<{^_^}>
swaywm/sway#4727 (by RPigott, 1 year ago, merged): render: use nearset neighbor scaling for outputs with an integer scale
<samueldr>
implemented there
<samueldr>
so seeing it's just sent to a GL surface without much care explains why
<samueldr>
not sure how they went... how many years... working on Xwayland and scaling stuff for wayland and never stopped to add 1+1=2 rather than 1+1 = ~2
<samueldr>
my inexperienced self assumes it should be part of the... thing that renders... to somehow be able to handle how things get scaled
<samueldr>
wait... so a pixel art game would be scaled nearest too?
<drakonis>
yes
<drakonis>
linear scaling is what you want i think?
<samueldr>
yes
maljub01 has quit [Ping timeout: 268 seconds]
<drakonis>
since it draws all the pixels without colors merging at the pixel boundaries
<samueldr>
but how is it not like... a basic component of wayland???
<gchristensen>
one thing that comes to mind is zstd work has been discussed for years on the openzfs mailing list
<samueldr>
also fun fact
<samueldr>
phosh's lock screen when rotated doesn't fit
<matthewcroughan>
So in terms of process, what is the raw difference between an RFC and what ZFS did with ZSTD?
<samueldr>
so the unlock button is rendered outside of the display
<samueldr>
good thing a sliver shows up
<matthewcroughan>
Is an RFC in some way supposed to make things happen faster and be more "mature" and "corporate" ?
<matthewcroughan>
Linux happens via a mailing list, right? What is it that RFCs are good for?
<samueldr>
an RFC is just rules pre-agreed upon to make big changes
<samueldr>
and "an RFC" is diminutive
<samueldr>
the RFC process for project ______
<matthewcroughan>
Yeah, sorry, English is my first language, which means I'm destined to be ignorant of it ;)
<gchristensen>
an RFC is a way to socialize changes, agree to large changes collaboratively, and give adequate time to involve the very many stakeholders of the project
<drakonis>
an rfc, depending on the project, can range from being improductive bikeshedding to yielding a common direction for something
<gchristensen>
all of these things happened in the zstd / openzfs example over several years
<matthewcroughan>
gchristensen: right.. makes sense. And ZSTD isn't worthy of an RFC because it's not changing something fundamentally, just adding a feature.
<gchristensen>
openzfs has its own process
<matthewcroughan>
It works really well, whatever it is. From a layman's perspective (me).
<matthewcroughan>
I feel I can review changes without any actual connection to the project.
<matthewcroughan>
Nix is something that one can only review if they're familiar with the Nix language, and you're plugged into Git.
<drakonis>
samueldr: i dont think it is within wayland's scope because it supposedly sets a standard for how wayland consumers communicate
<gchristensen>
this is dangerously on topic of course
<samueldr>
drakonis: isn't asking for 1:1 pixels part of things that should be communicated?
<drakonis>
it doesnt talk about how it should be rendered
<drakonis>
hmm, that's something only the wayland folks would be able to answer
abstrn has joined #nixos-chat
<samueldr>
like, the first moment I'd see fuzzy pixels I would figure out a way that it's opt-in/opt-out
<samueldr>
either way
<matthewcroughan>
gchristensen: yeah, not something I want to debate or do long-form on. I just wondered what you thought, so thanks!
<gchristensen>
ZFS is a very different project in many ways, though. for one example, it is a large C codebase and mistakes are very costly
<samueldr>
I don't think any app should be fuzzy outside of any control
<samueldr>
that's like... basic decency
<gchristensen>
in Nixpkgs etc. mistakes are much less costly, and it is perhaps easier to go do some wild stuff in a weekend
<samueldr>
it looks amateurish, stuff you'd see on Windows
<drakonis>
i think it is a kneejerk reaction to how x permeated everything in the display stack
<andi->
reading the last few lines interleaved makes it sound wierd...
