<elvishjerricco>
Maybe uh oh... I've got that new "magsafe" wireless charger for iPhone 12, and I apparently let an old, presumably dead, AAA battery get attracted to the magnet. They've probably been in contact for a little less than a day, and now there's a bunch of corrosion on that battery.
<elvishjerricco>
I'm *guessing* this hasn't damaged the charger, but I do not know how these things work. The charger was slightly warm when I decoupled them, but otherwise seemed fine, and still at least charges the phone
<samueldr>
yikes
<gchristensen>
I thought they did a handshake ... weird :/
<samueldr>
but if it's Apple-branded Qi, it's double-yikes!
<joepie91>
things are going really great here in NL
<joepie91>
rioters tearing apart cities
supersandro2000 has quit [Disconnected by services]
supersandro2000 has joined #nixos-chat
<elvishjerricco>
samueldr: magsafe is a 7W Qi compatible wireless charger that also has proprietary apple 15W wireless charging. The unique feature is that it has builtin magnets to align the charger with an iphone 12 correctly, because the phone also has magnets. Solves two things really nicely: You can hold the phone while it wirelessly charges, and you don't have to worry about whether you aligned it properly.
<samueldr>
yeah
<elvishjerricco>
And yea it (poorly) recycles the old magsafe name from macbooks
<elvishjerricco>
It's not really a safety feature so I dunno why they called it that
<samueldr>
I was thinking it was those "magsafe" (not magsafe) bits you pop in the connector and use a spooky usb cable to plug into
<elvishjerricco>
I will say, for all its problems, this charger is really quite nice.
<elvishjerricco>
I just leave it on my desk, so my phone ends up charging most of the time I'm at the desk. It's really easy to attach and detach. And it charges reasonably fast
<elvishjerricco>
Of course this is my first wireless charger so maybe I'd feel the same way about android qi charging, but I do really like the magnets
srk has quit [Ping timeout: 268 seconds]
srk has joined #nixos-chat
rajivr has joined #nixos-chat
cole-h has quit [Ping timeout: 246 seconds]
<siraben>
Does anyone know of a good plain text email client for iOS?
<ashkitten>
i wonder if there's an implementation of xdg-desktop-portal screencast for x11
<ashkitten>
not that it's needed, but maybe in the future programs will only need to support that
<ashkitten>
it kinda interests me to have sort of a backwards wayland compatibility layer, which pretends to be a wayland compositor for an x11 server. that way new things will only need to support wayland.
<ashkitten>
and it could theoretically work better than xwayland does, since wayland is more restrictive
lunc has joined #nixos-chat
<ashkitten>
,locate libspa-0.2.pc
<{^_^}>
Found in packages: pipewire.dev
endformationage has quit [Quit: WeeChat 2.9]
spudly1 has quit [Ping timeout: 265 seconds]
spudly1 has joined #nixos-chat
waleee-cl has quit [Quit: Connection closed for inactivity]
cole-h has joined #nixos-chat
kalbasit_ has quit [Ping timeout: 265 seconds]
cole-h has quit [Ping timeout: 244 seconds]
<philipp[m]>
I miss my palm pre and it's magnetic wireless charger :(
<adisbladis>
philipp[m]: Heh :) I got one of those and modded my phone at the time to get wireless charging
<philipp[m]>
I'm looking forward to the day when android finally manages to copy it's multitasking completely.
<jD91mZM2>
I just killed my DigitalOcean droplet by accident, oops
<LinuxHackerman>
was it an important one?
<jD91mZM2>
Tried to use morph on it and selected the harddrive /dev/vda1 which for some reason didn't want to boot
<jD91mZM2>
Well, it hosted my email server but I was gonna move that anyway
<jD91mZM2>
Also my password manager... I have the /var/lib backed up so I think I can recover
<__monty__>
jD91mZM2: Meant to ask you, why do you want to migrate from Bitwarden to pass?
<jD91mZM2>
__monty__: I don't, actually. But there's a voice inside me telling me I should do it because all the cool kids use pass lol
<jD91mZM2>
Pass is cool because it works with everything and feels more minimal. But I think it's just not convenient enough for me yet, and I have trust issues with my gpg keys for some reason.
<__monty__>
Which things does pass work with that Bitwarden doesn't?
<jD91mZM2>
I... uuhhh... No idea
<jD91mZM2>
Oh shit, my bitwarden chrome extension just logged me out
<philipp[m]1>
Does somebody here have some knowledge about the linux IP stack and can tell me how exactly linux decides which IP it receives from a AAAA request it uses?
<jD91mZM2>
If my data inside the downloaded /var/lib is for some reason insufficient, I am dead
<philipp[m]1>
I have heard rumors about it prioritising IPs that are next to it in terms of subnets but have found nothing concrete.
