Your twitter archive

end me

 Tue Jun 01 15:40:23 +0000 2021

funny how "stack overflow developer survey" lists "HTML/CSS" as a programming language, but doesn't list roff. mark of the times!!

 Tue Jun 01 19:14:56 +0000 2021

doesn't

 Tue Jun 01 19:17:28 +0000 2021

there's an ethnicity question and i misread one checkbox as East Sussex and i cannot measure the psychic damage i took

 Tue Jun 01 19:23:33 +0000 2021

work laptop btw, and the fans are Going

 Wed Jun 02 12:10:22 +0000 2021

happy pride month to all australians, i applaud your perseverance in face of your vowels being in what society would call the absolute wrong places

 Wed Jun 02 14:15:01 +0000 2021

me

 Wed Jun 02 18:26:20 +0000 2021

a C++ programmer wrote this template

 Wed Jun 02 21:25:09 +0000 2021

DESSICANT
SILICA GEL
DO NOT EAT

DANGEROUS
THROW AWAY

 Thu Jun 03 20:44:23 +0000 2021

babe wake up, new dessicant messaging dropped

 Thu Jun 03 20:44:42 +0000 2021

VGA GPU (400W)

 Thu Jun 03 20:49:23 +0000 2021

dont tell seasonic, but i plugged it into a 5700XT which doesnt have a vga output (but it does have the six-pin hole spaced to fit an eight-pin, which I thought rather neat)

 Thu Jun 03 20:51:38 +0000 2021

nieblo*d

 Fri Jun 04 08:52:19 +0000 2021

girls on this website, or something, dunno

 Fri Jun 04 22:37:44 +0000 2021

My Cat Smells Of Cat Spit! And Other Unexplainable Mysteries

 Fri Jun 04 23:05:59 +0000 2021

three-way multitenancy on the gameing motherboard, what a time to be alive

 Sun Jun 06 01:00:04 +0000 2021

FreeBSD (Free Berkeley SystemD)

 Sun Jun 06 08:46:34 +0000 2021

unbothered, moisturised, in his lane, well-hydrated, flourishing

 Sun Jun 06 09:27:46 +0000 2021

anyone ever get an urge to just write coreutils

 Sun Jun 06 13:57:17 +0000 2021

busybox is more than fine for what it is, but it also sucks ass for interactive use for that very reason, and GNUwno coreutils have several outstanding problems that are simply unsolvable because you can't patch them (and also they're GNU, which is reason itself to berid thereof)

 Sun Jun 06 14:00:02 +0000 2021

like we would all be better off if someone forked NetBSD coreutils and provided a mostly-drop-in-equivalent interface for GNU coreutils

 Sun Jun 06 14:02:48 +0000 2021

like, the very ability to have an rm(1) with error reporting as good as the Macintosh one without needing to send feet pics to stoolman would be good enough for me

 Sun Jun 06 14:04:46 +0000 2021

i mean, how hard would it really be to get NetBSD coreutils to be drop-in Policy-compatible, really, and to backfill the rest with simple C++

 Sun Jun 06 14:08:52 +0000 2021

yes I've seen the Rust coreutils, they were annoying and Rust doesn't (didn't? last time I tried I had to port half my dependency tree and LLVM generated broken/missing relocations) have a competent x32 back-end

 Sun Jun 06 14:10:58 +0000 2021

like, dpkg -L coreutils | grep bin/ is just 105 lines and half of that is shit no-one has ever used before and ls with presets

 Sun Jun 06 14:19:14 +0000 2021

i may be most a two-liter bottle of harslevelu in, but I still oddly think that's not that outrageous a concept, really

 Sun Jun 06 15:43:29 +0000 2021

x32 is an x86-64 ILP32 ABI; in the context of Linux support, yes, you do enable it at configuration time on an x86-64 kernel like i486, and if Rust were Go, the story'd end there,

 Sun Jun 06 19:32:06 +0000 2021

but because it isn't, like with any ABI, you also need a userland to match, and for environmental reasons (openssl, libssl, libgit2, &c.), you end up with pretty much 90%, if not more, of a normal system for any non-trivial program

 Sun Jun 06 19:33:35 +0000 2021

so, yes, you can run normal amd64 Rust binaries on an x32-capable kernel just fine (and if you're on Debian, you already are! x32 syscalls are just a cmdline switch away), but on native x32 systems using Rust like that in any real capacity means a full second set of lib*:amd64

 Sun Jun 06 19:41:34 +0000 2021

(hope that answers your question?)

