<unnamed>

Public Key

npub1k6ykawdchynr4qhzka4wk3qpj09g0p2kgcprd4y20mgrnlxxmucq5mm3fx

Profile Code

nprofile1qqstdztwhxutjf36st3tw6htgsqe8j58s4tyvq3k6j98a5pelnrd7vqprpmhxue69uhhyetvv9ujumt0d4hhxarj9ecxjmntzulgmt

Publishing to

relay.momostr.pink

Last Notes

2026-08-30 23:10:41 CEST

- reply

Sure, but they aren't different from their owners, and people are morally responsible for who they willingly give money to.

2026-08-30 22:32:57 CEST

Quoth Nussbaum at #Debian:
“I view the spread of AI we've seen over the past two years as a revolution similar in impact to the generalization of Internet access to the wider public since the mid 90s.”
I don't remember the head of AOL publishing any weird manifestos about how [your] democracy is incompatible with [his] freedom.
I don't remember the head of CompuServe performing any “Roman” salutes at any government functions.
I don't remember the Internet being meant to replace all humans.

2026-08-23 23:12:50 CEST

- reply

Bill C-34 https://www.parl.ca/documentviewer/en/45-1/bill/C-34/first-reading will give a monopoly on all Internet-based speech to a handful of tech giants, almost all of which are US-based, and shut down all other social sites located in Canada, including the one you're using right now.
No offense to you personally, but your government's show of defiance is purely theatrical.

2026-08-19 19:26:06 CEST

- reply

Isn't it a vineyard now? https://en.wikipedia.org/wiki/Bliss_(picture)

2026-08-17 06:47:25 CEST

New Jersey has passed a law regulating social media use by children. https://www.njleg.state.nj.us/bill-search/2026/A4015 It has been billed as a “big tech law”.
Surprisingly, this one actually is a big tech law. It covers only providers that make $25M revenue and/or process personal data of ≥25k people in one year.
Even the biggest Fedi site makes nowhere near $25M, and no Fedi site processes personal data (as defined by the law) at all.
It also does not require #ageVerification beyond what a platform already does.

2026-08-16 22:27:11 CEST

- reply

No sins can be hidden from the great code reviewer in the sky.

2026-08-15 23:02:27 CEST

- reply

> The data disclosure issue is not any less present in monoliths or microkernels unless said microkernel makes liberal use of virtualization and cleans up everything properly between every context switch.
Is that not what virtual-memory operating systems already do?
I was under the vague impression that process-level isolation is adequate to mitigate it, but language-level security is not. But understanding the details of Spectre is way above my pay grade. 🤷‍♂️
@npub1e8t…n3ds

2026-08-15 21:35:38 CEST

- reply

Didn't Spectre nail the coffin of language-based security models?
@npub1e8t…n3ds

2026-08-15 21:33:46 CEST

- reply

Computers don't spend all of their CPU time on the file system. They do have other work to do. The more CPU time they spend on the file system, the less other work they can do.
@npub1e8t…n3ds @npub1ngr…gmse

2026-08-15 21:32:54 CEST

- reply

Compare a FUSE implementation of a file system with a kernel implementation of that same file system. That's the penalty.
@npub1326…gzqx @npub1e8t…n3ds

2026-08-15 21:28:23 CEST

- reply

“Modern computers are fast enough that the overhead doesn't matter” has never been true.
Because yes, the hardware is faster, but the workload is also bigger.
Modern operating systems are more efficient software-wise than older ones (e.g. io_uring vs blocking system calls), despite running on far faster hardware, for this reason.
@npub1e8t…n3ds

2026-08-15 21:22:21 CEST

Suicide-prevention crusaders aren't in it because they care about people.
If they cared about people, their crusade would be against poverty, oppression, and disease, without which no one would feel any need to commit suicide in the first place.
No, suicide-prevention crusaders are the sort who pull the wings off insects for fun. They're in it because they enjoy torturing the helpless and they don't want us to die before they're done having their way with us.

