Notices by LisPi (lispi314@udongein.xyz), page 2
-
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Monday, 27-Apr-2026 05:00:14 JST
LisPi
@enigmatico I think the argument is silly though.
Why does knowing reality is false matter at all?
The question should be whether you like it or not. -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Saturday, 25-Apr-2026 03:58:29 JST
LisPi
@icedquinn @mia @navi > dhall wanted to use it to avoid being reliant on github, but claimed latency was problematic. they went to github dependence.
That's kind of a mismatch of purpose. Uneven zero-guarantees seeding should be used for high-latency/delay-tolerant purposes more so than anything that requires immediacy.
> buttcoiners dropped out of ipfs when the NFT astroturfing failed.
That's good they're not welcome.
> its still a neat piece of tech but something about the tree model needs to be refactored for it to be useful (there are some attempts to make a new ipfs that does that; i'm not convinced they will succeed.)
I see. -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Saturday, 25-Apr-2026 03:43:09 JST
LisPi
@mia @navi Last I checked IPFS' main problem was the lack of an anonymization layer.
Has something annoying happened to it? -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Saturday, 25-Apr-2026 02:53:00 JST
LisPi
@ska @dalias I'm getting the vague impression you mean something different by "crowbar" than the prying tool I'm thinking of. -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Friday, 24-Apr-2026 13:10:03 JST
LisPi
@radmin @Suiseiseki @prettygood @dps910 Cooperating or collaborating with authoritarians, being one of the "good ones", doesn't work.
Anyone with basic skill at pattern recognition would've noticed it across the decades. -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Wednesday, 22-Apr-2026 16:11:13 JST
LisPi
Sometimes my brain has funny ideas.
For about half an hour the other day I had the amusing misconception switching endianness would require full mirroring of the bytes.
I'm also almost certain I've either posted or read something similar (of someone having this experience) on Fedi. -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Wednesday, 22-Apr-2026 16:11:12 JST
LisPi
While that's wrong and it's solely relevant for multi-byte values, I'm now wondering how weird it would be if that was actually the case and what could've led to such a thing happening for real.
(The order is arbitrary, after all.) -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Wednesday, 22-Apr-2026 11:23:11 JST
LisPi
@dalias I'm thinking of say... a component with 8-bit or 16-bit wide access to another.
The ordering of the pins which represents 0 and which represents 7 or 15 seems not to vary.
Is that purely a consequence of most early designers speaking English and the whole left-to-right ordering progression? -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Wednesday, 22-Apr-2026 10:55:20 JST
LisPi
@dalias CPUs all seem to order their words the same way, which probably matters as far as mapped registers & electrical interconnection goes.
But surely there were some early attempts that did not coordinate and ordered it differently?
Or did pre-existing electrical standards lead to that just not happening in the first place? -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Wednesday, 22-Apr-2026 10:51:41 JST
LisPi
@dalias I'm vaguely curious why there was unanimity over word ordering. -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Wednesday, 22-Apr-2026 02:52:33 JST
LisPi
@dalias Isn't that something compiler-checkable in C now? -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Wednesday, 22-Apr-2026 02:27:35 JST
LisPi
@dalias > contractual constraints that analyzers can't see.
How does that work? -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Wednesday, 22-Apr-2026 01:58:06 JST
LisPi
@dalias @burger @masklayer Generally one should always use such cables if it's an option, for exactly that reason. -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Tuesday, 21-Apr-2026 07:54:50 JST
LisPi
Has there been a reckoning for the employees, pilots & so on that knew and did nothing?
-
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Monday, 20-Apr-2026 11:06:48 JST
LisPi
@dalias @navi On regular files, yes.
But if one builds FUSE filesystems for purposes similar to sysfs endpoints, that analogy breaks.
And the index limit for mmap is just a uint64 (this lacks equivalent infinity semantics).
One write use-case that comes to mind is a hash/checksum FUSE filesystem, where the input can indeed be of any length, from zero to infinite (though the trigger write to finish the digest will never happen in that last case).
Such "special" files are typically non-seekable, too.
(Yes, the ability of FUSE to serve as a poor man's 9p IPC interface is something I care about.)
-
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Monday, 20-Apr-2026 10:36:40 JST
LisPi
@dalias @navi Does mmap cope with infinite streams?
I'm just trying to think of reasons why read/write may be necessary. -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Monday, 20-Apr-2026 10:24:59 JST
LisPi
@navi It really is. Most of that stuff has no place in the kernel.
For performance FUSE also have io_uring support, though it seems unavailable in Debian.
-
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Saturday, 18-Apr-2026 07:14:21 JST
LisPi
@icedquinn @zzt @lanodan @mia @navi DIY supports autonomy, can't have that.
Of course the only people that learn how to do things for themselves are <insert favorite horseman of the infocalypse>.
-
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Friday, 17-Apr-2026 12:35:18 JST
LisPi
@fluffykittycat @mttaggart @cwebber Unsurprisingly, abusers don't like it when they can't isolate their victims anymore. -
Embed this notice
LisPi (lispi314@udongein.xyz)'s status on Wednesday, 15-Apr-2026 02:57:21 JST
LisPi
@dalias What I see is that they still failed to armor the cabin properly so the users will still die horribly the next time a deer or moose decides to throw itself in front of them.
The parts they armored are all the parts that don't matter because they were already engineered as a safe crumple zone.
What they did do is increase its lethality against children and pedestrians. A complete failure of a design.