<pie_>
Amazingly, the x-hack could be used to work around certain bugs all the way up until 2015, seven years after StackOverflow wrote it off as an archaic relic of the past!
<pie_>
The bugs are of course increasingly hard to come across. The Zsh one only triggers when comparing left-paren against right-paren, as otherwise the parser will backtrack and figure it out.
<pie_>
[...]
<pie_>
In all cases, the x-hack is effective because it prevents the operands from being recognized as parentheses.
<eyJhb>
supersandro2000: Ended up with freeing 200 GB
<eyJhb>
Wonder if it should be more than that.
endformationage has quit [Ping timeout: 265 seconds]
<ashkitten>
:( this is the only level 2 doesnt work in ruffle
<ashkitten>
i'm pretty sure i know exactly why
<ashkitten>
there's a level where you have to open a second tab with the game to flip a switch and open the door in the first tab
cole-h has quit [Ping timeout: 252 seconds]
__monty__ has joined #nixos-chat
<arcnmx>
would anyone be able to kick me from #nixos?
<Taneb>
That's an odd request
<eyJhb>
Taneb: don't kink shame \s
<__monty__>
arcnmx: If this is because you can't renick because of a previous log in you can use nickserv's ghost command.
kini has quit [Remote host closed the connection]
kini has joined #nixos-chat
<hexa->
likely because matrix-appservice-irc.
* arcnmx
nod
<arcnmx>
hoping it might help the server or bridge realise I'm not supposed to be in the channel .-.
kini has quit [Remote host closed the connection]
kini has joined #nixos-chat
endformationage has joined #nixos-chat
rj has joined #nixos-chat
kini has quit [Remote host closed the connection]
kini has joined #nixos-chat
AkechiShiro has quit [Quit: WeeChat 2.9]
rj has quit [Ping timeout: 240 seconds]
MichaelRaskin has quit [Ping timeout: 240 seconds]
rj has joined #nixos-chat
rj has quit [Remote host closed the connection]
rj has joined #nixos-chat
MichaelRaskin has joined #nixos-chat
AkechiShiro has joined #nixos-chat
<eyJhb>
Debugging why the hell wireguard won't work, and you forget that networking.firewall.enable is true by default.
<eyJhb>
Damn it
waleee-cl has joined #nixos-chat
<Church->
heheheh
<Church->
Yeah that'll getcha
<gchristensen>
unless you set up both peers with direct IPs and persistent keepalive
<eyJhb>
Church-: And you need to reboot, if you want it to be gone. :/
<eyJhb>
gchristensen: Well, they still won't be able to connect if you don't open the UDP port :D
<clever>
eyJhb: outgoing udp packets punch a hole in the firewall for the reply
<clever>
and udp is connectionless, so if the 2 ends agree on ports ahead of time, they can talk thru those holes
<gchristensen>
yep
<gchristensen>
as long as both sides are trying to talk to each other at the same time, the firewall opens it up
<Church->
eyJhb: Really? Bah
<Church->
Okay there is no reason this should coredump
<Church->
I could just run it on a cron
<Church->
But I am stubbornly going to get this systemd service working and upstrea it.
<pie_>
eyJhb: did that yesterday with "why arent the dhcp packets hitting dnsmasq"
<pie_>
eyJhb: to my defence, the containers where i was doing stuff before werent firewalled
kini has quit [Remote host closed the connection]
kini has joined #nixos-chat
kini has quit [Remote host closed the connection]
kini has joined #nixos-chat
h0m1 has joined #nixos-chat
cole-h has joined #nixos-chat
kini has quit [Remote host closed the connection]
kini has joined #nixos-chat
Dotz0cat_ has quit [Ping timeout: 240 seconds]
<eyJhb>
Church-: iptables is... Hell.
<eyJhb>
pie_: WELL! That is the same for me. Or.. Something along those lines :p
<Church->
Heh it can be
<MichaelRaskin>
gchristensen: I thought that with proper abuse of TCP simultaneous-syn you could do the same with TCP, it is just not done (and VoIP wanted UDP in the first place)?
<gchristensen>
I don't think so, because TCP sends back "uhh.. no" packets
<cransom>
i would think that would make stateful firewalls very upset.
<gchristensen>
which?
<cransom>
trying to do a simultaneous syn thing.
<gchristensen>
oh
<MichaelRaskin>
Collision-course-SYN
<gchristensen>
OH
<MichaelRaskin>
Because of course it is legal and of course firewalls do not care about standard
<sphalerite>
Forgive me, Father, for I have synned.
<cole-h>
sphalerite++
<{^_^}>
cole-h: 16 hours, 43 minutes ago <elvishjerricco> ping
<{^_^}>
sphalerite's karma got increased to 122
<cole-h>
elvishjerricco: pong
<cole-h>
elvishjerricco: if we miss eachother again, feel free to dm on twitter
rajivr has quit [Quit: Connection closed for inactivity]
<elvishjerricco>
And I'm pretty sure I reattach the GPU correctly too
<elvishjerricco>
Considering the display manager comes up fine
<cole-h>
but you just can't switch to e.g. tty1, tty2, ...?
