Your twitter archive

S/FTP (Shielded/Fluorinated Toilet Paper)

 Fri Jan 01 13:26:52 +0000 2021

he's learning :)

 Fri Jan 01 20:56:01 +0000 2021

i interrupted brushing him to shoot and now i will never recover

 Sat Jan 02 01:20:39 +0000 2021

getting really into vore, not for horny reasons, but for the FDA to classify me as food and put me on the unimportables list

 Sat Jan 02 02:18:52 +0000 2021

wild to me that we hit the four-cpu mark in second-gen intel core i in 2011 and nothing has really changed for the consumer until, like, last year

 Sat Jan 02 11:26:10 +0000 2021

:)

 Sat Jan 02 13:38:13 +0000 2021

i wonder how difficult it would be to make an SD card that exposes a fake MMC backed by, like, a blockdev it gets over, like, iSCSI

 Sat Jan 02 13:50:05 +0000 2021

SDIO exists, right, so there /are/ Wi-Fi SOCs with antennas that fit into SD form factor (or, well, just flat ribbon for Ethernet), so I expect this to be both fully possible and desirable, if only to avoid the dogshit flash inside SDs, so?

 Sat Jan 02 14:04:43 +0000 2021

oh, to be him, just for a fleeting moment!

 Sat Jan 02 14:40:11 +0000 2021

wasn't that like a citrus?

 Sat Jan 02 18:34:37 +0000 2021

originally prompted by a friend complaining "my SD card died mid-[3D-]print", but, like, I have always-on SBCs whose SDs have just worn out and it's a PITA to dump whatever's left, find another card, reimage, &c.

 Sat Jan 02 18:35:45 +0000 2021

also i barely ever shoot outside nowadays, and it's a /process/ to kill the camera, pop the card, find a dongle, plug all three togeter, and unwind, each time, similarly for a scanner; having these interface directly to an iSCSI-or-equivalent target would just work so much better

 Sat Jan 02 18:41:46 +0000 2021

its a new day! he's reading the paper to find out what's gone on

 Sat Jan 02 19:02:15 +0000 2021

when the 386 did it it was "a long time coming", and "allowed you to unlock the true potential of your PC", but when I do it, I'm "disrespectful", and "the principal needs to see me" (adopting a flat addressing scheme)

 Sun Jan 03 01:16:21 +0000 2021

this has haunted me for weeks and i have no clue how to phrase it for it to be funny

 Sun Jan 03 03:39:27 +0000 2021

allowing flash and pressing play on the presentation(?) hard-yeets the firefox process. magnificent

 Sun Jan 03 19:43:12 +0000 2021

reject modernity, embrace tradition https://twitter.com/nabijaczleweli/status/1173260806957469696

 Mon Jan 04 03:03:54 +0000 2021

so i was wondering why absolutely trivial shit is just plain unresolvable over in netdata land and how the turnover is as bad as it appears to be and apparently there's thirty one million dollars of venture capital pumped into the damned thing and that explains it all tbh

 Mon Jan 04 12:12:38 +0000 2021

Previously:
* installer broken if cc = clang
* i open an issue with a solution inside
* employee opens a PR with the solution and assorted clean-up
* another contests this
* the opener says "if you want to go to the mat over this"(??)

then half a year passes

 Mon Jan 04 12:17:18 +0000 2021

* a third employee comments on my issue saying that PR man has (a) left and (b) deleted his fork so (c) his changeset is lost forever
* I have to first explain then /link/ <PR URL>.patch because they cannot figure it out, apparently(?)
* netdata now builds when cc is clang

 Mon Jan 04 12:20:54 +0000 2021

Currently:
I opened a PR yesterday that stops the diskspace plug-in from instantly mounting all autofs nodes, and got this. HOW do you write so much, yet say so little. This is literally fucking worse than empty, the man that got tagged is code owner and already got notified.