<drakonis>
so they went for minimal
<matthewcroughan>
gchristensen: what's the worst thing that's ever happened to you in NixOS/Nix?
<andi->
nixpkgs is amateurish, you can hack on it during the weekend :D
<gchristensen>
but the purpose of RFCs is the same purpose as the very many meetings and conference talks and etc. etc. etc. of zstd
<matthewcroughan>
Maybe it'd work better if there were more regular events eh? :)
<samueldr>
andi-: better than C project being basic decency
<gchristensen>
this was probably the worst thing: 2016-01-15 19:56:56 --> gchristensen (~gchristen@unaffiliated/grahamc) has joined #nixos
<matthewcroughan>
Physical ones too. Can't wait.
<gchristensen>
(I kid)
<drakonis>
samueldr: its left up to the compositor i think
<matthewcroughan>
I don't trust anybody who has sub-second access to their IRC logs.
<drakonis>
and libraries that can be used to implement compositors like wlroots
<gchristensen>
hah
<samueldr>
drakonis: yeah, and to me that sounds like a big deficiency
<matthewcroughan>
gchristensen: What do you use as an IRC setup? Are your logs plaintext?
<gchristensen>
head -n1 .weechat/logs/irc.freenode.#nixos.weechatlog
<samueldr>
I should try and dig the tweets from flibitijibibo
<andi->
is that zstd compressed tho?
<gchristensen>
lz4 andi-
<matthewcroughan>
gchristensen: ah sweet, I use quassel because it's the only thing with a decent mobile client, and notifications without gcm that actually idles properly.
maljub01 has joined #nixos-chat
<andi->
matthewcroughan: wrong :P
<matthewcroughan>
which means I'm quite tied to their framework, which uses a database.
<drakonis>
i just use weechat and ssh into my server
<andi->
weechat android is top notch
<drakonis>
its pretty sweet
<gchristensen>
I specifically *don't* want IRC on my phone :P
<drakonis>
i havent figured out how to achieve a relay on my vps, sadly.
<drakonis>
i did it once but it doesnt seem to work now
<matthewcroughan>
You don't want to get pinged by someone asking how semicolons work whilst camping?
<andi->
I get nice notifications to my cheap chinese bluetooth LE "smart" watch (OLED display + BT)
<gchristensen>
people on IRC don't pay me well enough to have that level of SLA on me
<matthewcroughan>
Ah, well I'm a sucker because I enjoy helping people when I should be working.
<gchristensen>
yeah me too, no doubt
<samueldr>
I want IRC on my phone, but not freenode notifications from it :)
<matthewcroughan>
I can't believe you manually define your sha256sums instead of using zero trust content addressed concensus. Your build system is trash.
<samueldr>
oh, I also really don't like how the keyboard can't be configured system-wide
<drakonis>
this guy is one of the early linux focused game devs
<drakonis>
he did the humble indie bundle linux ports
<samueldr>
gchristensen: he's basically THE wizard that ported the most games to Linux
<drakonis>
the ones from way back
<drakonis>
that and icculus
<samueldr>
his portfolio of ports is amazing
<drakonis>
he's also cool
<samueldr>
icculus built good foundations
<samueldr>
but in sheer amount of ports, flibitijibibo trumps him :)
<drakonis>
oh yeah
<matthewcroughan>
I've been using raw wayland and sway for a while now, and I haven't noticed any real issues other than QT5 not knowing how co-ordinates work.
<samueldr>
but yeah, apps not being able to ask/request (might be optional even) for how to scale the surface is awfully rude imo
<drakonis>
ethan did VVVVVV's linux port too...
<drakonis>
one of the true luminaries of linux game dev
<andi->
it is made for a world where you just run one windowing toolkit and that one comes bundled with a compositor that is part of a huge set of systemd services ...
