<unnamed>

Public Key

npub1887fe0nagc0c5dctz6t7w70sd36k96zka4lcwga8lmfn97wct02qgjglkl

Profile Code

nprofile1qqsrnlyuhe75v8u2xu93d9l808cxcatzaptw6lu8ywnla5ejl8v9h4qpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dspt5xqp

Publishing to

relay.damus.io

nos.lol

Last Notes

2026-05-14 17:44:27 UTC

imagine a porn-free internet, a beautiful place where you could look up fanart of your favorite characters without having to bleach off your eyes afterwards

2026-04-25 12:30:43 UTC

hey, remember anubis? that anubis that blocks your connection if you don't let them waste your cpu cycles as a measure to purportedly prevent slop bots? that same anubis encourages the use of slop bots in the repo: https://github.com/techarohq/anubis/blob/main/AGENTS.md

2026-04-23 17:48:20 UTC

it's crazy that the "send me 25 dollars and i'll send you 2000 back" scam still works and that i sometimes have to explain to people that they shouldn't send money to a stranger of the internet who calls themselves "thomas james" and it's also crazy how many times my explanation fails and they do it anyway

2026-04-18 14:56:17 UTC

nothing is more temporary that a permanent solution

2026-04-15 18:17:45 UTC

is github broken again? they're not really good at keeping up those 9s

2026-04-14 16:21:49 UTC

fedi user whose profile picture is permanently broken (it might or might not be you)

2026-04-13 06:45:02 UTC

- reply

it's 2.4 gigahertz, i'm not sure whether the access point can do 5 or if it's just disabled, and if it can then i'm not sure the server's wireless card supports it, and even if it does i'm in the next room over. there's many access points belonging to the neighbors.
sometimes it works quite fast though and sometimes it doesn't.

2026-04-13 06:24:34 UTC

- reply

just usual home server stuff, ssh, http, sftp and such
i think it's a crappy wifi problem since out of 250 pings i get 1-3 retransmission when close to the access point but 400+ when i'm in the next room (but only on a specific device, the one acting as a server)
i don't think the regular pings i did are actually helping, i just need to move the server closer

2026-04-12 17:07:18 UTC

why does zen browser, often recommended as an alternative to firefox, pre-enable widevine drm by default? even firefox doesn't do that on linux. i guess foss stops mattering when a fraction of the userbase is mildly inconvenienced by clicking on a popup once so just shove a proprietary blob into everyone's browser instead.

2026-04-12 16:45:51 UTC

i've tried ssh, it's super slow
i've tried http, it's super slow
now i'm trying ssh again but i'm tickling the server every 5 seconds with ping and hoping that it will make the controller not go to sleep (if that's why it's slow anyway)

2026-04-05 19:06:29 UTC

- reply

right, maybe 3 year olds are already too old for lua. maybe lua is best enjoyed between 1-2 years of age and at the kindergarten we move on to c89.

2026-04-05 19:01:44 UTC

- reply

do you think kindergarteners would understand that? maybe we should leave that one at least until the first grade of primary school.

2026-04-05 18:56:07 UTC

they should teach lua in kindergarten as a mandatory subject

2026-04-01 20:53:44 UTC

"set -euo pipefail" is a horrible way to start a script, whoever decided to teach it to beginners needs to be thrown into the sewers

2026-03-27 23:20:12 UTC

- reply

even if there are no blobs in source, i feel like a program flow that one does not understand (especially the abundance of crypto functions with 30-line ifdef chains and various variables with matrixes that one might not be familiar with) leaves enough room for something sketchy to be going on eventually. these projects are maintained but it's hard to say whether the maintainers are playing the long game. i'm not saying that bento4 is sus, it very well seems legitimate, but i can't be sure because of the complexity.
i could have a vm that receives data and then sends back the result, but if the program was malicious is there anything stopping the program from making the output malicious too? maybe it's a bit too paranoid but it seems like in a scenario where bento4/gpac were attempting malicious stuff, they could do something to malform the video into output that triggers when ffmpeg/mpv/vlc plays it. i guess that applies to any file online though, as i don't know how it was created.
for my use case the mp4box from gpac was enough as it can do decryption too. the documentation just kinda sucks.
i went to a mailing archive on debian when researching gpac and the consensus was "it's vulnerable, let's drop it", so no patent issues:
https://lists.debian.org/debian-lts/2022/04/msg00018.html
https://tracker.debian.org/pkg/gpac