 Mon Jan 04 12:25:39 +0000 2021

look, I'm all for funnelling venture capital into workers' hands as quickly as possible. only thing it's good for, really. the harder you scam rich people the more morally pure you are. but why must i suffer for this

 Mon Jan 04 12:30:10 +0000 2021

hm, I've been able to use the vendor executable to upload vendor firmware, and flash․py to upload peadee firmware, but fw_pad.py was detrimental to peadee firmware, and no amount of dicking about let me upload vendor firmwares via flash․py; guess I'll just wait for the Article

 Mon Jan 04 15:50:25 +0000 2021

SH has 135 instructions and 54264 encodings, SH-2 144 and 52680, mostly because they don't have FPUs

 Mon Jan 04 19:33:34 +0000 2021

this is official J-core promo material, wild shit

 Mon Jan 04 21:18:04 +0000 2021

lmao apparently http://FPSCR.SZ being set or not changes the instruction decoding for FMOVs. what the hell am I supposed to do about this

 Mon Jan 04 21:54:17 +0000 2021

i.e., depending on architectural state, 1111_0101_1010_0110 is either FMOV.S @(R0,R10),FR5 or FMOV @(R0,R10),XD2; cool, great. lovely!

 Mon Jan 04 22:38:49 +0000 2021

this fucking sucks, I'm actually writing a decoder instead of generating nine megabytes of code and indexing a 256KiB array

 Tue Jan 05 00:11:51 +0000 2021

enterprise devops, or "using 2kw for something you could do less annoyingly on a core 2 duo"

 Tue Jan 05 14:22:45 +0000 2021

yes im using aws again and yes im angy again

 Tue Jan 05 14:23:01 +0000 2021

upgraded to zfs 2.0.0-1~exp1, switched from compress=lz4 to zstd[-3], funie stats:
* clang-10 executable (2.4G apparent): 1.5G -> 1.1G
* stripped (238M): 91M -> 59M
* llvm-toolchain-9 partial build tree (734M): 312M -> 247M
* llvm-project-10 build tree (22G): 9.6G -> 6.1G

 Wed Jan 06 03:15:59 +0000 2021

uh, hm, I haven't managed to measure any significant difference on a Core 2 Duo E6400 @ 2.14GHz, but this system is horrifically starved for I/O, so dunno how representative this is; both give me roughly 20-25% when reading and 10-15% when writing

 Wed Jan 06 03:32:26 +0000 2021

* nextpnr+icestorm+prjtrellis build trees (2.6G): 834M -> 471M
* another LLVM build tree, but this time without any of the source (19G): 8.0G -> 5.1G

it also does slightly better at a 1.6G mailbox than compress=gzip[-6]

 Wed Jan 06 04:25:07 +0000 2021

overall im Slightly surprised by how well zstd does in this application but not That Much, but also I've doubled my free space on that pool and it's measurably faster by up to 15-25MB/s; for Free™!

 Wed Jan 06 04:27:49 +0000 2021

bruh why do i supposedly have to pay at least one thousand dollars per year to the SD Association for a full SD spec, but apparently a "simplified" variant is free? what are they hiding

 Wed Jan 06 16:39:39 +0000 2021

I have since yeeted the binary so no way to check, but I appreciate the explanation nonetheless, thanks :)

 Wed Jan 06 16:48:40 +0000 2021

coming soon to yaxdis in /your/ area! (hopefully)

 Wed Jan 06 18:19:30 +0000 2021

what have we let the world come to? (referring to having just sent a PR over twitter DM)

 Wed Jan 06 18:30:03 +0000 2021

hm, i don't intend to do drm or interface to the ass daemon, so that'll be fine i think; thanks!

 Wed Jan 06 22:18:12 +0000 2021

okay, the dates on the MMC spec are between 2001 and 2004, but like, this would look like shit even on A4

 Wed Jan 06 22:32:03 +0000 2021

NEW JEDEC JENDERS, WHO UP??????

 Wed Jan 06 22:35:44 +0000 2021

this "Personal" sexion consists of: Salutation, First Name, Last Name, Company Name, and Country – the personallest of all the database fields

 Wed Jan 06 22:38:36 +0000 2021

yee

 Thu Jan 07 00:15:11 +0000 2021

oh lmao thats just my last name field, but i put in ha ha ha yee after the iconic video; class

 Thu Jan 07 00:17:51 +0000 2021

🅱️oneless

 Fri Jan 08 14:01:44 +0000 2021

my fans have been asking me, for the longest time – are OneMix laptops truly free of bones? – it's time for me to finally publish a verdict.

