<infinisil>
elvishjerricco: What was the problem with it before that?
<elvishjerricco>
samueldr: That tweet contains confusing wording so I'm not sure what they're saying :P
<samueldr>
elvishjerricco: inscrutable I said
<elvishjerricco>
infinisil: Couple of things. Primarily, it lacked simd acceleration. But it also used ccm by default if you just set encryption=on, which performs worse
<elvishjerricco>
All because the kernel GPL'd the FPU functions
<samueldr>
I think they meant that `csrutil` is going to be how you'd "bless" something to boot, rather than using the lower level `bputil`
<infinisil>
I see
<elvishjerricco>
samueldr: Ah, interesting
<elvishjerricco>
Of course it won't matter if we can't make GPU driver for these things
<samueldr>
people eventually will
<samueldr>
but even without gpu driver, you can get by with a simple framebuffer driver I would bet, like you can on iPhone
<samueldr>
as long as the firmware initialized it
<samueldr>
though, you say "it won't matter", I think it will if you can run bare-metal KVM to *then* launch macOS
<samueldr>
(without graphics)
<elvishjerricco>
samueldr: That's a fair point. There's gchristensen's cool stuff for booting macos over a Linux ZFS volume.
<elvishjerricco>
Which of course may require a driver for that custom SSD controller...
<samueldr>
exactly what I'm thinking about; hydra builders
<abathur>
my 2013 macbook air would estimate over 21h with wifi off and ST3 and a few terminals open
<__red__>
Should an updage to glibc for nixos-20.09 go into staging or staging-next?
<__red__>
for 20.09
<gchristensen>
-dev :) too off (on) topic for this chanel
<gchristensen>
hmmm this nixos iso doesn't see the wifi adapter
<samueldr>
new_kernel?
<samueldr>
dmesg | grep -i firmware # or maybe fw
<gchristensen>
I just fetched whatever I got from the website
<samueldr>
so, no
<gchristensen>
:D
<gchristensen>
no big deal, I was going to erase the disk anyway
<samueldr>
y'all should stick with ~2014 era laptops, works just fine here
<gchristensen>
lol
<__red__>
hah - okay :-P
<infinisil>
Hm, would it possible to hold lockdowns every other week to slow it down, but not drive everybody insane
<samueldr>
it takes around 2 weeks for an "event" to show its full potential in the community, COVID-19-wise
<samueldr>
so e.g. you do lock down, it will take at least two weeks until the new cases _stop rising_
<samueldr>
so it's not really a realistical thing to hope for, it might slow down, but keep climbing exponentially :(
<infinisil>
Like, holding weekly lockdowns should at least help
<infinisil>
(assuming social distancing is still done in the other week)
<infinisil>
In comparison to just social distancing all the time
<infinisil>
s/weekly/other-weekly
<infinisil>
And it's a state that might be maintainable over a long period of time
<infinisil>
In comparison to e.g. a single 6 week lockdown
<samueldr>
but when a lockdown is in place, at least here, it's because social distancing is not cutting it
<samueldr>
because there is no situation where there is no social distancing
<infinisil>
I guess the main point I'm making: full lockdown is too harsh, but only social distancing doesn't really cut it, so how about a middle ground
lopsided98_ is now known as lopsided98
<samueldr>
there is no middle ground :/
<infinisil>
I just proposed a middle ground..?
<gchristensen>
anyone else have a knowing chuckle when they typo nix-hell?
<samueldr>
AFAIUI your proposal will give the same results as no lockdown, but later
<samueldr>
yeah, things suck, they might and will get worse, then they'll get better
<infinisil>
samueldr: Suppose somebody gets infected in the non-lockdown week but they don't realize it yet. During the lockdown week they do, which leads them to stay inside for that week and the following ones
<infinisil>
I can imagine that this would happen a lot
<cole-h>
What if they don't realize they're infected until after the next non-lockdown week
<cole-h>
Now they've infected everybody they came into contact with, no?
<cole-h>
Or risk doing so, anyways
<samueldr>
I don't know when the infectious phase of this virus is, is it before or after symptoms start?