 Sun Jun 06 19:41:50 +0000 2021

musl does also support x32, but i haven't seen a quad to match, yeah

producing the binaries is also just half the battle, IME; when i did get gnux32 builds, most nontrivial programs would sigbus or segfault because the codegen Wasn't Very Good (improved in newer LLVM, i think?)

 Sun Jun 06 19:57:47 +0000 2021

you're in her DMs. I'm scanning her argument vector. we are not the same

 Sun Jun 06 23:13:12 +0000 2021

uhoh

 Mon Jun 07 15:23:02 +0000 2021

waagh https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=901636#40

 Mon Jun 07 16:46:40 +0000 2021

has anyone done cppreference as man3/ yet?

 Tue Jun 08 10:39:15 +0000 2021

think I just found a bug in FreeBSD's uname(1) lol

 Tue Jun 08 17:02:23 +0000 2021

waaagh, it strips the whitespace differently than uname(3)

 Tue Jun 08 17:04:27 +0000 2021

lol https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=256486

 Tue Jun 08 17:26:05 +0000 2021

[becomes hexavalent]

 Wed Jun 09 01:48:48 +0000 2021

yo, when they do butt implants do they just.. stuff a bunch more poop in there?? or what. any1 know?

 Thu Jun 10 01:10:23 +0000 2021

holy fucking shit Windows has now decided that it will not, in fact, be displaying the Notification Area at all past very early login; what exactly is the point of a commercial system that's literally more broken than early X

 Thu Jun 10 11:09:42 +0000 2021

toggled a Widget, it reappeared, minimised total commander, its back babey

 Thu Jun 10 13:30:00 +0000 2021

not only does netbsd unexpand(1) somehow destroy the glibc allocator so thoroughly that the only way to get it to not assert in malloc is to run it under valgrind, i also think its terminally broken for sone input patterns

 Fri Jun 11 00:23:00 +0000 2021

i mean i cant really fault it, log indicates it got some fixes in 2009ish and i dont think the manpage has changed since 4.3BSD so all documentation is so out of date to be 100% meaningless

 Fri Jun 11 00:31:59 +0000 2021

plus its already annoying enough to test even with a solid baseline, i cant imagine writing a remotely comprehensive suite of test data by hand. hell, i had to compromise on /my/ suite since its extremely superlinear and grows to hundreds of megabytes very easily

 Fri Jun 11 00:38:19 +0000 2021

even then, some 512B random hunks and nonobvious -ts shuffled into 3MBish of fuzz reveal that, e.g., [2 spaces, "bread", 10 spaces, "whatever"] with -at2,26 yields [tab, "bread", tab, "whatever"], but it should be " whatever", since the w is 1 column beyond the stop, not on it

 Fri Jun 11 00:44:28 +0000 2021

im quite sure because ive spent like 10 dry hours rewriting this and have 4 different spellings of unexpand that arrive at the same mismatches and i feel like im gonna throw up

 Fri Jun 11 00:48:33 +0000 2021

pinnacle of consistency

 Fri Jun 11 09:53:22 +0000 2021

not true actually! turns out my test dataset was unrepresentative! this is still fucked for some inputs!

 Fri Jun 11 10:15:04 +0000 2021

фуцкин кмс дуде

 Fri Jun 11 20:45:05 +0000 2021

genuinely the worst time ive had in a long time

 Sat Jun 12 12:27:44 +0000 2021

three fucking days to get a sum +102 to behave right. just mercy-kill me at this point

 Sat Jun 12 12:56:52 +0000 2021

just realised one of my friends has turned 21 this week. chilling

 Sun Jun 13 21:15:41 +0000 2021

in no small part because that means that so will I this december, but in so many others because all I've done for, well, ugh, years, now, is stare at a screen and be depressed and/or angry, which is hardly of note, let alone import

 Sun Jun 13 21:25:12 +0000 2021

i think instead of overloading operator[], std::vector should just have operator+(size_t) -> iterator and operator[] should remain *(vec + off) always

 Mon Jun 14 11:23:19 +0000 2021

this goes for all standard containers like std::map, too

 Mon Jun 14 11:23:59 +0000 2021

so you're telling me an electrical machine discharged this feature?