* __monty__
crosses their fingers
<pie_>
jD91mZM2: yeahhhhhh back up your password managers
<jD91mZM2>
btw __monty__ that's one reason, in pass (and KeePassXC, previous favorite of mine), you have more transparacy over your files
<pie_>
jD91mZM2: preferably dont require the password db encrypted on the drive, to decrypt your drive, when you forget the drive password
<jD91mZM2>
pie_: I did, but to access the backup I need to login to services. I managed to do that before it died, just hoping my backup was enough
<pie_>
I did that at work once.
<jD91mZM2>
IT WAS ENOUGH
<jD91mZM2>
YESSSS
<adisbladis>
jD91mZM2: Pass has a pretty sweet git integration
<jD91mZM2>
Currently I'm not encrypting my borg backup of /var/lib, I noticed. I guess the pw manager should have all files encrypted anyway
<etu>
jD91mZM2: Now migrate it to pass and sync that git repo to all the places :)
<etu>
jD91mZM2: You don't need the private key everywhere of course :)
<adisbladis>
I wish pass didn't leak as much metadata as it does
<jD91mZM2>
I need to access a subset of my passwords on a stupid windows work computer
<__monty__>
This sounds like fanboyism to me, just fyi.
<adisbladis>
I guess maybe you can circumvent this with git-crypt?
<adisbladis>
But...
<adisbladis>
That feels pretty dumb
<__monty__>
adisbladis: There's also pass-tomb.
<etu>
adisbladis: Sure, but I still prefer that over... some obscure lockin-database format where it's a single file.
<__monty__>
etu: Bitwarden doesn't lock you in. There's plenty of export formats.
<adisbladis>
etu: Sure, but I think a slightly more complicated file scheme is better than leaking this much metadata
<adisbladis>
__monty__: That looks pretty OK-ish
<adisbladis>
Tbh I should probably write my own pass implementation
<etu>
__monty__: Possibly, have been using pass for like 9 years, so yeah... Not planning to move away :)
<adisbladis>
Like.. Why can't I use my phones enclave as a key and get a push notification there to unlock a password?
<adisbladis>
I like "pass the interface", I'm not a huge fan of pass itself
<__monty__>
The files v. database debate (unrelated to password managers specifically) I don't really understand. Both have disadvantages.
<adisbladis>
As long as the schema is open it's whatever
<adisbladis>
It doesn't matter
<adisbladis>
Imo
<__monty__>
Exactly.
<infinisil>
adisbladis: The advantage of pass's exposed metadata is that you don't need to enter the password just for searching what passwords you have
<__monty__>
infinisil: That's literally also the disadvantage though : )
<infinisil>
I guess :)
<etu>
Yes, it's a matter of opinion if that's a good or bad thing
<adisbladis>
infinisil: You could have files named `hash(read(".gpg-id"), "user@host")` and still get the same nice UX
<adisbladis>
I guess?
<adisbladis>
Hm, probably not the nice browserpass and completion interfaces
<__monty__>
With some veneer over another password manager you could have it sync all the password names or whatever to another password vault with a super weak easy to type/remember password, even one hardcoded in the veneer : )
<etu>
adisbladis: No fuzzy matching that way
<adisbladis>
Lol, this feels like crowd sourcing a password manager design
<adisbladis>
I'll write my own password manager, with blackjack & ....
<__monty__>
Maybe you can back it with Vault so you can have it do fancy ephemeral ssh key issuance and stuff.
<adisbladis>
I'd probably use crypto primitives instead
<__monty__>
For ephemeral ssh keys?
<adisbladis>
I would like to be able to use my phone as an internet connected smartcard
<__monty__>
But you need PKI.
<adisbladis>
Not really
<__monty__>
Distributing a key to check signatures with counts as PKI.
<adisbladis>
I wonder what I even want
<adisbladis>
I want something that doesn't require gpg at least
<etu>
hmm
<__monty__>
But sequoia's gonna save us all!
<talyz>
etu: KeePass' db may be obscure, but it's definitely doesn't lock you in - it has multiple implementations and allows you to export your database to other formats :)
<adisbladis>
etu: And honestly, git is a pain on mobile
<etu>
adisbladis: Maybe it should be possible to decrypt just the metadata to allow fuzzy matching and that that is one event that has to be "approved" somehow, using a password or token somehow.
<adisbladis>
etu: But at that point you might as well unlock the entire db
<etu>
adisbladis: I use the password store app and do "git pull" to get updates there, it's a menu option.
<jD91mZM2>
Oooh, tomb looks really nice though
<adisbladis>
As you're already having access to the key material
<adisbladis>
etu: But I don't even feel comfortable with having ssh keys to pull git on my phone
<adisbladis>
Which is yet another key not in an enclave
<__monty__>
jD91mZM2: Tomb's cool.
<adisbladis>
etu: I think something eventually consistent is a far better fit than git is
<__monty__>
adisbladis: And gpg-as-ssh keys aren't an option?
