dunno if it's just me or a general indictment of the Java everything-ORM-always-everywhere model, but it's somehow infinitely easier for me to look around around a text dump of a top-level document than it is to attempt to traverse sixty different files with fucked up accessors
Replying to @nabijaczleweli
maybe im just spoiled by languages with enums to Be Able To Get a sensible document tree
waagh, openssl contains seemingly-fully-capable blake2 implementations /and/ an EVP facility to set the output hash size, but..... ignores it entirely
Replying to @nabijaczleweli
thank fuck for sublime's seemingly-infinitely-deep undo stack, I guess, but still
Replying to @nabijaczleweli
this shit again
oddly, I don't know a Team Team
Replying to @nabijaczleweli
replied and got 554 User unknown - arslan.security.researcher@inbox.eu, which, lol
following last night's ИИ (Игристое Инцидент), I feel qualified to speak out against the blatant mislabeling of Советское Игристое as "carbonated fruit wine drink, contains sulfites", seeing as it's much better described as "hard soda, also stinks"
Replying to @nabijaczleweli
no, the label of this supposedly-fruit-wine does /not/ say what fruit it is. nor can I.
this is also the first time I've smelled anything anything that reeked as strong as a bag of pyrotechnic sulfur without actually being fine sulfur
y'know what? fuck you. *signals inside your band*
Replying to @__phantomderp
i /will/ sue for brand dilution
praise bunix
Replying to @nabijaczleweli
having a "testsuite moment" and launching 92k 7-process groups that live for 0.02ms each
Replying to @nabijaczleweli
also happens on .TS requests (hehe), and unlike .EQ I actually need them :v
Replying to @nabijaczleweli
clown moment https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=901636#52
Replying to @nabijaczleweli
how the fuck was it 3 hours already
"open source", but with the same implication as "open wound"
thanks, Dennis!
Replying to @nabijaczleweli
more thought was put into locating this than into safety handling of most software you can get from your distribution today
Replying to @nabijaczleweli
i want to say this is mildly excessive, but i know it very much is not
Replying to @nabijaczleweli
i was told UNIX did not have file types, what's this then
Replying to @nabijaczleweli
god, simpler times, wannit
Replying to @nabijaczleweli
what fresh fucking hell
Replying to @nabijaczleweli
tm – meditate
Tm does not use any files, from external influences and distractions.
Attempts to use flog(1) on tm are invariable counterproductive.
NOTE:
This page was copied from PWB/UNIX Release 2.0 (1H) and brought to you for your amusement.
Replying to @nabijaczleweli
dae double-free in 31-year-old code
Replying to @nabijaczleweli
the only problem with this is that it apparently exhausts FreeBSD's process table lol
No Sharing. No Copy/Paste. Print once. More information.
hog (!!)
Replying to @nabijaczleweli
while i had been passingly aware of some lengths to which /bin/sh was actually half the system, it's inexonerably wild to me that the /only two/ builtins in the V2 shell were chdir and login
Replying to @lambdakitten and @leftpaddotpy
this stanza is still there in V4 and V6; as of V5 (the earliest for which the TUHS has a source dump), (a) sh still does this (1) and (b) login still exits on EOF (2) (and is then restarted by init according to /etc/ttys, this is consistent since V1)
Replying to @lambdakitten and @leftpaddotpy
reading through V5 login.c again, I think this is just because of accounting, cf. (1) – t/utmp.tty are the current terminal (ttyn(0)) – if you didn't destroy the session by execing, the child user would be listed as logged in for your session after it finishes
Replying to @lambdakitten and @leftpaddotpy
a far cry from process control indeed
Replying to @nabijaczleweli
but I have to admit, it was a rewarding exercise, and I did get got by a fifty-year-old manual again
thank, s
i never understood the phrase "friends with benefits" until I started leeching off my friends' IEEE subscriptions
Replying to @_m0nt3
i mean, enterprise IEEE subscriptions have landed me a few kilobux (if bought individually at consumer rates) of standards that don't exist anywhere (I looked), so dunno about this
