpacks the deb and installs the files the “nix” way.
It isn’t recommended, but dpkg will install it if you really want to. You just need to handle dependencies manually.
But it’s a pretty rare issue. If something isn’t available in the official repo, AUR probably has it.
The problem is that 3rd parties are doing the packaging both on Snap and Flatpak whereas if they had followed proper security practice ONLY THE REAL DEV should ever be allowed to package their app as a Flatpak or Snap.
This would ensure security, as well as a proper functioning flatpak/snap and also all feedback would be directed to the Dev.
I’ve never liked the fact that Canonical and whoever can make Snaps and Flatpaks of other people’s software. There is zero security guarantee, zero guarantee they’ll update it and zero guarantee it will work.
Just because Snap and Flatpak exist doesn’t mean just anyone should be able to just make them.
If Valve only chooses to make a deb then so be it! It’s their product!
For security reasons the packaging of flatpaks in flathub is done by flathub, whether they are devs or third parties they just write the manifest. Although I seem to remember there are some exceptions, such as firefox.
Ah, I was not aware of that. Ok, that’s good to hear because it potentially adds a layer of security.
Any idea whether they vet the code as well?
Not that I know of. But you may be interested that it requires prior authorization to modify manifests.
Excellent. Thank you
deleted by creator
How so? How does ensuring they only the real dev of the app is also the only one allowed to package it hurt desktop adoption.
It’s very easy to enforce. Flathub need to verify the identity of the person submitting the Flatpak to make sure it’s the app’s dev uploading it and not Joe Smith or nsa.gov…
The problem is that 3rd parties are doing the packaging both on Snap and Flatpak whereas if they had followed proper security practice ONLY THE REAL DEV should ever be allowed to package their app as a Flatpak or Snap.
Says who? If it were the case, Linux would either be a nightmare of fragmentation or become centralised on one distribution. Distros need to be able to package their own software, and these are kind of like distributions. Also since we’re talking about proprietary software here, is it really any better security practice if the “real dev” packages it or somebody else, they both could contain malicious code.
Valve are not going to put malicious code on their app. Neither is VLC or any other FOSS developer.
The distros should stick to packaging their repo apps and leave the Snap/FlatPak tech as an alternative to the original dev if they decide they want to use that.
We can’t have Bob from nowhere packaging Valve, then not updating it or patching it because he doesn’t have time. Or 5 Bob’s all doing the same thing with 5 copies of Valve on the Store.
It’s crazy. This is what causes fragmentation. Flathub should vet every app and if you are not the dev of the app, you may not host it on Flathub. You’re still welcome to make a Flatpak for home use on your own pc but not for wide distribution.
isn’t that kind of what AUR is, and exactly what people love about arch ?
Yes but if you use an Arch distro like Endeavour, they won’t support you with issues caused by AUR apps. Because of these reasons I mentioned.
How is “the dev of the app” defined, exaxtly?
The official Developer of the app. E.g. the official dev of Blender is blender.org. The flatpak people give them a line of code to embed in their website and they use that to verify that the dev really is blender.org and not a malicious actor.
Ah, that makes sense. Inwas hung up on the word, interpreting it as a single guy, not an entity. Thank you.
Valve are not going to put malicious code on their app. Neither is VLC or any other FOSS developer.
How would you know that? It’s not like it’s something that doesn’t happen.
Or 5 Bob’s all doing the same thing with 5 copies of Valve on the Store.
It’s crazy. This is what causes fragmentation.
I don’t know what snaps are like but that’s clearly a non-existent problem on Flathub.
Flathub should vet every app and if you are not the dev of the app, you may not host it on Flathub. You’re still welcome to make a Flatpak for home use on your own pc but not for wide distribution.
I don’t know why you feel like there’s permission involved. You don’t have to use Flathub, therefore Flathub can have what ever policies it likes. Users can set up a different flatpak repo if there’s a need.
That’s not my point. I use Flathub but I try to only use verified apps which were packaged by the actual dev.
