Replying to
@davidgerard@circumstances.run I've been calling them "the OpenBSD of android ROMs" for a while now and I do not mean that as a compliment.
Remote
Zack Weinberg
@zwol@masto.hackers.town
mastodon 4.7.3Nerd of all disciplines. Online just barely long enough to remember USENET.
Analog interests include pottery, candlemaking, board games, science fiction and fantasy, photography, and cooking.
New follows welcome, but I probably won't follow back because I can't keep up as is. Also, if you are on a large instance please fill out your profile first so I can tell you aren't a spambot.
844 Followers
222 Following
50 Posts
Joined April 12, 2025
🦉:
physiologically cismale | politically agender | he/him preferred
🌎:
pittsburgh, pennsylvania
🤪:
all ages welcome, occasional swears
Replying to
@davidgerard@circumstances.run @thesupertechie@lgbtqia.space the deranged internet cult leaders of the present day do seem to be better at recruitment than those of the 1990s
(sociology research question: are they genuinely _better_ at it or do they just have access to a much larger pool of poorly socialized people who are desperate for human connection and a sense of in-group membership?)
So it turns out that if you decide your keyboard keys are filthy and put them through the dishwasher, an unfortunate side effect may happen
(zoom in on individual keys; the gray blobs and splotches weren't there before)
anyone know offhand if there's any way to undo it? It doesn't come off with water, but it looks more like residue than bleaching to me
Replying to
@joe@f.duriansoftware.com @eevee@slime.global no joy
socketpair(AF_UNIX, SOCK_DGRAM, 0, [3, 4]) = 0
write(4, "\0\0\0"..., 1024) = 1024
shutdown(4, SHUT_RD) = 0
write(4, "\0\0\0"..., 1024) = 1024
close(3) = 0
write(4, "\0\0\0"..., 1024) = -1 ECONNREFUSED (Connection refused)
(and if I _don't_ close the read end then we're back to "data hangs around consuming potentially unbounded amounts of RAM")
Replying to
@eevee@slime.global hm, yes, you're right, i was mixing up memfd_create with open_memstream (which is, somehow, not the same thing as fmemopen)
behaving like a regular file is a deal-breaker for the actual application that wants this, though; i really do need data written to the fake /dev/null to be *thrown away*, not to hang around consuming an unbounded amount of RAM
Replying to
(2/2) Under these conditions, is it possible to create an open fd that behaves like it's open O_WRONLY on /dev/null, and _keeps_ working that way after you exit?
("After you exit" is what that one open local socket is for: this is a thought experiment, but if it was actually a coding challenge, you'd prove you've completed the challenge successfully by sending the fake /dev/null fd to whatever's at the other end of that socket.)
Does anyone make MX-compatible mechanical keyswitches that have a tactile bump, but right after that (well before the plunger bottoms out) the resistance gets *much stronger* ?
*Ideally* what I want is for the resistance past the operating point to increase in inverse proportion to the remaining distance to the bottom stop, but that might be difficult to achieve mechanically 😜
Replying to
@thesupertechie@lgbtqia.space @davidgerard@circumstances.run Well, if you're familiar with GrapheneOS's "we think we're more secure, for certain narrowly defined threat models, than any other Android respin, and we're going to be smug and insufferable about that to the point of driving away people who ought to be excited to work with us" attitude, OpenBSD is exactly the same attitude in BSD format.
Dear Wikimedia Foundation, I will not be giving you any money until you quit it with the union-busting, kthxbye
Replying to
@davidgerard@circumstances.run @thesupertechie@lgbtqia.space
(n.b. all such personal encounters occurred more than 15 years ago, when I was still a core GCC developer, and that might have predisposed them to not like me - but if that was the issue, that still says bad things about their ability to work well with others)
Replying to
@davidgerard@circumstances.run @thesupertechie@lgbtqia.space Meh. I can't really imagine Theo going slopbrained -- he's far too much of a perfectionist -- but every time I've actually had to talk to the OpenBSD people they've been every bit as smug, condescending, and inflexible as the Graphene people are. Arguably _more_.