absolutely fucking wild shit
Replying to @nabijaczleweli
id probably rest easy just having read the PDP-11 assembly, but it's quite something to (be able to) validate that read against an actual V5 instance within seconds
Replying to @nabijaczleweli
writing an essay about sum history in the manpage again
Replying to @nabijaczleweli
while not necessarily a bad thing, it also proves the POSIX cksum non-normative section and most sum manpages wrong
Replying to @nabijaczleweli
but then again it is sum, so
Replying to @nabijaczleweli
anyway, i think this is a complete history of sum(1)
Replying to @nabijaczleweli
as with many things, this would be better if it weren't exactly what it is, but the same concept applied to something useful. alas
[devolves into self-referentiality]
Replying to @lambdakitten, @Annatar34381343 and @leftpaddotpy
(the V5 one doesn't (chdir, login, shift, wait, : (the latter 3 of which appeared in V4)), so previous ones likely don't either)
Replying to @lambdakitten, @Annatar34381343 and @leftpaddotpy
also, at least in V5, you /can/ fork login off and it does work, it just fucks up login accounting (note who output), which is suboptimal
Replying to @lambdakitten, @Annatar34381343 and @leftpaddotpy
same thing happens on the V2-on-V1 thing (https://code.google.com/archive/p/unix-jun72/)
Replying to @lambdakitten, @Annatar34381343 and @leftpaddotpy
yeah, this feature is new in V5 (undocumented, but V6 says (1), V4 lists it in BUGS as wanted), the question marks are unprintables, and I'm not sure what PID 0 is even supposed to be, I assume this is some random chunk of memory coming from the unused zeroth kernel _proc entry?
Replying to @the6p4c
buying silicon from tek and copper from holtek. real nominative determinism hours on da market to-day
kinda fucked how pretty much every alcohol and light fuel just tastes like spicy water!
Replying to @nabijaczleweli
the only thing gasolines, kerosenes, and the like have going for them is that they're a bit oily at the end. i don't really understand the point of having so many when they all taste the same
it's 3pm on a wednesday. my bank's website has given me timeouts with "try again or call support on the phone :)" three times in the past ten minutes of trying to do a single thing. how is this possible
you're the austin group chair and get this mail. wyd?
Replying to @nabijaczleweli
its here lads
Replying to @nabijaczleweli
22-hour payment-to-courier-on-my-doorstep turnaround is quite something
Replying to @nabijaczleweli
🥺
Replying to @nabijaczleweli
this photo doesn't do justice just how absurdly sharp the text is, in both senses
Replying to @iximeow
Ghostrunner does this, but all the MODULE_LICENSE stanzas I saw were Dual MIT/GPL, leading me to believe that they've thought about this, which I don't know if is worse or not?
Replying to @nabijaczleweli
glam shots
Replying to @nabijaczleweli
this may look large because the laptop is small, it's just 9.5x149.5 (3.5x143.5 + 3 padding)
Replying to @nabijaczleweli
either way, here's a poast https://nabijaczleweli.xyz/content/blogn_t/006-UNIX-r-ATT.html and source material if you wanna print your own https://lfs.nabijaczleweli.xyz/0009-UNIX-r-ATT
Replying to @nabijaczleweli
or, seeing as I have 160 of these now (80 on foil, 80 on paper, both mat), DM me if you want a few
fun fact: GNU tsort is terminally broken lol
Replying to @nabijaczleweli
the literal example from the standard (unchanged since 1987) produces fucking gibberish. makes you proud to be british
Replying to @nabijaczleweli
absolute fucking clown shit https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=990854
Replying to @nabijaczleweli
XPG2 may've come out in 1987, but it originated in 32V, /which does this correctly/, in 1978. how do I know? it fucking runs on contemporary systems
Replying to @sehetw
I mean, I read this as ordered according to the tree then broken according to the input (so tree stable over last unique of input), as does every other implementation, so dunno
Replying to @nabijaczleweli
wow this is literally me rn https://twitter.com/nabijaczleweli/status/1408845771609456642
saw a beverage that very loudly advertised "Zero Sugar!" on the paper label. sickening.
something wrong with these dogs
1978!
