robryk @robryk@qoto.org

@grrrr_shark

Huh. I thought this is a universal property of humans (for basically any axis of capacity I can think of).

@koakuma

By "not seeing hazard pointers used outside of theory" I meant that in places which would use them I'd see RCU used instead (with the thing that deallocates -- which might be a thing to the side of everything else -- waiting on the RCU "lock").

@koakuma

I was under the impression that ll/sc in hardware is terribly rare, so everyone makes do with CAS. They then use of the tricks to prevent ABA in CAS-based setups (either counters we trust to never wrap around, or pointers to objects that the actor who does the CAS keeps alive[1]). If we had ll/sc those tricks would be unnecessary.

So, I think that people don't use that property of ll/sc because they don't use ll/sc, because it's very rarely available.

[1]
roughly:
- read ptr from X
- increment refcount on *ptr
- read X again, verify it's still equal to ptr (if not: decrement refcount, start from scratch)
- ...
- CAS X from ptr to something else
- decrement refcount on *ptr (and potentially handle the case when it decrements to 0)

This guarantees ABA-freeness of the CAS if each potential target of the pointer in X can become it only once in its lifetime (which is pretty common in structures such as queues, stacks, approximate queues, ...).

This is a reasonably common pattern in lockfree data structures that need to do their own memory management (i.e. for environments without a garbage collector) and that don't want to use thread IDs (if they can, refcounts often get replaced with hazard pointers; that said, I've never seen hazard pointers used outside of theory).

@LukaszOlejnik

There was a case 10~20 years ago of a teenager who started triggering emergency stop for fun using self-modified radio equipment. I was actually surprised that no vandalism-leaning person did something similar in a massively disruptive way (by hiding multiple transmitters and doing so in a well-chosen location), so I wouldn't exclude the vandalism hypothesis yet.

BTW. The physical proximity requirement is not that strong: the range is limited by terrain and R^2, and fixed transmitters of rail dispatchers have a range of iirc small tens of km. I would wager that using a transmitter with ERP of 100W or so from a good vantage point would give you at least as much.

@rysiek

OT: co dokładnie znaczy ogólnobojowy?

@rysiek

Istnieje stary odpowiednik takich działań: poszukiwanie "dźwiękowego sobowtóra", który umie naśladować głos naszego "celu". Czy wiesz może o przypadkach, kiedy ktoś w jakimś podobnym kontekście takowego w przeszłości używał?

@ftdl

Serwujecie na pol.social media z niewłaściwym Content-Type. Np takie https://pol.social/system/media_attachments/files/110/935/884/311/554/703/original/be66df9224d6be69.webp ma `Content-Type: text/plain`, a jest obrazkiem. To powoduje, że linki do nich (np. te w klientach innych instancji) otwierają stronę robaczków miast obrazek.

@kravietz

Thanks for making me realize that there were two attacks (that the S-400 AD system was attacked earlier).

@lutzky @delroth

If you want to understand the mechanics of this in Fedi, there is no single description I know of, but reading https://www.w3.org/TR/activitypub/#inbox-forwarding helped me understand how the heck all of this is not magic.

@delroth @lutzky

One set of relevant situations is what delroth described above. Another situation is posting something and using your effective moderation power over responses to create an impression of a wide agreement (note that the more egregious variant of this, where you call someone out whom you've blocked, was impossible on Twitter, because you couldn't mention people you've blocked).

There are multiple approaches to dealing with those issues. One of them is what Twitter has (used to have?). Another very similar one is the G+ approach of making top-level posts very different from comments, and explicitly giving moderation power over comments to post's author (this differs insofar comment authors do not get to moderate responses to their comments). Another one is what Fedi would approximately have if all servers were single-user: you see responses that the post's author approves of (passively), and responses of everyone you follow. (Multi-user-serverful Fedi expands this to responses of everyone anyone on your server follows.)

I think that the distinction between post and comment is something that's badly missing in Fedi, and is probably the source of most of the nonmisguided[1] requests to defederate from a server A due to that server not defederating from B, whom we consider bad.

[1] Quite often they come from some misunderstanding on how silencing and defederation work, which is not surprising given that most of this appears to be emergent properties of a system.

@madargon I wonder whether the extensions that allow one to somewhat limit agent forwarding can be helpful here. (After all, if one limits agent forwarding to only be used to authenticate to host foo, then something somewhere has to be able to evaluate this predicate.)

@lutzky

I'm not sure if we're talking about the same situation.

Let's say Alice blocks Bob and posts something. Now it's impossible for Charlie to see Bob's replies to Alice. Note that Charlie's client is not aware of Alice having blocked Bob, so even if Charlie was willing to cooperate, they wouldn't be able to reproduce the same effect client-side.

@rysiek

flightradar24 claims that there's ~1 plane/hour within ~30km of Ozyorny (I didn't try to figure out more precisely where the crash was) during daylight hours. That coupled with differences in size, speed, and altitude between drones and private jet aircraft make it IMO very unlikely that it was accidental.

@delroth Did you happen to try measuring its resonance(s)?

@lutzky

I thought that people you blocked couldn't reply on your posts (or, expressed in observable terms, that neither you nor unrelated 3rd parties would see their posts in your replies).

@janl @grrrr_shark

In my experience they heat up just as slowly. Is it also yours?