2026-08-15 20:01:06 CEST

- reply

Possible solution: design all file system drivers such that they can be compiled for either user or kernel mode. That way, every file system has a FUSE equivalent with the same code.
Not sure how feasible this is, though.

2026-08-15 19:55:58 CEST

- reply

Problem: user-mode file system drivers are slow. So are microkernels in which all drivers are user-mode.

2026-08-14 07:32:00 CEST

PAX BROUGHT TRIBBLE ABOARD. TRIBBLE ESCAPED. PAX TRIBBLE AND 2 FA KILLED BY KLINGON PAX. NOT GETTING PAID ENOUGH FOR THIS
#LCARSDrama #StarTrek

2026-08-14 07:08:40 CEST

PORT NACELLE OUT. MAX WARP 2.3 W/ 1 NACELLE. PAX XPORTED TO NCC-1701-E BUT UNABLE TO REPAIR NACELLE. REQUEST TOW TO STARBASE.
#LCARSDrama #StarTrek

2026-08-12 07:00:36 CEST

- reply

Strange. I'm slightly obsessed with Star Trek and I've never heard of “punk on a bus”.
@npub1d58…wmvf

2026-08-11 17:53:56 CEST

- reply

Speaking of which, does your screen reader have trouble with hashtags that appear in the middle of a sentence?
@npub1d58…wmvf

2026-08-10 18:15:34 CEST

- reply

That keyboard looks cooked, though. Does it, like, actually work at all?
@npub19wj…g8gs

2026-08-10 04:07:28 CEST

- reply

Yeah, I generally don't post photos I took myself. I've heard that AI can figure out where a photo was taken just by looking at it, so if I did post any photos, I'd be narrowing down my location quite a lot.
Although of course you also have to ask yourself what anyone would do if they did find out where you're located. If the answer is “nothing”, or possibly “have a very short and very unpleasant conversation with my .45”, then I guess post away. But I don't own a .45, so… 😅

2026-08-10 03:53:05 CEST

- reply

You don't *have* to post your life's happenings here. You can also chat about whatever's on your mind, or look for discussions that you might have something to say about.
Some people here do reveal information about their real-life identity. Others, like me, don't. Like, I have no idea who, where, or even *what* @npub17sg…xt22 really is, but seeing their posts is still fun.
I don't know how it is on Facebook, but on here, there's something for everyone.

2026-08-06 22:25:49 CEST

- reply

I've heard that even the banks that do offer other options are planning to phase them out in favor of proprietary phone apps only.
@npub1afm…5gp4

2026-08-06 20:28:22 CEST

- reply

Yes, and that is an abomination. But those laws do not require the use of a proprietary device. You're still in control of what you reveal, when, and to who.
A proprietary device could be recording and reporting everything you say within range of its microphone, and you'd never know because the public can't see the code. As horrible as US social media age verification is, that's even worse.

2026-08-06 20:19:25 CEST

- reply

It's my understanding that the EC also intends to impose age verification on most of the Internet, also to be conducted through a phone app (EUDI Wallet) that requires an unmodified proprietary operating system.
This would make open-source phones effectively illegal. You can have one, but it can't access anything.
Without open source, there is no privacy.

2026-08-06 20:11:57 CEST

- reply

One of the main complaints about open-source phone operating systems is that banking apps don't work. I always found this confusing, because I just ignore the app and use the bank's website, but I learned recently that most Europeans aren't allowed to do that. The app is mandatory.
I too had been under the impression that privacy was a big deal in Europe. I've been repeatedly shocked and horrified over the past few months by revelations like this.

2026-08-06 17:30:57 CEST

- reply

At least in the USA I can *use* my bank without needing a creepy proprietary phone app. I just need a browser, a password, and either an email or a phone, and if I choose the phone option then even a 50-year-old land phone will suffice.
I don't need a proprietary app to ride a bus, either. Cash or a bank card will do. I think there's a city or two where you don't even need that.
Don't act superior. You've got bigger problems than we do. We still have some privacy; you have none.