Replying to @nabijaczleweli
R. Morris, K. Thompson, Password Security: A Case History Encryption Computing, April 3, 1978, and this is already largely a retrospective; the searches are on a PDP-11/70. this really hasn't improved, has it
Replying to @nabijaczleweli
is "quantum overflow" just time, meaning "higher priority restored when syscall is made" or?
Replying to @nabijaczleweli
nice(2) has the literal worst documentation in existence I think
Replying to @nabijaczleweli
just count the bugs lol
Replying to @nabijaczleweli
yes i did spend the past two hours scouring fifty-year-old manuals and kernel sources
Anarchist Win? This Woman Calls Her Girlfriend 'Bro'
Replying to @nabijaczleweli
please, my execution, she is very sick
fun fact: that bundle of wire that was glowing red not five seconds ago? still fucking hot!
Replying to @nabijaczleweli
literally burned out a few layers lmao holy shit
Replying to @nabijaczleweli
thanks to that, it's no longer painful at least, but i also no longer have feeling in it
also it's noticeably stiffer, probably because its fucking cooked
Replying to @marcinskarbek
never wrote the driver for the fingerprint reader in my laptop, so dodged that bullet
vapourised holes in my finger notwithstanding, how are we feeling about this
Replying to @nabijaczleweli
⌀0.5mm titanium is always surprisingly graceful to work with
absolute fucking clown shit. no explicit label on /tmp, GNU coreutils leaks the previous context with -Z (bottom) vs reverting to the default (top)
Replying to @nabijaczleweli
https://bugs.debian.org/990962
how do i have a gold vetusware membership
Replying to @nabijaczleweli
im very much not complaining, indeed i'm grateful for't as I need to pull some SysV dumps, but it's as surprising to me as anyone
at da poast office, its pissing it down, im 3 minutes after shutter
Replying to @nabijaczleweli
slavs posting their Ls online
Replying to @nabijaczleweli
this does mean that if you'd like a few of these I'll poast them tomorrow https://twitter.com/nabijaczleweli/status/1413198119320395781
Replying to @__phantomderp
recursive constexpr templates on the relay plane when
Replying to @iximeow
i mean, beside it sounding bad, it's not really a /tremendously/ bad idea, right? GID 0 gets, realistically, nothing much with a rootfs prepared with a default 022 umask, I wouldn't think?
why mine bitcoin when we could've run, cumulatively all-time, a big ole 6MW train for a year and nine months straight instead?? missed opportunity
Replying to @iximeow
bragging rights at the -o bsdgroups dinner table
yoo, mans do be dripped tf out
pronouncing AT&T as @ ref T, accidnetally reinventing APL in the process
ok me
not a place of honour, &c.
Replying to @nabijaczleweli
Replying to @nabijaczleweli
i have a dense 600 words about the worst utility program in history (df) and SysVr4 rewrites it entirely once again
Replying to @nabijaczleweli
i no longer question why POSIX washed its hands entirely of this shit (funnily enough, the output it notes as matching by default with -P – 4.2BSD – is wrong: neither the heading nor the block size match (and rather obviously at that, since 4.2BSD switches to "kbytes"))
Replying to @nabijaczleweli
why is svr4 like this
Replying to @nabijaczleweli
yes, it fork/execs filesystem-specific helpers, too. it's 1000 lines long. the no-argument output is empty (spot the bug lol)
Replying to @nabijaczleweli
i thought the SysIII version was bad (300 lines of oddly interwoven flags and way too many options), but this is just incomprehensible
Replying to @nabijaczleweli
not to mention that there's no on-line documentation in the source or any of the distribution tapes I've seen. this is the fucking usage string. what are you supposed to do with this.
some magick indeed
ok bad post
Replying to @nabijaczleweli
if only there'd been some way to wrap this that didn't gigablow!
damn.. says a lot about society!