<adisbladis>
__monty__: Not on android
<adisbladis>
Android friggin sucks
<adisbladis>
I've gone down this rabbit hole before on android
* infinisil
is still waiting for the Librem 5
<siraben>
infinisil: you ordered one?
<siraben>
all the pinephone activity is interesting, also NixOS mobile
<infinisil>
Yup, should arrive soonish!
<__monty__>
Doesn't git's history make compromise easier? Let's say you change a password because it was leaked. Whoever has that password now has a known-plaintext to brute-force your private key on because git keeps history.
<__monty__>
infinisil: Looks like a chonker : ) Any project ideas in mind or do you intend to daily drive it?
<etu>
infinisil: Any year now?
<etu>
infinisil: Were you in the first crowdfounder?
<infinisil>
Not the first batch, but not long after it
<infinisil>
I ordered it over 3 years ago, I think October 2017
<infinisil>
__monty__: Just daily driver, my current phose is slowly falling apart :)
<infinisil>
Having NixOS on it would be awesome though
<etu>
Just the dock itself is worth the extra money
<etu>
Then you get extra ram as well iirc, maybe extra emmc as well?
jD91mZM2 has joined #nixos-chat
<etu>
siraben: I've tried with a thinkpad to display image, worked there, so yeah, it can be used for other things than the phone :)
<siraben>
Oh cool, how much do docks like that usually cost?
<siraben>
Are you satisfied with the PinePhone?
<talyz>
siraben: I bought the convergence package since you get 1GB extra ram and twice the emmc :)
<siraben>
hehe, convergence it is then
<talyz>
the dock is a nice extra, though
<etu>
siraben: I have a braveheart edition, I tried it very early on, it wasn't usuable at all at that point. But I haven't had time to play with it at all since then and I know they have done a lot of progress on many fronts.
jD91mZM2 has quit [Ping timeout: 264 seconds]
<siraben>
What edition is it now, Mobian?
<gchristensen>
is there a way to jump to a section of a man page? specifically, bash's SHELL BUILTIN COMMANDS
<elvishjerricco>
gchristensen: I just use forward slash like a noob
<gchristensen>
same, and they mention it like a dozen times making it super annoying :D
* etu
thinks there's no better way...
<etu>
Since man opens it in the pager, which is less
<etu>
By default
<etu>
So the searching is done by less, which really doesn't have any tie in to man
<gchristensen>
quality software
<gchristensen>
I get the elegance of reusing less but c'mon
<etu>
One could use another pager
<elvishjerricco>
Man pages are kind of annoying
<etu>
gchristensen: env PAGER=emacs man man
<elvishjerricco>
They always have references to other pages but there's no way to actually jump to them
<__monty__>
gchristensen: Maybe open it in info?
<__monty__>
I share the pain though.
<etu>
and then emacs doesn't read stdin
<elvishjerricco>
`DOCUMENTATION=foobar(6)` but how do I look at foobar(6)?
<gchristensen>
elvishjerricco: you man up! put on your plaid suit jacket and jump between pages like your father in the 70s!
<elvishjerricco>
Lol
<etu>
gchristensen: Now you make me wish I had a better pager that could get the context of man pages to navigate it better :p
<__monty__>
Still sounds like info.
<__monty__>
I've been down this rabbit hole before. Haven't been able to find a better answer than info.
<etu>
__monty__: Does info also read man files or are they provided as separate files for packages?
<etu>
Just by doing some comparisons I find the man pages to be much more useful... Say "man ls" vs "info ls". The synopsis and the description of the flags are upfront in the manpage. Info tells me more what ls is for, which I already know.
<etu>
info may be a better "browser" and "reader"
<__monty__>
Info falls back to man pages if there's no info page.
<etu>
That's good :)
<__monty__>
`info ls` gives me the exact same content as `man ls` for example.
<__monty__>
Since info pages can be provided separately I guess there's no guarantee they're formatted as nicely. Though they could also be formatted *more* nicely of course.
cole-h has joined #nixos-chat
<V>
supersandro2000: sure, IRC sucks
<supersandro2000>
but everything else sucks, too but different
<V>
yup
jD91mZM2 has joined #nixos-chat
<jD91mZM2>
I just killed my server again, but this time it's on a different host that let me upload the NixOS recovery iso and fix it
<jD91mZM2>
phew
<adisbladis>
[irc sucks]: Huh, what's the context?
