Your twitter archive

klapki v0.2.0 released, with a fix for some firmwares that would never settle, a manual rewritten in mdoc, and a proselytisation sexion!
Announcement mail: https://lists.sr.ht/~nabijaczleweli/klapki/%3C20211201003913.6oqmyp7covieu6qj%40tarta.nabijaczleweli.xyz%3E; full changelog: https://git.sr.ht/~nabijaczleweli/klapki/refs/v0.2.0; x86 Debian sid+bullseye packages: https://debian.nabijaczleweli.xyz/README

 Wed Dec 01 00:42:07 +0000 2021

compelling point! :0:0

 Wed Dec 01 01:00:17 +0000 2021

(but also, you seem to've a "<author><name></name></author>" stanza, too, somehow? (and <title type="html">, which is just… odd)) (also also, going by the incredible speed at which im managing to suck off the feed and images alike, is this coming directly from AU-grade dial-up?)

 Wed Dec 01 01:06:06 +0000 2021

in many ways this is true, but I'd nevertheless be remiss as an WG21 NB sitter to not insist that just because it's "valid" doesn't mean it's not "odd" (indeed, in RSS as in C++, these are usually homotopic) or that it should be "acceptable societally"

 Wed Dec 01 01:15:48 +0000 2021

hm, on a time trial it seems to get a reasonable speed (wget peaked at like 8ish MB/s) but firefox seemed to be having a real bad time going past the tcp slow start speed on the car poast

 Wed Dec 01 01:24:23 +0000 2021

y'know what? fuck you. *shafts yuor vuwils*

 Wed Dec 01 02:50:39 +0000 2021

what fresh hell do these jpegs mean and why must i see them

 Thu Dec 02 12:22:09 +0000 2021

ah, that i have acquiesced my twenty-oneth. most regrettable.

 Thu Dec 02 15:34:21 +0000 2021

Amazon.‌co.‌jp <support-amazon.jp@r6pk6wij.cn>
Amazon
24
Amazon
Amazon
:
24 Amazon
Amazon

 Thu Dec 02 21:05:49 +0000 2021

i think it reflects badly on me that i knew what the mail said when it was rectangles. nevertheless, quite surprised there isn- oh there *is* a link, but only in the HTML version. outstanding.

 Thu Dec 02 21:10:20 +0000 2021

full text, see alt text for actual text

 Thu Dec 02 21:10:47 +0000 2021

it passed postgrey and there *is* a DKIM-Signature for the right domain, but no Authentication-Results from my MX, apparently "opendkim: key retrieval failed (s=default, d=r6pk6wij‌.cn): 'default._domainkey.r6pk6wij.cn' query timed out", which is quite amogus impostor

 Thu Dec 02 21:16:52 +0000 2021

doubly weird, since i can see it with dig, too, from the very machine that hosts my MX?

 Thu Dec 02 21:23:34 +0000 2021

following a restore from backup and full update, im doing another backup, and the disk is it's rattly as hell and achieves 115% utilisation and 9MB/s when its reading sequentially, usually closer to 3, both completely wild numbers to me lol

 Fri Dec 03 00:08:16 +0000 2021

hm okay why does ghostscript from FreeBSD pkg(7) pull in dbus, swtpm, tcsd, gnome components and fucking cups

 Fri Dec 03 14:01:09 +0000 2021

bonus points for xorgproto in the -base package which is explicitly No X as compared to the -x11 variant

 Fri Dec 03 14:03:29 +0000 2021

oh boy, serial copy actually saturates 1Gbps, I don't think I've ever seen something destroyed by fragmentation /this/ badly

 Fri Dec 03 15:10:38 +0000 2021

oh wow, it's actually, like, usable after copying it to a rectangle. this really do be what fragmentation does to a mf, huh

 Fri Dec 03 20:16:53 +0000 2021

in a welcome and not at all unplanned coincidence, we're also observing bandcamp friday – acquiesce my album(s) https://nabijaczleweli.bandcamp.com/

 Fri Dec 03 20:57:09 +0000 2021

Phoebe Bridgers' Motion Sickness is actually prospective-utopian/wish-fulfilment; this is evidenced by bar 15 (3 of the chorus) – "There are no words in the English language"

 Fri Dec 03 21:37:49 +0000 2021

FWIW, .patch works, too, and includes all commits format-patched in order

 Sat Dec 04 00:43:08 +0000 2021

unrelatedly, can some FreeBSD wiz explain how you're supposed to use FreeBSD's built-in ZFS if libzfs.h fails to find libnvpair.h (pkg(7) has "libnvpair" with the header but (a) what the hell and (b) it says it's unmaintained, and even after forcing it it's missing sys/mnttab.h?)