<drakonis>
its more of a "delegation of responsabilities" situation here
<gchristensen>
systemd-windowsd / windowsctl
<drakonis>
i'm sure it exists already
<drakonis>
ethan also helped with terraria on linux, fantcy that
<samueldr>
btw, my intent is not to belittle wayland... but if it's so ready to be used... why every time I try to use it things are just so... not there
<adisbladis>
samueldr: Same
<adisbladis>
I was excited a long time ago but have since given up on it completely
<samueldr>
things like 20fps non-tearing rather than 60fps tearing... I get it... I should get a new computer
<samueldr>
(not really)
<samueldr>
but it's a choice they made, and I can respect it
<samueldr>
but I'm not sure pushing all responsibilities into a fractal of implementations is a good choice
<Church->
samueldr: As Devault would say, you're just not using it right. Git gud.
<samueldr>
obviously I shouldn't be able to type in my native language
<Church->
And add more cursing and ad hominems.
<adisbladis>
samueldr: Non-english! Why would you want that?!
<samueldr>
adisbladis: you wouldn't get it
<samueldr>
only a more worldly person would ;)
<colemickens>
I'm such a noob, does CONFIG_VIRTIO_FS=m mean that it's built in or not?
<samueldr>
it can also tell you the evolution through versions
<samueldr>
e.g. from boolean to tristate
<samueldr>
so you'd know on an hypothetical option in 5.1-5.3 it was y/n only, but starting with 5.4 onwards it's y/n/m
slac91918 has quit [Remote host closed the connection]
<samueldr>
can PAM use two passwords?
<samueldr>
e.g. I'd have my normal password, and additionally a numeric pin
<samueldr>
so that I don't fail to sudo multiple times
<samueldr>
and can actually unlock the lockscreen
waleee-cl has quit [Quit: Connection closed for inactivity]
<samueldr>
:| a gnome app won't fit in width since the headerbar tabs absolutely need to be centered
<samueldr>
so it adds dead space to the left of them
<samueldr>
wait, there's even useless space between the header bars and the menu button
<samueldr>
so it's doubly uselessly too wide
<samueldr>
ah no, depends on the tab, on one tab there's one more button
<hexa->
samueldr: yes, there can be multiple pam hooks, that are marked as sufficient
endformationage has quit [Quit: WeeChat 2.9]
<samueldr>
not in theory
<samueldr>
in practice
<samueldr>
is there a way to attach a second password *right now* without having to implement a new thing :)
<hexa->
Check the security.pam Module 😄
<samueldr>
implementing a NixOS module and packaging a module would be fine
<samueldr>
I don't see, from the name of the options, something for a second password though
<hexa->
I guess the question is, what is the backend for your PIN?
<hexa->
for password it's /etc/shadow and pam_unix.so
<samueldr>
I don't really care, but it's also why I asked
<samueldr>
I assume there's no way for /etc/shadow to have multiple passwords
<hexa->
yeah, I don't think so, too
<samueldr>
so I guess that whatever method would need a second mean to store
<samueldr>
I'm thinking this is going to be a requirement for many mobile-first environment that assumes PINs are cool
<hexa->
honestly, I'd expect companies like purism to take care of that
<samueldr>
I think they're going with "your password is your pin"
<samueldr>
as is plasma-mobile
<hexa->
for security reasons I assume?
<samueldr>
no idea
<samueldr>
considering many mobile linux users will expose ssh with password auth enabled
<samueldr>
how long is it to enumerate all combinations?
<hexa->
Of numeric Pins? Not long
<samueldr>
yeah
<samueldr>
so if your password is your pin
<samueldr>
and it has to be numeric
<samueldr>
and ssh is enabled with password auth enabled
<hexa->
But you'd use a different Pam config for SSH and lockscreen
<samueldr>
your *password* is numeric
<samueldr>
so that wouldn't change a thing
telnetkitten has joined #nixos-chat
<telnetkitten>
eyes emoji
<telnetkitten>
using telnet for irc means you're a cool kid right?