<V>
adisbladis: -dev
<V>
was getting a bit off-topic
<adisbladis>
V: Thanks
<adisbladis>
<supersandro2000> [15:55:26] IRC can't even do multi line messages
<adisbladis>
That's a feature
<adisbladis>
<supersandro2000> [15:55:29] or pics
<cole-h>
tbh agree
<gchristensen>
it is? :P
<adisbladis>
Sure it can, that's a client concern
<V>
tell that to IRCv3
<cole-h>
there's a balance though
<adisbladis>
gchristensen: I think so, but I can see why others don't
<cole-h>
ikd how to express eloquently, but I like that IRC (when familiar with its limits) forces users to make their message more concise
<adisbladis>
Having multi line changes the conversation style I think
<cole-h>
s/ikd/idk/
<V>
It impresses me how much people are willing to defend a deficiency in something that only exists due to how old it is
<adisbladis>
cole-h: Exactly this
<cole-h>
I know I definitely ramble a bit in e.g. GitHub issues when I'm writing multiple paragraphs
<cole-h>
Not to say I don't ramble here, but it's a different kind of rambling
<adisbladis>
But either way multi line or not is not a big deal
hexa- has quit [Ping timeout: 240 seconds]
<adisbladis>
And pictures is _not_ a protocol concern at all
<adisbladis>
There are irc clients that show pictures inline
<adisbladis>
I have used erc like that
<adisbladis>
It was horrible :P
<V>
OTOH they also provide zero means of uploading images
<V>
none of this is integrated
<adisbladis>
Hmm, valid point :)
<gchristensen>
I find it sort of annoying that `foreman` doesn't have a better way to run a subset of services than just commenting out the lines:
<adisbladis>
gchristensen: What's your desired UX?
<cole-h>
"Not suck"
<cole-h>
:D
<adisbladis>
I meant the filtering in particular, but sure :P
<adisbladis>
I'm tempted to write a patch, because I've wanted the same before
<betawaffle>
a rust security fob? sounds neat
<adisbladis>
It's pretty nice to just reference something in the procfile in ci
<gchristensen>
`foreman start foo bar baz` starts the foo: bar: and baz: lines
<adisbladis>
Should be simple enough :)
<gchristensen>
neat :D
<cole-h>
What about the "foo bar:" line
<cole-h>
/s
<adisbladis>
cole-h: Just quote the argument as "foo bar"
<cole-h>
waow
<cole-h>
👏
<cole-h>
(hehe, sorry)
<adisbladis>
10/10 trolling
<gchristensen>
I guess y'all really don't see multiline messages as a feature
<adisbladis>
Would be trolled again
<cole-h>
Thanks, I tried my best
<cole-h>
I'm neutral on it, in all honestly. Them being implemented (properly) wouldn't make me irrationally angry, but I'm also perfectly content with "the way things are"
<adisbladis>
this ^
<__monty__>
gchristensen: I'm more afraid of the spam potential.
<__monty__>
Images would be even worse.
<__monty__>
That's why I said I consider their lack a feature for large communities.
<joepie91>
__monty__: image spam is a minor annoyance but fairly easily counteracted by the ability to remove spam
<adisbladis>
gchristensen: I take it __monty__ isn't talking about you
<adisbladis>
But about multi line
<adisbladis>
Tbh I think the entire Nix community should move over to Matrix with IRC bridging
<adisbladis>
With all that such an effort would entail
<V>
bait
<adisbladis>
V: I'm serious
<adisbladis>
I think a lot of the community fragmentation is because of IRCs bad UX
<adisbladis>
I don't think there would be an unofficial Discord if we were using Matrix
<adisbladis>
For example
<V>
there would
* joepie91
loosely agrees
<adisbladis>
And Matrix much more closely aligns with open source values
<adisbladis>
joepie91: I would absolutely love your take on this
<V>
there's a discord for almost all the major distros IIRC
<bbigras>
cole-h: thanks! damn those keys are going fast
<betawaffle>
cole-h: I just did the 4x solo v2
<adisbladis>
Since you have experiences with similar things
<joepie91>
I do think there are still remaining UX issues on the Matrix side, mainly Synapse performance is a big one (and E2EE unreliability, though that would not be an issue here, really)
<cole-h>
bbigras: And earlier, it was only 100 units (of the 1x "super chear")
<adisbladis>
Those solos look interesting... Too bad I'm homeless and have no shipping address
<bbigras>
oh nice
<joepie91>
but overall.... I still definitely consider it a UX upgrade
<V>
people will always want a community on the platform they're most comfortable/embedded into, and there are people for whom discord is that platform
<eyJhb>
adisbladis: I can throw it and see if I hit egypt at some point :p
<cole-h>
I really, really, really hope it hits the 64k stretch goal
<cole-h>
"Reversible USB-A. Yes, that you can plug it both ways and it works! We designed it already, just need the numbers to make it."
<adisbladis>
eyJhb: lol
<joepie91>
cole-h: I've had those cables
<joepie91>
Blitzwolf sells them
<eyJhb>
I wonder how well it will work, if I am going to move in a couple of months? Can you just change your order details?