 Sat Dec 04 12:10:25 +0000 2021

like, is it just me or does FreeBSD have, despite (because of?) constant large-scale corporate promotion, consistently the lowest QoI of modern berkeley derivatives, by a pro-verbial country mile

 Sat Dec 04 12:13:12 +0000 2021

if you're reusing linux chunks, might as well also reuse the EFI strap so you don't need a bootloader at all

 Sat Dec 04 22:49:03 +0000 2021

(this is The One Thing Distributors Don't Want You To Know about. linux has a full-fledged and quite powerful EFI strap if you Just Invoke It™, and loads initrds/DTBs it gets in the cmdline – bitbang w/efibootmgr or something more sophisticated like klapki; OoB SecureBoot, too)

 Sat Dec 04 22:52:59 +0000 2021

pdf.js support pinch zoom challenge 2021 gone wrong

 Sat Dec 04 22:57:36 +0000 2021

i apparently dont have enough confectioner's sugar (substituted with commercial pink meringue) or pale oil. big mistake.

 Sun Dec 05 01:25:02 +0000 2021

eats alright though

 Sun Dec 05 13:18:46 +0000 2021

are we perfectly sure 1600x900 is the native resolution of a 1440p display?

 Mon Dec 06 12:59:10 +0000 2021

yes, 1600x900 is the resolution of the display windows has designated as "primary". simultaneously: who cares

 Mon Dec 06 13:00:13 +0000 2021

bitches really out here thinking they can pull a nice high-quality milf without an anomalous body plan. please.

 Mon Dec 06 15:06:13 +0000 2021

literally what the piss is this post

 Mon Dec 06 16:13:56 +0000 2021

bonne chance! 😈

 Mon Dec 06 18:27:44 +0000 2021

RT @__phantomderp: They're dating !

 Mon Dec 06 18:27:59 +0000 2021

fruit

 Mon Dec 06 23:56:34 +0000 2021

love too post a linux-man patch and need to have it posted to and opined upon by libc-alpha. love too struct timespec.

 Tue Dec 07 01:44:27 +0000 2021

hey @__phantomderp you're like easily-annoyable into yelling at WG14, right? here's a fun post: https://sourceware.org/pipermail/libc-alpha/2021-December/133702.html (linux-man side of the thread: https://lore.kernel.org/linux-man/e8fc5b91-d417-d388-5e6f-4ee02106ac0f@gmail.com/T/#t)

 Tue Dec 07 01:46:50 +0000 2021

:):):)

 Tue Dec 07 01:53:16 +0000 2021

timespec::tv_nsec is a long. *except* on x32 (where the syscall long is actually long long, because it's using the amd64 ABI, which is LP64). the headers already do a lot of fucked-up magic to make this the only issue, but it's still an issue

 Tue Dec 07 02:04:04 +0000 2021

*if* you defined this in terms of a nsec_t type, then it'd never strictly be any fundamental type, so you'd always have to make it concrete for formatting. this is superiour to it being long everywhere except on a fringe ABI where it's actually long long (so 8-byte instead of 4-)

 Tue Dec 07 02:06:14 +0000 2021

you always name the member anyway, the glibc headers use a bitfield for a reason (and it's phrased as "at least" so you're allowed to do it this way), but having the type fixed for a post-IP16L32-era (it's defined as 0..10^9! no reason for tv_nsec to be 8 bytes!) would be cool

 Tue Dec 07 02:10:19 +0000 2021

that being said, it manifests in a fringe way on a fringe architecture, don't worry about it too too much if you've bigger fish (also pushing this through the austin group and into downstreams would be a fun time™ prolly)

 Tue Dec 07 02:13:42 +0000 2021

ugh, fine https://srhtcdn.githack.com/~nabijaczleweli/wg14/blob/7b063d740fc1245d92fb1755511a44776c49231d/N%3F%3F%3F%3F.pdf

 Tue Dec 07 05:41:27 +0000 2021

>3h
uhoh

>6:41
uhoh uhoh!!

 Tue Dec 07 05:42:01 +0000 2021

RT @nabijaczleweli: @__phantomderp ugh, fine https://srhtcdn.githack.com/~nabijaczleweli/wg14/blob/7b063d740fc1245d92fb1755511a44776c49231d/N%3F%3F%3F%3F.pdf

 Tue Dec 07 05:43:13 +0000 2021

im in the polish WG21 NB but also id personally sooner die than talk to people

 Tue Dec 07 05:51:45 +0000 2021

would it surprise you i stole 99% of this (well, save for the actual -mom document) from some other paper lol

 Tue Dec 07 05:55:33 +0000 2021

https://srhtcdn.githack.com/~nabijaczleweli/wg14/blob/868d3e096db0fe4378d77cd67215bed56045196d/N%3F%3F%3F%3F.pdf

so i like mail them directly and say "hey i have a paper (see draft attached) pls give number, PhD will tell you its good", or?

 Tue Dec 07 06:14:13 +0000 2021

lmao dmk's mx has the AUDACITY to 550 my mail "because sender name appears forged" but his head ass is using a self-signed, wildcard, and wrongly filled-out certificate on his web server. "computing experts committee" my ass

 Tue Dec 07 14:51:38 +0000 2021

in a fun twist, tar lets you do this in two slightly different ways!