I’d rather get a deb from the official dev than a flatpak from flathub packaged by someone who is essentially anonymous and could easily inject malicious code.
If you think the dev himself could inject malicious code in the official app, then you should be super aware that an anonymous Joe can too, and is far more likely to.
Anyway flatpak ideally was supposed to save Devs the work of packaging for every distro so it makes sense that the real actual verified dev of the app would package the flatpak/snap himself
Wait until you find out how distro packaging works
I’m sure Canonical’s neverending death march towards Snap, along with the OS running outdated packages, is why Valve no longer uses Ubuntu for SteamOS development. The greatest April Fools was Ubuntu dropping Snaps because so many people were saying how they could go back to using Ubuntu again…then they noticed it was a joke and the sadness set in.
Why do people hate snap over flatpak? I feel like I’ve read a thread or two about it, but I haven’t seen an answer that was particularly satisfying (almost definitely for a lack of trying on my part, to be clear).
-
Proprietary on the server/distribution end
-
Controlled 100% by Canonical
-
Worse performance, particularly in terms of app startup times
-
Snaps are mounted as separate filesystems, so it can make things look cluttered in your file explorer or when you’re listing stuff with lsblk
-
Canonical often forces users to use Snaps even when users have explicitly tried to install with apt. e.g. you run sudo apt install firefox and it installs a Snap
-
It hasn’t gained traction with other distros like Flatpak has, and Canonical’s insistence on backing the “wrong” standard means Linux will continue to be more fragmented than it would be if they also went along with what has become the de facto standard
There are however benefits of snaps. It works for better for terminal programs, and Canonical can even package system stuff like the kernel as a snap - as you can imagine, this might be a very powerful tool when it comes to an immutable version of Ubuntu.
Snap startup times are awful, tens of seconds to open a simple text editor, even on an nvme ssd…
edit: Also it doesnt bother following XDG specifications, further cluttering our home folders.
Proprietary on the server/distribution end
Zoinks!
Snaps just act strange. They update in weird ways, it’s always automatic and it’s confusing how to keep something in a version that won’t auto update. It’s been a bad experience for me.
-
- Flatpak is open source, Snap isn’t
- Flatpak allows other repositories besides the official one, therefore having the ability to be decentralised, Snap doesn’t
- Canonical (the company behind Snap and Ubuntu) is hated for some past decisions they made with Ubuntu
- and more
(The only thing I really prefer Snap over Flatpak is that you need the whole package name in Flatpak (like com.valvesoftware.Steam for Steam) whilst you can simply use “steam” in snap but that’s due to decentralisation vs centralisation I guess and overall a minor problem for me)
deleted by creator
I was certain you had to be joking in this post, holy shit.
That’s gotta be the funniest backfire for an April Fools’ joke I’ve seen in a while lmao
And still is, as Google still has it on the first page of results for “Ubuntu without snaps”.
Don’t they own the Code?
Can’t they just cease-and-desist them if they cause them trouble?
They could and they would if they wouldn’t profit from this in the end,
for some reasons there is no official Flatpak and they don’t want to support a Snap package, they just say anything but the *.deb is unsupported, kinda weird because they use the Flatpak package on Steamdeck because that is Arch-based, i guess they are somewhat involved there.
As much as they do for the Linux movement, they should get their shit together when it comes to a cross-distro client, preferably Flatpak obviously.
Do they use the steam flatpak, or just allow flatpaks to be used? There’s a difference.
the latter
This is a big issue with Snap. It may be like Flatpak, allowing devs to set their own dependencies for ALL distros, but its poor uptake outside of Ubuntu’s ecosystem means that it’s no different to yet another distro repackaging system.
Flatpak, or even Nixpkgs, are the future because they allow devs to have control over the distribution of their software. Snap being such a closed ecosystem in comparison only means it will replicate many of the problems we’ve found with traditional (re)packaging systems.
I can’t speak for Flatpak as I haven’t tried it but nixpkgs are beautiful to work with and configuration of my system has become completely reproducible in a clean format.