2026-08-05 20:30:22 CEST

- reply

> Firmware anti-rollback is an industry standard practice.
Oh for goodness' sake, stop talking like that. All you're basically saying is, “You vill buy ze new phone ven ve say so. You vill run ze proprietary OS. You have no choice. It is ze industry standard.”
> Laptops and desktops generally have absolutely atrocious firmware security
They also don't self-destruct!
@npub1ze4…0y3x [@collective_truth](https://mastodon.social/@collective_truth ) @npub19cp…6and @npub1h54…c9uf

2026-08-05 20:25:10 CEST

- reply

> Breaking the device with firmware anti-rollback requires unlocking it and bricking it with unsafe use of fastboot commands.
So you've said, but are you sure you aren't assuming that the SSD doesn't flip any bits? Because they do flip bits sometimes, and as far as I know, the flash cells storing the bootloader aren't special magic bit-flip-proof ones.
@npub1ze4…0y3x [@collective_truth](https://mastodon.social/@collective_truth ) @npub19cp…6and @npub1h54…c9uf

2026-08-05 20:19:30 CEST

- reply

I guess so, but it's not so simple with the anti-rollback feature we were originally discussing. Concerns about its security efficacy aside, anti-rollback also potentially destroys the device and frustrates the device's owner's software freedom. That's crossing the line.
@npub1ze4…0y3x [@collective_truth](https://mastodon.social/@collective_truth ) @npub19cp…6and @npub1h54…c9uf

2026-08-05 19:50:54 CEST

- reply

If keys are not stored in the “secure element”, and it is merely being used to derive keys based on some input (e.g. user's PIN) with a rate limit, then what's stopping the CPU (or some other device's CPU) from doing the same thing without rate limiting?
@npub1ze4…0y3x [@collective_truth](https://mastodon.social/@collective_truth ) @npub19cp…6and @npub1h54…c9uf

2026-08-05 19:38:46 CEST

- reply

> GrapheneOS and iOS both quite successfully protect against data extraction by state actors
By storing the key in a “secure element” designed by a company that is under the thumb of those same state actors? Surely you can't be serious.
@npub1ze4…0y3x [@collective_truth](https://mastodon.social/@collective_truth ) @npub19cp…6and @npub1h54…c9uf

2026-08-05 19:27:02 CEST

- reply

Because it's futile.
Well-resourced attackers with physical access should be assumed to be able to extract any bit from any memory cell anywhere in the device. Most such attackers are in contact with hardware manufacturers, after all.
Those who can't do that—common criminals and pranksters—can't dump the RAM in my laptop while it's powered, either.
If physical attack is likely, powering the device off is the only way to be sure.
@npub1ze4…0y3x [@collective_truth](https://mastodon.social/@collective_truth ) @npub19cp…6and @npub1h54…c9uf

2026-08-05 17:14:05 CEST

- reply

If a remote attack can trigger my phone to self-destruct beyond repair, then…I'm not sure why you think that's a feature, because that sounds like a catastrophic, lawsuit-worthy design defect to me.
As for physical access, no device an attacker had physical access to can be trusted to be secure any longer. Hardware fuses don't change that. The attacker could've replaced the entire motherboard, and what good would the fuses do then?
@npub1ze4…0y3x [@collective_truth](https://mastodon.social/@collective_truth ) @npub19cp…6and @npub1h54…c9uf

2026-08-05 07:59:35 CEST

- reply

I'm not sure it *was* a Google update that bricked my phone. @npub1kwa…e0nj thought it was a hardware failure.
I still find the anti-rollback feature's existence disturbing, though.
@npub1ze4…0y3x @npub19cp…6and @npub1h54…c9uf

2026-08-04 17:30:00 CEST

- reply

I already tried flashing stock Android. Didn't work. Can't imagine trying to flash GrapheneOS would end any differently.
@npub1kwa…e0nj @npub1h54…c9uf