 Wed Dec 08 00:01:23 +0000 2021

(it makes sense, of course, since true tar doesn't have -O and this is pretty useful, esp. for actual tapes, but this is also The reason tar doesn't just abort processing when it matches all listed filters)

 Wed Dec 08 00:03:13 +0000 2021

in many ways this is internet-over-TV-coax

 Wed Dec 08 00:07:26 +0000 2021

avoiding the famed cpio trailer moment – love to see The Industry learning from its mistakes

 Wed Dec 08 00:09:20 +0000 2021

cpio has a "TRAILER!!!" file at the end. two problems: (a) you can, obviously, have a "TRAILER!!!" file in your archive, but this aborts it. (b) you can concatenate archives. all you need to do is ignore the "TRAILER!!!" (this historically meant you get garbage instead).

 Wed Dec 08 01:02:46 +0000 2021

fun fact: unless you're running gentoo and booting straight to your rootfs, you've actually done the latter, because that's how you do early microcode loading. you can also append another (potentially-compressed) archive and it'll Just Get Unpacked to the initial /

 Wed Dec 08 01:05:04 +0000 2021

think its winter, lads

 Wed Dec 08 02:34:28 +0000 2021

febug v0.2.0 released, with manuals that aren't grating (and the aforementioned Macintosh port)!
Announcement mail: https://lists.sr.ht/~nabijaczleweli/febug/%3C20211209152451.hxdncexd3pdc4svu%40tarta.nabijaczleweli.xyz%3E; full changelog: https://git.sr.ht/~nabijaczleweli/klapki/refs/v0.2.0; x86 Debian sid+bullseye packages: https://debian.nabijaczleweli.xyz/README; pkgsrc: https://git.sr.ht/~nabijaczleweli/febug-pkgsrc

 Thu Dec 09 16:25:27 +0000 2021

in many ways this is some of my finest work

 Sat Dec 11 02:30:23 +0000 2021

are smartcards, like, actually networked? i was under the impression they just do the auth exchange/network preferences for the baseband? right??

 Sat Dec 11 02:57:41 +0000 2021

is being depressed the same as or opposite to being hard-pressed?

 Sat Dec 11 13:33:10 +0000 2021

hit/miss if you click off your monitor it truly turns and falls off vs if it's Still There as an output but doesn't render what you send it; first mode (OP) will cause your WM to reflow windows across extant screens, latter (cintiq) won't, but you won't see your windows either

 Sat Dec 11 17:53:53 +0000 2021

in light of the third-wave log4j discourse im unfortunately seeing i want ten bucks from everyone running the ZED

 Sun Dec 12 00:46:04 +0000 2021

https://paypal.me/nabijaczleweli lmao

 Sun Dec 12 13:25:10 +0000 2021

96 commits,
> 112 files changed, 446 insertions(+), 330 deletions(-)
and all of lib/ builds clean with -Wall -Wextra, too.

big mistake.

 Sun Dec 12 17:34:17 +0000 2021

97, -330, and include/ is too

 Sun Dec 12 17:50:19 +0000 2021

wo fucking wee my mans really be putting his phone numbers in the signature (cropped out! obv.) like its 1986, holy shit

 Sun Dec 12 21:51:02 +0000 2021

RT @nabijaczleweli: @__phantomderp wo fucking wee my mans really be putting his phone numbers in the signature (cropped out! obv.) like its…

 Sun Dec 12 21:51:16 +0000 2021

https://srhtcdn.githack.com/~nabijaczleweli/wg14/blob/e07288f51d8155c2e485f0303a4b9652f8d2275c/n2878.pdf its done. i have poosted. https://twitter.com/nabijaczleweli/status/1468034232589111296

 Sun Dec 12 23:32:18 +0000 2021

could the mRNA vaccine technology be used to like become immune to salmonellosis :0

 Mon Dec 13 00:32:38 +0000 2021

the mad lad actually did it oh my god

 Mon Dec 13 00:40:20 +0000 2021

also does it apply to cats, one of my m^Wfew ongoing fears is that one of the funny guys i know will succumb to salmonellosis because these mfs love poultry and haven't the gift of patience nor metallurgy

 Mon Dec 13 00:53:13 +0000 2021

degen cat :)

 Mon Dec 13 12:22:52 +0000 2021

in the periodic "C a[x] is *(a + x) sugar" moral panic people forget that in pre-K&R C, val.field and (&val)->field are, literally, *(decltype(field) *)((char *)&val + offsetof(field)). with all the implications thereof.