<elvishjerricco>
cole-h: I can switch to a blank black screen with ctl+alt+f1
<cole-h>
and you can get an actual tty like that if you don't boot into the vm first?
<cole-h>
(e.g. no black screen, but getty splash screen)
<elvishjerricco>
Yea I can get a tty with getty via ctl+alt+f1 if I haven't used the vm this boot
<cole-h>
hmmmmmmm
<cole-h>
I don't have that issue
<cole-h>
do you restart the display manager?
<elvishjerricco>
Yes I do in the ExecStopPost of the service that starts the vm
<cole-h>
what if you manually restart it?
<cole-h>
I'd also check `systemctl status getty@tty1.service` to see if there's something fscked up
<elvishjerricco>
cole-h: Hm not sure why manually staring the display manager would work. Isn't that independent of the Getty ttys?
<cole-h>
no idea -- I don't use a display manager :P
<cole-h>
just trying to explore
__monty__ has quit [Quit: leaving]
<elvishjerricco>
I know no services are failed when I do this (my shell tells me any failed ZFS pools or systemd units when I log in)
<elvishjerricco>
You don't use a display manager?
<cole-h>
no
Dotz0cat has joined #nixos-chat
<pie_>
joepie91: UX circlejerk: if people spent 5 minutes to an hour actually thining about CLIs from user perspective...
<joepie91>
pie_: unfortunately there seems to be a heavy bias in UX-blind people working on all the CLI stuff
<pie_>
blind or just did not spend 5 minutes actually trying to design soething
<joepie91>
pie_: UX-blind in the sense that they do not recognize the importance of UX as a thing
<pie_>
im probably the asshole here but
* pie_
victim shames himself :P
<pie_>
ok </ux circlejerk> thanks
<joepie91>
lol
<samueldr>
but ugh
<samueldr>
CLI don't have UX
<samueldr>
duh, there is no GUI for the UX
* samueldr
smears the /s heavily
<joepie91>
samueldr: without the /s you sound just like every other terminal fetishist :P
<samueldr>
I know, my imitation is perfect
<samueldr>
and yes, please, think more than 5 minutes about the experience of the CLI tools you build!
<samueldr>
and there's probably (sadly) no one-size-fits-all answer here
<samueldr>
but really, an important thing I guess would be: EXPLAIN
<samueldr>
explain why the thing failed
<samueldr>
-EBADTHING is not helpful
<pie_>
i mean, i can accept that the design space is big and hard and sometimes the problem domain is just complicated
<pie_>
ok so here's the actual thing that im wastin way too much time being angry about
<samueldr>
if you have an error message, write "what" exactly in the error message
<pie_>
theres this thing in networking where apparently you can make differnt routing tables
<samueldr>
and try to make unique error messages so grepping your source doesn't reveal multiple location with the same error message
<joepie91>
if the design space is too big for you to deal with, then err on the side of information vomit, not on the side of "shit's fucked lol"
<pie_>
and they can have names, but the kernel only actually sees id numbers
<pie_>
which is fine i guess
<pie_>
but ip lets you use names
<joepie91>
people can extract meaning from information vomit with sufficient effort
<joepie91>
a generic "shit's fucked" however is a guaranteed dead end
<pie_>
except the default names are hardcoded and theres no interface for adding names, you have to add stuff to some config file
<pie_>
its just an id - name mapping
<pie_>
the file doesnt exist by default
<pie_>
why couldnt they just add an ip rule create table <name> command so i dont have to figure any of this out
<pie_>
AITA?
<samueldr>
I don't know about your specific case
<samueldr>
but at face value it sounds right
<pie_>
it would have been "add a named routing table"
<pie_>
but $person told me they just use the id number and i guess ehh why not
<MichaelRaskin>
samueldr: if -EBADTHING is the same as syscall returned and you print it right afterwards, then it is more helpful than what people thinking about the topic come up with
<MichaelRaskin>
Because -EBADTHING I will search in strace log.
<MichaelRaskin>
And error messages are not always better than strace output
<pie_>
so say "Syscall returned EBADTHING"
<samueldr>
it can say things
<MichaelRaskin>
That's surely better
<samueldr>
like what it was trying to do when it failed
<pie_>
though tbh ididnt think of that till you brought it up
<samueldr>
*and* transmit that error
<samueldr>
and that's assuming here it's about syscalls
<samueldr>
but I've seen tools reuse those error codes
<samueldr>
for their own uses
<samueldr>
so that wouldn't hold water here
<MichaelRaskin>
Yeah, that's bad
<MichaelRaskin>
Also requires thinking
<MichaelRaskin>
So maybe thinking for 5 minutes about what the hell you are doing is not always making things better