How to improve Python packaging, or why fourteen tools are at least twelve too many
Still relevant.
- Posts
- 21
- Comments
- 418
- Joined
- 3 yr. ago
- Posts
- 21
- Comments
- 418
- Joined
- 3 yr. ago
Opensuse doesn't have rpm-ostree. Their immutable offerings are just snapper/btrfs snapshots before changes to the system.
Such a setup is nowhere near as powerful. rpm-ostree can rebase itself based off of a container/oci image. It can layer images on top of eachother. Rather than just tracking when changes happened, it can also track what change happened, in a git style setup.
The whole point of it is that in a truly random system all known patterns should eventually emerge somewhere within it.
So pi (probably) has this property. There are some joke compression programs around this (they don't really work because it takes up more space to store where something in pi is, than storing the thing itself). But it is funny, to think that pi could theoretically hold every past, present, and future piece of information within those digits after the decimal.
https://github.com/philipl/pifs
https://ntietz.com/blog/why-we-cant-compress-messages-with-pi/
A new k8s cluster was created and planned to replace the current setup. Instead of ad-hoc scripts, conventions and associated documentation, it relies on a declarative description
Gitops!
It seems that they are using fluxcd, just like I am, to manage their kubernetes cluster.
I really like it as a solution, as you just edit configuration files, push then to git, and then your kubernetes cluster changes. Deploying an app is as simple as adding a file, and deleting an app involves deleting that file.
FYI, DeVault’s Stallman Report explicitly attacks the FSF as well as Stallman.
I read the report. You are free to show me exactly where it criticizes the FSF beyond their interactions with Stallman.
I took it mean that the report was such a deceitful and irrational work, presented in such a duplicitous manner as to constitute an attack on the senses of the community.
I'll just copy my older comment, and put it here
Begin quote:
Stallman doesn’t seem to get that pedophilia is wrong because of the hierarchy of power, and the power imbalances between older/younger people, not because of some inherent wrongness about being attracted to a prepubescent person. This is shown by how he condemns some pedophilia, but is accepting of 12+/past puberty. (I despise this logic, because it would also make gay sex and sodomy wrong, as well).
I find this deeply ironic, because his primary issue with proprietary software is the way that it gives developers levels of power over users. From his article Why Open Source Misses the Point
But software can be said to serve its users only if it respects their freedom. What if the software is designed to put chains on its users? Then powerfulness means the chains are more constricting, and reliability that they are harder to remove.
You would expect someone who is so in tune with the hierarchies that appear with software developers, publishers, and users, to also see those same hierarchies echoed in relationships between people of vastly different ages, but instead, we get this. I’m extremely disappointed.
These failures to understand hierarchy and power, are exactly why Stallman shouldn’t be in a position of power. Leaders should continually prove that they understand hierarchy and the effects of their actions on those below them. Someone who doesn’t understand how their power could affect another, shouldn’t be a leader.
End quote.
And I'll add onto this a little bit: Although Stallman seems to have redacted his earlier claim about pederasty, continuing to defend the legality of the possession of CSAM (beyond safe harbor/hospitality provisions), is very problematic, and clearly shows that he hasn't learned his lesson. CSAM ownership should be heavily disincentivized, to disincentivize the selling/buying of CSAM, as that's one of the most effective ways to stop more CSAM production.
I don't view pointing out that Stallman is not fit for a position of leadership to be an "attack" on the FSF or the free software community. And although the information gathering of the linked post is very, very impressive, it doesn't really invalidate what was said in the Stallman Report, or the Stallman Report's core points.
the report was an attack on the FSF and the free software
Stallman is not the FSF. The FSF is not the free software community. The free software community is not free software.
I'm so tired to people presenting these things as if they are equivalent.
Yeah. this was in high school, in my math class, and we were playing a math game.
The way it worked, was that every table was a team, and each team had a "castle" drawn up onto the whiteboard. A random spinner was used to determine a team, who would then solve a problem the teacher assigned. If you successfully solved the problem, you could draw an X on another teams castle. 3 X's mean that you are out.
My team was out. But, since this was a class, we could still solve problems, and still draw X's. Our table got selected to solve a problem, and I did successfully. I looked at the board, and realized that only two teams had a single X, every other team had either two or three. In other words, I could choose who won the game, even though I could not win.
So, I started trying to get bids. I tried to get real money, but someone tried to scam me with some "draw the X first" nonsense. But, the other team offered to pay me four of the school's fake money, and I accepted that and allowed them to win.
I may not have won the game, but I certainly felt victorious that day.
Is installing a package as simple as $ nix install vscode ? And would it “just work”?
You fool. You absolute buffoon. You're doing it wrong. That's the wrong command, and even if it was the right command, this method could cause issues down the road. It goes against the declarative philosophy of nix. You're supposed to refer to packages by their attribute, rather than name. If you launch the installed app that way, it won't have any hardware accererated graphics becuase hardware accelerated graphics are impure, don't you know this?
You need to read the nonexistent documentation to find all of this out, and then select a declarative solution, like home manager or nix-shell or nix develop, to install software, and nixgl to get hardware accelerated graphics for nix packages on non nix systems.
(/s (ish) over)
Yeah. Nix has some issues. If you just want more packages available, I would recommend distrobox.
It's a very powerful tool, and I use nix to manage all of my development environments, but it has some severe limitations that only have hacky workarounds. I could not get hardware accelerated nix packages to have a working dekstop entry. So instead, I have to type
nixGLIntel gzdoomin my terminal when I want to play doom.In addition to that, the documentation is very poor.