 Mon Dec 13 21:35:14 +0000 2021

ever wondered why all struct stat fields start with st_ and similarly for most others that still exist from ancient UNIX? yeah. wanted to swap bytes in an int?
swb(i) {
struct { char h, l; };
char tmp = i.h;
i.h = i.l;
i.l = tmp;
return i;
}

 Mon Dec 13 21:38:22 +0000 2021

this is not a made-up example. i literally saw this in a distribution V4 game. it's a wild time out there.

 Mon Dec 13 21:39:33 +0000 2021

hm, my family uses 7 as an otherwise-concrete but here-irrelevant "some", but for Big Number, the first 50/500/&c. that is obviously hyperbolic, ig? usually also over-bastardised regionally into "pisiont", "pincet", "pińć koła", &c.

 Mon Dec 13 23:18:43 +0000 2021

haveing a normal one today lads

 Tue Dec 14 01:03:19 +0000 2021

honestly disgraceful that after a year i still consistently place no lower than a high fourth place even in matches otherwise-full of DPSes

 Tue Dec 14 01:07:49 +0000 2021

http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2878.pdf :0:0:0

 Tue Dec 14 12:39:06 +0000 2021

even a cursory glance at the document log (http://www.open-std.org/jtc1/sc22/wg14/www/wg14_document_log.htm) by a not-very-observant but otherwise trained in the script native to over a third of a billion people spectator reveals that it isn't just excessively capitalised into Наб, but is in fact ha six

 Tue Dec 14 12:41:46 +0000 2021

y'know, like the literal thing a latin-reading child would read when presented with наб in printed cyrillic. why a child? the н is the wrong fucking size to be a H. it's SO obvious

also, the б is vastly different but you need to've read a line of cyrillic before to note this, so

 Tue Dec 14 12:45:21 +0000 2021

ALSO, it's literally on EVERY page and you can copy it from EVERY fucking page. for the love of piss!!

 Tue Dec 14 12:46:26 +0000 2021

наблюдателей eater

 Tue Dec 14 19:08:20 +0000 2021

benjamin merged the hid-tools tablet testsuite baseline so i updated the kernel on my laptop for the first time since february (5.10.0-3 -> 5.15.0-2) thereby replacing an extraneous iwlwifi message with an nvme error lmao

 Wed Dec 15 17:51:53 +0000 2021

a newer libinput also presents a good opportunity to demonstrate the weirdo reporting of Win8 styli (this is mainline, so with patches 1-3): notice how when doing the same movements with the pen down (pressure >0), the cursor stops moving and there's no longer a tool in proximity

 Wed Dec 15 18:08:10 +0000 2021

fun fact: smbd is vastly superlinear with directory depth. consecutive cds 35 levels deep:
real 0m5.954s
real 0m17.381s
real 0m21.389s
real 2m20.870s
real 5m48.622s

 Wed Dec 15 20:20:55 +0000 2021

i think this (well, Sublime indexing this directory tree) is also how I've been crashing smbd quite regularly

 Wed Dec 15 20:21:36 +0000 2021

oh hey another level done, 7m55s. the smbd process for this connection holds a steady 50+ish% CPU for the duration, this is the only thing it's doing

 Wed Dec 15 20:27:57 +0000 2021

lol. lmao.
real 9m52.979s
user 0m0.007s
sys 0m19.134s

 Wed Dec 15 20:35:44 +0000 2021

rebases cleanly, it seems like there were appx. three non-single-device commits to hid-{input,multitouch}, and the tip actually becomes down (I didn't even notice this in the mainline recording among the sea of text because it didn't move), remains in-range, and draws as expected

 Wed Dec 15 21:46:43 +0000 2021

attempt #2; 140 levels: 55m37s
and the smbd process is spinning at a solid 90+% (it was only 50% above because i was also building linux on all cpus)

 Wed Dec 15 22:43:21 +0000 2021

so in a way i guess this is a shitty DoS (you get one process per connexion, i think you can do this on all of them). if the loop gets to >PATH_MAX I guess we'll see if you can get it to crashloop into root mail (and 30-350M cores in /var/tmp but i forgot if thats the default)

 Wed Dec 15 22:47:32 +0000 2021

(its a shitty DoS because it consists entirely of syscalls so it yields quite well, as seen above. nevertheless, very funny)

 Wed Dec 15 22:51:34 +0000 2021

lol. lmao.

 Thu Dec 16 01:40:20 +0000 2021

the smbd process is now at 3h44 total CPU time lol

 Thu Dec 16 01:41:33 +0000 2021

bar potential DE-specific hacks, yeah

 Thu Dec 16 11:55:19 +0000 2021

new meaning to nightly tests!

 Thu Dec 16 12:13:41 +0000 2021

tracing find(1) reveals that it's the request path that takes a long-ass time each time, and subsequent single-level opens/stats execute near-instantly. also: didn't crash it.