2026-03-27 19:43:12 UTC

- reply

of course i'm not qualified enough either, so if someone jia-tanned the source then i'd not be able to notice. but maybe going through the source code for even a super slim chance is better than running it blindly, isn't it? if thousands of people glance at the source code a bit, one of them might catch something.
the bento4 library i mentioned (for mp4 file manipulation when ffmpeg failed) seems to include its structs from their own headers in the crypto part which would probably make patching the codebase extremely painful. maybe the crypto part is needed, since it does decryption schemas that ffmpeg can't seem to handle?
either way i did replace it with gpac which as far as i know is completely filled with vulnerabilities (and that's why it got dropped from some distros) but it did the job at least (i ran it in a live session, lol)
the syntax is awful and instead of arguments it reads data from an xml file so i did a few wrappers for that.
generally feels like any tool that touches video is a hot mess and ffmpeg might be the most user-friendly of them.

2026-03-27 19:14:09 UTC

- reply

i mostly do the same, if a popular distro i would use uses the source then it's probably fine, otherwise i'll read the source myself. i checked some that were packaged just because they seemed like high-risk pieces of software but the software maintainer are likely doing a way better job and i obviously found nothing.
i'm wondering a bit how viable checking all the sources is though, because even after reading the full source i sometimes feel like i didn't understood it enough. for example bento4 took me several hours to read and contained a ton of cryptography so if something nasty was hiding there, jia tan style, how likely would i be to notice? it's so troublesome with tools that are complex, useful, but not popular enough for a capable maintainer to pick up.

2026-03-27 18:24:28 UTC

i'm curious, fedi. when you're compiling software from outside your distribution's repos (so from codeberg, github, someone's personal site) do you always read the full source code for the program you're building?

2026-03-23 10:29:58 UTC

can we please just get an ordinance against sound trucks

2026-03-22 13:37:41 UTC

- reply

i might have to get some spicy peppers and experiment a bit

2026-03-22 13:32:06 UTC

- reply

i guess that makes sense.

2026-03-22 13:28:54 UTC

- reply

do people actually feel it when taking a shit? i've never felt it on the way out and i like to eat spicy stuff

2026-03-22 12:28:44 UTC

having a large pet snake would fix all my problems

2026-03-15 11:36:14 UTC

six sevennnnn
https://media.blackstar.cafe/media/4569b355c42a391b3948aff46874ddc3f1528d58bd8edc89dff9bb4fa7e56b99.png
#note1qm6…gerr

2026-03-09 19:41:59 UTC

#blackrockshooter #animeart :brs:
src: https://www.pixiv.net/artworks/138980341
https://media.blackstar.cafe/media/6d7936f8bcf0d782076f8f670960a273c2088e8f27ada9d60181edc495f92cd1.jpg

2026-03-09 13:37:44 UTC

- reply

thanks, i'll give it a try. i'll very much miss the vim keybinds though.

2026-03-09 12:47:23 UTC

- reply

well, with vim getting ensloppified by the maintainers it might actually be high time to give emacs a try

2026-03-09 12:41:46 UTC

- reply

doesn't that stand for eight megabytes and constantly swapping?

2026-03-09 12:27:24 UTC

vim is taking too much of my processor, help

2026-03-02 12:42:39 UTC

- reply

that might be a part of the problem, but i remember the performance being unoptimal before making any modifications to the system.

2026-03-02 12:31:41 UTC

- reply

i run my home services on bare metal - they don't all run all the time and those that do are quite lightweight - but sadly need a vm for some classes where we run specialized linux distros or are required to do stuff that compromises a system's security. so i don't need huge speeds but it's annoying when i find my vm stopped or when input lag is more than a second.