2026-08-04 06:48:54 CEST

- reply

I have my old Pixel 3a working until I can afford a new phone, but I'm feeling rather naked running a phone this old…
I see that GrapheneOS no longer supports this phone for lack of firmware updates, but the GOS website is pretty vague, so I'm hoping you can fill in some details for me.
If I were to install LineageOS or something on this old 3a, what kind of attacks will I still be vulnerable to?
@npub1h54…c9uf @npub1ze4…0y3x

2026-08-04 00:12:07 CEST

- reply

You said it draws the line at the lockscreen / home screen UI. Does that mean anti-rollback doesn't get incremented until *after* I unlock my phone following a reboot?
Because I never got that far. I left the phone to boot and do its thing, and when I later picked it up to check if it was ready to unlock, it was dead.
@npub1h54…c9uf @npub1ze4…0y3x

2026-08-03 23:25:44 CEST

- reply

If that were to happen, would I still be able to do the power+volume-down thing to get into fastboot? Because that isn't working; the screen remains black.
@npub1h54…c9uf @npub1ze4…0y3x

2026-08-03 23:23:55 CEST

- reply

The only early symptom I saw was that, the first time I tried installing Android 16, just before I could tap the reboot button, it said there was a problem. Didn't specify what was wrong.
I told it to start installing the update again, and that time it seemingly succeeded…but died at some point after attempting to boot.
I suppose the SSD might have died, but in that case, wouldn't I have seen more severe symptoms before rebooting, like apps freezing?
@npub1h54…c9uf @npub1ze4…0y3x

2026-08-03 23:22:48 CEST

- reply

This was stock Android, not GrapheneOS, just so you know.
I wasn't watching the phone while it was booting, so I'm not sure if it ever made it to the home screen. I do know that the boot-time Google logo showed up, which is when I took my eyes off it.
And that was the last time I saw my phone alive. 🪦
@npub1h54…c9uf @npub1ze4…0y3x

2026-08-03 23:11:44 CEST

- reply

I don't know for sure, but that's what Google AI says, and I saw a bunch of articles warning people that it would hard-brick their phones.
Super not amused that the phone I paid good money for is literally designed to destroy itself if an update goes sufficiently sideways.
@npub19cp…6and @npub1h54…c9uf

2026-08-03 21:41:14 CEST

- reply

Are you a bad enough dude to save the Fediverse?

2026-08-03 20:24:44 CEST

- reply

I've also learned that #Google #GooglePixel phones have an “anti-rollback” feature that will permanently hard-brick the device if tripped. Which it can be, if the bootloader gets corrupted. Nothing anyone can do.
In other words, I was sold a self-destructing #phone.
Fantastic.
https://share.google/aimode/O8eWGurPxT7kNG8s4
@npub1ze4…0y3x

2026-08-03 20:24:03 CEST

- reply

Google official tool doesn't work. Usbdl asks for my specific model, and Pixel 6 isn't one of them. Looks like it needs some kind of ROM images or something.
Apparently Google keeps the necessary files behind lock and key, and thousands of phones have been bricked by updates like mine was. <https://issuetracker.google.com/issues/507968853?pli=1> <https://issuetracker.google.com/issues/402455330#comment119>
@npub1ze4…0y3x

2026-08-03 08:10:05 CEST

- reply

…Or do you disagree? You seem to know more about this than I do, so…
@npub1ze4…0y3x

2026-08-03 08:09:19 CEST

- reply

That too. I figure I should try Google's official tools first, though, yeah?
@npub1ze4…0y3x

2026-08-03 08:08:05 CEST

- reply

Not out of the woods yet! That post mentions a program called tensor-usbdl, but it apparently doesn't work on the Pixel 6. https://github.com/JoshuaDoes/tensor-usbdl/releases
I wonder if Google's “Pixel Update and Software Repair” site can work with the phone in this Pixel ROM Recovery mode? https://pixelrepair.withgoogle.com/ Guess I'll find out tomorrow.
@npub1ze4…0y3x

