Yet Another internet Page

Age Verification is a mess

2026-03-15T14:49:52+00:00

init=/bin/bash</code>, need I say more? 

California is having bad ideas and everyone else is likely to follow suit — leading to a global outburst of poorly written age gating legislation that probably does not protect children because evil people on Discord will in fact just pretend to be children.

The number of teens that use Linux is low at best, but yet California really wants the big distros to require age verification. Not like:

D-Bus intentionally lets literally anyone implement an interface — we call it «Open Standard», something MS clearly hasn’t heard of and absolutely should, considering Windows is “The most open operating system”</a></li> The …3 if you squint hard enough…teens on Linux and other less busy users are way too obsessed to use a normal distro (I use Alpine, by the way</a>)</li> The general concept of age gating is utterly silly, if not backed by ID — a controversial and major violation of various places’ privacy laws.</li> This legislation was built with commercial and locked-down systems in mind, two words that most definitely conflict with both “Linux” and “Desktop”, the latter sometimes also referred to as “a pretty much user controlled environment” by practically no one except me — though it’s accurate, I’d say.</li> </ul> Yes I am still a flaky at best blogger. Oh well. </aside>

A Smaller Approach to Blogging 2025-11-11T18:51:58+00:00 An update and a pology I haven’t blogged in a while. Don’t know what I did in all that time. Probably having busier times come up and actually having to do things for once. Expect the content rate to stay similar to this ^-^ Now, why I’m actually writing this: What if I built a private Microblogging system to make quick and dirty posts to that either pile up, or just might eventually become real things I said in fully written out form. The only problem is that I have better things to do than pressure myself even more to say things. On an unrelated note, I’m compiling AOSP. If you plan to roll your own boot animations, stop, it’s hard. Sometimes SEL isn’t happy about the audio, other times /sbin/bootanimation segfaults, and when that doesn’t happen, it’s Magick optimizing out your image contents because you really wanted dynamic_color</code>. More on that once I got somewhere. Becoming the very thing you already destroyed 2025-08-28T18:35:45+00:00 A rewritten version of an article where I got a little emotional This article probably got split into three, as I tried to talk about three things at once, and readability suffered. →</a> The Fax</h2> If you either scroll down sometimes or have a non-negligible memory, you may remember that I wrote about the Play Integrity changes a few…months(?)…ago? If you have ever argued with an iOS user, you probably know that Android — mostly GMS’d Android — globally runs about 70% of the market, and I’d presume it’s in part popular due to its openness, which Google seems to be plugging, for good — or really, for bad — if there’s no backlash. I am apparently not so great at blogging regularly :) </aside> Google is restricting sideloading a lot,</a> dare I call this iOS-level. In case you prefer reading at the source, it’s also written down on Android Developers</a>, which may be less or more neutral about the subject, but probably more one-sided. Only on devices with GMS/GSF, but that’s most of them </aside> →</a> The Considerations</h2> While being able to know exactly who’s distributing malware sounds cool, we need to remember that not everything sideloaded is malware. For example: AdAway</a>, which isn’t on Google Play because Google wants ad revenue </li> F-Droid</a>, which can’t be on Google Play because it’s an app store…and because it’s kind of silly to download an app store that doesn’t track you over a platform that tracks you </li> Various, really</a> interesting</a> root solutions</a>, that Google will tolerate, as long as they’re not accessible to your average Joe:anne </li> ::: aside Not a real bullet point but how many times did I link to Magisk yet? ::: </li> Projects with privacy conscious developers</a> that don’t really want an ad company to know their driver’s license </li> Termux</a>, which violates W^X by design and thus can’t update the Google builds anymore. </li> Hardened Fenix forks</a>, which don’t distribute there because of privacy either </li> </ul> I presume this list will age like milk, esp. if I put many things in it </aside> With this new policy, Google would know exactly who makes what applications, and can choose to censor any of these from most of Android, like they did to YouTube Vanced. →</a> The Speculation</h2> Even if Google says something along the lines of ‘You need to calm down everyone, turns out we’re not doing that, promise’ the precedent of we do what we want to your phone is still there. Apple has this tight of a grip on the ecosystem, or used to, and look what they did to my boy Firefox! They forced him to WebKit. A KHTML-based Mozilla product, blasphemous! The above is very obviously sarcasm. But Apple did use to enforce that, until governing bodies told them to, you know, stop being anticompetitive pieces of- it’s fine, there’s nobody to demonetize me. Except my employer, maybe. </aside> And it’s not just internal pressure that could make Google do things. An oppressive government could force Google to do anything, and since Google is in it for the money, I don’t see them above bending to the will of dictators. Or maybe they are. They used to have a motto of ‘Don’t be Evil.’ →</a> The Brainwashing</h2> Even if you may not use non-store apps, some people do. Maybe Google eventually pulls some private messengers in response to the recent EU Chat Control discussion. Maybe you use that messenger to stay connected with friends and family. Maybe you want to sideload it. Maybe Google banned it. Do you now wish to rely on Feta(formerly Makebook)? Even if Google says they won’t release the data, they will still have it. It’s about data and control. It seems harmless today, but it’s a slippery slope. Do you want to let Google track more than they need to, to sell you more with ads? I doubt it, let’s not ignore this. (You’ll see this one 2 more times, in regard to Chat Control and the Steam payment processor debate.) </aside> Modern Laptops are really Splash-Resistant 2025-06-06T20:46:11+00:00 I accidentally sent mine bathing in milk Don’t you just hate it when you read a newspost and accidentally spill your drink? Don’t you just hate it even more when the table isn’t set and you have to clean it immediately? In my case, the laptop was in splash range. Lesson learned — The hard way. On the flip side, I am writing this post. The only broken hardware are the touchpad click buttons. I’ll either get used to their absence or try to fix them. Google is tightening PI 2025-05-22T15:40:13+00:00 I am not happy, who would’ve guessed? So, who here uses rooted Android? Oh, right, literally just me, Cloudflare says I have basically no legit visitors. If you are a legit visitor, you’re definitely not the one-hundredth one — but since you’re already connected to the internet — might as well drop me some mail. did I use those em-dashes wrong? </aside> Back to what I was actually writing about, Play Integrity: For the unfamiliar, Google wants to make sure apps can know if a device is in an out-of-box state or actually usable. Tampered with, as Google likes to call it. The former option is the 255th layer of hell in most cases. The latter is what I’m using. Google decided to bump the requirements for Android 13 and newer, as the integrity level for older, unmodified devices was easily attained with Magisk</a> / APatch</a> / KernelSU</a> modules like Play Integrity Fix</a> — PIF for short. This isn’t an ad, I just like those projects. </aside> The new ones are pretty hard to get. I can’t even get the basic attestation anymore, which is rather annoying. Unrooting would be an option, but Magisk is just so useful… Ever tried a DSP app? You either get system-wide audio effects via those that just ask Android, or near-global support with things like RootlessJamesDSP. The best option? Fully rooted JamesDSP — no compromises. </aside> Have I mentioned Google uses proprietary byte code running in a quite secure environment for this? Huge props to people like chiteroman</a> and osm0sis</a> for dealing with this stuff. They may not make all of the world go round, but they sure do make mine. I hope somebody finds a way to get just basic or device attestation again at some point. Until then, I’m using a keybox that will get revoked as soon as I look the wrong way — because some apps — especially the ones that don’t need it — really want integrity, whether I like it or not. PS: First blog post where I asked an LLM to proofread. Hope you don’t mind. If you do, your problem. Follow-Up 2025-05-18T12:14:00+00:00 The past is r/o As hinted at, I tried partclone</code> first, but that ended up with a corrupted superblock, so a 1:1 dd && btrfstune -u</code> it is. No Blogpost today :( 2025-05-18T12:10:00+00:00 Partclone won’t work on my Btrfs Unfortunately can’t do a proper blogpost today, my main machine is really busy copying a partition to a bigger drive — Problem is that it’s copying my rootfs, which means booting from a thumbdrive, so I can’t use my shellscript to add timestamps and wrap my post in <article></code> tags. QoL thing, and I won’t write a full-blown post without it now. A Quick and Dirty List of Shell-Scriptable Tools You'll Rarely Ever Need 2025-05-12T20:10:03+00:00 …and indicate you should’ve used Python →</a> jq - JSON processing as a CLI</h2> Grepping and cutting within minified JSON APIs is…not fun, to say the least. What if you had a way to parse it properly? jq</code> lets you do just that — and everything else you could possibly want to do to arbitrary structured data — by either vibe coding with your LLM of choice, querying a search engine you still haven’t decided on, or by reading the man page, like a nerd. From my experience it can be a little quirky regarding piped output. Put a while read</code> loop around it, if you really need to stream the data. Should work if the input is minified. 