 Mon Jun 14 12:14:01 +0000 2021

daemon (aff) vs god (pej!)

 Mon Jun 14 22:20:07 +0000 2021

passed on to its next owner today – same description, photos, and price – as is tradition

 Tue Jun 15 09:08:36 +0000 2021

i recommend simply becoming part of a standardisation body and getting standards to opine on for free

 Tue Jun 15 12:02:35 +0000 2021

well, rather, this worked for me and WG21, I assume it'd be less easy for something like SD

 Tue Jun 15 12:03:52 +0000 2021

also the packaging, which was also very clearly used before prior to me getting it, heh

 Tue Jun 15 18:56:02 +0000 2021

warning!!!!! hog :)

 Tue Jun 15 22:08:54 +0000 2021

this is also bollocks, seeing as it's been here for like a week(? reportedly) and I've last passed a full-size pig on the side of the road no more than a month ago

 Tue Jun 15 22:11:51 +0000 2021

indeed, there's a multitude of pigs that just live here; i can't imagine what point this is supposed to make https://twitter.com/nabijaczleweli/status/1251224980366471171

 Tue Jun 15 22:14:30 +0000 2021

splotchy cat in da flannel! unthinkable

 Wed Jun 16 08:35:55 +0000 2021

need a girl to take me here 🥺

 Wed Jun 16 16:55:12 +0000 2021

it's genuinely fucking insane that you can just download a 4.1BSD tape, but simultaneously it really shouldn't be an "it popped up on bitsavers" affair

 Wed Jun 16 17:00:27 +0000 2021

(i really should publish the AMIX font extraction at some point, I even got it into a sensible format for Linux to use for VTs at one point, but then I upgraded from an 1024x768-in-normal-size to a 2K-in-hand-size uhidpi laptop which pushed that back a bit)

 Wed Jun 16 17:03:19 +0000 2021

anyway I'm writing an env(1) page and apparently literally nobody notes when the single dash option happened and just says it's obsolete; so far, lessing over raw tapes, I think it split off printenv(1) in 4.2 and had it innately

 Wed Jun 16 17:06:21 +0000 2021

The UNIX Heritage Society's 4.3BSD-Tahoe/{usr,src}.tar.gz are broken :/

 Wed Jun 16 17:55:03 +0000 2021

apparently this is a known issue and there just isn't a base system of 4.3BSD-Tahoe in existence :(

 Wed Jun 16 17:56:19 +0000 2021

the official canonical source of the CSRG ISO, which purportedly contains the source set is very dead, but http://archive.org has it, but it's 60kB/s and 560M. im going to die

 Wed Jun 16 18:07:16 +0000 2021

currently in 2006

 Wed Jun 16 19:33:36 +0000 2021

design language of a warez site but its entirely software you didn't think about since 2003

 Wed Jun 16 19:34:35 +0000 2021

if ATT-SYSIII.7z is correct, env(1) existed in that form in SysIII; this is less helpful than anticipated, considering it's not in V7 and SysIII is derived from about seventy other different unixes

 Wed Jun 16 20:11:57 +0000 2021

WARNING
Over-use of this program will cause it to go away.