The other comments in this thread elaborate on these, and a few other issues.
It's not. Noncommercial limitatation makes it also not open source.
I also think they don't allow forking.
Is there a way to do something like this on KDE? I'm considering streaming soon, and I want to be able to share some windows, but not all, or only share a workspace/virtual desktop with my stream.
As a someone who has used both Arch, and Debian, neither has less or more bugs.
Debian has the same bugs, over the period of their stable release, and Arch has changing bugs (like a new set every update lol).
Yes, Arch is going to get a lot more features. But it comes at the cost of "instability". Which is not so much a lack of reliability but instead, how much the software changes. I remember a firefox bug that caused a crash when I attempt to drag bookmarks in my bookmarks bar around, which lasted for like a week — then it went away.
The idea behind projects like Debian, is that for an entity that needs stability, you can simply work around the bugs, since you always know what and where they are. (Well, the actual intent is that entities write patches and submit them to Debian to fix the bugs but no one does that).
Another thing: Debian Stable has more up to date packages than Ubuntu 20.04, and Ubuntu 22.04. This happens because Ubuntu "freezes" a Sid version, and those packages don't get major updates for a while. So often, the latest Debian stable has newer packages than the older Ubuntu releases.
Linus complains the author didn’t submit the patch to some places for public comments and testing BEFORE requesting a merge.
Although a reasonable expectation, I can't find anything about this on the kernel.org docs for posting patches. They seem to imply that you just check and verify your patch before submitting it on the kernel mailing list, but that's it. I didn't see any mentions of mailing lists explicitly for feedbacks or other conventions.
And before you start whining - again - about how you are fixing bugs, let me remind you about the build failures you had on big-endian machines because your patches had gotten ZERO testing outside your tree.
As far as I know, the Linux Foundation does not provide testing infrastructure to it's developers. Instead, corporations are expected to use their massive amount of resources to test patches across a variety of cases before contributing them.
Yes, I think Kent is in the wrong here. Yes, I think Kent should find a sponsor or something to help him with testing and making his development more stable (stable in the sense of fewer changes over time, rather than stable as in reliable).
But, I kinda dislike how the Linux Foundation has a sort of... corporate centric development. It results in frictions with individual developers, as shown here.
Over all of the people Linus has chewed out over the years, I always wonder how many of them were independent developers with few resources trying to figure things out on their own. I've always considered trying to learn to contribute, but the Linux kernel is massive. Combined with the programming pieces I would have to learn, as well as the infrastructure and ecosystem (mailing list, patch system, etc), it feels like it would be really infeasible to get into without some kind of mentor or dedicated teacher.
Linux mint debian edition is not based on testijg, but rather on stable*.
This misconception may be caused by the fact that the latest debian stable, has newer packages than many of the older-but-not-ancient ubuntu releases, which were originally based off of debian sid.
*I cannot find a first party source for this, only third party
Linux Mint Debian Edition 6 hits beta with reassuringly little drama. Think Debian 12 plus Mint's polish and a friendlier UX for non-techies
https://www.theregister.com/2023/09/13/linux_mint_debian_edition_hands_on/
Sorry. I meant if you wanted to use only packages from one set of repositories/one distro, for if you were looking for lower level packages like the kernel or desktop environment to be updated.
No, I think if you're using the nextcloud all in one image, then the management image connects to the docker socket and deploys nextcloud using that. The you could be able to update nextcloud via the web ui.
https://github.com/nextcloud/all-in-one?tab=readme-ov-file#how-to-update-the-containers
Because forgejo's ssh isn't for a normal ssh service, but rather so that users can access git over ssh.
Now technically, a bastion should work, but it's not really what people want when they are trying to set up git over ssh. Since git/ssh is a service, rather than an administrative tool, why shouldn't it be configured within the other tools used for exposes services? (Reverse proxy/caddy).
And in addition to that, people most probably want git/ssh to be available publicly, which a bastion host doesn't do.
So based on what you've said in the comments, I am guessing you are managing all your users with Nixos, in the Nixos config, and want to share these users to other services?
Yeah, I don't even know sharing Unix users is possible. EDIT: It seems to be based on comments below.
But what I do know is possible, is for Unix/Linux to get it's users from LDAP. Even sudo is able to read from LDAP, and use LDAP groups to authorize users as being able to sudo.
Setting these up on Nixos is trivial. You can use the users.ldap set of options on Nixos to configure authentication against an external LDAP user. Then, you can configure sudo
After all of that, you could declaratively configure an LDAP server using Nixos, including setting up users. For example, it looks like you can configure users and groups fro the kanidm ldap server
Or you could have a config file for the openldap server
RE: Manage auth at the reverse proxy: If you use Authentik as your LDAP server, it can reverse proxy services and auth users at that step. A common setup I've seen is to run another reverse proxy in front of authentik, and then just point that reverse proxy at authentik, and then use authentik to reverse proxy just the services you want behind a login page.
Oops... my bad. In my earlier comment I assumed that this would be a Fedora/Ublue based distro, rather than an Arch one. Arch doesn't have RPM ostree either (which makes me dislike it as a choice for an immutable distro).
But, it's highly likely that with the steam deck and other projects, there is already an ecosystem for immutable Arch, and a minimal base system to start is advantageous, as Possibly Linux said.