2026-07-21 18:30:07 +00:00</time> : So basically if you pass --stream</code> it’ll actually line buffer by itself… </aside> →</a> XMLStarlet - Doing absolutely everything you can imagine to well-formed XML</h2> Similarly to jq, xmlstarlet</code> also deals with structured data — Because apparently we can’t agree on a one-size-fits all standard. Aside from escaping and unescaping streamed data, it can also edit, beautify (format, as it calls it), extract data from, list the filesystem as, canonicalize — an euphemism for ‘fix your weird misuses of insignificant whitespace’ — validate and do magic on pretty much arbitrary XML, as long as it’s valid…enough. →</a> zopflipng - For when you have too much time and a high-fidelity scan</h2> Don’t you just hate when you have to downscale the document you scanned at 1200dpi — for no reason at all — to something sane because it’s over five megabytes? Then zopflipng</code> can’t help you, unless you actually have way, way too much time on your hands. If you don’t, it’s still useful for shaving off a few bits from lower-resolution icons. (Of course, you can still waste time by spamming -m enough times.) With great power comes great responsibility, and — if overdone — great electricity bill. You can probably get away with the defaults. Bumping up the settings is just candy on top. </aside> →</a> ImageMagick/Graphicsmagick - when GIMP is too heavy</h2> Two related tools for image conversion and procedural(?) editing. gm</code> is slightly less powerful but should be nicer on licensing, while magick</code> is the upstream and has a less stable interface. Use what you prefer, or use neither. I primarily use it for when I need to operate on a ton of images — mogrify</code> is your friend, although nowadays it should be invoked as (magick|gm) mogrify</code> — I’m sure they’re not deprecating the legacy variant any time soon, though, too much breakage. Site Maps 2025-05-11T00:00:00+00:00 Or “Why I solve everything with shellscripts” Sometime ago I started doing Search Engine Optimization — and before anybody flames me for it, I won’t add analytics beyond what Cloudflare ships, those are helpful for preventing bot spam. — which also meant making my <head> much bigger by including a lot of metadata. Another fun thing was adding a humans- and a robots.txt file. Spent the last thirty-or-something minutes writing a sitemap generator in bash, with a little help from ChatGPT for finding the right tools for the job. Is XMLStarlet</code> overkill? …Probably. Does it work? Until you try to change something. Am I running it way too often? Arguably. But hey, the only dependency on Starlet is a single function that can be replaced quite easily. Feel free to look at the output yourself under sitemap.xml</a> As for why everything I do involves way too many shell scripts, I only know Python, Bash, that weird scripting language Plymouth uses and traces of C, but I’ve almost perfected the second one to the extent of an art form, so I’ll just use it until I have to do things properly. Permanently borrow my code</a> I forgot to write down when exactly I wrote this. Oh well </aside> Rest in peace, my unifying receiver 2025-05-09T12:39:00+00:00 The older ones were better Today is a tragic day. One of my permanently borrowed Unifying Receivers died — I now have one left that went bathing. I think it’s been jammed too hard into my laptop when putting it in its case. That USB port is fine, so at least the damage kept itself to a receiver… …for a deprecated protocol replaced by Logi Bolt. This is where Logitech went downhill, killing off support for older mice on newer receivers. I know that supporting old protocols is annoying, but Unifying devices are still pretty popular One can argue about the correctness of mice as a plural of computer mouse, but you know what I mean. And besides, mouses sounds horrible </aside> Another day, another debugging session 2025-05-07T19:33:00+00:00 Not the best way to start a blog, really Having a working page is cool. Having Cloudflare error out isn’t. That’s why one should listen on the right port, with the right protocol, and with a proper cert. I was going to put a blogpost here, but a problem with the web is that you need content if you want to make pages. And a lorem ipsum is not the answer. So, I guess this will have to do for today. Building for the web is annoying, but at least I have something like a portfolio now.