what's the deal with "kosher salt". salt's pareve? how could it become /not/ kosher
so you're telling me an adversary generated this network?
im separate im pure im segmented im fixed-stack im word-swapped im small model im executable im V2.3 im V3.0
Replying to @nabijaczleweli
the man set just has catpages, there are C, CP, CT, DOS, F, HW, M, S sections, corresponding roughly to what you'd call 1, 1dev, 1text, 3dos, 5, 7hw, 7, 2/3
also spot the missing mdoc macro lol
Replying to @nabijaczleweli
Intel is distributed by Intel Corporation
Replying to @nabijaczleweli
here's a suggestion: don't heat up tooling for (what I assume is) the sole purpose of getting it to an oxide purple if you can't make it stay vaguely straight lol
Replying to @nabijaczleweli
anyone else's @sqfmi watchy completely refuse to meet tolerance against the top case before filing away the PCB flare?
Replying to @sqfmi
okay so i clicked firmware update, connected with bluetooth, and uh I'm not quite sure what I'm expected to do. w.s.c doesn't have a prominent button and Docs doesn't mention OTA; I guess it's supposed to be Watch Faces -> Try it! on one of them, except it does literally nothing?
Replying to @sqfmi
oh, apparently it uses navigator.bluetooth, so when it says "with a Bluetooth enabled device" it means "with chromium". cool. could at least show something instead of quietly tossing TypeErrors
Replying to @sqfmi
you can do OTA with base16 ROM | fold -bw1024 | sed -e 's/../& /g' -e 's/ $//' | awk '{print "write \"" $0 "\""; system("sleep 0.5")}' > fifo & { echo menu gatt; echo select-attribute 86b12866-4b70-4893-8ce6-9864fc00374d; cat fifo; } | bluetoothctl actually
Replying to @sqfmi
but you'll doubtless drop packets regardless of the delay and it'll take 30 minutes because the firmwares are 1.7 megs, this really does need to listen to the change events
Replying to @sqfmi
anyway, it'd be cool to actually be able to upload a firmware image without downloading hundreds of megabytes of python and building world
Replying to @sqfmi
also, for something with actual literal physical clicking buttons, the click detection is atrocious
Replying to @nabijaczleweli
okay I lied it's actually fine (by accident), but it's space-indented and therefore unseeable on narrow tabs
Replying to @nabijaczleweli
in piracy as in intellectual property theft – always successful
Replying to @nabijaczleweli
okay, I admittedly write sparse mdoc, but I skipped most of the insanity, and have 545 lines. df.cpp is 465
Replying to @nabijaczleweli
3.5 full A4 pages of just STANDARDS and HISTORY lol
Replying to @the6p4c
is dative for like cute dog girls or
Replying to @infra_naut
if you mean piracy – yes; if not – i don't know what you mean
Replying to @infra_naut
don't know per se, outside it being the big red guy from (1) https://twitter.com/nabijaczleweli/status/1304138312744153088
im in "the big leagues" now (got mail from a .edu address)
Replying to @eng1nqueer
don't think so, no
Replying to @eng1nqueer
omg it'd be an honour
LX-branded timezone
requisite "got ownt in the group chat" poast but im cropping it worse than usual in hopes this drives engagement
no, okay, but I'm gonna; what fresh hell?
Replying to @nabijaczleweli
if you care to mime with me:
1. argv[argc] = 0
2. argv[argc - 1] == -1
3. to turn into -2
4. chroot perrors but chdir doesn't
5. the error path /doesn't actually rebind stdout/
Replying to @swagick
it did, yeah
do electronic versions of the X/Open Portability Guide exist before vol. 4? the earliest useful document i could find is the notes to the XPG3->4 changelog, but it's just sus
Replying to @nabijaczleweli
AFAICT there's an obviously-sold-out listing on amazon and there's one copy of vol. 2 in a library in greece, and I only know the latter because google books suggested it; which, again: do these, like, exist?