<joepie91>
both reversible A and micro B
<cole-h>
joepie91: The thing I'm more interested in is how Conor created this one himself
<joepie91>
adisbladis: that having been said... a) it's probably a good idea to wait for Spaces to land before moving over a multi-channel community, so that they can actually be meaningfully grouped together, and b) one thing to consider is that it will *slightly* worsen the experience for IRC users if the Matrix room were to be the canonical room; for example, people are more likely to use code blocks (which get bridged as
<joepie91>
markup-less text pastes) rather than syntax-highlighted pastebins, and they are more likely to use replies, which currently bridge poorly
<eyJhb>
^ If one wants to see what a Matrix room looks like from IRC, just join any #matrix room on Freenode. It is hell with all the text links
<eyJhb>
But still nice
<joepie91>
I don't think those things would be significant roadblocks (especially if a fix can get merged for the reply bridging), but they are things to keep in mind
<eyJhb>
But if it is selfhosted joepie91 , that could be changed?
<joepie91>
eyJhb: if the bridge is self-hosted, yes, you can change the reply format - however, you will then have to arrange IP limit exceptions with freenode
<joepie91>
and they will expect you to set up a weird ipv6 thing where each Matrix-side user gets their own IPv6
<eyJhb>
Well okay, not it
<f0x>
nah
<eyJhb>
BUt I assume that is for LARGE communities
<eyJhb>
Not like Pixietown
<f0x>
they don't require that, it's just nicer anti-abuse wise
<Peetz0r>
<joepie91 "I don't think those things would"> Yes, the reply formatting is especially ugly. Look at this!
<joepie91>
f0x: oh, I thought it was a requirement
<eyJhb>
f0x: You are always silently watching, aren't you? :p
<V>
adisbladis: also, remember how mozirc went. there were still stragglers
<Peetz0r>
Hey, is this #revspace-matrix or what? :p
<f0x>
it might've been oftc that did require it but my memory is a sieve
<f0x>
and all the bridges have the setup now anyways hah
<eyJhb>
joepie91: Do you know when spaces will land?
<Peetz0r>
soom(tm)
<f0x>
eyJhb: one of the reasons I like modern chat UI is message previews in the room list, so i can see when interesting topics come by :)
<eyJhb>
I cannot read that without thinking of Zoom
<philipp[m]1>
whenitsdone
<joepie91>
eyJhb: "this month", in valve time
<joepie91>
(it will not be this month)
<betawaffle>
what's spaces?
<bbigras>
cole-h: oh they just bumped it to 300. I hope they do this because they are surprised by the success and not because they trick people into buying one impulsively like I did.
<eyJhb>
As long at it's not Half-life 3 time, it's OK
<cole-h>
:P
<joepie91>
betawaffle: community mechanism, think Discord 'servers' or Slack 'workspaces'
<cole-h>
bbigras: And, if you're interested, the project description says they'll ask if you want a reprogrammable key ("Solo Hacker") at the end of the run
<f0x>
hmm the current "sent a long message" is kinda meh anyways maybe I should write something that properly pastebins/highlights stuff in it
<f0x>
.. next week
<betawaffle>
and nixos is switching once spaces is a thing?
<joepie91>
Peetz0r: the spec is not settled on yet, FTR
<joepie91>
betawaffle: nonono
<betawaffle>
unrelated to nixos?
<joepie91>
I'm just saying that if NixOS were to plan on switching to Matrix, it would be good to wait for that :)
<betawaffle>
ahh
<Peetz0r>
It's going to be way more flexible that Discords stuff btw. Rooms can be in multiple spaces and spaces can be public or private, so you can make categories of channels however you like and such.
jD91mZM2 has joined #nixos-chat
<bbigras>
cole-h: I head about this version before, and I'm looking at the website right now but can't figure out quickly what is does.
<cole-h>
basically, you can reflash it
<__monty__>
Are these Solo keys just open source yubikeys?
<cole-h>
I'm sure there's more to it than that, but I'm no expert
<bbigras>
thanks
<cole-h>
bbigras: So you can (theoretically; never had one to test on) make your key only accept firmware signed by your key, and then play around with it however you want
<cole-h>
"signed by your key" being some other secret key and not the Solo itself :P
<bbigras>
oh. I thought maybe it was for people wanting a custom firmware.
waleee-cl has joined #nixos-chat
<cole-h>
I mean, it is
<cole-h>
Provided you write / build the firmware yourself ;)
<Peetz0r>
About matrix, I just got myself the domain glitch.im :D
<gchristensen>
nice!
<philipp[m]1>
Hahahaha! A project lead pinged me last night in a channel that I just idle in and have nothing to do with.
<Peetz0r>
And a nice hardware upgrade with stuff I already had but didn't realise how much better it is than what I was using before
<gchristensen>
HAH.
<gchristensen>
oops, wrong channel
<joepie91>
Peetz0r: how on earth did you get a domain like that
<Peetz0r>
happened to be, eh, wel, available
<Peetz0r>
paid €13,33 for it (sadly not €13,37)
<philipp[m]1>
They told me that something that looks like it processes the result of their ci pipelines throws 502s now and I just claimed it probably was that recent gitlab update.