2026-03-02 12:12:30 UTC

- reply

i might try to check out if something is causing too many writes but i doubt it. there's basically nothing running besides a window manager and it still feels bad to use. but maybe i just have too high expectations about what a virtual machine on a mid-range laptop should feel like.

2026-03-02 11:51:00 UTC

- reply

oh, i am using kvm too or at least i believe it's set up correctly. it's "usable" but still very much an unpleasant experience.
i wondered at first if the vm stops itself because of dpms but that shouldn't be it because it's still on after the screen goes black and stopped even if the screen was on.

2026-03-02 10:01:35 UTC

- reply

the host is arch linux with 16 gb of memory and 4 cores, hypervisor is qemu. an even worse experience was with ubuntu and virtualbox. i've never had a good experience with vms.

2026-03-01 20:20:53 UTC

- reply

yes please, i want a system that does not work. yes, i'd like an underpowered machine that lags via ssh. i'd love some input lag too. please also make it run on a daemon that halts itself after 10 mins of usage. if possible, maybe add a kernel taint on the host too and permission errors at rumtime.

2026-03-01 20:18:11 UTC

i hate virtual machines and everything about them so much

2026-02-27 13:35:41 UTC

- reply

i guess the assumption is that you'd see that in your own dms

2026-02-24 11:46:51 UTC

- reply

back in the 11th century when dragons walked the earth, the most evil and twisted wizards gathered to conjure such an operating system

2026-02-24 11:06:36 UTC

- reply

welcome back, how was the middle ages?

2026-02-23 09:48:11 UTC

- reply

2026-02-23 09:45:51 UTC

can we please ban java and make everything procedural

2026-02-22 18:00:50 UTC

if people accept proprietary software maybe they should accept proprietary food too
i baked some cookies that i want to sell, no you can't know what's inside them, testing them for arsenic is illegal, i can take them back after you paid for them and i'll sue you if you bake some cookies too

2026-02-18 10:09:11 UTC

honestly, ircv3 is kinda disappointing to me
maybe people will someday understand that not everything needs to be discord

2026-02-08 22:55:13 UTC

if i get a scam email i tend to reply from my dedicated mailbox and drag out the scam to get the banking/payment info and try to shut the scammer's account down, but it's getting increasingly rare that i get any cooperation from the bank. many times banks will only accept calls, or have their own custom forms that disallow anything besides letters and spaces, or don't accept messages beyond a specific character limit, or just don't accept anything at all. i'd say the success rate of actually getting a single message through is ~25% (that won't mean that the scammer's account will be taken down)
on the other hand, knowing which banks are hard to reach with abuse reports i also know which banks to use if you wanna scam someone.

2026-01-25 22:33:08 UTC

- reply

it should be okay if you avoid the brain and spinal cord

2026-01-25 12:12:40 UTC

- reply

the slow drive device is the third one i have, the slow rsync is the second one. my setup is probably confusing, but the one i'm downloading from with rsync has a good drive so that's not an issue.
i guess i'll start trying the rsync download while every other device is offline to check if that gets a better performance.

2026-01-25 12:05:17 UTC

- reply

i don't think it's throttling because i do get regular speeds, but sometimes it drops to kilobytes or just lags altogether and then errors out. it's on the local network and i didn't set up any throttling. downloading via a http server works well enough, so i suspect it might be something with the ssh server.

2026-01-25 11:55:14 UTC

- reply

the locking is happening on the third device, not because of rsync, so it won't help much there. i already fixed it my simply making the necessary changes in tmpfs.
i could use partial, but i don't want unfinished files. i want it either to transfer or not. i know it will pick up again if i re-run rsync, but still i'd be rather that a file is not broken if it fails.
the problem seems to be on the system i'm downloading from (the second one, not the one that's locking up) - if i download to it then it's quite fast, but download from it is slow and often interrupts. the ssh session is also noticeably more laggy, so the problem very likely isn't in rsync.

2026-01-24 23:30:02 UTC

- reply

i am also pushing files in a loop instead of all at once, so the drive becomes locked with all the dirty writes.
this is a disaster

2026-01-24 23:26:39 UTC

rsync is now super slow, the source device often disconnects and the download interrupts and the file is lost, another device that i will be pulling from is under constant load average of ~6 at 4 cores
this is a disaster

2026-01-15 21:48:07 UTC

- reply

it does not, and it's also unrelated to my post.