 Fri Jan 08 14:14:43 +0000 2021

holy fucking hell, Windows iSCSI Initiator program will refuse to connect with this bull shit when the target name has /capital letters/ in it. for fuck's sake

 Fri Jan 08 18:52:27 +0000 2021

just got hit with the "are you part of a team" and im infinitely old now; i have aged a thousand years

 Sat Jan 09 11:12:05 +0000 2021

135cm (well, 25+55*2) plain fails to initialise with an EILSEQ; just 80cm tries a bit, but still ultimately fails. truly shocking

 Sat Jan 09 13:24:20 +0000 2021

today in Banking: i got an (1) e-mail with the summary for december, (2) attached as a PDF, (3) "secured" with my PESEL, (4) the PDF just says "attached to this document"(?), is a (5) .htm, with (6) an unfilled "Mail reference number: UNIQUE_MAIL_XREF." at the end

 Sun Jan 10 11:13:31 +0000 2021

personally I had no fucking idea you /could/ attach to PDFs, and uhh, out of the four readers I've since tried, only firefox actually indicates that there is an attachment. I had no idea with my default one

 Sun Jan 10 11:15:08 +0000 2021

yeah, live with two https://twitter.com/nabijaczleweli/status/1339901978441342976

 Sun Jan 10 15:23:22 +0000 2021

perfectly ripe and brimming with juicy seed

 Sun Jan 10 17:52:56 +0000 2021

delicious, i can smell the caustic smoke already

 Sun Jan 10 18:00:06 +0000 2021

we must cut down on food-based waste, just look how much packing i had to strip to get to the rich, calorie-dense core

 Sun Jan 10 18:01:46 +0000 2021

the lettering says "13312B2 99 1E MHP-TA 15-9-85", which appears to be a Littelfuse(!) MHP-TA15-9-85, now MHP-TAM15-9-85 thermal 9V 15A 85°C fuse; cute!

 Sun Jan 10 18:11:58 +0000 2021

yeah, %PROGRAMFILES% and %PROGRAMFILES(x86)%

 Sun Jan 10 18:59:59 +0000 2021

it took me one whole month to think this through and come up with a solid reply, and yet my post still ends with being a coward. never not on brand (unfortunately, this time)

 Sun Jan 10 20:00:47 +0000 2021

doing "journalism" to be "forcibly silenced 😩"

 Mon Jan 11 01:56:45 +0000 2021

you have the right to copy any of my work unless you like the foundation or think it's in any way good, then you can copy this boot in your ass

 Mon Jan 11 02:13:40 +0000 2021

i would like to retract my earlier statement — apparently you hafta be a nazi for that — but it ain't my fault that I'm out here getting loose, gotta blame it on the goose (half a cup of gin, technically, but who's counting), gotta blame it on the juice, baybee

 Mon Jan 11 03:11:24 +0000 2021

According to https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/5/html/virtualization/ch33s08 at least, this means that the guest can run on all currently installed CPUs

 Mon Jan 11 20:36:03 +0000 2021

this is a certified ZFS moment !

 Mon Jan 11 22:12:42 +0000 2021

I ain't been slept on this much since none of y'all fuckers listened to my album

 Tue Jan 12 10:24:25 +0000 2021

ohshite(dpkg)

 Tue Jan 12 17:08:04 +0000 2021

days since crashing clang:

 Wed Jan 13 20:18:16 +0000 2021

fixed in latest snapshot :)

 Wed Jan 13 20:20:12 +0000 2021

anyway, here's user-space debugfs

 Wed Jan 13 21:27:51 +0000 2021

honestly, twitter makes a damned hash of this, but I personally find live inspection of std::sort (with sleep(1) in the comparator) pretty damn cool https://youtu.be/HTubYyhcCLk

 Wed Jan 13 21:35:33 +0000 2021

anyway, here's user-space debugfs; you can build it, run it, and debug your applications with it (if badly): https://git.sr.ht/~nabijaczleweli/febug

 Thu Jan 14 02:39:38 +0000 2021

and they say machines can't break your heart

 Thu Jan 14 05:47:42 +0000 2021

this machine shipped with DOS

 Thu Jan 14 12:50:57 +0000 2021

this looks fake as fuck. I mean, it's literally true, but it looks like a shitpost

i was gonna try to do NVMe-oF boot support for the shits, but, astoundingly, you just copy /sbin/nvme, /etc/nvme/, nvmf-autoconnect.service, and modules to the initrd and it'll Just Work. wild

 Thu Jan 14 13:05:48 +0000 2021

who did this?