 Thu Dec 16 12:57:03 +0000 2021

hid-test includes a usb_crash module, and, surely, it manages to generate a null pointer deref in the kernel. what i don't necessarily get is the point of this, though

 Thu Dec 16 22:36:57 +0000 2021

it doesnt, like, hard-lock the system or whatever (indeed, that's a screenshot from the laptop), but I couldnt unrfkill wlan, PAM stopped working, and systemctl reboot killed all sessions but never actually started to go down and I just got a blank console with a cursor

 Thu Dec 16 22:49:32 +0000 2021

i filtered out the "crash me" module, and here's a fun question: why does the hid-tools testsuite fail on both (3) mainline (1) and the version with my patch (2) in the same otherwise-unrelated place?

 Thu Dec 16 23:47:01 +0000 2021

lets fucking gooo, 9 months in the fucking making https://github.com/systemd/systemd/pull/19006#event-5784788214

 Fri Dec 17 15:12:43 +0000 2021

fuck sake, I need accurate events for the testsuite; the modules with the logging i have saved are all dated Feb11-Feb16,Feb28

-- Journal begins at Mon 2021-03-08 18:13:55 CET, ends at Fri 2021-12-17 18:48:48 CET. --

 Fri Dec 17 17:55:08 +0000 2021

i explicitly turn off all log rotation on my dev systems for this precise reason. didn't think I'd get bitten by fucking sd-journald rotation of all things

 Fri Dec 17 17:58:54 +0000 2021

thankfully I did actually save them, so it's solely a problem of booting into that kernel, loading the modules, and generating 1.9MB of dmesg over the course of like a minute (again) lol

 Fri Dec 17 18:21:23 +0000 2021

ive gone through a lot of very cringe captions for this and yet none manage to express how FUCKING pogged i am

 Sat Dec 18 00:34:53 +0000 2021

I've had this beast for a bit over a year and each time I'm shocked by it being under-3W at functional-idle with instant 15-30W peaks when under load

 Sat Dec 18 00:41:46 +0000 2021

Maybe I've grown jaded (I have) and distrustful of most battery life projections, but when at 60% battery it says 8h40, experience says that it's 100% right. Truly insane shit.

 Sat Dec 18 00:43:39 +0000 2021

FUCK yeas baybee

 Sat Dec 18 02:53:00 +0000 2021

https://gitlab.freedesktop.org/libevdev/hid-tools/-/merge_requests/125

 Sat Dec 18 03:11:32 +0000 2021

randos that "review" with FUD spew shut the hell up for once challenge

 Sat Dec 18 13:32:55 +0000 2021

aaaa i shouldn't have engaged at all but i noticed its a rando after first two replies and now its too late

 Sat Dec 18 13:47:48 +0000 2021

this is like reply guying but worse

 Sat Dec 18 14:03:07 +0000 2021

i think it's cool when i post shots from an uhidpi screen because big font prettie :0 https://twitter.com/nabijaczleweli/status/1472002394548584448

 Sat Dec 18 19:17:07 +0000 2021

are we sure its not "💍{2,}"?

 Sat Dec 18 19:37:16 +0000 2021

if the double circles mean optional, then it's "💍+", yes; if theyre double for decoration only, then its "💍{2,}", cf. regex(7), below.

 Sat Dec 18 19:43:42 +0000 2021

right, so its just "💍💍*"/"💍+" yeah

 Sat Dec 18 20:10:07 +0000 2021

remound by a "friend" of my "second furriversary"; suffer as I have. https://twitter.com/Prince_Flamingo/status/1207466391713923073

 Sat Dec 18 22:00:47 +0000 2021

im sorry i cant be her (hog with miku hair)

 Sun Dec 19 21:16:50 +0000 2021

y'know what? fuck you. *restores you to your former glory*

 Mon Dec 20 02:16:46 +0000 2021

man, i miss inty

 Mon Dec 20 04:03:54 +0000 2021

given how "downloading more RAM" is the exemplary pure computing scam, absolutely hysterical that zram is literally that, it actually works, and is a real-life panaceum

 Mon Dec 20 11:57:05 +0000 2021

Your message has been sent to the package maintainer(s):
APT Development Team <deity@lists.debian.org>

ah, yes.

 Mon Dec 20 15:29:05 +0000 2021

normal day. https://bugs.debian.org/1002021

at least it only shows it for the native architecture

 Mon Dec 20 15:33:43 +0000 2021

haha fun fact: tar -x that has a lib/... directory WILL replace your /lib -> usr/lib symlink with that directory

 Mon Dec 20 17:45:24 +0000 2021

another fun fact: you can't invoke the dynamic linker by hand with LD_(LIBRARY_PATH|PRELOAD)=... /usr/lib/.../ld-?.??.so /usr/bin/ls

 Mon Dec 20 17:47:17 +0000 2021

thank fuck i figured out the iDRAC 6 wrapper thing because i do NOT have enough energy to throw a monitor and shitty keyboard into the attic

 Mon Dec 20 17:50:00 +0000 2021

spot the difference lmao

 Mon Dec 20 17:52:13 +0000 2021

GNU tar has a --keep-directory-symlink that follows them instead apparently, but guess when i thought to look lol

