Show newer

@eloy Okay, but how isn't this already the case for 404 errors today?

hell yeah, love it when moderation actually happens. got a report of a remote user, opened it right away, and the remote instance had banned them before i could even view their profile lmao

@eloy Okay, but putting aside abstract notions of 'correct' for a moment, how would this technically improve things?

Because as far as I can tell, you'd just be introducing a new HTTP status code that everyone now needs to *also* support and treat like a 404, without adding any new functionality (because it's still just "not found"), so then what is the practical purpose of the new ambiguous status code?

re: matrix.kescher.at shutdown 

@kescher Re: purging, do you mean in the sense of "users automatically leave a room if their instance has been unreachable for X time", or something else?

Re: defederating instances, I assume you mean on a server level rather than on a room ACL level? For what reason(s) would they be defederated? (As it's a technical impossibility to literally defederate without breaking rooms, and workarounds are potentially viable but need a more precise problem description to get right)

Re: banning individual known-bad users, I assume this goes beyond room bans, and you (conceptually) mean something along the lines of "defederating between your homeserver and a particular remote user" instead?

@eloy The general recommendation is to just use 404 for that, making it deliberately ambiguous - because any HTTP code that reveals that something is being intentionally hidden (as opposed to not existing) would defeat the point of hiding its existence

re: matrix.kescher.at shutdown 

@kescher Tangential question: what moderation features are you missing that you feel Matrix should have had?

(Asking because I'm still working on a protocol fork, and I'd like to get this right from the start)

like trans folks, there are way more plural folks around than you realise

like, actually, but also as a joke since we're kinda by definition more folks than you realise

re: cryptocurrency, but more generally 

@AFriendlyBeagle (Also, none of this is *really* relevant to the point I was making originally, anyway.)

re: cryptocurrency, but more generally 

@AFriendlyBeagle I am well aware of all of this - and I have been around usecases like this (often activism-related) for a long time.

The thing that often gets missed, however, is what we had *before* Bitcoin. Because before Bitcoin, there was already an industry of anonymous/"risky" payments that *didn't* use cryptocurrency. Often it was more accessible to people than cryptocurrency is today.

Some examples included UKash, PaySafeCard, Liberty Reserve, and so on. These were absolutely not without their problems, and they were absolutely run by sketchy people, but they fulfilled the exact same requirements as Bitcoin, just without the destruction of climate and communities.

But the hype around Bitcoin ate basically all of them, and now you need to deal with privacy-invasive cryptocurrency exchanges instead of just buying a giftcard in a local shop, like you used to be able to.

Anonymous payments did not start with cryptocurrency, and I would argue that the diversity of them was actually another victim of the cryptocurrency hype.

So this is what I need to find:

- an on screen keyboard that works with Arrow Key + Enter navigation (to be used with a USB Remote without an airmouse.)

- a wifi configuration utility that is entirely navigable with arrow keys + enter + escape (with OSK for text entry.) Ideally something that can run in a terminal.

- ... I think that's it. If I can find solutions for those two things, I can probably whip up a prototype this afternoon.

Show thread

cryptocurrency, but more generally (2) 

"But then how do we make sure that people can eat and don't become homeless, if we can't do monetization?"

Hi! Welcome to the anti-capitalist movement! Let's get to work.

Show thread

@JennyFluff Someone was definitely looking for an excuse to use the phrase "spontaneous combustion" there :D

GenAI, NaNo meta 

A quick thought on the "criticism of generative AI is classist and ableist" bs, and why it's so fuckin' insidious:

Yes, those are genuine issues. AI is not a meaningful solution to *any* of them.

The solution to "there aren't enough non-English speaking/disabled/working class voices being published" is not and *never will be* "here's an ethically abhorrent tool that fundamentally reshapes marginalised voices into the statistical mean voice of those already published"

cryptocurrency, but more generally 

Now that the cryptocurrency hype has largely died down, maybe this will finally be the right time to say this:

The lesson from cryptocurrency ruining everything it touched, wasn't just "cryptocurrency is bad". The other lesson is that making something *about* money, of *any* kind, is the fastest way to stamp out any kind of healthy community dynamics and turn it into a race-to-the-bottom.

If you start your project by focusing on "monetization" as the goal, it will never become an enjoyable place for people to be. You cannot "monetize" your way out of a capitalist society. That people need to pay the bills, doesn't change this.

@cynicalsecurity short answer: yes
daemonising is, as a concept (forking into background), essentially incompatible with go runtime model, which implements its own M:N threading and uses OS threads rather loosely, and it's trivial to end up in a situation where the process would already have some threads started before it would reach your daemonizing code.

long answer: still yes, but daemonizing is bad anyway.
as a preface, the following is coming from being burned in many ways by processes attempting to drop privileges and daemonizing on their own. most often by silent failures with nothing on stdout/stderr/logs; but sometimes by leaking/retaining elevated privileges when they weren't supposed to.

self-daemonising is surprisingly difficult to do properly in general, arguably maybe even impossible if your code is anything but a statically linked executable directly interfacing with the kernel syscall interface (not even going through libc) because of how many things happen before "your" code is reached in process lifetime.
i've seen services dropping privileges improperly too often to trust just about any service to do so, regardless of what programming language they're written in, and instead i strongly prefer to have a service manager that would setup proper environment (privs dropped, etc etc) first, and then start the service.
if nothing else, there's less security sensitive code to audit, and it's in just one place, instead of having a myriad variations, with every service author implementing their own slightly different way of doing things.

And we're talking a fairly significantly-sized company here, this is absolutely not a sole developer or even a small team

Show thread
Show older
Pixietown

Small server part of the pixie.town infrastructure. Registration is closed.