 Wed Jun 16 20:15:12 +0000 2021

love too types

 Wed Jun 16 20:29:24 +0000 2021

it also uses =+, =-, =| and declares initialised data like this, which is truly enlightened

 Wed Jun 16 20:31:12 +0000 2021

it's also written for a 16-bit int and too broken to link after you patch these all away such that a modern C compiler accepts them, unfortunately

 Wed Jun 16 20:32:19 +0000 2021

oh my god

 Wed Jun 16 20:32:56 +0000 2021

very much not so lol, needs =s

but a global =? -> ?=, this file, and booki() was pretty much all I needed to fix, which is astounding in its own right

 Wed Jun 16 20:36:24 +0000 2021

"cbunix_man3_02.pdf", it can't possibly be, right? oh it is. the way it was meant to be read

 Wed Jun 16 20:38:52 +0000 2021

i'm not sure what i expected from SourceCode/cbunix*.pdf, but it was /not/ filthy scans of pr(1)ed src/

 Wed Jun 16 20:43:26 +0000 2021

🥺

 Wed Jun 16 20:44:46 +0000 2021

we NEED to go back to the times when this was pertinent information

 Wed Jun 16 20:45:33 +0000 2021

i'm gonna be honest, I didn't expect to (be able to) read an actual literal UNIX manual

 Wed Jun 16 20:49:14 +0000 2021

like, ever

 Wed Jun 16 20:49:30 +0000 2021

there's just something about the register (and presumed epistemological context to a significant but less obvious degree) of UNIX manuals that you can't get anywhere anymore

 Wed Jun 16 20:53:37 +0000 2021

so true, bestie!

 Wed Jun 16 20:56:24 +0000 2021

NOTE:
This page was copied from PWB/UNIX Release 2.0 (1H) and brought to you for your amusement.

 Wed Jun 16 20:58:40 +0000 2021

"No shell", "cannot open password file", "no directory": consult a UNIX programming counselor.

 Wed Jun 16 21:03:05 +0000 2021

the last two pages of the first PDF are.. completely different? cpio(1) from UNIX 5.0, which I assume means SysV, from a line printer (or photocopied, I guess, but it looks pretty clean? but I don't think the original can be anything else), and someone scribbled over them!

 Wed Jun 16 21:06:53 +0000 2021

the second PDF starts with the much shorter cpio(1) from CB-UNIX, which likely means that that site had upgraded to a cpio(1) from a SysV distribution

similarly, 45(!) pages of the second PDF are csh(1) – native to 2BSD – printed on a dot matrix(!)

 Wed Jun 16 21:15:17 +0000 2021

a true delight

 Wed Jun 16 21:20:40 +0000 2021

> Getty also understands the "standard" ESS protocols for erasing, killing and aborting a line, and terminating a line.

this sounds much more insane when you remember ESS is the Electronic Switching System. standard protocols for killing and aborting lines on a telephone switch

 Wed Jun 16 21:23:04 +0000 2021

oddly, that's also the diagnostic /I/ issue when my input is garbage or out of range!

 Wed Jun 16 21:24:52 +0000 2021

aaaand here it is!

 Wed Jun 16 21:26:52 +0000 2021

out of SysIII's parents listed on Wikipedia (Version 7 Unix, PWB/UNIX 2.0, CB UNIX 3.0, UNIX/RT, UNIX/32V): V7 doesn't have it (well, one tape does, but in usr/lbin, and no manpage), PWB doesn't (neither the Spencer nor Bostic version from the TUHS – the intro for both says (1)),

 Wed Jun 16 21:34:06 +0000 2021

UNIX/RT isn't in the archive and there's one »"PDP-11 UNIX/RT ldp" files ?« alt.folklore.computers post, but it's narkive so it might as well not exist, and UNIX/32V doesn't have it either

 Wed Jun 16 21:39:33 +0000 2021

> Downloaded: 2 files, 608M in 2h 35m 36s (66.7 KB/s)

aaand it just has 4.4. time well spent

 Wed Jun 16 21:44:56 +0000 2021

and the winner is... 4.3BSD-Reno!

this is also congruent with the other tidbit i saw that people weren't happy about Reno doing POSIX conformance, hence Quasijarus et al.

 Wed Jun 16 21:48:46 +0000 2021

i just wanted to not lie in my manpage and somehow ended up downloading 2.7G of UNIX source and deadmen

 Wed Jun 16 21:54:26 +0000 2021

that being said, if anyone /does/ have a 4.3BSD-Tahoe source tree, I'd be ecstatic to have a copy

 Wed Jun 16 21:55:13 +0000 2021

Give tape to SYSTEM GURU.