2026-08-03 06:03:49 CEST

- reply

Holy shit! Sign of life!
If I plug it into my PC while holding down all three buttons, it *does* show up on the bus!
@npub1h54…c9uf graciously provided a link https://social.highenergymagic.net/@freya/117029554424881416 to a post explaining how I might get this phone back up and running from this point.
Tomorrow I will investigate!
[@collective_truth](https://mastodon.social/@collective_truth ) @npub1ze4…0y3x @npub15gl…flgs @npub12l0…85j6 @npub1pmk…k8up @npub15r3…r92k @npub1m95…ehng

2026-08-03 05:44:40 CEST

- reply

Just for completeness' sake, I asked Google's AI to help troubleshoot, seeing as it's a Google phone. Same outcome: phone pronounced dead.
The bot says it's a hardware failure that the operating system upgrade exposed. Seems unlikely, but it's as good a guess as any.
I wonder if the upgrade installation process trashed the bootloader. It's my vague understanding that phones don't have actual ROM like PCs?
[@collective_truth](https://mastodon.social/@collective_truth ) @npub1ze4…0y3x @npub15gl…flgs @npub12l0…85j6 @npub1pmk…k8up @npub15r3…r92k @npub1m95…ehng

2026-08-03 03:36:44 CEST

Welp, I took my #Pixel 6 #phone to a repair shop, where it was pronounced dead.
Cause of death: fatal complications of a botched software update.
It will be missed. 🪦
I don't have the budget for a new phone right now, so this forces me to use my old Pixel 3a, which still works. But the stock OS insecure af now, so…what OS should I install on it? Preferably something that Just Works™ once installed.
[@collective_truth](https://mastodon.social/@collective_truth ) @npub1ze4…0y3x @npub15gl…flgs @npub12l0…85j6 @npub1pmk…k8up @npub15r3…r92k @npub1m95…ehng

2026-07-22 20:31:35 CEST

Does #SNES music count as a #chiptune?
That machine arrived in the late chiptune era, when FM synthesis was the norm, and it seems to have some PSG-like functions (fire and explosion sounds in SNES games often sound like the NES), but for the most part it is a sample-based synthesizer.
Wikipedia says Fast Tracker makes chiptunes <https://en.wikipedia.org/wiki/Chiptune#Contemporary_chiptune_music>, but unless I'm mistaken, FT uses the CPU for software synthesis, not a PSG or similar, so… 🤷‍♂️
#retrocomputing #retrogaming

2026-07-22 06:52:04 CEST

I just found out that @npub1aey…y08m #Borg #BorgBackup is vibe coded now. That's a tragedy. Really upsetting, to be honest.
I've been safeguarding priceless data with Borg for years, but I obviously can't entrust something so important to code that wasn't written by skilled hands and reviewed by skilled minds.
What full-system #backup software for #Linux do people recommend instead?

2026-07-19 19:55:04 CEST

- reply

You all realize this meme is 27 years old, right? You realize how old this means we all are? 👴 We're so old that the porn is old enough to watch porn! 😂

2026-07-18 05:52:19 CEST

#MSDOS #retrocomputing #shitpost
https://mastodon.sdf.org/system/media_attachments/files/116/938/944/335/738/614/original/0f4ca9113d176564.jpg

2026-07-16 04:46:25 CEST

The invention of mirrorless cameras was well received by the general public, but vampires were quick to criticize the new device.
#shitpost

2026-07-02 19:19:08 CEST

- reply

Passwords: problematic
Password managers: solve the problem
2FA people: “Solving problems is NOT ACCEPTABLE!”

2026-06-28 06:11:11 CEST

- reply

Little-known fact (so little-known that I didn't know it myself until I looked it up just now): HPV also causes cancer in men. Mostly in the genital region. Ouch. Best get vaxxed, people.
@npub16kg…afn8

2026-06-26 19:12:27 CEST

- reply

If “no address associated with hostname” then it's probably DNS.
Although that isn't saying much, because it's always DNS. 😛

2026-06-25 19:40:52 CEST

- reply

Wait, converting PDFs to presentation slides? PDF *is* a format for presentation slides.

