Open Source Software Community - it's about ethics in Code of Conducts

I had an idea while I was in the shower for a new, improved programming language. It's a simple idea, but I think it could be revolutionary.

It's a fork of rust. But it will be called, "tranny rust". And the other big change, will be instead of having the unsafe keyword. It will be called jeet. So instead of unsafe code, it will be jeet code.

Obviously it will need to use the +nigger license.
 
I learned a lot from Jeff Atwood's blog back in the day, he was always a corpo liberal faggot but I wanted to believe he was the harmless nerd type. It's been sad to realize/watch his decline into complete commie lunacy. His meltdown about DHH clowning on windows (A) was at least funny.

Admitting microsoft pajeetware is the worst choice
1763097891663.png

In hindsight he sounds exactly like rust troons, although I was too naive/young to pick up on this. And it was just the one guy saying this, afaik people weren't forming mainstream cults to screech this shit nonstop back then. Ruby/rails were rightfully mocked for the hipster fadware culture around them, but Atwood's "nazi bar" rhetoric hadn't yet caught on.
1763097929317.png
 
Anyone know how Servo compares to Ladybird? Haven't had time to take a close look, but the latter seems much further along. They have a JS engine and everything. Plus it uses C++/Swift instead of Rust.

(Although Swift is a kinda weird choice IMO. I remember Apple being very willy-nilly with breaking updates in the early days. Maybe it's stabilized since then? Hopefully they're not trying to get acquired by them, that would be turbo cringe.)
 
Around 17 minutes of compile time in total for 920 crates (I think they're crates? I don't actually know how to Rust), compared to the 30 or so minutes to compile the Linux kernel. I've never compiled a web browser before, is this good?
Your Firefox builds faster than mine. But Firefox builds fast. Chromium takes me a few hours. I'm still on an aging Ryzen 5700G tho, so that explains it. But yeah, normal experience. Chromium requires like 30-40GB of space to build.
 
Anyone know how Servo compares to Ladybird? Haven't had time to take a close look, but the latter seems much further along. They have a JS engine and everything. Plus it uses C++/Swift instead of Rust.

(Although Swift is a kinda weird choice IMO. I remember Apple being very willy-nilly with breaking updates in the early days. Maybe it's stabilized since then? Hopefully they're not trying to get acquired by them, that would be turbo cringe.)
Ladybird could already play HTML games like Cut the Rope, source: Lunduke.
1763146680586.png
 
Anyone know how Servo compares to Ladybird? Haven't had time to take a close look, but the latter seems much further along. They have a JS engine and everything. Plus it uses C++/Swift instead of Rust.

(Although Swift is a kinda weird choice IMO. I remember Apple being very willy-nilly with breaking updates in the early days. Maybe it's stabilized since then? Hopefully they're not trying to get acquired by them, that would be turbo cringe.)
I posted an interview somewhere on the site with the dev from ladybird explaining why they went with swift.


He said he got his people to try a bunch of languages. And had them decide what they liked the most for the job. He mentions them trying rust. But it felt awkward using it for web stuff, because of it's functional nature (that's my recollection at least). I'm pretty sure they tried go, and a lot of others. But apparently they all found doing the job with swift was really nice. So that's what they chose.
 
But it felt awkward using it for web stuff, because of it's functional nature (that's my recollection at least)
What he specifically said is that the web standard is written with object orientated design in mind and ladybird tries to follow the standard very closely which makes it easier to map the code to the standard. IIRC they also have comments in the code for each standard paragraph and then the code that follows the standard. Chromium does it somewhat similarly, but not everywhere.
 
Anyone know how Servo compares to Ladybird? Haven't had time to take a close look, but the latter seems much further along. They have a JS engine and everything. Plus it uses C++/Swift instead of Rust.
All I know is that I can actually compile and use Ladybird. I'm not even going to bother trying with Servo.

(Although Swift is a kinda weird choice IMO. I remember Apple being very willy-nilly with breaking updates in the early days. Maybe it's stabilized since then? Hopefully they're not trying to get acquired by them, that would be turbo cringe.)
Yeah, I'm not won over by that either. Andreas Kling said it was either Swift or Rust so I think we should be grateful. I looked but couldn't find the tweet right now. I have to say though, I don't know (nor much care) what his actual politics are but I appreciate his rejection of letting himself or Ladybird get dragged into it:
1763148574532.png

EDIT: This is what happens when I take more than ten minutes to finish a reply! :D
 
I remember Apple being very willy-nilly with breaking updates in the early days. Hopefully they're not trying to get acquired by them, that would be turbo cringe.
And on this point: IIRC, Swift is no longer Apple-managed and split off into its own foundation, at least in theory. In practice it's probably just like Go/Google and .NET/Microsoft, though.
 
1763158639618.png
1763158645444.png
Today in FFmpeg land, the main bodyguard of ffmpeg has fixed almost 3000 security issues. How much did Google reward him for such a herculean task? $7560, before tax, of course! It's almost as if most of these security issues in ffmpeg did not actually matter, huh...