<Ke>
:asterisk:) using socat and xxd on matrix
<cole-h>
hehe
telnetkitten has quit [Remote host closed the connection]
<samueldr>
only if you hand type TLS
<hexa->
sure
<hexa->
I speaks ASN.1 in my sleep
<hexa->
much simpler than imagining sheep jumping over a fence
<Ke>
samueldr: implied
<Ke>
I really don't know how to send :asterisk: on matrix
<Ke>
maybe it's because I am not using socat and xxd
<cole-h>
✳
<cole-h>
?
<cole-h>
:P
<samueldr>
oh, cat butt!
<cole-h>
.....nice
<Ke>
✳✳
<eyJhb>
cole-h: Did you ever setup a private club penguin server?
cole-h has quit [Ping timeout: 252 seconds]
<eyJhb>
Okay then, didn't mean it like that
lunc has quit [Ping timeout: 265 seconds]
<LinuxHackerman>
samueldr: you could use pam_exec as a sufficient auth backend for the services you want unlockable by PIN and handle PIN storage in your script
<samueldr>
interesting
<LinuxHackerman>
I saw something for that a while back, done by some nixos person and implemented in haskell. Can't remember who or where though :/
<LinuxHackerman>
or it might have been a PAM module directly written in Haskell, not a pam_exec tool
<LinuxHackerman>
Found it! 2020-11-27, lassulus was the author
<colemickens>
yall how am I getting an error -8 frm my custom init
<colemickens>
that doesn't even make sense, I'm building on x86_64 and running with qemu-system-x86_64 :S
<samueldr>
ask the author of that init
<LinuxHackerman>
colemickens: this sounds familiar, hang on…
<samueldr>
"error -8" is reported how?
<samueldr>
ENOEXEC 8 Exec format error
<samueldr>
hmm
<LinuxHackerman>
oh wait no, the problem I had back then was actually that my init was built for the wrong architecture (little-endian vs big-endian MIPS)
<gchristensen>
the gist of it is without dio all your writes spend less time in user, but slightly more time in sys. if you use dio, the writes spend more time in user but less time overall
<gchristensen>
like 2% less
<hyperfekt>
so the lesson basically is "if your cpu is the bottleneck for io, use direct io"
<hyperfekt>
like i see little reason to avoid buffered io by default unless you have a write-only workload that is larger than memory
<hyperfekt>
Ke: that's an interesting point
<hodapp>
I'm still trying to comprehend it
<hyperfekt>
hodapp: with direct io the buffer that is to be written to the device would get read twice: once to calculate the checksum, and once when issuing the write. and there's nothing preventing changes to the buffer between those reads, which would lead to a checksum mismatch
<hodapp>
what would produce those changes though?
<hyperfekt>
in buffered io the data is copied out of the hands of userspace so it can't modify it anymore
red[evilred] has joined #nixos-chat
<red[evilred]>
,space # ULA launch 1 hour
<hodapp>
right
<gchristensen>
, launch :)
<{^_^}>
Invalid command syntax
<cole-h>
red[evilred]: It's ,launch :P
<red[evilred]>
thanks
<red[evilred]>
,launch # ULA 1 hour
<{^_^}>
# ULA 1 hour: Ping for space stuff (edit this command to add yourself, see ",help"): infinisil Taneb ldlework etu philipp[m] eyJhb gchristensen __red__ red red[evilred] risson aaronjanse Church-
<__monty__>
Recommended streams?
<hyperfekt>
maybe direct io could be modified so it switches the containing pages to read-only
aocusr has joined #nixos-chat
aocusr has joined #nixos-chat
aocusr has quit [Changing host]
<hodapp>
hyperfekt: so, because the filesystem code in the kernel (presumably where this COW FS is implemented) reads this memory twice, the userspace code could conceivably modify it in between those two reads, thus putting checksum and data out of sync?
<hyperfekt>
yes. although i think technically the second read may not be done by the kernel but by the disk but that doesn't really matter for the problem
<hodapp>
how might this happen - with another thread doing something while the syscall is in process?