 Thu Jan 14 19:38:00 +0000 2021

she

 Thu Jan 14 21:01:10 +0000 2021

now is a time for healing (thinking why you told me to let you out, got snowed on, then sprinted back inside, soaked through)

 Fri Jan 15 00:23:26 +0000 2021

now with ergonomic C (so long as the compiler isn't dogshit, at least) and C++ APIs! https://git.sr.ht/~nabijaczleweli/febug/tree/dc50f05bffbb2a834f79f67b58f82ea66d23b088/item/examples

 Fri Jan 15 00:35:33 +0000 2021

Only one of these cats is ready for work in winter conditions; are you?

 Fri Jan 15 13:45:11 +0000 2021

this is what the city took from you. from us all

 Fri Jan 15 13:50:20 +0000 2021

it seems that ronn-ng won't let me do the formatting I want, nor does it handle cmsg(3)-style multi-name pages gracefully, so i think its (unfortunately) finally time for plain roff. OTOH, this does mean I can finally just use the C preprocessor, so

 Sat Jan 16 04:06:38 +0000 2021

RT @iximeow: /!\ new yaxpeax-dis /!\

this is one (0.2.3) i'm very excited for: @nabijaczleweli let me know they wrote a yaxpeax-superh, su…

 Sat Jan 16 04:06:49 +0000 2021

building new zfs package for the 3rd (third) time because i've managed to forget about it and, hence signing timed out 2ce (twice)

 Sat Jan 16 15:17:05 +0000 2021

four times

 Sat Jan 16 15:33:36 +0000 2021

five fuckign times . im done. im fucking over the god damned gag; praise be cache invalidation

 Sat Jan 16 15:51:02 +0000 2021

so mister Fed Ex (who I assume delivers all parcels) has bestowed unto me my crowdsupply order, and uh.. it's in (dogshit machine-translated) Polish (they failed to print)? I was billed from the US, and this left from the US, one product is from HU, another from CN, wh-

 Sat Jan 16 23:39:53 +0000 2021

i redacted away the bit where they printed an Ę in my shipping address as AE. not Æ, which Ę does collide on some common codepage with. AE

but it's fine in the footer?

 Sat Jan 16 23:43:05 +0000 2021

I had ":)" prepopulated after pasting in the first screenshot, but im no longer ":)" because i lost the last five hours to setsockopt(2) level being 0 for NetBSD's LOCAL_CREDS (contradicting unix(4)) vs SOL_SOCKET for Linux's SO_PASSCRED, failing silently (errno=0), doing nothing

 Sun Jan 17 05:37:52 +0000 2021

downloading source: five seconds
autoconf: sixty-ninety seconds
building package: five-ten seconds
installation: one second

someone who's good at packaging please help me budget this, my NetBSD port is dying

 Sun Jan 17 15:39:11 +0000 2021

that being said, is there any reason autoconf couldn't do 90% of its checks in parallel to one another? (besides having to send toe pics to stallman to be graciously granted the rights to touch foundation code)

 Sun Jan 17 15:45:42 +0000 2021

when uhhh.. when you yeah.

 Sun Jan 17 17:01:14 +0000 2021

uhh is NetBSD poll(2) broken? if I kill a client via SIGINT or the client just closes its end of the socket, I never receive POLLHUP and keep getting POLLIN instead?