<LinuxHackerman>
whaaaat :o
<Peetz0r>
But hey, now I can *own* the "it's a glitch in the matrix" joke
<philipp[m]1>
They actually bought it and are looking into it now :D
<Peetz0r>
Now I need to fork Element, rebrand it with the joke, and host that at glitch.im
<Peetz0r>
and if I also add in my custom css, it might not even be crazy
<Peetz0r>
About the hardware, I basically realised that the RockPro64 is a lot faster than the Celeron N3050 (while probably using less power at the same time)
<joepie91>
is your I/O going to be good enough though
<joepie91>
because I/O is a common bottleneck
<Peetz0r>
It... has an nvme drive on it :p
<Peetz0r>
(a cheap chinese one, but still)
<Peetz0r>
the rockpro64 has an actual pcie slot, and it can boot from nvme (if you flash a recent enough uboot to the SPI flash or SD card)
<Peetz0r>
I might do some disk benchmarks before I install my stuff. Any suggestions?
<LinuxHackerman>
Can matrix IDs contain zalgo?
<f0x>
not ids, but display names can
<Peetz0r>
I think the ids can't even contain uppercase
<Peetz0r>
but yeah, display names can even contain (or be) emoji
<Peetz0r>
(I do hope they filter out stuff like LTR)
<V>
BIDI overrides >:D
<philipp[m]1>
Ooooh boy. The customer doesn't want the URL to 404. They want the URL to not exist.
<gchristensen>
??
<philipp[m]1>
Yes, that's all of us right now.
<philipp[m]1>
I'm considering to send them a copy of "The Physicists".
<cransom>
'the client would like 7 red lines, all perpendicular, some with green ink, and some with transparent ink'
<gchristensen>
lol
<philipp[m]1>
So, good people of #nixos-chat, tell me: what exactly are the necessary and sufficient conditions for the existence of an URL?
<philipp[m]1>
Is it that at least one webserver answers with some http status code that doesn't begin with 4 when asked for it?
<V>
philipp[m]1: that it shows something, presumably
<V>
you could make it such that the 404 just returns a blank page or so
<philipp[m]1>
V: Yeah, that's my suspicion, too. But I'll get to the bottom of this with an open mind.
<cole-h>
Man, readline binds have ruined me
<cole-h>
I keep trying to ^W in my browser, which ends up closing my tab and interrupting my typing :(
<V>
same :(
<philipp[m]1>
Ahaaa! There is still a google index of that page because it used to exist. That is their problem.
<V>
time to purge
<philipp[m]1>
I unironically love my job for this.
<LinuxHackerman>
philipp: I heard that all URLs exist, as long as they're well-formed.
<LinuxHackerman>
Whether they're accessible is another matter.
<LinuxHackerman>
philipp: but what we want is one that _doesn't even exist_
<lukegb>
Just add the URL to Google Safe Browsing /s
<V>
loll
<joepie91>
lukegb: that'll stop working soon
<joepie91>
:P
<joepie91>
(apparently Safe Browsing is among the APIs that Google will cut access to for Chromiums, soon)
<lukegb>
what, even Safe Browsing?
<lukegb>
I mean, Firefox uses that API too
<joepie91>
so I just got a pillow in the mail, that I ordered online
<joepie91>
... the box was full of padding
<joepie91>
I have questions :)
<philipp[m]1>
Linux Hackerman: The minimum I'm willing to go down to is "the URL needs to be well-formed and either been imagined by some human being or exist in some sort of memory device in a recognizable fashion".
<cole-h>
joepie91: inb4the padding is more comfortable than the pillow
<philipp[m]1>
If the Universe is infinite though, we'll have a problem because then all URLs should be encoded ind some sequence of something somewhere...
<joepie91>
cole-h: crumpled up paper, so nope, unfortunately not :P
<joepie91>
but also the pillow is nice
<joepie91>
I suddenly remembered the existence of buckwheat pillows
<hexa->
Sadly we are short on Firefox maintainers these days
das_j has quit [Quit: killed]
<__monty__>
joepie91: Buckwheat's great. I do remember liking spelt similarly though.
das_j has joined #nixos-chat
spacekookie_ is now known as spacekookie
Dotz0cat has joined #nixos-chat
<betawaffle>
philipp[m]1: omg, that's so funny
cole-h has quit [Quit: Goodbye]
cole-h has joined #nixos-chat
buckley310 has quit [Quit: Connection closed for inactivity]
<eyJhb>
I really don't get some designs, e.g. CTFd when you load a challenge, will send you all the information nicely (XHR + JSON), and have a pretty little template that can be used to insert the stuff (HTML...). BUT! Instead, the XHR request also returns a "view" field where everything is just smashed together and will be inserted directly into the modal.
<LinuxHackerman>
eyJhb: performance.
<LinuxHackerman>
why have the client do extra parsing and rendering when it could also not?
rajivr has quit [Quit: Connection closed for inactivity]
<lovesegfault>
rmcgibbo[m]: Posted one last review comment on your code. Looks good to me overall, hope I wasn't overbearing with the reviews!