Also, doesn't matter much but:
>Kieran made bad comments about google which I don't agree with
>Also his comments finally got me paid
So...did he send the money back if he disagrees with the comments that got him paid or how does it work?
 
No one will get paid in free/open source until they have a license that demands they be paid .... of course by that point the google system lords will probably be convinced the combination of jeets and stochastic parrots can write all their code for them so they'll tell FOSS developers to go pound sand. Lesson learned: corporations are the real thieves, and stealing from thieves isn't theft, so do it sometime.
 
when it comes to open source the closest you will get to payment is freelance work or people who feel your work was good enough to donate for.
 
Anyone know how Servo compares to Ladybird?
Servo is far easier to install, there's pre-built binaries and even a windows installer. Compiling Ladybird is quite something, mostly because Google is involved. I actually gave up and just used the AUR. I'm not sure that I trust the maintainer Ali Mohammad Pur <totally@fakegmail.ch> all that much, but it looks fine.

For Servo, YouTube doesn't play videos (videos don't seem to work at all) and the search function is broken, but a direct video link will load the thumbnail, comments, and recommendations (eventually). Ladybird can search YouTube, but the video page does not load properly, though it can actually load some local video files. Both seem to handle wikipedia and reddit ok, aside from videos, with some degree of javascript support, just not enough for YouTube.

But, more importantly, Servo managed to pass kiwiflare quite quickly, and the site seems to work ok, though none of the custom icons show and themes do not work. Ladybird took significantly longer to pass kiwiflare (I thought it was stuck in a loop) and just having the site open uses an entire CPU core. The page is 5 times as long as it needs to be, but I can change the theme when I finally reach the bottom of the page. (I don't know if logging in/posting works because I can't remember my password at the moment.)

Servo works fine for browsing simple HTML pages with some images like wikipedia. Ladybird is much more developed, but it still needs a lot of work. I think Servo has the best chance of going somewhere as "Electron 2" if they get the "embeddable" aspect right. Even now, it might do for a lightweight offline information kiosk type situation, depending on how much runtime it requires.

https://github.com/google/skia is a good example of why open source is not good enough.

The classic ./configure; make has served us well for thousands of years, even CMake followed the same pattern; cmake .; make. But, this is too easy. If this sort of thing is allowed to continue then people might get the idea that they can compile their own software, perhaps after modifying the source to suit their personal requirements. Thankfully Google is here to save the day.

Google code is built using ninja, and the ninja files are generated with a tool called "gn" which stands for "gay niggers"; The geniuses at Google named it after themselves, of course.

Running "gn" complained about missing files, so I figured they were using git submodules, but no, Google is too good for git submodules, they have a Python script (without its extension, for some reason).
Code:
$ ./tools/git-sync-deps
Which, like Rust, downloads the entire fucking universe, because there's no way that I could ever have software installed on my computer that other software could use. Dynamic linking is unsafe, you know? There were quite a few "429: Too Many Requests" too, when I ran Google's script to fetch repositories from Google's servers.

Finally building was familiar, if not needlessly verbose.
Code:
$ gn gen out
$ ninja -C out
This took an hour to compile, and now my CPU fan makes a rattling noise when it spins up. Thanks Google.

Now, for the next dependency. https://github.com/google/angle... What makes it even worse, is that this repo is completely different. At least the README is more than just a link to the website which has no useful information. No ./tools/git-sync-deps this time though, git submodules are good enough for Google now.

After downloading the entire universe once again, including LLVM and Rust this time;
Code:
$ ./build/install-build-deps.sh
install-build-deps.py [ERROR]: lsb_release not found in $PATH
install-build-deps.py [ERROR]: try: sudo apt-get install lsb-release
I'm on Arch (btw). It's not just a lazy error message either, the script only works on Ubuntu and Debian, which would have been useful information to put in the README instead of just referring to "Linux".
Code:
$ ./build/install-build-deps.sh
subprocess.CalledProcessError: Command '['sudo', 'apt-get', 'update']' returned non-zero exit status 1.
And if you're looking at that error message and thinking to yourself that it looks an awful lot like a Python error message to have come from a shell script, well, this is the shell script.
Code:
#!/bin/bash -e

# Copyright 2012 The Chromium Authors
# Use of this source code is governed by a BSD-style license that can be
# found in the LICENSE file.

exec "$(cd $(dirname $0) && pwd)/install-build-deps.py" "$@"
I wonder why Google is so ashamed of using Python that they had to hide it. twice.

If this is anyone's primary experience with building software, I can see why they would welcome something like cargo. I'm giving up on this one for now.

Your Firefox builds faster than mine. But Firefox builds fast.
Sorry, I meant in terms of running, "more responsive" might have been a better choice. I think compiling a real browser might actually set my computer on fire.

aging Ryzen 5700G
My i5-1035g1 says hello.

I don't know if web browsers are far more complicated than I thought, or if the kernel is simpler, but maybe 30 minutes doesn't seem so bad after Ladybird, especially when chrome looks like it will still be an overnight job for me.
 
when it comes to open source the closest you will get to payment is freelance work or people who feel your work was good enough to donate for.