<infinisil>
That also happens with no lockdown anyways. I'm saying that you can at least prevent a portion of people realizing it earlier, before spreading it more
<samueldr>
my gut feeling really is that any saving it does, the equivalent opposite effect is likely to happen a week later or earlier :/
<samueldr>
you've said it, that's going to happen with no lockdown anyways, which is why there is a lockdown
<infinisil>
I mean, from a purely mathematical perspective, the lockdown <-> no lockdown is a spectrum, and the spread of the virus is pretty much continuous along that line
<samueldr>
so yeah, it's going to _slow_ down propagation, but it's still going to propagate
<infinisil>
So assuming no human strategizing is in play (which in reality we can't ignore), having other-weekly lockdowns *has* to help
<elvishjerricco>
samueldr: Slower propagation is valuable
<elvishjerricco>
Lessening the stress on the healthcare system is important
<samueldr>
yes
<infinisil>
samueldr: Yeah, but we just need to keep the R_0 or whatever it's called as low as possible
<samueldr>
but as far as my gut tells me, the only thing the proposal does is delay the same end-result
<elvishjerricco>
infinisil: Never seen the term R_0. What is that?
<elvishjerricco>
samueldr: which is still useful
<elvishjerricco>
Delaying eases stress on healthcare
<samueldr>
elvishjerricco: yes, but I don't think you get where I'm coming from
<samueldr>
elvishjerricco: the proposal from infinisil is to do lockdown one week in alternance
<samueldr>
elvishjerricco: of which I think: that's just delaying things that will happen, probably not enough to matter; the full lockdown is the way
<infinisil>
Of course a full lockdown would be best to prevent spread. But economy and the psychology can't handle that
<elvishjerricco>
samueldr: Yea I don't have the knowledge to argue the matter. But I would be surprised if the alternating plan would slow the spread measurably
<samueldr>
elvishjerricco: so I think we agree
<infinisil>
I think this calls for a simulation!
<elvishjerricco>
Would not* sorry
<infinisil>
Hehe
<samueldr>
my uneducated gut feeling also thinks this would not slow the spread as much as anyone would hope
<samueldr>
uh
<pie_>
i need more ram...i cant migrate tabs to my wiki fast enough...
<elvishjerricco>
I'll copy and fix my message to be clear :P
<elvishjerricco>
9:44 PM <elvishjerricco> samueldr: Yea I don't have the knowledge to argue the matter. But I would not be surprised if the alternating plan would slow the spread measurably
<infinisil>
I'd like to summon 3blue1brown for simulating this
<samueldr>
my gut feeling tells me it'd give you about a week in offset
<elvishjerricco>
Yea I don't have the knowledge to make a claim about it
<samueldr>
coming from the fact that things are definitely not instant
<pie_>
i summon the fifth piece of exodia and activate my blue eyes white dragon in trap mode
<pie_>
check mate gandalf
<elvishjerricco>
Lol
<samueldr>
where I'm from, things got heated in a day, and it took two weeks until we hit the tip of the spike in new daily cases, from basically about one bad instance of ~50 cases in a day
<pie_>
pikachu never saw it coming
<samueldr>
that's with "red zone" measures in place
<samueldr>
not total lockdown, but basically no gatherings at all
<samueldr>
5 weeks after those measures started, we are _just_ about to get back to the numbers from that moment
<samueldr>
pie_: bars
<pie_>
samueldr: you can close the bars separately
<abathur>
kids, parties
<samueldr>
that's what is used here
<samueldr>
uh, not what is used
<samueldr>
but that's the reason I understood from that
<pie_>
abathur: hmmm well ok i was going to say they can do that anyway but yeahhh i guess youre right
<samueldr>
but yeah, curfew doesn't, in my mind, seem to be a good solution
<samueldr>
only if it's in a no-lockdown and gatherings allowed situation maybe
<abathur>
they can, but if you've got a curfew you don't have to play games with asking why they're out
<elvishjerricco>
samueldr: curfew is better than nothing
<elvishjerricco>
Not great though yea
<abathur>
and them lying to you about how they're going to get some soda
<pie_>
well
<infinisil>
Oh also, with a low number of infections left, other-weekly lockdowns probably make contact tracing a lot more effective
<samueldr>
I think it's useless if you're in a situation where you can't have gatherings anyway
<pie_>
people run into eachother during the day anyway so
<pie_>
like, curfew but not full lockdown
<samueldr>
I wonder if anywhere else they have the measures we have here
<elvishjerricco>
samueldr: Well those infected appear to be twice as likely as others to have gone to bars and restaurants.