magnus
Replying to @winocm
classic "PATH_MAX isnt, actually" trick. love too see it
Replying to @winocm
systemd pervasively strdupa()s most paths w/o a glibc-style fallback for long ones, too; I can't wait for the poettake on this
Replying to @nabijaczleweli
SysVr3.01(?) notes that the -1/-2 thing "catches potential problems in old 16-bit implimentations(sic)" and uses a /third/ error reporting method; the why remains unknowable
Replying to @nabijaczleweli
in what can only be described as a historic blunder, I've managed to get the flag backward /while looking at it/
Replying to @nabijaczleweli
apparently argv ended with -1 not 0 before V7? this excuses the argv[argc]=0 stanza, but argv[argc-1] is a valid argument, and -2 is exactly as outside the process's address space as -1 is
UPU? more like uwu
Replying to @nabijaczleweli
i swear i had a catgirl poaster read where this was funny
Replying to @mycoliza
having read the V7 bootstrap and breeding document, as well as the porting memoranda, I would say that yes, it is
Replying to @mycoliza
that this applies to UNIXes of today, with their sophisticated memory management schemes (difficult to breed) and being 90% device drivers by volume (dominant behaviour), I would hesitate to agree with necessarily, but the NetBSD rumpkernel scheme is a step in the right direxion
have you ever used, or seen used, the multiplication (as in `[ioc]bs=69x420` for 28980) flag format for dd
Replying to @di_v_erge, @mycoliza and @hikikomorphism
(even worse, POSIX clearly and broadly states that it makes literally no guarantees on output from df that's not df -P [path]... (512) or df -Pk [path]... (1024), and pleads to the implementers, noting that df -P[k] is the only portable usage. GNU coreutils fuck /even this/ up.)
Replying to @lambdakitten
it's in V5, where dd first appears, so it was probably use(d|ful) then, what with the shell being what it was (plus, there's no expectation of being able to use non-decimal numbers in any utility I think, so that's not really a shock to the user)
this can be certified as a definitive lily moment!
Replying to @infra_naut
okay, I'll bite, what does agf mean here; is this an american thing?
Replying to @infra_naut
oh. bad /bad/ post
Replying to @infra_naut
i am very unfortunate to be aware of old blanchie's antics
it's folded, tempered, and it does its job
literal wood-side concrete wall of the military base btw
television is worse now; this is apparent. why? RGB. NTSC colour television works like the human eye. RGB is for computers. simple stuff.
i went outside in shoes not sporting a heavy vulcanised sole for the first time in years, and to quote a modern-day literary masterpiece: damn bitch, you live like this?
we need to return to the essence
Replying to @nabijaczleweli
note how (a) even V5 has no fucking clue about the dd ASCII/EBCDIC conversion tables and (b) POSIX blatantly plagiarises the conversion, to the word. truly magnificent
the thing i don't get with eARC is why a TV would be running a ZFS fork
Replying to @iximeow
among others, like the frame buffer data height, it describes a full n-dimensional frame buffer data space, the thorough understanding of which will be sure to help you take your gaming to the next level with alienware and gavin newsom
are the yellow minions with cylindrical eyes and fake glasses, like, a pre-amogus amogus?
Replying to @nabijaczleweli
pictured: a middle-aged mother in an unfulfilling marriage she can't leave because England doesn't recognise no-fault divorces
*inflates your metatext, making you big and perceivable*
m4 is just lisp but the function name is to the left of the paren
stooped so low, opening non-critical groff bugs :/ https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=991633
Replying to @nabijaczleweli
in which I advocate for .At 32v to produce "AT&T UNIX/32V" instead of "Version 32V AT&T UNIX"; this is the pettiest bug I've ever opened, I think, and yet it's one I feel very strongly about
Replying to @nabijaczleweli
more than half of SVr3's dd 1600-line implementation consists of option parsing. that's more than the entirety of my implementation
becoming a bug girl, call that arthropomorphisation
are there any белаз girls out there? 🥺
love chess. great iteration on the rogue-like concept
Replying to @oculisprimed
classic careful what you wish for moment