supersandro2000 has quit [Disconnected by services]
supersandro2000 has joined #nixos-chat
<colemickens>
ashkitten: / sphalerite : was it one or both of you have yubikey gpg flakiness issues? Do you happen to use pcscd as well? I've found that just restart pcscd is a lighter "fix" than what I was previously doing. I'm thinking of adding a udev rule for it.
<ashkitten>
i don't use pcscd
<ashkitten>
but i have to replug my key whenever i reboot or log out
<colemickens>
hm ok
evanjs has quit [Read error: Connection reset by peer]
evanjs has joined #nixos-chat
rajivr has joined #nixos-chat
cole-h has quit [Quit: Goodbye]
cole-h has joined #nixos-chat
cole-h has quit [Quit: Goodbye]
cole-h has joined #nixos-chat
cole-h has quit [Client Quit]
cole-h has joined #nixos-chat
cole-h has quit [Quit: Goodbye]
<abathur>
anyone know much about binary analysis? wondering how feasible it is to do something like taint checking to determine which syscalls anything from argv makes it into? (more to the point, if anything in argv is passed to a list of syscalls)
<gchristensen>
*perks up*
<abathur>
oh, yes, it is your fault :P
<gchristensen>
:D
<abathur>
I need to do <refactor I haven't quite found enough lightbulbs/motivation for>, so instead I'm procrastiworking at the bottom of this well instead
omnd has quit [Remote host closed the connection]
<gchristensen>
I love it when things like this are my fault
<abathur>
with the caveat that I don't ~know anything here; I stumbled into something the other day that makes me think we can use `nm` and some grep/filter on an unstripped binary to figure out if it can exec*
<abathur>
and so from there my question is, yes, but can it exec arbitrary arguments
tilpner_ has joined #nixos-chat
<abathur>
it's also not very helpful on stuff like the coreutils single shared binary
<infinisil>
Oh wow, only just now learned about the boom
<infinisil>
A broom, vroom and a boom in a single day!
tilpner has quit [Ping timeout: 256 seconds]
tilpner_ is now known as tilpner
<abathur>
though I do have a more immediate use-case this time; picking at how resholve should handle some more complex command-running commands (ex: sudo, find, xargs)
<abathur>
but I'd really rather try to keep it from needing to just understand the syntax of a large number of long-tail commands
<abathur>
I had a small lightbulb the other night that I might be able to flip the problem around: if resholve can determine whether the command can exec arbitrary arguments with tolerable accuracy, when it lacks "hints" to tell it a command/arg is actually fine, it can make the monkey triage
skrzyp has quit [Ping timeout: 276 seconds]
sterni has quit [Ping timeout: 265 seconds]
<V>
abathur: uhh I'd opt to insert the tracing as a compiler pass, it's probably simpler
<V>
at least if you're using LLVM
<abathur>
I don't want to trace it
<V>
tracing in the sense of tracking which data gets into the exec call
<V>
you lose information once it's been compiled
<V>
so static analysis is better here
sterni has joined #nixos-chat
skrzyp has joined #nixos-chat
<abathur>
I don't know exactly what that would entail, so I don't want to over-assume, but it'd be out-of-scope for resholve to recompile a dozen packages to resolve paths in a shell script or something
<abathur>
so if it was going to work that way around it'd need to be more like something reusable/queryable in the nixpkgs soup
slack1256 has joined #nixos-chat
Sigyn has quit [*.net *.split]
AMG has quit [Quit: Quit]
dingenskirchen has quit [Ping timeout: 246 seconds]
<DigitalKiwi>
i'm pretty sure a lot of business insider articles are written by bots lol
<eyJhb>
You get extra points in buzzword bingo if you use ML/AI to write your stuff!
<philipp[m]1>
DigitalKiwi: They prefer the term "drones".
ece has quit [Read error: Connection reset by peer]
<DigitalKiwi>
>>/say /home/kiwi/.glirclogs/#haskell-offtopic/2019-04-30.log:[20:34:00] <DigitalKiwi> businessinsider once had an article about how someone had train ed a neural network on news articles and it was 'too good' to be released; ironic since i've often suspected businessinsder articles of being writte n by AI
<{^_^}>
Starlink launch in 4 minutes https://www.youtube.com/embed/d5DzoKuhdNk: 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
<philipp[m]1>
liebach: That thing doesn't even go to space! Boooooring! There are other concepts for nuclear powered rocket engines that don't create nuclear wastelands everywhere they go though.