2026-06-25 17:43:52 CEST

They're not shitposts. They're Poop-Toots!
#shitpost
https://mastodon.sdf.org/system/media_attachments/files/116/811/508/893/926/772/original/b61644faec0fe7cb.jpg

2026-06-21 00:31:37 CEST

- reply

The obvious answer is to embed CEF and use it to generate PDF as an alternate report format.
But I'm hoping to hear that it isn't actually common for company email servers to block HTML attachments. Then I won't need to go to such lengths.
I suppose I could release with HTML reports only and see if anyone complains…

2026-06-21 00:29:08 CEST

- reply

You might have misunderstood. We don't provide email service to our customers, nor do we intend to.
We provide a desktop app that works entirely offline. It can generate reports, but if our customers need to send those reports to their colleagues, they must use their existing email infrastructure to do so.
My question is whether their email servers are likely to block sending of those reports because they're HTML and not PDF, and if so, what I can do to solve that problem.

2026-06-20 23:51:22 CEST

- reply

What do you mean? What bad practice?

2026-06-20 23:26:53 CEST

- reply

Because this app does not have any server from which to do so. It runs offline on the user's desktop.
My org is not interested in the cost and legal liability of storing confidential customer data on our servers, and I like to think our customers appreciate the privacy/security benefit of their data never leaving their own computer.
But that means we can't send the report to others on their behalf, like your typical SaaS could.

2026-06-20 23:09:20 CEST

- reply

Go on?

2026-06-20 21:18:47 CEST

So, I'm writing a reporting system for my employer's app. It generates reports in #HTML, which the user can look at, print, and save. They look great—modern #browsers are quite good at #typesetting.
But it occurs to me that some users might want to send these reports as #email attachments to colleagues.
To the #IT people on Fedi: is that going to be a problem? Is it common for email servers to block HTML attachments? If so, what do you suggest I do to work around that?
#programming #webdev

2026-06-16 19:00:15 CEST

- reply

Did you have an alternative in mind?
My org may or may not need something like this soon, so I'd very much like to know what we should be using instead.
@npub1evj…970v

2026-06-15 03:54:33 CEST

The existence of Brooke Shields implies the existence of Brooke Phasers.
#StarTrek #shitpost

2026-06-13 21:02:23 CEST

- reply

I don't know what that site is, but it is super weird that it doesn't even say what it is before demanding I agree to a giant wall of legalese. (I declined.)

2026-06-12 07:28:18 CEST

- reply

That could be anywhere from 427°C to -193°C. With no atmosphere or liquid water distributing heat around, Mercury is a place of extremes in both directions.

2026-06-05 05:34:18 CEST

All these supply-chain attacks in npm are making me awfully nervous.
I need to use Sass and PostCSS in a project, and I think I'm going to copy an old package-lock.json from another project that uses them.
That's not a long-term solution, though…
And I'm worried the same thing is going to happen to crates.io at some point. 😬
#programming #security #infosec #cybersecurity

2026-06-03 21:24:44 CEST

whirrrrrrrrrr 🛢️
grind grind 💾
grind grind 💾
beep! 🔊
clickityclickityclickclickclick 🛢️
Modern #computers just aren't noisy enough.
#retrocomputing

2026-05-26 17:15:02 CEST

- reply

Noooo! I'm using Barcode Scanner! 😱 It's one of the coolest apps I've seen.
#FDroid

2026-05-20 00:10:43 CEST

Did some reading about #WebAuthn / #Passkey.
The specs and promotional materials make much ado of hardware tokens, TPMs, and smartphones, but completely fail to mention how this might be used with software password managers like #KeePass / #KeePassXC.
Hardware tokens are great in corporate environments but horrible otherwise because you can't back them up.
TPMs and smartphones are notoriously insecure. Storing secrets on them is laughable.
#security #cybersecurity #infosec

2026-05-14 03:44:01 CEST

Me: goes to drop some garbage down the chute
Chute door: locked
Me: goes downstairs to check the garbage compactor
Some idiot left the compactor door open! This prevents the compactor from compacting, and also locks all of the chute doors.
Me: 😡
Me: closes compactor door, goes back upstairs
Me: goes to drop some garbage down the chute
Chute door: locked
Me: FFFFFFFUUUUUUUUUUUU-