2026-01-15 21:39:50 UTC

- reply

no one's getting raped on a public train, even on the evenings there's people around, and no one's starting a fight either. if anything, there's emergency signals basically everywhere.

2026-01-15 09:57:44 UTC

a few times i've seen a young woman in public transport be approached by someone who is pushy and probably under influence and starts flirting with her despite her being uncomfortable and not being interested at all and right away a guy nearby approaches them to kick the weirdo out... but so far in all instances the guy's girlfriend stops him from helping, because "don't get involved with that" and "you're embarrassing me in public" and "let's just leave them alone"
like... i thought that there'd be at least a bit of solidarity between women in such situations and that one should be proud of their boyfriend who steps in to help someone who needs it but apparently not

2026-01-01 23:57:57 UTC

- reply

well i actually changed the hostname for the sake of my ssh prompt or else i won't know where i'm running commands

2026-01-01 23:51:08 UTC

- reply

the network was functional and i don't use mdns and refer to the device by its ip address anyway

2026-01-01 21:30:44 UTC

is it somehow an issue to have two devices on my network have the same hostname? it seems to work but i only had it that way for like two minutes so i'm not sure if it actually breaks something meaningful

2025-11-26 00:13:49 UTC

- reply

while that's technically true, in most cases it poses no benefit for the attacker so i imagine there's a much bigger market for attackers who want to get into your bank account compared to attackers that want to spoil your milk and eggs

2025-11-19 19:20:34 UTC

yo you should talk to me on whatscrapp, i've heard it's like super secure, it's even got encryption too so it must be great! source code? what's that?

2025-11-11 20:53:13 UTC

- reply

i change it up occasionally but always keep the same character, sometimes i find some art that's way too good. i might not recognize myself for a few posts though, almost replied to my own post once.

2025-11-11 20:31:35 UTC

- reply

the hair color is a bit different so i thought you were a different person for a few posts

2025-09-04 02:14:30 UTC

- reply

yeah true, but i'd assume that they'd pack up for the night and return in the morning as pretty much every other construction team does

2025-09-04 01:54:59 UTC

the construction site is *still* going after 13 or so hours, do these guys never take a break?

2025-09-03 20:13:13 UTC

totally mental :ado_load:
https://media.blackstar.cafe/media/598f8d1f9cca753f0e6e46e7df5bdc34c89887268c04a408f77e1eef2d10a7e9.png

2025-09-02 03:39:22 UTC

- reply

the fediverse glows bright tonight, be careful :blobcatglowie:

2025-08-23 02:02:29 UTC

- reply

it's true that many people use phones nowadays, but it defeats the purpose of the privacy software if i have to subject myself to tracking to use it. post quantum end to end encryption is useless if the messages are sucked up by the keyboard and used for personalized ads before the user sends them.
i'm not asking for the devs to stop developing for phones, but it would be nice to see actual good clients for their software on the desktop; one would think that if privacy is really your goal then you'd put at least some effort into it, and not put on a pseudo-privacy front a la nordvpn or protonmail.
if your desktop client is a mess then it drives away the people who would like to use your program for real privacy.

2025-08-23 01:47:36 UTC

- reply

because it's the modern times, no one knows real programming languages anymore and javascript is all the rage, the developers who write it don't care about the performance on user's side because it presumably runs okay for them.
signal-desktop is 427.12 megabytes on arch (not counting dependencies), which is almost double the size of firefox and around thrice the size of the default kernel.

2025-08-23 00:04:37 UTC

what's even the deal with all these so-called "privacy" chat software being mobile-first? mobile devices are notoriously privacy disrespecting and using one as a main way to use your software seems silly.
signal, session, deltachat all have mobile clients where the effort goes and for computers there's a half-broken ugly electron client that's apparently only an afterthought.