Replying to
@jwz@mastodon.social @jspath55@chaos.social that may well be the oldest occurrence still to be found *on the web*, but I'm certain it was already making the rounds on USENET at least five years before 2003
thanks for trying, though! at least that copy's got *someone's* pseud attached
Replying to
this poll brought to you by the knob and tube wiring and the absence of subfloors in my 90 year old house
CPU architecture in which all instructions that access data in memory *require* you to specify the byte order of the value in memory
CPU architecture in which all instructions that access data in memory *require* you to specify the byte order of the value in memory
Replying to
@darius@friend.camp I also read McSquared's Calculus Primer at about age 12. Sill the most lucid presentation of epsilon-delta proofs I have ever seen.
(however, it spent *so much time* on continuity that it didn't really have enough space left over for differentials; when I got to Calculus A in high school we breezed past continuity in a couple weeks and spent months on what I'd thought was far less important)
Does anyone make a mousing device that's in two separate pieces, one piece either a trackpad or a trackball, and the other piece has a set of mouse buttons? So you can hold buttons with one hand while dragging with the other.
Replying to
@ifixcoinops@retro.social can the handle be removed and replaced with a longer pole acquired in some raccoonish fashion?
Replying to
My "something else" is the windows installed by the previous owner, which look nice when new and are decent insulators, but according to the guys I hired last spring to clean them all, something about the surface coating makes them attract dust which is then really hard to clean off
dishonorable mention to all the frickin' glass block (but that's really about prev owner's questionable opinions about where ventilation is not needed)
Anyone happen to know who originally wrote the checklist response to poorly thought out antispam measures, easiest found today at https://craphound.com/spamsolutions.txt ? I remember seeing it circulating on USENET back in the nineties, but today I cannot find any version with attribution.
Let's Encrypt was down for three hours and apparently Caddy doesn't implement exponential backoff when it starts getting 503s from its ACME server, nor does it have the sense to remember the failure across vhosts
so I have received 102 copies of the same error message from my log scanner
Replying to
So uh, anyone know a fedi client for android that either does boost and like notification aggregation, or can mute selected notifications for specific posts? I am delighted that yinz love my casual photo of a car in my neighborhood but this is getting ridiculous
Replying to
@david_chisnall@infosec.exchange Because of this, my thought is we bake into machine language the principle that registers *have no byte order*, byte order is a thing that happens when data moves in or out of memory, and you have to make an explicit choice every single time you do that. And then that gives us the leverage to insist on the same principle at all higher levels.
It's a rebuild-the-world idea but that's where all these polls are anyway.
CPU architecture in which each register is extended with the standard set of arithmetic-result flags (zero, negative, carry, etc), which are set by anything that writes that register; conditional instructions can read these flags from any register; thus splitting the difference between "one flags register" and "predicate bits" architectures
CPU architecture in which each register is extended with the standard set of arithmetic-result flags (zero, negative, carry, etc), which are set by anything that writes that register; conditional instructions can read these flags from any register; thus splitting the difference between "one flags re
CPU architecture in which there are no asynchronous interrupts, only hardware tasks that can be suspended waiting for an external event
CPU architecture in which there are no asynchronous interrupts, only hardware tasks that can be suspended waiting for an external event
WD-40: machine lubricant or prophylactic?
Replying to
@megabyteghost@masto.hackers.town @hugo@social.treehouse.systems @unsafelyhotboots@sharkey.world further research indicates they _are_ road legal in many US states but it's kind of a mess, see e.g. https://keitruckconnect.com/us-states-you-can-drive-kei-trucks/
Replying to
@SpiritualTechie@weirder.earth me and @haloedrain@toot.cafe are still driving a 2002 Mazda Protege and plan to keep driving it until the engine gives out
Replying to
@sarajw@front-end.social I love the idea and if you had one of those retro 88x31 gif badges I would put it on my website
Replying to
@keithpjolley@discuss.systems @unsafelyhotboots@sharkey.world I'm a ride height extremist. I think you should have to have a commercial driver's license to drive anything whose front grill is tall enough that, if there were a small child standing directly in front of the grill, you wouldn't be able to see them from the cab.
(I don't know if CDL training covers that, but it _ought_ to.)