2026-05-07 23:08:30 CEST

Oh good, another high-severity #Linux #security vulnerability that somebody botched the disclosure of, turning it into a high-severity zero-day.
Because #CopyFail wasn't bad enough. Now we've got #DirtyFrag too.
Can #cybersecurity people please stop botching vulnerability disclosure? Thanks.
https://github.com/V4bel/dirtyfrag/blob/master/README.md
#security

2026-05-05 17:28:31 CEST

Imagine being such a control freak that you prohibit your children from talking to anyone.
And moreover demand that the government demolish any place where people could potentially talk to each other, just to be sure that your children can't talk to anyone.
Because I sure can't. That's insane. Absolutely bat-guano insane.
And yet that is the reality in which I find myself.
They have yet to invent words to adequately describe how disappointed and frightened I am by my fellow adults' behavior.

2026-04-30 06:52:50 CEST

- reply

Server people don't care what CPU architecture is popular. They'll compile their code for whatever hardware is best.
So if the performance of such a machine were only slightly worse, then I really can't imagine why people wouldn't be using it by now. They've already accepted performance penalties for Meltdown mitigation, and those penalties were not slight.

2026-04-30 06:45:39 CEST

- reply

Yes, because most people aren't using hypervisors for what we currently use processes for. If the hypervisor becomes the new process, that will surely change.

2026-04-30 06:43:31 CEST

- reply

That's a very extraordinary claim. It requires very extraordinary proof.
If there were a way to achieve performance on par with what we have now, yet do so without the security risk and extra circuitry required for speculative execution (and the security/stability risk of ambient-authority monolithic kernels), wouldn't we be using it already?

2026-04-30 06:33:25 CEST

- reply

Sure, until people start demanding that hypervisors expose these sorts of APIs for whatever reason. Linux didn't always have this, you know.

2026-04-30 06:24:31 CEST

- reply

You mean like Microsoft Research's Singularity operating system?
I vaguely recall that that sort of security model is insecure because of the Spectre vulnerability. Not sure about the details, but the hardware would also need to lack speculative execution, and nobody's going to accept such a severe performance hit.

2026-04-30 06:21:16 CEST

- reply

Well, per the writeup, Copy Fail is the result of a silly optimization case. Linux certainly could adopt a policy of being more paranoid about isolating user-space data from memory belonging to other processes or the kernel, speed be damned. Don't need to hypervisor everything just to accomplish that.

2026-04-30 06:11:28 CEST

- reply

In short: Qubes is only right until people start actually using it en masse.

2026-04-30 06:09:38 CEST

- reply

That's the whole point of processes and memory protection: to virtualize the CPU, memory, and other shared hardware resources.
Running everything in a hypervisor is just processes with extra steps. And when everybody starts demanding all the fancy features and performance tricks you'll currently find in the Linux kernel, which you know they will, we'll end up right back here, having expended titanic effort and accomplished nothing.

2026-04-27 07:19:09 CEST

- reply

As far as I'm concerned, the Year of the #Linux Desktop was a long time ago, and now I'm just waiting for everybody to get the memo.
Not, like, Debian 1.2 kind of long time ago. It was very rough back then. Could explode your screen if you weren't careful, and that's only a slight exaggeration.
But this is 2026. It's been solid for ages (and so have screens).

2026-04-27 07:18:38 CEST

Somebody on YouTube posted “DEBIAN 13: I could actually use it as my desktop, now!”
Heh. I've been using #Debian on desktop since version 1.2. Or maybe 1.1. I forget.

2026-04-25 03:33:39 CEST

- reply

Let's not. Kids are people, not property to be muted at a whim. If they have something to say, the government has no business stopping them from saying it in the place where the general public will hear them, and in 2026 that place is social media.
Frankly I'm shocked at how quickly freedom of speech flies out the window. When I was in school, I was taught that freedom of speech is sacred and our ancestors died for it! Am I the only one who took it to heart?
@npub1vh5…r3th

2026-04-21 05:02:54 CEST

- reply

34 paragraphs on one page? What is this, a font size for ants?