<lovesegfault>
Was mostly trying to steer you heavily into best practice since you mentioned not having written much Rust before :)
<rmcgibbo[m]>
lovesegfault: hehe, thanks a lot for all of the effort. the end result is drastically improved, and hopefully it will be more robust and maintable for others, especially if it follows the language conventions better. (in my moments of frustration i sometimes think the iterator stuff is a bit code-golf-y, but :P)
<lovesegfault>
rmcgibbo[m]: Rust iterators definitely feel "but why" at first, but IMHO it pays dividends as the code grows in complexity, or as you go to write more complex code elsewhere :D
<lovesegfault>
But figuring out all the one zillion different iter methods when you're getting started is a pita
<lovesegfault>
(the last one can be used outside of an iterator, but if you have iterator-stream code it's a one line change to use it)
<eyJhb>
LinuxHackerman: Seeing is the clients are users and not "my own" resources, then I would rather offload it to them. But atm. it is shitting on both sides, as it wasted bandwidth on both sides :p
<eyJhb>
ALso, now I have to do HTML parsing in BeautifulSoup...
* cole-h
wishes it was possible to delete journald entries for a specific unit only
<joepie91>
eyJhb: do your sanity a favour and use pyquery :)
kalbasit_ has joined #nixos-chat
<cole-h>
Phew, solo V2 reached 64k
<cole-h>
(Reversible USB-A goal)
<bbigras>
I wonder if I'll take a usb-c just to make it more future-proof. but in a sense it's not expensive to get a new one later.
<cole-h>
I don't have a front USB-C, otherwise I'd do the same lol
<joepie91>
local contractor wheeled in a couple of trailers of wooden boards
<joepie91>
and everyone's been boarding everything up all day
<eyJhb>
joepie91: For what? :o
<eyJhb>
HTML parsing in Python?
__Sander__ has joined #nixos-chat
<joepie91>
eyJhb: yeah
<philipp[m]1>
joepie91: Surreal.
<joepie91>
indeed
<bbigras>
cole-h: yeah I don't have any USB-C too. except my phone but the key will have nfc for that.
<cole-h>
Same sitch here.
Guest88372 has quit [Read error: Connection reset by peer]
Guest88372 has joined #nixos-chat
jared-w has quit [Ping timeout: 272 seconds]
jared-w has joined #nixos-chat
<eyJhb>
I want to cry. Blender crashed on me
<eyJhb>
I just wanted to measure a wall
<__monty__>
Blender doesn't sound like the right tool for that job : p
<etu>
:D
<__monty__>
What's your take on the rioting, joepie91?
<joepie91>
__monty__: it's dumb and needs to stop, not really much more to say about it unfortunately
<joepie91>
there's no activism or anything here, it's just rioting for the sake of rioting
<__monty__>
No insight into how/why this started?
<joepie91>
__monty__: a lot is still fuzzy, it seems like curfew protests were used as a flimsy pretense, though they gave up on that by the second night
<joepie91>
it's quite possibly a consequence of people not doing well as a result of the constant lockdown measures dragging on and on
<joepie91>
but the rioting certainly isn't with the intention to stop that
<eyJhb>
__monty__: Well, trying to make a sketch of new apartment + furniture etc. :p So just a nice little tool for it
<samueldr>
since I usually care only about footprint on the floor, I use inkscape
<samueldr>
it's been invaluable in the past to fit theorized but tight layouts
<samueldr>
so I can understand how blender can help
<__monty__>
I don't understand they're targetting businesses and that hospital though. I could understand smashing some windows of a government building but hurting the people who are among those that have been hit the hardest?
<bbigras>
you could use freecad too for 2d
<samueldr>
but that would mean having to learn a new tool!
<LinuxHackerman>
yessss I used inkscape for layout out my old flat as well
<samueldr>
I also used it with measurements on store pages to theorize layouts for stuff I didn't own yet!
<bbigras>
eyJhb: nice. a 3d plan is probably useful for some usecases.
* samueldr
thinks
<samueldr>
I *could* import the house layout from SVG as a blender layer
<bbigras>
is blender easy to use as a cad software? is there a way to show the distances
<bbigras>
like a mesuring tool
<samueldr>
but eh, I don't really want to fight with that new tool
<eyJhb>
bbigras: Yeah there is a measuring tool! Like I noted above, but save before use. It did crash Blender :p
<bbigras>
oh yeah. thaks
<bbigras>
thanks*
<eyJhb>
But I don't think it is "easy" as such, but I have a friend who is it for 3d printing etc. :p
<eyJhb>
But it requires a little extra
<eyJhb>
But 2.9 is SO nice compared to the old versions
<eyJhb>
The main limitation in what I am doing atm. is that the floor plan SUCKS.. I don't have enough measures to actually do it. E.g. there is a staircase, but I cannot place it, as there is no placement for it
<eyJhb>
So time for some guessing :D
tilpner has quit [Remote host closed the connection]
tilpner has joined #nixos-chat
<joepie91>
__monty__: right, that's my point, it's not a protest riot :P it's just smashing things
lunc has joined #nixos-chat
<__monty__>
joepie91: I was mainly wondering about your POV because that's also what the opponents of most riots say. Like when the George Floyd protests developed into looting.