Or working for a company that contributes to open source.

....or treating your code as a loss-leader for the sale of the physical object it's embedded in, the same way it used to be before software could be copyrighted. This, of course, assumes you develop something that somehow can't/won't be immediately stolen by chinks, jeets, and jews. Good luck with that. Better to just develop things that wreck the economy of thieves instead. If I can't profit, they shouldn't profit either.
 
Whenever I see people talking about monetizing free software I often feel like I'm taking crazy pills. It's all "donations this", "support contracts that". Free software by definition has basically no distribution value once it's already publicly-available. So instead of shrugging and saying "guess I'll never make money", why not just shift the extraction of value from distribution-time (the way proprietary software sellers do it) to publication-time and just sell the product once like God intended?

Here's a more concrete example: You write some program P. You create an escrow account with the terms "once this account gets $X, I will release P as free software and take the $X. Anyone can add however much or little they want to the account." Buyers want to evaluate P to ensure it does what you say it does? Good thing you wrote P and can therefore make it available under whatever terms you want, including a restrictive evaluation license or contract. Sure, the compiled version might not be redistributable under those terms, but someone going to the trouble of evaluating the software can probably just compile it themself. Of course, ideally the escrow provider would offer their own verification of P against the documentation and claims, but that seems like it might be a bit of a niche thing.

Basically, a "group buy" model. Sort of like crowdfunding except it involves a trade of actual things instead of hopes and dreams.

Am I just retarded and missing something obvious? Requires too much upfront investment? Requires that evaluators don't just break the contract and post it as a torrent (doesn't seem to stop proprietary software, but then again they're not distributing the source)? Requires good advertising? None of those seem individually to make it infeasible, is it just a combination of factors? Does using contract law to make a restricted evaluation version available upon request prior to the publication as free software just feel too gross?

Has something like this been tried before? If so, why did it (presumably) fail?
 
Basically, a "group buy" model. Sort of like crowdfunding except it involves a trade of actual things instead of hopes and dreams.
No, that's exactly crowdfunding, which succeeds only when you treat it as the patronage system that it is, and people actually want to be your patron. PROTIP: 99% of you will never have a patron.
Has something like this been tried before? If so, why did it (presumably) fail?
Yes. Someone tried something like that and it failed precisely because everyone wanted to freeride and nobody wanted to be a patron. You don't write software and then seek to build a patronage network. That's simply ass-backwards.

There are ways to make money from software, but it involves actually demanding corpos pay for what they use and suing them when they refuse. This requires you to use a license that doesn't involve free as in freeloading. I think we're finally at a point where dot.communists, combined with their glorious corporate overlords (and their perchant for replacing us all with jeets, chinks, and stochastic parrots), have pissed off enough people that they're willing to accept the idea that maybe "free software" has failed even harder than communism, and it's time to make corpos pay what they owe. I eagerly await the dawn of the Fascist Software Foundation and the glorious Datanacht to come.

Also, ich frage euch: Wollt ihr den totalen Krieg?
 
Whenever I see people talking about monetizing free software I often feel like I'm taking crazy pills. It's all "donations this", "support contracts that". Free software by definition has basically no distribution value once it's already publicly-available. So instead of shrugging and saying "guess I'll never make money", why not just shift the extraction of value from distribution-time (the way proprietary software sellers do it) to publication-time and just sell the product once like God intended?
There are some more methods. What Aseprite does is selling "official" binaries, while still having the source code available on GitHub, Some open source projects also sell their applications on stores like Steam, Microsoft Store, Google Play. These methods are not perfect but at least rake some money.

I believe the only, and best, way to fund open source software project is by government grants. Force everyone to chip in a little using taxes. Taxes already pay for physical public infrastructure, why not open source software? Right now a large portion of your taxes for software go to kikes at Microsoft and AWS. In Europe many governments want to move away from US software companies, preferring open source locally hosted systems. They do give out grants on a small scale, but it's nowhere near what it should be. To blame are those in control rather funding bureaucratic money guzzling non profits in Burundi, that are supposed to teach children about AIDS-free gay anal sex. And also the intrinsic lack of a formal organizational structure in open source projects.

the idea that maybe "free software" has failed even harder than communism
TRVKE
 
Today in FFmpeg land, the main bodyguard of ffmpeg has fixed almost 3000 security issues. How much did Google reward him for such a herculean task? $7560, before tax, of course! It's almost as if most of these security issues in ffmpeg did not actually matter, huh...
I suspect that he fixed them because he wanted to and has a desire to make their software better. It's a good thing. But it needs to be tempered with being mindful of your time. Some engineers are horrible at this which is why bug intakes have someone who looks at the requests and assigns work out as needed based upon severity, or just straight up filters out noise with E_WONTFIX.

Yet another problem AI is going to cause. Finding all of these stupid little bugs with small impacts based upon niche circumstances. Going to generate a ton of busywork for no real gain.
 
Back
Top Bottom