<philipp[m]1>
At least during nominal operation...
<liebach>
Well, with this one the fallout was a "desirable" effect.
<liebach>
"It was proposed that after delivering all its warheads, the missile could then spend weeks flying over populated areas at low altitudes, causing secondary damage from radiation."
<liebach>
Extreme evil.
<ehmry>
the weird thing is the project was directed my merkle, the father of merkle tree merkle
<pie_99>
elvishjerricco: did you ever get around to implementing this?: <elvishjerricco> jasongrossman: For backups, I'm looking at getting proper logarithmic snapshotting, rather than the dating crap that most systems use
<pie_99>
elvishjerricco: on an unrelated note, we seem to have convergetntly found the same solution to getting grub on zfs, by using a fake partition to get it to generate stuff and then using moving/mkimage to put the right stuff in the real place
<pie_99>
grub-install is so inflexible i hate it :P
<eyJhb>
pie_99: New name? :o
<pie_99>
eyJhb no i just havent set up my quassel after getting my machine booting again
<pie_99>
im on webchat xD
cosimone has joined #nixos-chat
cole-h has quit [Ping timeout: 256 seconds]
<ehmry>
has anyone done any packaging of software to run in wine or dosbox?
betawaffle_ is now known as betawaffle
<DigitalKiwi>
i have a nix-shell shebang script that runs a program with steam-run (i think that uses wine) does that count?
<pie_99>
ehmry sort of
<pie_99>
ehmry funny that i happe to be looking at chat :P , some other people have also aske dbut i dont know wo or if tey got anywhere
<DigitalKiwi>
wait now my head hurts; if steam uses wine, why's it work with linux programs lol
<ehmry>
it would be nice to do "unattended" installs of old windows stuff from original install media, not sure if that is practical
<ehmry>
also, I've never understood why installers were a thing in windows
<ehmry>
if someone has a copy of windows 3.1 and could PM me a link, it would be greatly appreciated
hax404_ has quit [Quit: WeeChat 2.9]
hax404 has joined #nixos-chat
<ehmry>
or maybe this is actually on archive.org…
Hurttila has joined #nixos-chat
<andi->
Is there a way to get an increase API ratelimit for GitHub? I'm constantly hitting it... I do one API call for every email thread that they send me but that seems to be too much :(
<etu>
andi-: Is the same data available through RSS, github has RSS feeds for lots of different things?
<etu>
s/RSS, g/RSS? G/
<andi->
I need to query private repos/orgs as well :(
<andi->
I am annotating all my mails with github PR/issue status (and synchronize labels between GH and notmuch)
<DigitalKiwi>
what api endpoint do you call
<andi->
DigitalKiwi: webhooks wouldn't cut it unless I reconstruct the state of PRs on some server of mine..
<andi->
DigitalKiwi: mostly just the pull/:number and issues/:number
<andi->
that is all
<andi->
Even with using an API keys I regulary trigger "anti malicious" pages whenever I open GH in my browser now... And I do about ~1000 API requests a day..
<supersandro2000>
rotate multiple API keys possible from multiple accounts?
<DigitalKiwi>
just don't let them know that if they're free accounts lol
<andi->
I am trying to get rid of more GitHub workarounds in my life don't want to add even more /o
<DigitalKiwi>
or multiple IPs that you use unauthenticated ones for the public repos
<andi->
Sure, just tell GitHub to roll out IPv6 ;-)
<DigitalKiwi>
just burn a bunch of digitalocean ips with nixops-digitalocean lol
<YoDawg>
Yeah, I haven't tried that yet, I'll give it a shot. I think the main issue is how bazel fetches external dependencies, it doesn't seem to be compatible at all with the way NixOS handles things.