<joepie91>
__monty__: yeah, makes sense. but no, in this case there really is no indication of it being about anything activist-y
<bbigras>
protests = good. riots = bad
<bbigras>
whatever the cause
<joepie91>
the calls for riots are very explicitly "come riot with us at <place>!"
<cole-h>
(Which is already packaged and has a module :D)
<cole-h>
That said, the bug has been fixed and is working its way through channels :)
<joepie91>
doas is also C
<bbigras>
we should get it before easter when nixos-unstable updates
<cole-h>
lol
<bbigras>
doas is from the openbsd people. it should be well written. and I guess maybe has less old legacy code that sudo might have. maybe
<samueldr>
what's better, software that has more people trying to attack it because of its user base, or one that is seldom looked at because it's not as popular?
<bbigras>
one written by paranoid (openbsd) people?
<samueldr>
it *can* be that software with more past CVEs in more secure because it's more audited
<monsieurp>
the former
<samueldr>
paradoxically enough
<averell>
smaller user base and much smaller code base probably
<samueldr>
yeah, not a unique point to consider
<samueldr>
but something to add to the pile of considerations
<monsieurp>
how about framing the removal differently? something along the lines of "we didn't have enough devs to support and audit this part of the OS properly so we decided to remove it"
<monsieurp>
I'm fond of OpenBSD though but this concept of turning everything off by default to make the OS more secure is patently wrong
<monsieurp>
anyway </rant>
<cransom>
i worked with a guy who was ambitious enough to recompile BPF out of the kernel to make them more secure and no one could tcpdump. eventually he realized there were things he really wanted to tcpdump and stopped doing that.
<cole-h>
Why is there no easy way to convert a symlink to the underlying file :(
<lovesegfault>
cole-h: realpath?
<cole-h>
`cp (realpath file) file` fails because they're the same file
<joepie91>
bbigras: honestly I don't really buy the "OpenBSD, therefore better" thing
<lovesegfault>
ah, yeah, you gotta remove the symlink first
<lovesegfault>
or use `--no-dereference`
<cole-h>
--no-deref gives same file error still :P
<lovesegfault>
huh
<joepie91>
bbigras: like, C is notorious for being extremely difficult to write correct code in even if you know what you're doing, because it's so easy to make critical mistakes that aren't obvious... and so that automatically makes me distrust anything in C, and "we know what we're doing" does not really change that :P plus I consider tool choice a very relevant factor in determining whether someone is good at building
<joepie91>
secure/reliable software
<bbigras>
joepie91: yeah I think we need to move to "safer languages". Didn't Microsoft released a study saying that the vast majority of security bugs are silly mistake that those languages would help with?
<joepie91>
bbigras: not sure. I'm generally a bit skeptical of quantative studies on software safety though, their methodology is almost always severely flawed
<bbigras>
joepie91: what do you propose instead of C then?
<joepie91>
oh, hold on, this is on internal software, now it's interesting
* joepie91
goes skim
<samueldr>
but yeah, removing a footgun from the arsenal sure can't hurt
<bbigras>
yeah
<samueldr>
but believe me, I can produce great bugs in a totally type safe manner!
<samueldr>
memory safe too!
<joepie91>
bbigras: yeah okay this is actually plausible data
<joepie91>
"As was pointed out in our previous post, the root cause of approximately 70% of security vulnerabilities that Microsoft fixes and assigns a CVE (Common Vulnerabilities and Exposures) are due to memory safety issues. This is despite mitigations including intense code review, training, static analysis, and more. "
<samueldr>
and if I can
<samueldr>
so can you
<joepie91>
"we've tried to fix this in every conceivable way and it's *still* dominating the charts" is a much stronger case for memory safety than typical quantitative research from academia :)
<joepie91>
samueldr: I find typescript code to be very telling here
<philipp[m]1>
Which is another problem of sudo. It's complex enough to have logic bugs.
<joepie91>
samueldr: in my own experience, typical TS code causes an outsized amount of weird logic bugs compared to typical JS code, despite being statically typed
<joepie91>
even when it's properly typed with no escape hatches
<joepie91>
this is interesting because people tend to grab TS because they don't want to learn how JS works
<gchristensen>
I'm skeptical of your research :)
<joepie91>
... and then proceed to get screwed by the behaviour of JS, which remains unchanged
<joepie91>
(which in some sense accounts for "people who are very careful in writing their code tend to go for statically-typed languages, which does not seem to be the majority of the TS crowd)