 Mon Dec 20 18:06:05 +0000 2021

correction: you *can* invoke the dynamic linker (guess i fucked up the env vars?), but in a nonunforseen twist it's not suid so you can't elevate

 Mon Dec 20 18:10:50 +0000 2021

click here to cop: https://twitter.com/nabijaczleweli/status/1391228278170660867

 Mon Dec 20 19:08:58 +0000 2021

tzpfms 0.3.0 – integrates with zfs-mount-generator(8) – datasets on z-m-g-enabled pools should Just Work (just don't untar it over / – https://twitter.com/nabijaczleweli/status/1472986504272261124)! Release mail: https://lists.sr.ht/~nabijaczleweli/tzpfms/%3C20211220202718.kqhiaojx3qjwdww4%40tarta.nabijaczleweli.xyz%3E; full changelog: https://git.sr.ht/~nabijaczleweli/tzpfms/refs/v0.3.0; x86 packages: https://debian.nabijaczleweli.xyz/README

 Mon Dec 20 20:33:52 +0000 2021

APT dialect kinda sucks btw. not to mention the atrocious formatting that is purportedly clang-formatted but, uh, isn't.

yes, that's three-space indent with three-space continuations and 8-space tabs also sometimes, no clue when (UseTab: Always apparently??)

 Tue Dec 21 03:14:04 +0000 2021

it's not every day you manage to come up with something actively worse than emacs indent!

its like unusable, and the correct way to format it is to pipe git diff into clang-format-diff, it sucks SO much, no wonder people dont write patches :v https://bugs.debian.org/1002021#10

 Tue Dec 21 03:18:52 +0000 2021

also std::endl. std::endl!!!!!!!!!!! its simultaneously optimised out the shitter in every way possible and somehow also babby's first C++ porngram

 Tue Dec 21 03:21:32 +0000 2021

bro have you seen computing. its all a cargo cult ritual. nobody knows why or how this shit works 90% of the time

 Tue Dec 21 03:26:41 +0000 2021

also, this is the only file i touched - a .cc that ends up in http://libapt-private.so - and every time i rebuilt all 187 C++ objects. a real CMakeMoments.txt list

 Tue Dec 21 03:42:57 +0000 2021

also also, every single one has at least two warnings from a common header, most have many more. its a fun soup to wade through when ninja stops 273 lines after the error in the single file im interested in

 Tue Dec 21 03:46:13 +0000 2021

your girl says gn but you see her aptposting a half-hour later. wyd?

 Tue Dec 21 03:50:32 +0000 2021

consider this frenchman. he installed systemd/zram-generator yesterday and not only is he speaking human tongue ahead of schedule, he's also downloaded an additional 22.5% of RAM into his laptop

 Tue Dec 21 15:05:02 +0000 2021

 Tue Dec 21 17:09:34 +0000 2021

> Measurements with Sun Accelerator 4000 shows that 512-byte jobs or smaller are better handled in software. There is room for refinement here.

 Tue Dec 21 18:48:07 +0000 2021

this is the Sun Crypto Accelerator 4000 page – Gigabit Ethernet(!) NIC with IPsec and SSL(!) offloading. what a beast.

the copyright date is 2010 Oracle, but this is not to be fooled by, the board 1.1 relnotes are dated 2004 sunmicro. very cool and relevant in 2021

 Tue Dec 21 18:51:10 +0000 2021

i haven't moved in the over seven solid hours since i started this, but
> 144 files changed, 651 insertions(+), 769 deletions(-)

and I've evaluated every variable in .data and globals in .rodata in the zfs .kos and moved them accordingly

 Wed Dec 22 02:01:18 +0000 2021

zfs.ko can eat my ass. https://github.com/openzfs/zfs/pull/12899

 Wed Dec 22 02:33:46 +0000 2021

these keep taking me all around the pro-verbial world (this one especially) and my verdict is that the overall code quality is absolute, total, and utter fucking dogshit

 Wed Dec 22 02:35:57 +0000 2021

not that this is avoidable (it isn't, this is a miracle of engineerding as-is) but by GOD if you ever want to see some industry-proven code that no-one has had the time or resources to ever properly up-keep or snap to any sort of rigour in 30 years, just pick any file -mindepth 3

 Wed Dec 22 02:41:28 +0000 2021

my favourite is the entire Solaris loadable kernel module framework (syscall arguments, 17 module types, versioning, &c., &c.) straight-imported and embedded into the cryptography module, only shimmed out at the /absolute/ edges: https://github.com/openzfs/zfs/pull/12899/commits/025710a5ab516fd368ac182be54de2e367990696?authenticity_token=WpP3ZvK0MXj70oPfkyFO8MNvOUEM69C5orxjQOq%2FK5hDFGH82j1848aF0X0UFUS3OZZ%2BRKMzZTLM1fJzeaDrBw%3D%3D&file-filters%5B%5D=.h#diff-9d94388022ab7547146a2c997da7a66abf9b2c0804799629219dcbdd297ce934

 Wed Dec 22 02:48:33 +0000 2021

the annoying bitch in me says CC0, the diminutive rest says 0BSD

 Wed Dec 22 13:04:23 +0000 2021

<- this user is an anti-Foundation hard-liner

 Wed Dec 22 13:05:06 +0000 2021

-805; modctl.h, modconf.c, and edonr_mod.c are gone. there was NOTHING there.