<samueldr>
yep
<elvishjerricco>
So a curfew would quell the bars
<samueldr>
yes
<pie_>
i live in a dorm
<samueldr>
but e.g., here, we're in a "red zone", there is no bars open, no restaurants
<pie_>
there is no way they ar enot massive spread centers lol
<samueldr>
no gathering
<samueldr>
no visits either
<samueldr>
but not a lockdown either, as shops are open, takeout is possible
<pie_>
but yeah hungary is just now extending measures again
<samueldr>
you're free to stroll along outside, but not gather
<pie_>
dorms emptying, uni is going remote
<samueldr>
and that, I think, is better than a curfew, actually forbidding _at all time_ those activities targeted by a curfew
<pie_>
im confused because they seem to be doing both here
<samueldr>
we tried making bars close earlier
<abathur>
probably
<abathur>
it depends on how you are able/trying to enforce it I think
<samueldr>
and that's where most of our new cases started from anyway
<samueldr>
in superspreader events :(
<abathur>
i.e., if you can't legally close the bars, you might have to do a curfew to try and limit how many people risk going
<samueldr>
here, there is nothing more enforceable in a curfew than just disallowing gatherings
<pie_>
on the upside i can now sit naked in the dorm room all day because my roommate moved home
<pie_>
:(
<abathur>
you might also have jurisdictional overlap, where one entity can close the bars, and does so
<abathur>
and another entity can impose a curfew but not close bars, and does so to make sure it's obvious they're doing something :)
<samueldr>
yeah, luckily here it's all on the same level, without that many fine-grained levels anyway
<samueldr>
the provinces are given full authority, and there is no cities, "counties" or regions that have jurisdiction that matter here
<abathur>
we've had a lot of jurisdictional stuff here
<samueldr>
technically a city could have a mandate for _some_ activity sectors that goes further than the provincial government, but thankfully the provincial government "is doing too much" [according to some people]
<colemickens>
You can type three paragraphs out but you can't make the damn horse actually read it.
<colemickens>
also, I sitll have absolutely no idea how to get gpg-agent to recreate it's socket files under /run/user, if, for example, you replace them while sshed.
<colemickens>
okay, you can manually stop and then start the sockets, I guess that works.
<__red__>
Just checking
<__red__>
19.09 is EOL right?
<__red__>
We don't have any crazy LTS releases or anything
<etu>
Yes
<etu>
We don't do LTS
<__red__>
thanks
<__red__>
didn't think we did
waleee-cl has quit [Quit: Connection closed for inactivity]
cole-h has quit [Ping timeout: 240 seconds]
<eyJhb>
srhb: 18GB left now. Think it is time to cleanup/rethink my folder structure + backup plan
<eyJhb>
Atm. I have backups of random stuff, that I move into my project folder and then delete again (e.g. 20 GB of data from a simulation for B.A.T.M.A.N. V)
<eyJhb>
etu: you use fish right? How do you deal with having fish history persistent? Atm. I use the persistence module, but hm wants to write to the history file as well...
<etu>
eyJhb: I symlink .local/share/fish to a filesystem
<eyJhb>
etu: Hmm, yeah, that might be better. I only do it with the one file
inlaw[m] has quit [Quit: Idle for 30+ days]
<eyJhb>
WELL! Okay, now I have 0 disk space left
<ashkitten>
hm, at some point i'm gonna have to figure out how to move away from protonmail
<ashkitten>
nextcloud has a nice email frontend and having that integration especially with the new nextcloud dashboard would be very good
<eyJhb>
I am getting so screwed over atm. by my ZFS
<eyJhb>
Tried to mv from one dataset to another, and I ran out of disk space
<eyJhb>
Which somewhat makes sense, but... :(
<tilpner>
eyJhb: Yeah, I just walked into spa_slop_shift reserving 232.64GB again
<tilpner>
(You could check how much it reserves for you, and whether that would make a difference, and then figure out exactly why you would not want to adjust it)
<eyJhb>
tilpner: For some reason `rpool/safe/persistent 13.7G 29.8G 3.14G legacy`, so it uses 13.7, but the real size of the dir is 2GB
<JJJollyjim>
oh, i meant in a nixpkgs stdenv.mkderivation
<sphalerite>
ooooh
<tilpner>
eyJhb: nixpkgs.obs-studio
<JJJollyjim>
i'm not sure how ordering works? is it only defined by the subject lines? does git am sort by that?
<sphalerite>
me neither
<eyJhb>
JJJollyjim: you could just have a list, that you split and give to git am ind a mkDerivation
<eyJhb>
Else, if you give patches inside a mkDerivation, it is ordered by .. the list
<sphalerite>
you could git am, then use git diff to fold the whole series into one patch, but then you lose all the metadata
<eyJhb>
tilpner: Will try :D
<JJJollyjim>
git am will need the sources to be a git repo right?
<JJJollyjim>
(the thing im patching)
<eyJhb>
tilpner: After I have deleted all the empty dirs
<eyJhb>
Apparantly I have 2.8 GB of empty dirs
<tilpner>
How many did you have?