2025-08-22 23:57:58 UTC

- reply

okay scrap that, there seems to be no sane delta chat clients

2025-08-22 23:54:46 UTC

delta chat seems cool, too bad that i'm more likely to find a 900 kg chunk of pure gold than a delta chat user

2025-08-20 13:05:27 UTC

it once again happened that the screen of the device was showing something completely different than what was actually on there.
displayed was a normal chat window, but overlayed on top was an invisible gallery picker. clicking anywhere would have sent an image from my gallery. pressing back dismissed this picker, but nothing changed on the screen since the window wasn't visible in the first place.
stuff like this makes me not trust graphical software. it's the reason that when i click something and i'm unsure whether an action was taken i'll back out of the program completely and redo what i was doing because i don't know if no action happened or whether something happened but it's just not made visible and what i'm seeing on the screen isn't the actual thing that will be pressed when i do something.
does anyone know what's actually going on or did someone experience something similar? it's not device lag, or a device specific issue. i've seen it happen a lot on android.

2025-08-17 01:49:14 UTC

- reply

i don't interact with them, since it just feels pointless to point out that it's sus if they already warped the narrative. it just makes me wonder if people really believe bizarre stuff or boosted the post halfway through and didn't read the rest.

2025-08-17 01:34:57 UTC

- reply

i wouldn't even say it's people making disagreeable posts, but agreeable opinions that have specifically weird (and sometimes discriminatory) reasoning behind them. that makes me think that something's up, since it's easier to express an opinion without blaming a specific group of people if the opinion is universally agreed upon anyway as a general consensus.
it's usually a stray mastosoc user that finds their way onto the timeline.

2025-08-16 20:05:19 UTC

- reply

i'm convinced that at least half of the arguments that get popular on here are done in bad faith (or by someone who geniunely believes them which is arguably even worse)

2025-08-16 20:02:22 UTC

false dichotomy :ado_yabai:

2025-08-05 21:44:34 UTC

discord's maximum password length limit is 72, booooooo :akko_angrywhat:

2025-08-03 22:12:27 UTC

- reply

sure, but removing darkness from a survival game isn't one of them, the game mechanic exists for a reason (to encourage you to find lights or night vision)
https://media.blackstar.cafe/media/1754861e356786472fcabb627cb94e83efcc617b9b6ba2a865854e1d997c206e.png
https://media.blackstar.cafe/media/b489f3dcf8928b81e92bea111723bd83806c919e128d87fd25320d8207724d79.jpeg

2025-08-03 20:56:06 UTC

people are turning their gamma up with some nvidia tool in games and then justifying that "it's not cheating, if you don't do it you're at a disadvantage"
isn't that the very textbook example of cheating? you should let me use an engine in chess because others do it and then i'll be disadvantaged, you should let me use an aimbot in shooter games because if i don't then i'm at a disadvantage, you should just let me use external tools to bypass game mechanics because if i don't then i might have lower odds at winning

2025-08-01 11:30:28 UTC

- reply

that would work and would make them in order, but it would also make the timestamps inaccurate for the files that were downloaded hours earlier in the same command (which was cool to have but isn't that important)

2025-08-01 10:23:00 UTC

i want to somehow make a curl wrapper that downloads in parallel but the last bit of the file is only written after the previous file is saved so that saving is sequential

2025-07-29 21:30:39 UTC

#blackrockshooter #animeart :brs:
src: https://www.pixiv.net/artworks/131126245
https://media.blackstar.cafe/media/de22eee64280e80dbea86f902a9cd6e7be537079abe6270cea766c945779ff7a.jpg

2025-06-01 19:42:26 UTC

yeah, i use btrfs (bocchi the rock filesystem)

2025-04-22 22:05:14 UTC

- reply

thank you billionaires i appreciate your commitment to making the internet worse

2025-04-22 21:00:14 UTC

- reply

the official arch wiki now has the stupid anubis thing and you need to enable javascript and cookies and wait for the proof of work script to finish (unless you do not have the word "mozilla" in your user agent, so lynx/w3m are unaffected)

2025-04-22 20:52:30 UTC

thankfully the german and japanese arch wiki are still uncrippled, more of a reason to not forget the languages

