THE MARTYR OF BUTLERIAN JIHAD

We must negate the machines-that-think. Humans must set their own guidelines. This is not something machines can do. Reasoning depends upon programming, not on hardware, and we are the ultimate program! Our Jihad is a "dump program." We dump the things which destroy us as humans!

This is the tech account of @newt, formerly located at @pureevil. If you reached this place, it's your own fault. Don't blame me for any adverse effects!

Public Key

npub14zues5n3z7schvkwaausr8q2jsct8pzcncgkp0l5rye4t7ar5rvq397yp7

NIP-05 Address

icon_of_computational_sin_at_mstdn.starnix.network@mostr.pub

Profile Code

nprofile1qqs23wvc2fc30gvtkt8w77gpns9fgv9ns3vfuytqhl6pjv64lw36pkqpz4mhxue69uhhyetvv9ujumt0wd68ytnsw43qzxrhwden5te0wfjkccte9ehx7umhdpjhyefwvdhk6y8unwa

Publishing to

relay.mostr.pub

relay.noswhere.com

Author Public Key

npub14zues5n3z7schvkwaausr8q2jsct8pzcncgkp0l5rye4t7ar5rvq397yp7

Show more details

Last Notes

2024-08-16 09:34:20

- reply

I can also bet money that this thing runs some kind of BSD lol

2024-08-08 11:34:07

- reply

@npub1eeu…ap7h these are API errors

2024-08-08 08:36:13

an error has occurred
something went wrong
If you write error messages like these, you have too many fingers on your hands for your own sake.

2024-08-04 00:43:41

Every time C people discover that their language is insane and the compilers for it are written by deranged schizoids, it warms my heart. I love these texts!
https://blog.cr.yp.to/20240803-clang.html

2024-08-03 22:13:24

- reply

@npub15dr…er4n isn't Lobsters basically HN but with a different theme? They seem to public about the same content.

2024-07-31 22:34:58

Where do you get computer science news? Lambda the Ultimate has been comatose for some time now.

2024-07-30 13:08:33

Today marks the sixth time in my programming career when I used Comonads.

2024-07-26 15:07:33

- reply

@npub1guv…6hel did that. i still have to seed every repo manually (rad seed rad:...) from the seed node.

2024-07-25 15:17:21

- reply

@npub1qzy…95gn hey! I'm trying to be positive. Except computers don't make it easy for me and there is always something broken in them.

2024-07-15 22:10:27

- reply

@npub1gh2…la5f itanium was VLIW, not Alpha.

2024-07-15 22:09:53

- reply

@npub1q7h…23uz amazing!! Itanium was an extremely cursed piece of tech.

2024-07-15 10:15:50

DEC Alpha support was dropped from Linux 6.10. I find it fascinating that it managed to outlive support for Itanium, which was Alpha's replacement. VLIW is a truly cursed technology.

2024-07-14 22:53:20

- reply

@npub1sa2…jgrv no. Just differently. You should google typicalmacintoshuser.jpg and behold what you will eventually become, should you choose this path.

2024-07-14 22:52:01

- reply

@npub1sa2…jgrv oh don't you worry, it's fucked too.

2024-07-14 22:03:15

My favorite escape-chroot-as-root for new-ish Linux is setns(pidfd_open(getpid(), 0), CLONE_NEWNS).
This causes the calling process to re-join its own mount namespace, thereby resetting the current filesystem root to the root of that namespace.
Fuck off, L'Eunuchs!

2024-07-14 14:33:12

Turns out, GitHub now doesn't render an entire source file at once if it's larger than a couple of hundred lines, instead loading it as you scroll.
Fuck off, Github!

2024-07-11 21:20:11

Hey nerds!
I vaguely remember watching a talk from CCC some years ago, where a cute chick in a suit with a tie explained how Turing-completeness of Web APIs can be used for their exploitation.
Can anyone throw me a link? I can't find it.

2024-07-11 21:17:03

- reply

@npub1sw9…kar7
https://mstdn.starnix.network/system/media_attachments/files/112/769/933/121/657/350/original/128d47f4ad61d452.jpg

2024-07-11 12:26:48

Tech is full of oxymorons, because reasons. For example, in order to fully implement Simple Network Management Protocol Version 3 (SNMPv3), your implementation must conform to 29 different RFCs.