As a dev, you can just distribute a nixpkg with whatever build tool inside. That beats the current system of “native” packages where your software is repacked and then maintained by half a dozen teams for different distros that use different dependencies and update cadences.
Bottles has gone as far as to demand its fedora package be removed and now shows a warning if you’re not using the flatpak version because repackers just don’t properly test all their software (how can they? there are thousands of apps in these repos!)
Yeah there are some issues with compatibility, I’ve found a couple of apps that error on my Mac.
How does it compare to Flatpak?
nix is a “native” packaging format. Apps are compiled for your host OS and run in that environment with no restrictions, for better or worse.
Flatpaks are containers. They provide a virtual OS to the application such as the file system, and accessing host OS features is done through “portals” which just means you can give/revoke the ability of the app to access your host OS resources such as networking, file access etc.
Flatpaks are therefore much safer in theory. But Nix packages are lower overhead, and can interact like any built-in software binary that you’d have when you spin up a fresh install of, say, debian.
Nix packages are harder to use IMO thanks to their poor documentation and lack of GUI package manager support (not that it’s impossible, just that it’s been a niche system for most of its life) and since most people are accustomed to flatpaks and their permissions system (and the fact it comes preinstalled on most distros) so flatpak is still pretty ubiquitous, even for NIxOS users
Tbh i never found an app that runs better on snap than on deb
Same goes for almost anything like snap
Just to play devils advocate, why don’t they simply officially support the Snap store?
Because it’s the same story as with Mir or Upstart: it will die, because its half assed and tailored to Ubuntu, this time with dubious non-free parts even
The proprietary parts is what bothers me. Why would you make a foss OS/fork and put proprietary shit in? Its like taking something good and pure and instead of making it better or just different, make it worse.
Snaps just aren’t ready yet.
Bc/ they’re already packaging an deb-package. Why should the do that snap thingy?
- Nobody needs it
- It’s a lot of work
guys hear me out. steam os: debian edition
The original SteamOS is based on Debian, https://store.steampowered.com/steamos
I guess backporting everything was a pain.
SteamOS 2.0 which is used on Steamdeck (and only available on Steamdeck officially) has nothing to do with that.
TIL
This is not a problem with steam OS though
I’m really hoping this all forces Ubuntu out as the face of desktop Linux.
It’s been pretty low tier for years now, and Canonical just proves corporate backing doesn’t guarantee a good distro.
Snap is pretty garbage, default GNOME is horrendous, the repos break every other month, apt is still pretty lame despite being an user upgrade for apt-get, the packages are neither stable nor cutting edge, they change core OS backends like every update which breaks configs and makes documentation obsolete.
I’d like to suggest Fedora as the new goto, but I feel like it’s a bit too privacy and FOSS oriented which may scare away new users.
Debian is great but it doesn’t have latest packages which isn’t optimal as performance upgrades would take time to release or need to be manually installed.
unfortunately, industry loves shit like Ubuntu and RHEL because of their corporate backing. comps love having the insurance of someone to blame or somebody to fix their shit when things hit the fan. I’ve worked for many comps who choose RHEL for that alone. Should we choose the OS built by a bunch of randos in their basement, or something backed by Red Hat where I can just pay them money to handle my support tickets faster if shit blows up? or who tf do I have my cyber liabilities insurance guys sue if the OS has a huge fuckin problem? I want a company behind that shit.
Well, I’d prefer Canonical to fix their shit, instead of forcing immature products onto users. I’m not against snap per se, as there are valid reasons for sandboxing, especially for games (remember when Steam accidentally wiped some user’s home folders back in 2015? Sandboxing would have prevented that).
However, in its current state, snap causes just too much friction. For example Firefox can’t remember the last used directory for up/downloads, Steam snap will just create a new data directory (forgetting about the games already downloaded), there’s no way to allow additional folders (like /net from autofs) in snap apps etc. It’s just a myriad of issues which make working with the system unnecessarily complex and frustrating, and there seems to be little progress fixing those.
Fedora is the best!
(In my opinion)