 Wed Dec 22 14:42:27 +0000 2021

love a funny little gag! for fuck's sake.

 Wed Dec 22 14:48:10 +0000 2021

https://github.com/openzfs/zfs/pull/12902

 Wed Dec 22 18:31:19 +0000 2021

i am single-handedly bringing se^W the distinction between voiced and unvoiced ch and h back to Polish

 Fri Dec 24 00:20:13 +0000 2021

lads, serious question, how much does working at facebook suck?

 Fri Dec 24 01:35:19 +0000 2021

(this is 100% in relation to an e-mail from an @‌fb‌.com address in my mailbox)

 Fri Dec 24 01:37:11 +0000 2021

is dutch just uwuified germanic

 Fri Dec 24 02:16:19 +0000 2021

noooo dont uwuify the ubersprache your so sexy haha

 Fri Dec 24 02:20:25 +0000 2021

seeing a lot of faves, not a lot of takes. i require takes. please give me your takes.

 Fri Dec 24 12:41:50 +0000 2021

let's go

 Fri Dec 24 13:45:20 +0000 2021

bold statements from redhatters on the wg14 mailing list today

 Fri Dec 24 22:53:23 +0000 2021

just because fedora^WRH doesn't see fit to support the third x86 architecture doesn't mean nobody cares about it. by this metric you could discount approximately half of devices that are actively running a binux port

 Fri Dec 24 22:55:12 +0000 2021

lol. lmao. modhash is an entire, full, hash map implementation

 Sat Dec 25 01:32:21 +0000 2021

as i worried, then, on both fronts. thanks :/

 Sat Dec 25 02:50:48 +0000 2021

think I'm done. on a nice round -14000, too

 Sat Dec 25 03:40:58 +0000 2021

and to think that before this i was like +500 net on 44k baseline!!

 Sat Dec 25 04:07:35 +0000 2021

this is from a microSD card, always on for threeish years. incredible, being to see the rot so well and literally

 Sat Dec 25 15:39:33 +0000 2021

im with ma booooooooooys

 Sat Dec 25 16:47:54 +0000 2021

GOODRAM branding

 Sat Dec 25 20:20:33 +0000 2021

dae "find inc[cursor here] -type f -exec wc -l {} + | tail -1" and Tabbing errors with "bash: COMP_POINT: substring expression < 0"?

 Sat Dec 25 21:32:46 +0000 2021

the cryptography module (defined herein as module/[os/]icp and include/sys/crypto) reduced in volume from 45738 to 32551 lines (28.8%)

 Sat Dec 25 21:46:59 +0000 2021

fixed in upstream bash-completion apparently: https://bugs.debian.org/1002010?

 Sat Dec 25 22:14:48 +0000 2021

it's the register width on a pdp-11, you always by definition return one

 Sun Dec 26 18:43:18 +0000 2021

it's not a shock, really, int is the default, always, until specified otherwise. i mean, what else would you return? longs didn't exist until I want to say V6? V7?ish and couldn't be unsigned. a char makes no sense (you've already returned a full word). so: what else?

 Sun Dec 26 18:46:07 +0000 2021

remember that this is the language that doesn't have namespaced fields: https://twitter.com/nabijaczleweli/status/1470507631026577420

 Sun Dec 26 18:47:15 +0000 2021

haha! lol. lmao.

 Sun Dec 26 18:47:46 +0000 2021

some would say that's uintptr_t (that exists, and it's uintptr_t or [s]size_t, depending on your mood), others would continue in the 16-bit word tradition (this also exists, it's the WinAPI WORD typedef. it's also insane. the berkeley kernel has words & qwords. they're different)

 Sun Dec 26 18:52:00 +0000 2021

aave c compiler when

 Sun Dec 26 18:52:33 +0000 2021

unix

 Sun Dec 26 20:58:59 +0000 2021

caught a few more fields, 45738 -> 31869 for a nice 30.3%. the stripped size of the module went down by a total of 26%, too, from 256k to 188k. it's a wild time out here

 Mon Dec 27 02:29:31 +0000 2021

no talk me, in a primordial range

 Mon Dec 27 20:39:56 +0000 2021

I'd just like to interject for a moment. What you're referring to as the USPS, is in fact, UPU/USPS, or as I've recently taken to calling it, UPU plus USPS. USPS is not a mail delivery system unto itself, but rather another distributed component of a fully functioning UPU accord,

 Tue Dec 28 00:14:30 +0000 2021

have to reply to a friend on an IM so I'm writing my two-week mail backlog instead

 Tue Dec 28 01:34:23 +0000 2021

haveing a very normal one!!