Replying to
@keithpjolley@discuss.systems @unsafelyhotboots@sharkey.world TELO's still too tall. If they had a two-door model and they matched the ride height and roof height to the Mini Cooper SE that they like to compare to, then I might conceivably be interested.
Replying to
@eevee@slime.global anything done with those can't possibly keep working after the process that set it up is killed, can it?
Replying to
@davidgerard@circumstances.run @thesupertechie@lgbtqia.space I honestly think they would do that if they had a street team.
Replying to
@FrazzledWings@bark.lgbt the one in the photo is at least 30 years old and probably a stick shift (i didn't look at it that closely) but i'm sure they make 'em with automatic transmissions nowadays
Replying to
@david_chisnall@infosec.exchange I have the impression Spring's solution rests on "doors" being synchronous. If I designed a microkernel it would be strictly asynchronous: all cross-domain invocations "just" queue a message and the *only* blocking operation is "wait for next message". Backpressure is achieved by charging senders for all messages in flight. I *think* this environment could handle priority inheritance by tracking the task priorities of all unacknowledged messages.
Replying to
@david_chisnall@infosec.exchange I recall microkernels having similar problems with offering flexible scheduling policies, so that research seems like it might pay off in a bunch of directions.
Replying to
@eevee@slime.global @joe@f.duriansoftware.com I was hoping that there was something that had the right properties by accident, much like how you can get the same effect as an fd *reading* from /dev/null by creating a pipe and then immediately closing the write end
Replying to
@keithpjolley@discuss.systems @unsafelyhotboots@sharkey.world roof height is meh. it's ride height that really matters. (see other reply)
Replying to
@joe@f.duriansoftware.com @eevee@slime.global you know what that's worth trying, hang on
Replying to
@killick@dmv.community In Japan they're fully road legal including highways, so I would assume there's at least some level of crash protection in the design, but I do not know what the actual standards are over there
Replying to
@reykjalin@social.treehouse.systems @b0rk@social.jvns.ca Serious research question: suppose a search engine that uses an LLM to *interpret queries* but not to produce text; the output is the traditional set of documents that seem to be statistically relevant to the query.
The search engine's operators will need to train the model on the documents it can retrieve. How much *additional* natural-language text will they need to feed in to get something that does acceptably well at query interpretation?
possibly relevant to @marginalia@mastodon.social
Replying to
@david_chisnall@infosec.exchange It might help to understand where I'm coming from if you know that at my job (which is an almost exclusively Python shop and we deal with a lot of historical scientific data that was archived big-endian) it is regularly a problem for me that NumPy exposes the endianness of data on disk to the programmer *without* also giving them the necessary tools to handle it efficiently. You wind up copying entire multi-GB data blocks just to byteswap them.
Replying to
@drwho@masto.hackers.town @reykjalin@social.treehouse.systems @b0rk@social.jvns.ca @marginalia@mastodon.social I guess I was imagining that the LLM might be useful for handling things like, if I search for "convert A to B", I don't want instructions for converting B to A. Document retrieval really isn't my field though.
Boosted by @joe@f.duriansoftware.com
I can no longer count on getting trustworthy answers to this kind of question on Stack Overflow, so let's try here instead.
I need a *C language lawyer* to look at the functions `parse_exit_notif` and `marshal_exit_notif` at the bottom of this file and tell me whether the double casts are necessary.
Note that the portability baseline for this code is C1999 + POSIX.1-2001 so if any relevant rules changed after that point please be explicit about what changed and when.
Replying to
@david_chisnall@infosec.exchange I have had a *lot* of bad experiences with software that blindly assumes CPU endianness is interchange endianness, enough that I'm tempted to burn that bit of instruction encoding space, just to force low level programmers to make an explicit decision up front.
The real counterargument is that for this to help you'd also have to force people to annotate the endianness of every non-register variable in every language besides assembly.
And still I think it might be worth it.
Replying to
@marginalia@mastodon.social @drwho@masto.hackers.town @reykjalin@social.treehouse.systems @b0rk@social.jvns.ca I think that's the technique that I remember not being able to handle searches where the order of words in the sentence matters. If I search for "convert git to mercurial" I don't want instructions for the opposite procedure.