 Sun Jan 17 19:36:03 +0000 2021

turns out its just different from freebsd and linux and that's what it's Supposed to be doing, but i didn't fish it out from the blob of text in POLLHUP description. cool

 Sun Jan 17 19:49:03 +0000 2021

you may not like it, but this is what peak systems programming looks like

 Sun Jan 17 20:59:27 +0000 2021

it's important to always have a strong personal brand

 Mon Jan 18 05:01:51 +0000 2021

not every day you find a clang bug

 Mon Jan 18 18:41:32 +0000 2021

llvm-bugs@ doesn't let me post, and I have to beg admins for bugzilla access. pain

 Mon Jan 18 19:12:15 +0000 2021

friending a girl on discord to "play a game" together, but it's a screen of share battlestar(6) in an 80x25 QEMU window and a badly multiplexed pty

 Mon Jan 18 22:17:12 +0000 2021

man, this manpage is.. Quite something

 Mon Jan 18 22:21:49 +0000 2021

The "I wrote battlestar in 1979 in order to experiment with the niceties of the C language." bit does it for me. Rare Phrase. Plus it's literally twice my age.

 Tue Jan 19 09:57:20 +0000 2021

SO WORD OF FUCKING WARNING: if you're using ZFS-on-root with dracut on sid, then `ln -s /sbin/zgenhostid /bin/` and regenerate your initrd

 Tue Jan 19 13:34:17 +0000 2021

my only regret in life is that i cannot scream at maintainers retroactively, holy fucking shit

 Tue Jan 19 13:35:42 +0000 2021

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=980464

 Tue Jan 19 13:36:52 +0000 2021

say what you want about Lithuanians, you can't deny them the sense of style

 Wed Jan 20 00:16:50 +0000 2021

that's a cute RJ45 hole

 Wed Jan 20 00:21:04 +0000 2021

hi :)

 Wed Jan 20 00:22:11 +0000 2021

The goods in this invoice cannot be used for the military purposes and are not included in the multiple-purpose list.

thank you, American Empire®; very cool.

 Wed Jan 20 00:32:36 +0000 2021

languages stop putting their advertised-as-memcpy-equivalents' arguments in backward order to memcpy(3) challenge

 Wed Jan 20 06:25:49 +0000 2021

this has been a Rust call-out post

 Wed Jan 20 06:30:59 +0000 2021

the delicate downright erotic dance between MAKE_JOBS=$(sysctl hw.ncpuonline) and NetBSD/i386's three-gigabyte address space

 Wed Jan 20 19:13:27 +0000 2021

it turns out that this should not be considered a bug; indeed, "clang -E -" should fail or at least issue a diagnostic, but @Guriwesu is a coward and won't back me up on this.

 Wed Jan 20 19:33:34 +0000 2021

LET'S FUCKING GOOOOOOOOOOOOOOOOOO

 Wed Jan 20 19:47:40 +0000 2021

not pictured: a commented-out bit of the Makefile where some fucking brain genius set MAKE_JOBS_SAFE=no and having had to resize the pkgsrc disk because 5.5G was simply not enough (indeed, the workdir, which I've cleared beforehand, is now 8.6G)

 Wed Jan 20 19:54:59 +0000 2021

okay, so uhh, here's user-space debugfs with ports for Linux, FreeBSD, NetBSD, and OpenBSD and application programming libraries for C, C++, and Rust, all fully documented (I think!), v0.1.0: https://git.sr.ht/~nabijaczleweli/febug

 Wed Jan 20 22:05:46 +0000 2021

I am thinking about converting to
NetBSD. It will take me atleast 1 year of compilation. Thoughts ?

 Thu Jan 21 00:26:19 +0000 2021

guy who learns about the hyperinflation in Poland prior to 1995 and his response is "hot"

 Thu Jan 21 20:24:25 +0000 2021

forbidden USB

 Fri Jan 22 00:00:22 +0000 2021

bumping my cooperative debugging "frame work" for the final time because it took a lot of time and i don't hate how it turned out https://twitter.com/nabijaczleweli/status/1352014480960593922

 Fri Jan 22 01:44:29 +0000 2021

one day i will finally give into my urges (writing a soft mailing list ware), and then it'll be over for you fuckers

 Fri Jan 22 11:24:07 +0000 2021

i wget -red the 8devices wiki and what the absolute fuck is this i hate it i hate it i hate it i hate it

 Fri Jan 22 17:24:48 +0000 2021

hungy

 Fri Jan 22 23:19:37 +0000 2021

he has vored. he has chilled. one final item on the itinerary:

 Fri Jan 22 23:29:06 +0000 2021

we could've had this.