 Thu Jun 17 01:12:21 +0000 2021

dae

 Thu Jun 17 09:50:29 +0000 2021

one hundred and seventy three dollars for a PDF of a standard one-and-a-half-times my age, for which the current versions are free from the austin group

 Thu Jun 17 12:57:25 +0000 2021

h-hello? based department?

 Thu Jun 17 13:22:59 +0000 2021

pahz-itive is not how I'd describe reading most IEEE standards, generally

 Thu Jun 17 13:26:07 +0000 2021

at least it's got a text layer

 Thu Jun 17 13:27:17 +0000 2021

everyone's heard of -1 for errors, but have you heard of the -2 supererror?

 Thu Jun 17 21:25:03 +0000 2021

as of to-day, my personality is no longer the sole magnetic thing about me!

 Fri Jun 18 17:31:06 +0000 2021

this doesnt work as well when not forced by a selfie where I look deranged in front of a hospital, does it

 Fri Jun 18 17:32:50 +0000 2021

it was definitely a nice vaguely socrealist hospital, but it didn't look like one, and i walked 2.4km mostly uphill to get there and apparently summer has started so it was 30 degrees, still, and almost high noon; can't force a gag with emulsions that are not fit for broadcast

 Fri Jun 18 17:45:19 +0000 2021

was it not the rumour that mister pfizer (who i assume excretes all vaccines) had figured out a way to express magnetism in the inoculated?

 Fri Jun 18 17:46:59 +0000 2021

i think ive just found "the good shit"?

 Fri Jun 18 19:46:14 +0000 2021

all I really want is POSIX.2, which is 100% unobtanium. NIST published POSIX.1-1992 via FIPS, which means you can freely download it. IEEE still wants over a hundred dollars for the PDF of a thirty-year-old standard

 Fri Jun 18 19:51:23 +0000 2021

the interpretation (comments) is readily available, but it's (a) one PDF per page and (b) doesn't have the thing I want, even tangentially. pain

 Fri Jun 18 20:01:30 +0000 2021

what a deal!

 Fri Jun 18 20:10:05 +0000 2021

AFAICT, nobody /actually has/ POSIX.2. not IEEE – what's labelled "1003.2-1992 - IEEE Standard for Information Technology--Portable Operating System Interfaces (POSIX(R))--Part 2: Shell and Utilities" is just interpretation requests, not even NIST, which published it as well

 Sat Jun 19 10:06:00 +0000 2021

the best reference so far is FIPS publication 189, which is 14 pages, incl. blanks, and it has what I need now (tangentially, would prefer the actual standard wording)

 Sat Jun 19 10:09:20 +0000 2021

it also lists POSIX.2, but it's 7 pages and also published in 1994, which likely also makes it likely to be the interpretation requests

it's also in the NIST library on the main floor, so who knows

 Sat Jun 19 10:09:57 +0000 2021

proverbially this close to mailing the austin group chair with puppy eyes

 Sat Jun 19 10:10:30 +0000 2021

 Sat Jun 19 13:02:33 +0000 2021

I've now mirrored all FIPS publications, and there's the entirety of POSIX.1 there, but just notes on POSIX.2 and how obsolete options apply to federal computer systems :(

 Sat Jun 19 13:09:35 +0000 2021

gibberish, AFAICT, since https://dblp.org/db/journals/cacm/cacm4.html lists no such article

 Sat Jun 19 14:22:03 +0000 2021

interesting, cheers!

 Sat Jun 19 14:23:12 +0000 2021

holy shit i think I got it

 Sat Jun 19 14:50:19 +0000 2021

http://www.oldlinux.org/Linux.old/Ref-docs/POSIX/

 Sat Jun 19 14:56:20 +0000 2021

final version

 Sat Jun 19 15:27:50 +0000 2021

bitches b going outside dressed like this and still wonder why they're being homophobiad!

 Sun Jun 20 02:01:05 +0000 2021

lads, how are we feeling about this?