<YoDawg>
Thanks for the help
waleee-cl has joined #nixos-chat
feepo has quit [Ping timeout: 272 seconds]
feepo has joined #nixos-chat
cole-h has joined #nixos-chat
Jackneill has quit [Read error: Connection reset by peer]
rajivr has quit [Quit: Connection closed for inactivity]
<YoDawg>
I managed to get things working but I had to use buildFHSUserEnv to get go to work
YoDawg has quit [Quit: Connection closed]
Jackneill has joined #nixos-chat
Jackneill has quit [Ping timeout: 276 seconds]
Jackneill has joined #nixos-chat
Jackneill has quit [Ping timeout: 260 seconds]
Jackneill has joined #nixos-chat
Jackneill has quit [Max SendQ exceeded]
Jackneill has joined #nixos-chat
Jackneill has quit [Max SendQ exceeded]
Jackneill has joined #nixos-chat
Jackneill has quit [Ping timeout: 260 seconds]
<eyJhb>
Is there any tools that scan comments for `merge` and e.g. a link to a PR. That would be cool to automatically detect when something should be changed /fixed acording to the PR
Jackneill has joined #nixos-chat
Jackneill has quit [Max SendQ exceeded]
Jackneill has joined #nixos-chat
Jackneill has quit [Read error: Connection reset by peer]
Jackneill has joined #nixos-chat
<lovesegfault>
any `bup` users around?
slack1256 has joined #nixos-chat
Jackneill has quit [Ping timeout: 264 seconds]
slack1256 has quit [Remote host closed the connection]
<samueldr>
I've had seen a lot of places stating there are no downsides to them because of special-cased code paths
<samueldr>
ugh, looks like I'm having to review my usual spiel about swap files to: don't use them, linus torvalds states they're all weird and shouldn't really be used :((((
<gchristensen>
or do use them and force upstream to test them more =)
<samueldr>
>> because swapfiles just aren't normal
<samueldr>
that's an unfortunate phrasing, at best
<gchristensen>
not a profanity to be seen, wow
Hurttila has quit [Ping timeout: 240 seconds]
<samueldr>
>> Tags, unlike branch names, are a global namespace in git.
<samueldr>
I think it's time the linux kernel developers start working on their own source code versioning tool
<samueldr>
which works for their workflow
<gchristensen>
lmao
<samueldr>
>> Ubuntu's default install process creates a swapfile in the root directory. Apparently it is not a "normal" distribution.
<samueldr>
(comment from LWN)
<samueldr>
yay
<samueldr>
well I'll be... the "new" issue I was tracking through a bisect is actually the same issue I already tracked earlier on!
<joepie91>
gchristensen: Torvalds seems to have stuck to their promise
<samueldr>
(but actually it's the same root cause, but not the actual same surface issue, and both have to be solved differently)
slack1256 has joined #nixos-chat
<abathur>
samueldr: I enjoy searching/reading for a few hours only to dead-end at an issue or SO thread I opened and forgot about
<sterni>
samueldr: lennart poettering saved me from this situation by virtue of not supporting swapfiles apparently
<samueldr>
huh?
<abathur>
kinda like picking up a book and not realizing I've already read it for several chapters
<samueldr>
abathur: yeah
<ashkitten>
"swapfiles just aren't normal" i think is just literally saying that they are not the norm
<samueldr>
ashkitten: unfortunate phrasing still
<ashkitten>
meh
<samueldr>
ashkitten: and some distros and evangelists have been pushing swap files instead of partitions
<ashkitten>
i use a swapfile cuz repartitioning sucks
<samueldr>
(e.g. me)
<ashkitten>
so i just have an efi partition and a root partition on my ssd
<samueldr>
same~ish
<samueldr>
for me it's also because it's enrypted by LUKS for free
<ashkitten>
i wonder if it'd be possible to write a filesystem that dynamically allocates and deallocates space from a lvm pool
<samueldr>
LVM pools can be oversubscribed
<ashkitten>
oh huh
<samueldr>
I don't know what this cursed knowledge gets us
<samueldr>
but IIRC I was looking at it recently and was searching this specifically
<joepie91>
this feels like one of those "that was better left unmentioned" sort of things
<joepie91>
:P
<ashkitten>
so you could have a filesystem that takes up the entire disk and as long as you use discard and don't fill it all the way you could have a swap filesystem next to it?