2024-07-09 14:01:00

I would like to remind everyone again that the whole debacle around NAT66 is stupid as a retard monkey. There are several implementations and yet no RFC has been accepted, because IETF zealots keep religiously mumbling about IPv6 not ever needing NAT.

2024-07-09 05:29:43

Pretty cool paper on various manual optimisations used in HFT.
https://arxiv.org/pdf/2309.04259

2024-07-08 00:22:38

Mount UFS on L'Eunuchs challanged: FAILED

2024-06-25 10:14:18

- reply

@npub1lc8…92pf they look like pieces of clipped nails lined up.
https://mstdn.starnix.network/system/media_attachments/files/112/676/729/391/648/889/original/92b9e923598d64d2.png

2024-06-24 08:08:44

- reply

@npub1f8e…zu5u Scheme is LISP

2024-06-24 07:51:42

Lispbros, it's so over...
https://srfi.schemers.org/srfi-119/srfi-119.html
https://mstdn.starnix.network/system/media_attachments/files/112/670/506/591/368/180/original/3307cd06e6fb9bdf.png

2024-06-23 11:05:34

- reply

@npub1hwh…vepg I mean, it was clearly a shitpost on my part, but now I'm curious about your experience. Please, do tell.

2024-06-19 02:27:15

- reply

@npub1lc8…92pf ok, imagine an if-else, except the number of conditions for any given if is in the dozens.

2024-06-19 02:26:08

- reply

@npub1lc8…92pf the purpose here is to gather a list of predicates based on various conditions, combine them together, and evaluate the resulting predicate against certain inputs.

2024-06-19 02:19:46

- reply

@npub1lc8…92pf do you like C++? Hahaha

2024-06-18 22:27:25

- reply

P.S. it's still better than C

2024-06-18 22:25:40

- reply

ok i kinda do hate C++ now a bit
https://mstdn.starnix.network/system/media_attachments/files/112/639/967/077/852/975/original/1d8e53f76e209e08.png
https://mstdn.starnix.network/system/media_attachments/files/112/639/969/576/304/827/original/6c97e65f033c7300.png

2024-06-18 09:23:42

I don't know whether I should be mad at myself for writing this or at C++ for lacking this simple thing.
https://mstdn.starnix.network/system/media_attachments/files/112/636/894/656/879/183/original/ce50e6fcf49d2d29.png

2024-06-16 22:17:05

x86 is a bad architecture and I'm hopeful that it dies soon.

2024-06-10 12:30:57

ok this is rad
Source: https://feross.org/gcc-ownage/
https://mstdn.starnix.network/system/media_attachments/files/112/592/331/791/974/725/original/4ff96a413081c55b.png

2024-06-09 23:52:19

I LIVE DANGEROUSLY
https://mstdn.starnix.network/system/media_attachments/files/112/589/346/905/509/135/original/6b6e0cdabef9840d.png

2024-05-20 23:28:22

Sooo...
Does any of you nerds run Pleroma using NixOS module? Is it good? Or should I continue with my source-based install?

2024-05-17 14:46:14

Daily reminder that Mozilla today are a bunch of petty ass-backwards retarded clowns. And everyone responsible for its regression to the current state deserves to be pumped full of molten lead mixed with acid through a spiked urethra catheter.
https://github.com/dimdenGD/OldTwitter/discussions/752
https://twitter.com/d1mden/status/1783892349430460819
https://mstdn.starnix.network/system/media_attachments/files/112/456/964/638/344/518/original/b40dfab33d889bb9.png

2024-05-16 21:46:08

So, has any of you tried that Atom fork, Pulsar?

2024-05-09 18:51:49

Somehow I'm still amazed that TCP sucks so hard. There is pretty much no reason to use TCP these days, except that until recently it was the only way to handle somewhat reliable data transmission. Fortunately, QUIC is a thing today. Though, it would just make sense to throw out UDP and TCP entirely and just run QUIC over IP directly.
Inspired by this post: https://brooker.co.za/blog/2024/05/09/nagle.html

2024-03-12 14:53:13

- reply

@npub1l4z…cyh6 Guido didn't do anything for Lisps, because Python has nothing in common with Lisps whatsoever.

2024-03-12 14:26:10

- reply

@npub1f8e…zu5u ah damn.. stupid me!