 Sat Jan 23 12:39:14 +0000 2021

ubuntu users normally vs when you bad-mouth canonical

 Sat Jan 23 18:14:02 +0000 2021

so i copied the discs, naturally, and the plain Win2k one is 131M, the one with SP3 is 372M, and come out to 88M and 333M with compress=zstd, respectively; rather odd, since I'd expect the thousand archive files to compress worse than almost seven thousand plain drivers/help/&c.

 Sat Jan 23 18:34:19 +0000 2021

but, uh, what мore could there be in SP3 for it to be thrice the size? the base system can't've grown that much, surely?

 Sat Jan 23 18:46:56 +0000 2021

narkive stop being the only netbsd-users@ google result and only returning a PHP OOM error challenge

 Sun Jan 24 01:21:56 +0000 2021

what if you wanted to boot the networked berkeley system but the kernel said

 Sun Jan 24 13:27:54 +0000 2021

pogue

 Sun Jan 24 15:58:26 +0000 2021

this is a certified Big Endian moment!

 Sun Jan 24 17:05:11 +0000 2021

my favourite bit of this is that the web interface doesn't remotely match the system state and is very insistent that Wi-Fi is very much off and wholly unconfigured. I'm accessing the web UI over wifi

 Sun Jan 24 17:42:57 +0000 2021

takes one to know one

 Sun Jan 24 17:46:21 +0000 2021

"building OpenWRT with -j26", or "configuring every package for half a minute, then building it in under a second, one at a time" https://twitter.com/nabijaczleweli/status/1350830031191334913

 Sun Jan 24 19:00:15 +0000 2021

love too be able to run make clean to remove build artifacts and not have to resort to zfs rollback :)

why does every part of this suck so baaaaaaaaaaaad

 Sun Jan 24 19:04:58 +0000 2021

i,m on amd64??

 Sun Jan 24 19:25:02 +0000 2021

haha, of course this is a well-known and well-understood problem with a 2018 patch that applies with minimal fuzz. love too progaming

 Sun Jan 24 19:33:00 +0000 2021

i gave up and configured linux instead, time to expensive pendant draws ever closer

 Sun Jan 24 22:06:04 +0000 2021

!!!!!!!

 Sun Jan 24 23:12:12 +0000 2021

hi :)

 Sun Jan 24 23:35:04 +0000 2021

measuring git repository size by kWh per gc

 Mon Jan 25 00:32:06 +0000 2021

percussive maintenance but for software

 Mon Jan 25 03:40:06 +0000 2021

hm i posted this as a gag but i literally do this 100% of the time. if a soft ware requires me to scan through a list of more than five options, find an item in a menu tree, or doesn't expose a Thing in its front-end, it is getting lied to, repeatedly and severely, until it works

 Mon Jan 25 03:51:14 +0000 2021

CC is unset (clang), CC is clang, CC is gcc, CC is gcc (clang), CC is a wrapper around clang, and also I rollback the rootfs blockdev from the host at the right time, the output file is a FIFO catted into ssh. i give a /fuck/ about your build system telling me what I can't do.

 Mon Jan 25 03:56:15 +0000 2021

also the "bind-mounting /dev/null over /proc/swaps" bit to appease kubernetes (i will NOT be told what my deallocation semantics are supposed to be, fuck you) but that's so pedestrian it's lost its novelty to me

 Mon Jan 25 03:59:05 +0000 2021

new efi vars in gun-metal grey at great prices from lowe's

 Mon Jan 25 11:00:42 +0000 2021

uhhh is fuse3 just.. broken on i686? __USE_FILE_OFFSET64 is supposed to gate this, but doesn't seem to in practice?

 Mon Jan 25 15:43:42 +0000 2021

oddly, if unsurprisingly, x32 does build, so?

 Mon Jan 25 15:44:16 +0000 2021

record time https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=981012

 Mon Jan 25 15:57:37 +0000 2021

now with Debian and pkgsrc packages, and building on NetBSD's distribution compiler! https://git.sr.ht/~nabijaczleweli/febug

 Mon Jan 25 17:22:25 +0000 2021

apparently this is Known and upstream refuses to acknowledge this. cool

 Mon Jan 25 18:04:18 +0000 2021

if ONLY there was some tooling at hand that prints each step as it executes it, similarly to sh -x, mayhap, so a programmer could see what happens in real time and easily nail the offending action. alas!