<samueldr>
I'll be honest
<samueldr>
just that, I'll be honest
<samueldr>
ashkitten: that's something I had trouble finding confirmation about when I was looking
<samueldr>
but it felt like it
<ashkitten>
unfortunately i'd have to repartition to be able to do that so oh well
<ashkitten>
samueldr: ky0ko tells me that lvm will just spit out write errors if you try to actually write more than there is space for on the disk, but otherwise it works
<samueldr>
that's basically what I assumed
<samueldr>
because you can also AFAIUI provision it live with more disks
<samueldr>
so it kind of makes sense?
<samueldr>
still, thinking about it is a curse
<ashkitten>
hehe
evanjs has quit [Read error: Connection reset by peer]
evanjs has joined #nixos-chat
<ky0ko>
yeah, confirmed from manpages - all writes error when full
<ky0ko>
which can result in some fantastic data corruption
<samueldr>
fun!
<samueldr>
so yeah... not touching that
<samueldr>
considering even on a bog standard filesystem, I hit 100% filesystem usage and things go wrong in bad software badly written
<samueldr>
(hint: most)
<samueldr>
I would *definitely* hose my system monthly with an oversubscribed LVM
<ky0ko>
iirc in some cases on btrfs running out of space is an unrecoverable problem :D
<samueldr>
I said _bog standard_, but _bug standard_ ;)
<samueldr>
(tbf, I have nothing against btrfs)
<ky0ko>
i like it in theory
<ky0ko>
in practice it's eaten my data too much to be comfortable
<samueldr>
I only said that for the pun :)
<ky0ko>
i imagine it has gotten more stable since my usage of it in ~2015-2017 though
ky0ko has quit [Remote host closed the connection]
ky0ko has joined #nixos-chat
tomberek has quit [Quit: Connection closed]
<infinisil>
Argh, why does wireguard not automatically connect anymore
<infinisil>
I need to manually ssh from the client to the server for the connection to be initiated
<cole-h>
what are you trying to do?
<infinisil>
Wait, but znapzend does that anyways
<cole-h>
in my experience, wireguard only establishes a connection when it needs to
<infinisil>
One of my servers backups always breaks because of ssh: connect to host 10.99.2.2 port 22: Connection timed out
<cole-h>
so you'd have to do something to prod it, unless you're forwarding literally all of your interwebs
<infinisil>
The server is the wireguard server at 10.99.2.1
<cole-h>
infinisil: idk if it would do anything, but have you tried setting a persistent keepalive?
<infinisil>
Wants to back up to a wireguard client
<infinisil>
Yeah
<infinisil>
Hmm..
<infinisil>
Do I need persistentKeepalive on both server and client side?
<infinisil>
Because I only have it for the client
<infinisil>
"latest handshake: 1 day, 46 minutes, 55 seconds ago" on the server :/
<infinisil>
And "persistent keepalive: every 25 seconds" on the client
<infinisil>
Something's not right!
<LinuxHackerman>
Keepalives don't always cause handshakes.
<LinuxHackerman>
But yeah if the latest handshake is that far back there indeed probably something not right :)
<infinisil>
And wireguard doesn't have an issue tracker I could just look through..
<infinisil>
I would've loved to blame znapzend again for breaking my backups, but this time it looks like it's wireguard
<LinuxHackerman>
The classic on that is that you were in an ipv6-capable network when the tunnel was set up, so it used an ipv6 endpoint, then you moved to an ipv4-only network and whoops no more route for you.
<LinuxHackerman>
(if your endpoint is set as a DNS name)
Jackneill has joined #nixos-chat
<infinisil>
Hm no it's all ipv4, and all stationary
<infinisil>
Gonna try enabling debug logs, which is apparently done with `echo 'module wireguard +p' | sudo tee /sys/kernel/debug/dynamic_debug/control`
<infinisil>
As I discovered from a random internet page