Jackneilll has quit [Read error: Connection reset by peer]
<JJJollyjim>
i need someone who advocates for mailinglist-based development to point to the tools they use to actually make it manageable
<JJJollyjim>
cos it feels pretty hellish to me
<eyJhb>
tilpner: 55997 directories, 0 files atm. I think we were at ... 200000 before
<tilpner>
Sure, who doesn't have 200k empty directories... >.>
<JJJollyjim>
"Note, some patches are also depend to the code from GPIO fixes / for-next repositories. Unfortunately there is no one repository which contains all up to date for-next changes against GPIO subsystem. That's why I have merged Bart's for-current followed by Linus' fixes followed by Bart's for-next followed by Linus' for-next branches as prerequisites to the series."
<JJJollyjim>
screaming
<eyJhb>
tilpner: rsync will not remove the dirs, only the files :p
<JJJollyjim>
(there are also multiple linuses involved and i never know which one they mean)
<eyJhb>
Might be a indication of, htat I need to cleanup
<eyJhb>
etu: `Error installing file '.local/share/fish/home-manager_generated_completions' outside $HOME` I guess was the reason for not doing the entire fish folder
<{^_^}>
#94315 (by balsoft, 14 weeks ago, closed): MESA-LOADER: failed to open $driver
Jackneilll has joined #nixos-chat
<etu>
eyJhb: I see
leah2 has quit [Remote host closed the connection]
leah2 has joined #nixos-chat
<eyJhb>
tilpner: I got obs to work, it uses channels when i nix-shell, so the versions differ, and therefore will not work
<eyJhb>
srhb: Fun fact, if you start jupyter lab when there is a 3
<eyJhb>
srhb: Fun fact, if you start jupyter lab when there is a 3-4 GB .tar.gz docker image, then it will crash because it will apparantly try to read the file/parse it
<averell>
i'm guessing aarch64 has some growth ahead/recently?
<aminechikhaoui>
anyone remembers when wireguard was merged into linux ? I guess I wouldn't need boot.extraModulePackages = [ config.boot.kernelPackages.wireguard ]; for Linux 5.4.x ?
<aminechikhaoui>
gchristensen oh so only on >= 5.6.x
<sphalerite>
I feel like having logs.nix.samueldr.com on the first page of duckduckgo results for a non-nix-related search query might not be a good sign.
<gchristensen>
uh oh
<joepie91>
lol
<sphalerite>
(anyone here using journalbeat and have any idea why it's exiting immediately for me?)
<sphalerite>
(there it goes up the list of search results)
<joepie91>
you're making it worse!
<sphalerite>
I know!
<sphalerite>
OK, let me make it better! It works now that I defined journalbeat.inputs.
waleee-cl has joined #nixos-chat
FRidh has joined #nixos-chat
<jtojnar>
someone actually made a CMake fork that allows functions to return values :-O
<eyJhb>
`Someone has already added that SSH key.` arghhh!
<sphalerite>
eyJhb: ooh I like this DoS vector.
<eyJhb>
sphalerite: How come ? :p
<eyJhb>
Slow process?
kalbasit has joined #nixos-chat
monsieurp has quit [Remote host closed the connection]
monsieurp has joined #nixos-chat
monsieurp has quit [Client Quit]
endformationage has joined #nixos-chat
monsieurp has joined #nixos-chat
FRidh has quit [Ping timeout: 256 seconds]
FRidh has joined #nixos-chat
FRidh has quit [Ping timeout: 256 seconds]
FRidh has joined #nixos-chat
vika_nezrimaya has quit [Ping timeout: 264 seconds]
* etu
noticed that his blog was linked from the nixos.wiki about ZFS today
* etu
is a bit surprised but happy that people found it helpful :)
<aleph->
Okay stupid question because I can't remember. PHP app running over nginx + php-fpm, as a plain cli scipt takes about 18-20s to run like I want. Takes more like 1-2min~ over the local network which I assume is due to the amount of text the service is spitting out. Think increasing data buffers in nginx would help?
cole-h has joined #nixos-chat
<aleph->
sphalerite: I've fiddled a bit, can help yah later today assuming this isn't work. :)
rajivr has quit [Quit: Connection closed for inactivity]
<eyJhb>
I think etu is the PHP expert :D Not sure if that includes configuration
<etu>
eyJhb: I sure know a lot of PHP, but this doesn't really feel like a PHP question
<etu>
We had a cut down hash at some place that we returned to the clients, and somewhere we triggered some typecasting by accident so we got an integer overflow.