 Tue Dec 28 01:34:49 +0000 2021

I guess it's true what they say: Facebook by any other name /does/ smell just as sickly.

 Tue Dec 28 02:21:02 +0000 2021

i did it, i poosted. i hate it here. this sucks.

 Tue Dec 28 02:30:29 +0000 2021

bad time, all fronts; whats new

 Tue Dec 28 02:32:28 +0000 2021

Fastl

 Tue Dec 28 17:23:00 +0000 2021

note also that the install which made the Windows.old folder took like an hourish. how it's managing to project 16 hours is beyond me. it also took 20+ minutes to actually, like, count the files (for whatever reason)?

 Tue Dec 28 17:24:21 +0000 2021

the bottom panel is correct. the top one SOMEHOW isn't. what the fuck is this!!!!!! how does it happen!!!

(don't get fooled by the acronis installer, all changes made in this very diskmgmt.msc widow)

 Tue Dec 28 17:26:35 +0000 2021

one of the weirdest things in high-level american decision-making is how 9/11 was a Big Deal over there but their emergency number thing is also 911. what's up with that, why would you pick that.

 Tue Dec 28 23:04:47 +0000 2021

independently insane it's neither 112 nor 999/998/997 like literally everywhere else but that's the american baseline

 Wed Dec 29 01:33:05 +0000 2021

girls LOVE it when you tell them about the fucked-uppest parts of your favourite fucked up little unix

 Wed Dec 29 11:17:31 +0000 2021

holy shit they made

made tinned beans into a real thing

 Wed Dec 29 12:01:38 +0000 2021

grub.. shouldn't (I'd say doesn't but its grub) have any sort of GPU support? AFAIK it should just be efifb, right?

 Wed Dec 29 13:43:58 +0000 2021

(turns out I'm full of shit, there *are* video drivers but they all seem to just be for different platforms (fb, bochs, vga, &c.); OTOH videoinfo should list a video mode to set in gfxmode)

 Wed Dec 29 13:48:51 +0000 2021

so you can't select a GPU since the platform /is/ the GPU but you can probably coax it to pick a more sensible one from the list if the default gfxmode=auto doesn't

 Wed Dec 29 13:51:51 +0000 2021

tea towel set, "hermione's" & "hismione's"

 Thu Dec 30 00:42:56 +0000 2021

*advances your persistent threat like a 4X stat*

 Thu Dec 30 00:49:57 +0000 2021

christ i must've been poorly last night, what the fuck

 Thu Dec 30 10:55:52 +0000 2021

gave it a few rounds on a bank of neodymia, it's, hm. not necessarily what I expected?

 Thu Dec 30 14:54:52 +0000 2021

it seems to be as I'd thought: only actually moving it within the field does damage (eddy current moment), and statically being in a field does jack shit. output incinerated.

 Thu Dec 30 15:07:03 +0000 2021

why am i seeing blohajar vore algorecs

 Thu Dec 30 21:17:29 +0000 2021

certified uintptr_t moment

 Fri Dec 31 01:00:21 +0000 2021

void* is like char* but shit and uintptr_t is like char* but not a pointer anymore, but otherwise they're roughly equivalent, right? (if the standard underspecifies this as a "dont try it bucko!" then thats a "lol" moment, an architecture where it doesnt hold is inconceivable)

 Fri Dec 31 01:10:33 +0000 2021

(rather, I don't see the problem here; so long as [A, A+sizeof(T)) is mapped doing *(T*)A is, like, fine; so is memset((void*)A, 69, sizeof(T) and whatever)

 Fri Dec 31 01:14:02 +0000 2021

a pointer is just an integer with fucked-up arithmetic operators. this is like discussing if casting between one of the four different 32-bit integer types retains your final integer's kosherness imo

 Fri Dec 31 01:17:41 +0000 2021

as a wise man once said: "two integers" is just "one integer twice as long"

 Fri Dec 31 01:21:11 +0000 2021

(x86, at all compat levels, is segmented to right old fuck, you're probably sucking down those 4-level pages like it's no tomorrow /right now/)

 Fri Dec 31 01:24:55 +0000 2021

LEA moment

 Fri Dec 31 01:25:07 +0000 2021

she lea on my rip till i stall

 Fri Dec 31 15:24:58 +0000 2021

agree 100%

 Fri Dec 31 15:28:33 +0000 2021

gas:pclmulqdq.S:25:11: warning: unmatched lea/fea pair [-Wlost-effective-addresses]

 Fri Dec 31 15:42:47 +0000 2021

i thought we were getting shelled; apparently not so! truly joyous

 Fri Dec 31 22:33:06 +0000 2021