2024-03-12 09:55:30

AdaCore people were pushing for a C-like syntax for Ada. I found their post about it and I can't tell whether it's trolling or they are serious.
Link: https://blog.adacore.com/a-modern-syntax-for-ada
I write at around 10 lines of code per day. With this new syntax I save up to 30 keystrokes. That’s at a huge increase to my productivity!
we have created a font that allows you to display Ada code with the new syntax <...> The font contains other useful ligatures, like displaying the Ada assignment operator “:=” as “=”, and the Ada equality operator “=” as “==”.
What for the love of fuck is this?

2024-02-11 18:20:59

This is what cock shame looks like, boys.
https://github.com/coq/ceps/blob/coq-roadmap/text/069-coq-roadmap.md#change-of-name-coq---the-rocq-prover
https://mstdn.starnix.network/system/media_attachments/files/111/914/231/148/368/511/original/f0aedb2673b834af.png

2023-12-24 09:53:24

- reply

@npub1y9y…l6hs Wayland has already become a pile of hacks. The main problem is that it’s been designed by the same group of people who made Xorg a pile of hacks, representing the same culture and mindset. If you dare to look through Wayland protocol specs (https://wayland.app/protocols), the vast majority have Unstable or Staging or DE-specific statuses and may become broken or deprecated at any moment.
https://mstdn.starnix.network/system/media_attachments/files/111/634/775/297/090/513/original/c2566e12f5328188.jpeg

2023-09-27 08:37:53

It's utterly stupid and pathetic how GNU celebrates 40 years of GNU system, given that GNU OS doesn't exist, GNU as an organisation is at a mercy of corporate overlords, most GNU projects have either failed or taken over by the said corporate overlords, and Stallman eats stuff stuck to his feet like a dirty hobo.
But at least you can have a nice meetup and interject with each other.

2023-09-20 08:41:55

Choosing a programming language for a small thingie feels like:
Python - slow and total ass to build and launch
Golang - Google's a bitch and I hate them with all my soul
Rust - too much drama, most code is littered with .unwrap(), async implemented in the most horrifying way possible
Java/Scala - JVM doesn't fit into the memory budget
Haskell - no, I just want to get away from Haskell, even for a bit
C/C++ - C/C++
Zig/Nim/HaRe - I'll be crucified by my coworkers
This all sucks, you know :(

2023-09-13 05:10:03

Looking at async Rust makes me think that C++ (a restricted subset of it) isn't that bad.
Relevant article that I mostly agree with: https://bitbashing.io/async-rust.html

2023-09-12 05:11:34

- reply

@npub1mg0…r7j0 sure. When you have infinite time and manpower, you can make NPM builds reproducible or replace NPM with Nix. In practice, nobody has those resources.
In any case, Electron software is garbage and its use should be discouraged.

2023-09-12 00:34:20

- reply

@npub1mg0…r7j0 for one, many NPM builds involve downloading and executing binaries for whatever reason, including archiving and whatnot. If you run `npm build`, you may notice that 7z is being downloaded at some point, perchance even more than once.

2023-09-11 23:34:40

It is amazing how bad the modern build systems are. The worst offender is, of course, NPM. If you use NixOS, you may know that most Electron software in Nixpkgs isn't built from source. Instead, nixpkgs just re-packages binary releases from upstream, despite NixOS being source-based distro first and foremost. Even the founding project of Electron -- Atom, also known today as Pulsar -- is shipped this way.

2023-09-11 08:17:36

- reply

@npub1z4p…4u3u Messaging Layer Security (https://messaginglayersecurity.rocks/)
It became an RFC just a few months ago. Basically, it's an application-agnostic E2EE protocol.

2023-09-11 08:13:27

- reply

@npub1z4p…4u3u SMTP doesn't support encryption period. Hopefully, we might get email-like messaging based on MLS some time in the future.
Still, I don't believe P2P protocols are applicable in many cases. Here it would just lead to much worse experience.

2023-09-11 08:08:33

- reply

@npub1z4p…4u3u
>So what exactly prevents you from using a relay node on some cheap VPS somewhere for the uptime (as one of the many options, who knows, the datacenter could burn down)?
Because this is just reinventing email that already works.

2023-09-11 08:04:20

- reply

@npub1z4p…4u3u email works because the server is online 99.99% of time. With email, server downtime is incident, not a given.
Don't mention spam here, because with p2p systems spam an even worse problem.

2023-09-11 07:57:14

- reply

@npub1z4p…4u3u so, "reliable" isn't exactly a quality these systems possess. Which is sort of my point.