<hyperfekt>
there is also asynchronous io, which can be combined with direct io
woo has left #nixos-chat [#nixos-chat]
<hodapp>
but can this problem happen even with sychronous?
<hyperfekt>
hodapp: well, as you suggested, and kind of shared memory can also lead to it
<lovesegfault>
> On Guix System, you can run herd discover guix-daemon on to turn discovery on temporarily, or you can enable it in your system configuration. Opportunistic use of neighboring substitute servers is entirely safe, thanks to reproducible builds.
<{^_^}>
error: syntax error, unexpected ',', expecting ')', at (string):494:15
<{^_^}>
1 minute: Ping for space stuff (edit this command to add yourself, see ",help"): infinisil Taneb ldlework etu philipp[m] eyJhb gchristensen __red__ red red[evilred] risson aaronjanse Church-
<__monty__>
Liftoff!
<__monty__>
Oooo, cool how you can see the vapor on the nose disappear exactly when it breaches the sound barrier : o
<hodapp>
hyperfekt: what do you mean kind of shared memory?
<hodapp>
hyperfekt: and, whether async or sync, what prevents userspace from modifying it while drive or kernel is in the midst of a read? is that done atomically?
<philipp>
Thanks for the launch pings! Barely made it.
<hodapp>
Mic92: oh, man, has the FPGA toolchain situation gotten better than where it was circa 2015? (not with NixOS, just in general)
<hyperfekt>
well as you mentioned different threads may share memory that is submitted for direct io. those threads may or may not be part of the same process
<Mic92>
hodapp: Yeah. Very pleasantly surprised the xilins runtime is even open-source.
<Mic92>
Still 50GB of shit that needs to be installed
<hodapp>
hyperfekt: so said memory might be shared between multiple processes?
<hodapp>
Mic92: blugh
<hyperfekt>
hodapp: see shmget and shm_open
<hodapp>
lovesegfault: that's very slick
<hodapp>
hyperfekt: would the read otherwise be done atomically though, such that no write (whether from same process or different process in the case of SHM) would actually matter?
<hyperfekt>
otherwise? you mean with buffered io?
<hodapp>
if no checksums were being done, thus two reads were not necessary
<Ke>
just remember tweaking page tables anyway is costly
<hyperfekt>
presumably it wouldn't be done atomically, but it wouldn't lead to inconsistency
<hyperfekt>
so you can still mess up the contents of your buffer but the kernel is never going to know about it, be it during writing or reading
<Ke>
the problem is the checksum that needs to match with the contents
<Ke>
hmm this was mentioned, not sure what is the discussion anymore, should be sleep
<hodapp>
yes, I wasn't aware that two distinct reads had to occur of memory that userspace can still freely modify
<__monty__>
drakonis: Old comment and most of the advantages boil down to "It's Guile". UX and better bootstrapping excluded.
<__monty__>
Not sure whether you wanted to start some discussion.
<eyJhb>
Is there currently any projects to make the nixpkgs lib API better? ie. consistent functions, etc.?
<__monty__>
My impression is people think flakes will solve the problem.
<__monty__>
I haven't experienced the problem much tbh.
<__monty__>
Just the lack of documentation.
<gchristensen>
I haven't seen someone thinking flakes will improve the nixpkgs lib api
<__monty__>
The reasoning is that flakes will bring composability.