2026-04-18 05:04:41 CEST

#IBM: “Our big sophisticated computers are sometimes making mistakes! We can't figure out why! How will we ever fix this?”
This One Guy: “With capacitors. One on every chip, across power and ground.”
IBM: “…Capacitors?”
TOG: “Capacitors.”
[one capacitor-soldering spree later, the computer works fine]
IBM: “Well, that was anticlimatic.”
https://youtu.be/4X3wLqP11kQ

2026-04-17 02:41:21 CEST

- reply

Actor behavior can be scaled to whatever the game-logic tic rate is, by way of delta time. https://en.wikipedia.org/wiki/Delta_time This is how some games are able to have an adjustable tic rate.
So everything still works if the tic rate equals the frame rate, which for example it does in Descent if I'm not mistaken.
But yes, a high tic rate is hard on CPU and bandwidth, and game characters' behavior doesn't need to react that fast anyway, hence tic rates usually being lower than 60Hz.

2026-04-17 00:47:05 CEST

All computer games are turn-based. Some of them just take turns 60 times per second.
Seriously, read the source code of a “real-time” #videoGame some time. You'll pretty quickly observe that the game engine works by giving each actor a turn to decide what it's going to do next (move, shoot, cast spell, etc) and then do it.
Obviously most actions take more than one frame, and therefore in most cases the turn action will be “continue what I was doing last turn”, but still.
#gaming #gamedev

2026-04-11 22:27:24 CEST

- reply

Why the hell does a financial analytics and monitoring program need to store authentication tokens to a storage system for company documents?!
#Cybersecurity people are talking about hyper-paranoid zero-trust remote-attestation stuff, meanwhile companies are doing ridiculous crap like this.
Dear #infosec community: you know how the A-10 engineers were focused on armoring the parts of the plane that were actually getting hit and actually killing the plane? Maybe do the equivalent of that here.

2026-04-10 20:37:32 CEST

My cat is a cat who cats because she is a cat.

2026-04-06 05:19:12 CEST

- reply

I mean, there isn't really any natural law that says there can't be a #Wayland compositor that runs on #EGA hardware.
It would have the same chunky-to-planar-conversion overhead that complicated attempts to port #Doom to the #Amiga, but it's not impossible.
#retrocomputing

2026-04-04 01:43:56 CEST

- reply

I already did. I know what's wrong with that eye. Treatment options are not great.
The condition is, at least, not expected to get any worse than it already is. So, as long as nothing else happens to my good eye, I can live with it.
Here's hoping nothing else happens to my good eye…

2026-04-04 01:23:11 CEST

Just noticed something about my #eyesight.
One of my #eyes is messed up. It sees a blurry image, no matter the distance.
So I mostly use the other #eye to actually see things. That one still sees a good image.
But, if I close the bad eye, it becomes hard to properly focus on certain objects, like a digital clock across the room.
Apparently, even though the bad eye can't see very well, I still need it because it still gives usable depth information, which help me focus properly.
Fascinating.

2026-04-04 00:34:35 CEST

Can #browsers please stop sending such ridiculous user-agent strings?
Chrome is not a 25-year-old version of Netscape, for goodness' sake.
Are there seriously still sites that break if you don't pretend to be Netscape? And even if there are, can't we just put together a database of them and only spoof the user-agent for them?
And now I see that Chrome pretends to be Safari. Even when it's running on Linux. You gotta be kidding me.
#webdev

2026-04-01 17:16:26 CEST

Source: https://nisus.com/blogs/copybara-pasteybara/
#capybara #shitpost
https://mastodon.sdf.org/system/media_attachments/files/116/330/105/672/611/826/original/1fb03298d9f2213f.png

2026-04-01 09:03:59 CEST

- reply

The existence of addresses implies the existence of subtractresses.