2023-09-11 07:50:00

- reply

@npub1z4p…4u3u no, I assume an email-like setup where I have my email address published somewhere and people write me stuff.

2023-09-11 06:25:35

There is a certain notion among tech people that adopting a certain piece of tech or even an idea would solve all of the world's problems. For one, L'Eunuchs people believe that if all software was "free" :rms2: software, spying on people via their computers would be impossible because of course everyone is a programmer capable of removing the tracking code from every piece of software installed. Or rather, this was the case until Goolag proved them wrong with Chromium. Chromium is open source, yet it will track you til the end of the world, and there is nothing you can do short of leading an uphill battle againt both Corporate America and the Chinese Communist Party.
Another example is the fans of P2P networking protocols, who claim that every single networking application can be replaced with a P2P counterpart, including their own mother. Consider the following scenario: I decided to unwind, turned every single one of my electronic devices off, put all of them into a Faraday cage, and fucked off into the woods for three months. Three months later I return and check my P2P Email. Where exactly were the messages I was sent stored all this time? Yeah...
Just as with programming languages, there is no Silver Bullet when it comes to designing protocols and applications. And of course just putting smart license text and requiring everyone else to do the same won't save you. Neither will CoCs, by the way.

2023-09-09 14:20:59

- reply

@npub1qzy…95gn the source code for ld-linux.so might just help

2023-09-08 10:34:06

Public reminder:
This months we celebrate 15 years since Wayland was announced as the ultimate saviour of L'Eunuchs graphics. It is now legal to fuck Wayland in Germany. And that's basically the only useful thing you can do with it.

2023-09-04 12:23:40

Is there a website with programming languages news that doesn't suck? Lambda the Ultimate is almost dead. Places like HN or lobsters are just endless spam upon more spam and barely one in 50 articles there is worth reading. What am I missing?

2023-09-02 23:20:21

I would never think this was possible, but it happened. The proliferation of YAML actually made me miss XML. At least XML had schemas and you could validate your data without waiting for the shit code to crash.

2023-08-30 18:49:48

- reply

@npub17jq…v996 unicode exists

2023-08-30 17:02:05

The year is 2023. The C programming language still doesn't have a working strings library.

2023-08-26 18:29:26

- reply

oh wait.. there is no C support.
This abomination is truly cursed.

2023-08-26 18:26:43

I have decided to embrace suffering and will try to use Intellij for Haskell development.
https://mstdn.starnix.network/system/media_attachments/files/110/957/323/063/244/387/original/af93e8bc974e644c.png

2023-08-25 13:28:15

- reply

@npub1p5q…q5r0 no, you do it because you think it’s what’s right. I do it because I want fame, glory, chicks, and loads of money.

2023-08-25 11:28:51

- reply

@npub1yel…etuf
>In our projects we use a standard code style; even if we don’t agree with it, it’s what the standard dictates so there’s no room for discussion.
I love sabotaging projects like this! It is always entirely possible to come up with an insane ridiculous formatting that passes format check of whatever have you. And then I can sit back, relax, and enjoy autistic screeching in Slack.

2023-08-24 23:05:14

- reply

SO MUCH AUTISTIC SCREECHING

2023-08-24 23:02:25

My coworkers turn autistic screeching mode on when they review my code, because I format it a tab bit differently. Hence, I get lots of messages about BAD FORMATTING (which is actually good formatting, they just have no taste) and no real feedback.
:akko_giggle:

2023-08-22 14:06:06

- reply

@npub17eh…94rr thanks! I owe you one :2bcomfy:

2023-08-22 13:34:48

Scihub doesn't work for me today. If anybody can send me the PDF of this, I'll be grateful!
https://opg.optica.org/jot/abstract.cfm?uri=jot-72-6-464

2023-08-20 19:58:23

- reply

@npub14x7…wdx5 no, I was thinking about altering colours of sequences of pixels or perchance distinctly shaped blocks of pixels in order to save a short string of bits in the image itself. Something similar to that printer steganography thing below, but for images. I.e. something you will only see if you look really hard.
https://en.wikipedia.org/wiki/Machine_Identification_Code

2023-08-20 19:07:39

Hey lads!
Can anyone point me in the right direction to learn about modern image watermarking? Particularly, I'm interested in the kind that survives repacking, resizing, partial crops, or light colourgrading.
Repost plz

2023-08-20 19:06:00

- reply

@npub1gpc…3t70 lmao why? It's just a vacuum tube.