 Sun Jun 20 22:22:08 +0000 2021

you nicks

 Tue Jun 22 23:41:11 +0000 2021

in good company lol

 Thu Jun 24 21:13:01 +0000 2021

just had the worst hour of my life as my computer hard-reset and i seemed to've lost all 1191 tabs in my firefox session

 Thu Jun 24 23:00:33 +0000 2021

turns out, even when you can't get to it from within firefox because it's too big, 1.5G is a happy size for appending addon userdata, even if the actual data blobs are somehow compressed and/or scrambled

 Thu Jun 24 23:01:52 +0000 2021

praise fucking be firefox developers for making it tolerant of just yeeting 99% of the sqlite database

 Thu Jun 24 23:02:35 +0000 2021

anyway, and yes this is advertising but I don't care because its saved my ass (well, life in this case) twice in as many months, use tackup, kids

 Thu Jun 24 23:03:26 +0000 2021

i pressed C-A-D -> Switch User literal fucking minutes ago. please.

 Fri Jun 25 11:42:08 +0000 2021

scouring old manuals again, and it's very cute how up to V6(!) pages were officially numbered restor(VIII)

 Fri Jun 25 13:59:34 +0000 2021

ex/cuse/ me?

 Sat Jun 26 14:55:22 +0000 2021

TYPE is a string on one or more two-digit SPCS type numbers (e.g. "01" for No. 1 ESS; "01 02" or "0102" (the space is optional) for No. 1 ESS and No. 2 ESS.

 Sat Jun 26 14:56:28 +0000 2021

(sic! tremendous UNIX typists strike again)

 Sat Jun 26 14:56:36 +0000 2021

SCCS September 16, 1975 (!!)

 Sat Jun 26 14:58:14 +0000 2021

the oldest one in this set is, of course, BJ(1L) at March 1972

 Sat Jun 26 15:14:22 +0000 2021

fascinating how "file descriptor 1/2", repeatedly, instead of some variation on the standard output stream, was a Normal thing to write at some point in history

 Sat Jun 26 15:23:14 +0000 2021

a truly scathing review of IBM's cheap mainframe programming language from the late fifties

 Sat Jun 26 16:27:07 +0000 2021

not oft do you see an official signed AT&T Bell Telephone Laboratories memo stock

 Sat Jun 26 17:49:05 +0000 2021

Unfortunately, VAX 11/780 based UNIX systems seem to be naturally immune to all viruses that infect can produces. It seems that such systems are self-VAXcinated

BUGS (pardon the pun)

NOTE:
This page was copied from PWB/UNIX release 2.0 (1H) and brought to you for your amusement

 Sat Jun 26 17:53:01 +0000 2021

haveing a terrible urge to acquiesce a small-format "¹ UNIX® is a registered trademark of AT&T" sticker

 Sun Jun 27 02:27:03 +0000 2021

 Sun Jun 27 17:00:33 +0000 2021

 Sun Jun 27 17:07:44 +0000 2021

hm.

 Sun Jun 27 17:08:21 +0000 2021

how does font licensing work. do I technically need a blessing from Sun Microsystems and/or Commodore International (or, indeed, Commodore-Amiga, Inc.) for this or do the almost 30 years and the fact that none of them exist anymore pre-emptively exonerate me

 Sun Jun 27 17:18:06 +0000 2021

thank you groff, very cool

 Mon Jun 28 11:06:42 +0000 2021

oh you'd /love/ abigail output

 Mon Jun 28 16:27:57 +0000 2021

(cf. https://github.com/openzfs/zfs/blob/zfs-2.0.3/lib/libzfs/libzfs.abi, which is, yes, user and glibc code in XML with hard widths (and paths!), everything is type-id-num for thousands of lines, and it doesn't fold redundant consts together, so you can only generate these on gcc, and it's not compatible across versions)

 Mon Jun 28 16:34:03 +0000 2021

(this appears to be the peak of (semi)automatic ABI versioning)

 Mon Jun 28 16:34:39 +0000 2021

met uhh bright splotchy today

 Mon Jun 28 20:51:42 +0000 2021

day well spent

 Tue Jun 29 20:41:28 +0000 2021

so you're telling me this guy pre-dates pursuits?

 Wed Jun 30 01:02:55 +0000 2021

this but its tweets from people who are more online than me

 Wed Jun 30 16:50:56 +0000 2021