<eyJhb>
Even if flakes would do it, as it COULD allow for pulling in a small lib on the side. THat lib still needs to be made :):):) :D
gchristensen has left #nixos-chat ["WeeChat 2.9"]
gchristensen has joined #nixos-chat
<eyJhb>
^ get back here
<gchristensen>
whoops
<eyJhb>
Oh... That was quick
<eyJhb>
Glad to see you again gchristensen
<gchristensen>
:P
<joepie91>
I haven't followed the latest developments around flakes, but yes, I do feel that they can help utility quality if done correctly; because it significantly lowers the bar to trying out better APIs
<gchristensen>
keyboard shortcuts
<eyJhb>
gchristensen: Hate them, love them, can't live without them :D
<joepie91>
this is pretty much the exact same mechanism by which the JS ecosystem could very rapidly iterate towards much better-designed APIs than libraries commonly have in other languages
<drakonis>
when what you have is a language runtime and everything is built on top of it, its much easier to iterate
rajivr has quit [Quit: Connection closed for inactivity]
<joepie91>
"arguing that the stdlib should change in a certain way" is just something that is much more difficult and requires much more effort and commitment than "just writing the thing and seeing whether people are interested in using it"
<drakonis>
it also needs an wholy different culture to achieve it
<__monty__>
I do see flakes making it easier to use different lib implementations. I don't see this having a fast effect on Nixpkgs though.
<adisbladis>
I think according to current practices even after accepting a potential new flakes rfc and then stabilising the feature it would take at least 2 more years before nixpkgs could use external flakes
<joepie91>
I would expect roughly the same thing to happen as in JS: the best utilities exist stand-alone, and the mostly-decent designs make it into kitchensink libs with some latency (Lodash in the case of JS, probably nixpkgs in the case of Nix)
<hodapp>
what does flakes do to help composability? I ask this not knowing very much about flakes
<elvishjerricco>
Apparently Apple's new "Magic Keyboard with Touch ID" can be used with any M1 mac. It just acts like a an encrypted channel between a physical sensor and the M1's secure enclave. I was sort of expecting it to be perma-paired with the mac it ships with, like the sensor and enclave are on other apple devices with the sensor built in. But no, you can actually re-pair it with another mac (unclear if it can pair with ipads/iphones)
<elvishjerricco>
If only they would open source more of it...
<drakonis>
hodapp: i think it gets used for pinning multiple repositories
<aocusr>
iirc you can buy the keyboard without any device
<drakonis>
its kind of a weird band-aid honestly
<drakonis>
its not too different from proglangs with pinning based package managers
<aocusr>
even funnier, apple announced loseless apple music which non of it's headphone supports it
<drakonis>
it is a huge deal because it was supposed to deliver us from the dark ages of nix_path for using multiple sources
<MichaelRaskin>
None? Don't they have anything wired anymore with an adapter?
<drakonis>
aocusr: oh you're on #guix
<drakonis>
the name was familiar as heck
<aocusr>
rofl
Qwerky has joined #nixos-chat
<cransom>
MichaelRaskin: yeah, none. even if you plug a lightning cable into the brand new 600$ pair they debuted recently to your machine, it's not 'lossless'.
<MichaelRaskin>
Argh but how
<hodapp>
D:
<MichaelRaskin>
I mean, my opinion of Apple peaked right before the first iPhone and has never since trended up, but sometimes they still manage to surprise me
<hodapp>
why have a cable that can carry eleventy billion bits per second if it can't even carry 44,100 samples/second * 2 channels * 16 bits/sample?!?!
aocusr has quit [Quit: Connection closed]
<talyz>
eyJhb: [hiking] don't know any specific trails or such, but it's beautiful up north :)
<adisbladis>
talyz: I told him to go to Åre
<adisbladis>
Take the cabin lift up and start from there
<adisbladis>
It's spectacular
<talyz>
adisbladis: oh, yeah, Åre is great :)
<__monty__>
hodapp: My understanding of it is that flakes allow you to pin all your inputs without forcing your consumers to use the same pins.
red[evilred] has quit [Quit: Idle timeout reached: 10800s]
<hodapp>
__monty__: ahh, okay!
<eyJhb>
talyz: My biggest fear is if there is still snow/very cold there mid June
<__monty__>
hodapp: What I don't know is why doing the same with plain nix isn't equivalent. Since you can have your pins be default argument values.
<hodapp>
huh, it isn't equivalent?
kraem has joined #nixos-chat
<__monty__>
If it were equivalent then why would people hope it fixes the lib situation?