2023-08-19 21:21:21

I talked to an oxymoron today. Her official title is "Privacy Engineer" at Google.

2023-08-19 21:15:17

- reply

@npub1fdc…82j5
>An individual has a phone number
>Some people do not own phones, or do not wish to provide you with their telephone number when asked. Do not require a user to provide a phone number unless it is essential, and whenever possible try to provide a fallback to accommodate these users.
It's kinda ironic to read this from a company that requires a phone number to use their services. I love that you're defending them here.

2023-08-19 21:13:47

- reply

@npub1fdc…82j5
>I bet you check emails using regex considering you oversimplify and ridicule this problem, huh?
No, not really. I check that there is a @ somewhere in the middle and that the part after it looks like a domain name.
>fuck you
Well, fuck you to. Have a nice one!

2023-08-15 20:15:42

- reply

@npub1fdc…82j5
>People want to also use it in async contexts without having concurrency issues?
Wanna know the ancient secret of not having concurrency issues? It's called not having shared mutable data in the first place. And for a library that is supposed to provide pure functions, none seems required whatsoever.
>Also I love how you didn't engage with the rest of the arguments I laid out because it obviously makes you look dumb
Tbh the rest of your arguments make even less sense than the one about async above. All that data is just compiled overbloated XML that shouldn't be there in the first place. First of all, because there are more sane ways to specify all that data. And secondly, because most of it is redundant anyway.
So yeah, it's a typical overly bloated solution to a problem that shouldn't exist in the first place.

2023-08-14 16:58:22

- reply

@npub1fdc…82j5 why would a phone number parsing library need its own synchnisation?

2023-08-13 22:25:57

- reply

@npub1fdc…82j5 are there thousands of options?
Just fyi, the first linux kernel release 0.0.1 was around 10k lines of C. You could make a working Eunuchs clone with a shell and a text editor in 50k lines. If you want an example of software bloat, libphonenumber is it.

2023-08-13 22:22:36

- reply

@npub1fdc…82j5 sure. Does stripping two leading zeroes require thousands of lines of code? Probably no.

2023-08-13 22:20:24

- reply

@npub1fdc…82j5 if i'm gonna be a complete dork, 00 isn't a way to write +. 00 is the prefix signalling that the following number is an international format. All because those all pulse mode phones had only 10 possible signals (from 1 pulse for 1 to 10 pulses for 0). +, *, or # weren't a thing.

2023-08-13 22:09:44

- reply

@npub1fdc…82j5 ah cool cool... or you could just go with a drop-down list with country codes, add some digits to it, and call it a day.

2023-08-13 22:04:25

- reply

@npub1fdc…82j5 also, libphonenumber C++ code isn't all dedicated to parsing phone numbers. There also are locking primitives, singleton implementation, and what not. I'm sure what this thing parses is actually a kind of lisp in a phone number disguise and it's Turing-complete.

2023-08-13 22:01:56

- reply

@npub1fdc…82j5 0040764123456 doesn't even look like a phone number, so what is the point of parsing it as one? A similar sequence of numbers could be literally anything. Most software has dedicated fields with special syntax for phone numbers where input is a lot easier to parse, and parsing just arbitrary sequences of digits isn't the best of ideas.

2023-08-13 21:48:46

Google's libphonenumber contains almost 50k lines of C++ code and not that much less of Java and Javascript. All of that is needed to parse and validate phone numbers. Just think about this for a second.
Your input is as simple as "1 323 555 66 69". With or without a plus sign in the beginning. Maybe the country code is missing and passed explicitly. And you need 50k lines of C++ to parse, validate, and format this data.
This is real code, written by real programmers. Someone out there makes over $200k annually (i actually don't know how much people at Google make) for writing this. This is the most elegant solution one of the leading software companies could come up with.
I can almost guarantee than any single lad reading this post can do better. Not that I have high opinions of you, although you're probably fine, but because libphonenumber is an atrocious piece of shit. It must have taken a huge amount of effort to make it this bad.

2023-06-03 10:34:51

I didn’t know about this before. Someone has bolted on kinky Qt6 and Python fuckery on top of Emacs, allowing for embedding whatever you can wish for into your Emacs OS.
https://github.com/emacs-eaf/emacs-application-framework
Now, if only this allowed for having a decent UI in Emacs…