Archr

So fucking true. Once you get more than a hundred systems or so it feels like something new to resolve every day. Especially when your work let's people with no idea what they are doing have root on "their" servers.

I have dealt with issues ranging from someone uninstalling subscription manager (RHEL) and then clearing the package cache to people replacing chronyd with ntpd for.. Reasons? I guess?

I have thought about using bazzite but I prefer debian based as well as using more general purposed distros. So I am just on basic debian.

I just want to make sure that we do agree on a few things.

1. Requiring actual ID verification and/or face scans is bad and cannot be effectively anonymized.
2. That many of the current bills do not require ID verification or face scans. This includes the California one that the systemd merge request cites as well as the Colorado one that it mostly identical.
3. The laws in their current form are poorly written and clearly misunderstand how modern general purpose computers work and are referred to.

Given that, I think we can ultimately agree that the NY, UK, Germany, and I think also the Brazil laws are bad and cannot be fixed with simple updates to language.

So let's focus on the law's that do not require actual verification since that is what the systemd change cites.

What issues do you have outside of that they are poorly written and ineffective or that they are a slippery slope/frog in a pot/tip of the spear?

This is not about my comfort this is about what these laws actually require rather than some imaginary law that has not even been written yet.

I figured that someone might latch onto that "necessarily" and that's the great thing about open-source. If that distro/application/os does misuse your data then don't use it or fork it.

However.. ive read the associated analysis of the California bill that reads directly on legislative intent:

quoting he Cali Senate Judiciary Committee analysis : file:///home/jspaleta/Downloads/202520260AB1043_Senate%20Judiciary.pdf

Why are we listening to a person who tried to link a file directly from their downloads folder?

Also the original post that the article is referencing on the fedora forums is suggesting that we remove all networking support from baseline linux as some way to comply/circumvent the law.

I'm sorry, but I just can't take anything said in that forum post seriously.

Sure I can chime in here.

You did actually read the post correct? Not just the title? The original poster, Jef, is talking about implementing a Unix socket or a dbus protocol similar to what apple already has. They are literally just referencing their definition for a struct.

So no this will not be ID verification, it won't ask for face scans, and it won't necessarily send the data anywhere.

The article is just using the big A word as some boogeyman to generate clicks and further rile up the community.

The systemd change is benign and this is not proof of your slippery slope theory.

Edit: I swear literacy rates in the linux community must be dropping.

Ahhh. And then you have people like me who have been seeding for years but are way too socially anxious to even start any of that process. (:

So if you are a passing trans woman you either:

A. Risk retribution from the law and use the women's bathroom.

Or

B. Follow the law and get called out for being a women in the men's bathroom.

This law really just criminalizes using the bathroom if you are not 100% heteronormative.

Usually when I need to do something like this I use python and BeautifulSoup4. You basically get the content of the web page and use bs4 to parse it and pull out the correct link. You will need to look at the source of the page to understand their page format.

If python requests isn't able to get the right data then you might need to use selenium to use a full web browser to render the page and run any Javascript that might populate the page. Then you send that page content to bs4.

Edit: I know someone posted a link to archive but I figured some instructions would also be useful.

I'm not sure that I would recommend a newer user use sysrq. It is a very powerful tool that you definitely should not be blindly following from a random internet post without knowing what each command does.

In a truly frozen system then it can be good, but only as a final last resort. If the system can be unfrozen by other methods then that should be preferred instead.

There are a couple of things you can do. If it is frozen then try pressing ctrl+alt+1/2/3/4 to swith to a different terminal this will let you either restart your DE or reboot the system safely.

As far as debugging it I would typically start with looking at journal logs journalctl -b-1 should show you logs from the last boot.

Anything to replace firewall-cmd? (I know about ufw but it just feels too simple for my use case).

firewall-cmd commands are all fucked up and it feels like they use different verbs for very similar actions. Plus - - help is a few miles long which is not helpful. I just wish that it would have subcommands. Something like firewall-cmd zone public info

Hi. Are you a maintainer of one of the distros that might be affected by this law? If you aren't then you have no standing to blindly tell them that they should not follow the law and risk fines that would ruin the funding for their project(s).

Bringing up porn sites is a false equivalency. Many of these laws do not require verification of ID or face scans as some are incorrectly claiming. They require a birthdate be entered during installation. The laws surrounding porn sites required 3rd party age verification which many of these sites said would not only crater their traffic from these states but also introduce a privacy nightmare which would also work against their business interests.

Right. I thin you are ignoring some complexity here. This developer added a field to store some optional data in systemd. That code needs to be tested, reviewed, debated, and eventually needs to be merged in. Those merges, at least with large projects, don't typically get added directly to main they get added to a release branch. That release branch then needs to be completed and merged where it will then be packaged. Then different distributions/installers need to add that field as a requirement to their code which typically goes through the same process. Then all those changes need to be packaged for release by the distros themselves.

So I'll ask again. Assuming that distros do not want to risk being fined and financially ruined. What is a appropriate time before January 1st 2027 to open this pull request in systemd?

This would also assume that we would like to propose a solution (for the data storage) early enough that distros do not all come up with their own implementations and leave PII strewn across the system.