<__monty__>
Or helps fix at least.
<gchristensen>
like I said a while ago, I've never heard anyone suggest it would
<hodapp>
wasn't sure what you meant. something could result in an equivalent derivation - yet be much easier to work with
* colemickens
bought an aliexpression ryzen 5000 series laptop and yall its amazing and gorgeous and everything just worked out of box.
* hodapp
is delaying buying a Ryzen because of component prices being stupid, but is however enjoying his new M1 Macbook.
<hodapp>
...except that I want Nix on it and haven't gotten around to figuring out how yet.
* colemickens
sees his typo and imagines what business aliexpression would be in
<hodapp>
err, which typo?
<hodapp>
I saw 'aliexpression' and figured it was a real thing
<adisbladis>
hodapp: Declarative ecommerce
<gchristensen>
vaguely and sometimes completely wrong imitation expressions
<colemickens>
expression has multiple definitions too, just throwing that out there
* colemickens
slips out quietly to do battle with /init
Qwerky has quit [Remote host closed the connection]
h0m1 has quit [Quit: WeeChat 3.1]
h0m1 has joined #nixos-chat
<samueldr>
hmm.. been trying linux-zen on that bad atom tablet and it seems to help with responsiveness, not noticeably with phosh's, but with ssh'd uses
<samueldr>
e.g. opening a new tmux... "tab", and then cd-ing into the configuration and using `tig` is noticeably snappier
<talyz>
eyJhb: I can give you updates on the situation - one of my colleagues lives in östersund
<talyz>
eyJhb: i don't think snow's not very likely in mid-june, but who knows :p
<eyJhb>
talyz: The double negative `don't` + `not` seems to imply... There might be snow? :p
<adisbladis>
talyz: It's not unheard of iirc
<adisbladis>
Also the elevation difference between östersund and the mountains in åre is pretty drastical
<MichaelRaskin>
If not for the _book_, it would not even qualify as fiction!
abstrn has quit [Quit: later]
abstrn has joined #nixos-chat
<talyz>
eyJhb: oops :p
<adisbladis>
eyJhb: Might I recommend a warmer country if temps/snow are such a concern? :P
<talyz>
adisbladis: no, definitely not, just less likely :)
<philipp>
Wasn't there this cartoon show for kids a decent time ago that was supposed to teach them about their body where heavily anthromorphised cells were the characters?
<MichaelRaskin>
abathur: there is also the game of estimating the impact of the laws _forbidding_ wearing face-concealing masks in public, even during flu season. As we know masks slow down flu, maintaining normalisation of masks will save lives!
<lukegb>
there's a "show unavailable videos" option in the dango menu
<lukegb>
which add a bunch of "[Private video]" entries
<__monty__>
Decent show.
__monty__ has quit [Quit: leaving]
<philipp>
Jimbo nearly died because no adult could be bothered to keep track of their shots.
NinjaTrappeur has quit [Ping timeout: 250 seconds]
NinjaTrappeur has joined #nixos-chat
supersandro2000 is now known as Guest2135
Guest2135 has quit [Killed (orwell.freenode.net (Nickname regained by services))]
supersandro2000 has joined #nixos-chat
kalbasit[m] has quit [Ping timeout: 258 seconds]
Ke has quit [Ping timeout: 258 seconds]
ma27[m] has quit [Ping timeout: 245 seconds]
mjlbach has quit [Ping timeout: 258 seconds]
rmcgibbo[m] has quit [Ping timeout: 258 seconds]
puzzlewolf has quit [Ping timeout: 258 seconds]
arcnmx has quit [Ping timeout: 258 seconds]
veleiro has quit [Ping timeout: 245 seconds]
berbiche has quit [Ping timeout: 245 seconds]
DavHau[m] has quit [Ping timeout: 258 seconds]
aanderse has quit [Ping timeout: 258 seconds]
thefloweringash has quit [Ping timeout: 245 seconds]