 Mon Jan 25 21:53:11 +0000 2021

if i ever meet an openwrt enjoyer it's on fucking sight

 Mon Jan 25 21:55:01 +0000 2021

as of this morning, my shitty x32 sid router has better uptime than my ISP

 Tue Jan 26 10:21:51 +0000 2021

Arch Linux is a bleeding-edge distribution, not for the faint of heart

 Tue Jan 26 13:38:06 +0000 2021

i explicitly spec a new enough version of ronn in the README! it's one of the two(!) results of "git grep ronn"! this also proves too difficult for Arch users.

 Tue Jan 26 13:41:22 +0000 2021

if your salt isnt fuming purple DO NOT talk to me

 Tue Jan 26 17:10:34 +0000 2021

if kity not for kissing,, why have big fore head?

 Tue Jan 26 22:38:33 +0000 2021

real is wall-clock, user is CPU time spent in userspace, sys is in kernel, the more CPUs you have (and the smaller your HZ) the less they correlate; in this case you seem to be stuck in iowait and hence unscheduled, you might confirm this by correlating D status with open files

 Wed Jan 27 09:15:43 +0000 2021

that being said, unless you're under /heavy/ memory pressure, I'd expect most of your build source (assuming C and Many Small Headers) to be served out of page cache/ARC/whatever on subsequent runs, regardless of how starved for I/O you are

 Wed Jan 27 09:20:10 +0000 2021

this definitely says something about the field

 Wed Jan 27 11:12:25 +0000 2021

it appears that i will soon be vaccinated against pneumatic cocks. the relevancy of this appears pressing

 Wed Jan 27 12:13:01 +0000 2021

anno domini twenty fucking twenty

 Wed Jan 27 13:18:57 +0000 2021

This is the government healthcare website, "Oops.. Failure in accessing IKP. Remember, that the application doesn't allow working in multiple tabs - in that case return to the previous browsing screen. If you're working in one tab, refresh the page or log out and then back in."

 Wed Jan 27 13:27:03 +0000 2021

i put that in alt text but that is apparently fucking useless for the casual reader, in classic twitter form

 Wed Jan 27 13:27:37 +0000 2021

caught one

 Thu Jan 28 09:36:40 +0000 2021

graphics: weird, too close to cactus => not at all huggable
cryptography: pretty much just cryptozoology => very huggable

 Thu Jan 28 22:13:51 +0000 2021

caught another

 Fri Jan 29 14:18:31 +0000 2021

dont make me configure libinput i WILL cry

 Fri Jan 29 18:01:43 +0000 2021

this post brought to you by stylus distance entry/exit generating button 1 press/release, despite setting button masks and xmodmap pointer prop to all 0

 Fri Jan 29 18:05:12 +0000 2021

also by evtest happily listing pressure change events but neither Krita nor Inkscape remotely picking them up. oh apparently its also been like one and a half hours of this and I have gone backward. cool. great. lovely. amazing. outstanding.

 Fri Jan 29 18:10:09 +0000 2021

people on this web site be like "damn my ass fat, what's ma pronouns?"

 Fri Jan 29 23:34:02 +0000 2021

cat in da ventilation.. who up?

 Sat Jan 30 12:52:09 +0000 2021

into da filter

 Sat Jan 30 12:55:58 +0000 2021

аааА this is ABS_MT_DISTANCE not ABS_MT_PRESSURE why??

 Sat Jan 30 14:59:41 +0000 2021

i think im building qt. this is a "cry for help", i think they call it

 Sat Jan 30 17:53:58 +0000 2021

building Qt told me that libxcb doesn't return an AbsPressure valuator for this stylus. time to have some fun with a 45-megabyte strace dump and then report a Xorg bug :)

 Sat Jan 30 20:16:05 +0000 2021

you vs the input handler they tell you not to worry about

 Sun Jan 31 00:21:10 +0000 2021