2025-03-24 21:38:25 UTC

navigation to anywhere (when clicking on a link, page forward, page back, reload, etc.) should be disallowed when any text or image is selected
the browser should just ignore everything that results in the current page being unloaded and do nothing unless the selection is cleared

2025-03-24 21:27:47 UTC

#blackrockshooter #animeart :brs:
src: https://www.pixiv.net/artworks/128492844
https://media.blackstar.cafe/media/a9a78a470c05c53a089455a38779824c10a72b4e4ca83c32348df694ea797c8b.jpg

2025-03-24 21:25:16 UTC

i fully believe that the current landscape of technology and the "way things are done" in mainstream tech helps massively with the proliferation of scams and if the scum in big tech didn't insist on lining their pockets in any way possible then millions of grandmas (and anyone else) could be saved from falling for fraud

2025-03-21 01:00:34 UTC

- reply

no, the point of the rust coreutils is to be a replacement of gnu coreutils with the exact same functionality (and to fully replace {core,diff,find}utils according to them) but even then it's early and not every option and functionality is implemented and it's pretty tough to get every single detail to be 1:1 compatible. their repo states that "all programs have been implemented, some options might be missing or different behavior might be experienced".
that's why i could bet that this change in ubuntu can only result in more breakage with no upsides.

2025-03-20 22:53:13 UTC

- reply

i mean yeah, but there could also be unknown vulnerabilites in the rust implementation, it's a fact of how software works. maybe even a bigger chance than there being some in the c implementation that's been around for ages and that the majority of software uses under the hood.
what i could expect out of the situation is some scripts breaking since they expect the usual gnu implementation but instead get the rust one, something similar you can see when running scripts on embedded/mobile devices with non-gnu implementations.
even if they do eventually achieve complete feature parity i don't see much benefit for that effort since the implementation can either be equal or worse. but they're doing it anyway because we need to make it into rust, for some reason.

2025-03-20 22:13:34 UTC

- reply

gccrust is still in a very early stage and not suitable for real compiling, according to their repo.
i'm not against rust as a language but i see absolutely no upside here, there will be no performance benefits and it's not like there's any security issues with coreutils being made in c. the change is made purely because it's rust.

2025-03-20 21:18:42 UTC

https://www.phoronix.com/news/ubuntu-25.10-rust-coreutils
can't wait for the day when there's differences and breakage between normal coreutils and coreutils on ubuntu for absolutely no benefits but at least they're now blazing fast* 🚀🚀🚀 and memory safe 🦀🦀🦀🦀🦀
*blazing fast does not imply performance benefits

2025-03-16 20:59:23 UTC

- reply

absolutely not, if you're right and the other person is wrong it totally makes sense to just let them know that your opinion is the correct one so they can change themselves

2025-03-16 18:02:32 UTC

did you know that 90% of fedi users quit arguing right before the other person changes their opinion?
don't be that 90%, keep going, you've got this, just one more reply!

2025-02-28 21:21:52 UTC

you hate bill gates because he puts microchips into vaccines, i hate bill gates because he set my pet hamster on fire, we are not the same ^w^

2025-02-23 23:35:36 UTC

- reply

if you bundle libraries it's hard to update a single library as you now need to update every package that used it.

2025-02-23 23:30:22 UTC

- reply

you avoid overhead by not reading runtimes and using the system libraries.
flatpak is the issue there since it does not update the dependencies in a timely manner. the way that flatpak works inherently makes bug/security fixes more difficult.

2025-02-23 23:12:06 UTC

- reply

you have to read the runtimes from disk instead of loading the system libraries.
how do you know everything in the container is always using the patched versions of libraries? historically that was not always the case and software used vulnerable libraries.

2025-02-23 22:50:53 UTC

- reply

the launch time overhead and vulnerable runtimes would indeed exist there.

2025-02-23 22:50:46 UTC

- reply

does dpkg run what's inside as root? if so, that's bad.

2025-02-23 22:50:39 UTC

- reply

it's corpo software, they don't care about the drawbacks of containerization.
we're at the age where desktop software is written in javascript, don't expect sanity in developer who write such programs. it's the harmful "just get more resources" attitude.