<eyJhb>
That sounds like a fun thing to debug! :p
<etu>
And yeah, as you can expect the hash was something like 0e14123
<eyJhb>
Of course :p
<etu>
It's a bit stupid how it tries to auto typecast... to fail... overflow... return 500 server error rather than just... having it as a string.
<etu>
eyJhb: Most "security flaws" that people experience in PHP are easily avoided... such as prepared statements, don't deserialize data you don't trust, etc
<etu>
eyJhb: don't use eval
<etu>
etc
<gchristensen>
so many of the problems are just as common in other languages at this point
<eyJhb>
etu: but even when they are basic, people still do them! And they are nice
<eyJhb>
<?php include($_GET['page']); ?>
* colemickens
feels like he's stuck on "babies first socket app today
FRidh has quit [Quit: Konversation terminated!]
<eyJhb>
I still loved the, magic_espace_quotes or something like that, and then use it as $query = 'SELECT * FROM ... WHERE $var';, well.. I don't need to escape your quotes if you omit them
<gchristensen>
lol
<etu>
eyJhb: yes, don't do that :D
<gchristensen>
I feel like I'm back in 205
<gchristensen>
2005*
<etu>
gchristensen: People just love pointing out flaws of past PHP somehow
<eyJhb>
You should just do some google dorks for ?id=, and then have fun gchristensen ! (I do not recommend anyone do this)
<gchristensen>
I mean, I know
<eyJhb>
etu: I say that less than a year ago
<gchristensen>
but people still use pickle
<eyJhb>
The problem is, that PHP is easy to get into, and most pay for a design. So you get some designer, whe knows shit about code, making a barely working website that does very weird stuff
<eyJhb>
ANd then you end up with things that should be 2000 bugs, but are a 2020 bug
<gchristensen>
I'm sorry PHP is so successful
<eyJhb>
I saw* damn it
<eyJhb>
But I wish I was pointing out old bugs etu :( I fucking wish
<eyJhb>
A place I know had a SQLi on their card system, that made it possible to grant your access card rights to EVERYTHING
<etu>
eyJhb: Yes, but you can't always blame the language for people writing bad code.
<eyJhb>
No no, I don't blame PHP
<eyJhb>
I blame the people, but PHP is just popular
<eyJhb>
ANd easy to get into, etc.
<eyJhb>
I see stupid shit in Python as well :p
<eyJhb>
PHP is basically fun, because it is so popular like gchristensen also pointed out. You should know by now, that I love PHP as it was my first real true love <3
<gchristensen>
samj
<gchristensen>
e
<eyJhb>
I remember my first Paypayl IPN integration. Good thing people did not figure out the path to that thing.
<eyJhb>
`Ohh this curl thing here? That seems advanced, why do I need to make a request back to Paypals servers?`
* eyJhb
<-- was part of the problem
<eyJhb>
But everything sec related that people see commonly and maybe less common in some langs, that pose risks etc. I would fucking love snippets of that. Just saying!
<NinjaTrappeur>
I was about going to say that without php and easyphp, I wouldn't be reading this channel. These tools have been a superb all-in-one coding introduction for a lot of 2000s kids. But then I just realized easyphp is a very local French thingie that do not even have a English wikipedia page oO
<samueldr>
(it crossed the ocean in Québec at least)
<samueldr>
I think for me it might be PHP, maybe mirc scripting that got me hooked at first...
<samueldr>
oh, and the games factory and RPG maker
<gchristensen>
php then gamemaker for me
<srhb>
cole-h: Ha! The honor!
<sphalerite>
for me it was hypercard, then later my dad got me a C++ book.
<sphalerite>
Somewhere along the way I got pretty into PHP too
<sphalerite>
hypercard was well before anything else though, I didn't realise I was programming and it was quite a few years later that I asked my dad how programs are made
<pie_>
is facebook chat being broken for anyone else
nckx has quit [Ping timeout: 264 seconds]
<waleee-cl>
the site is a massive cpu hog, but the chat works
<pie_>
nvm turns out it was my local network being very weird
<eyJhb>
I did a lot of PHP and SourcePawn for Counter-Strike Source! (sourcemod), and whatever the hell es addons was coded in... :D Hell I miss doing SourcePawn!
Hm7000 has quit [Ping timeout: 260 seconds]
Hm7000 has joined #nixos-chat
nckx has joined #nixos-chat
<cole-h>
Aw yiss, my Rhythm of War signed copy will be "shipped out by the publication date of 11/17" :DDDD