<samueldr>
what about computer to the bluetooth speaker?
<samueldr>
no reason other than "what about?"
<samueldr>
I don't think it'll help much, other than seeing that it can or cannot do it as good
<ashkitten>
weird thing is, when i cover a certain place on my phone when it's connected to my desktop it totally drops out
<ashkitten>
but that doesn't happen with a bluetooth speaker
<ashkitten>
like, what's so different between connecting to these devices that the 100% guaranteed way to make the connection crap out on one doesnt do shit to the other
<ivan>
ashkitten: do you have USB 3 devices plugged into your computer
<samueldr>
ah, right 2.4 Ghz interference is possible
<ivan>
"USB3.x spreads RFI from zero to about 4.2 GHz or so. The exact spectrum pattern depends on the data rates in use, as the Intel paper Frennzy linked notes."
<ashkitten>
hm
<ashkitten>
is there no way to mitigate this
<ashkitten>
just have proper shielding ig?
<ashkitten>
seems difficult
<ivan>
hmm check if your adapter is even USB 3 first
<ashkitten>
yeah idk
<ashkitten>
either way i just unplugged all other usb devices and it didnt help significantly so shrug
<ashkitten>
bluetooth is weird and janky
<ashkitten>
i only use it bc my phone doesn't support pulseaudio network streaming
<ashkitten>
i wonder if theres a way to make it support pulseaudio over network with some root shenanigans
<ashkitten>
does anything besides pulse itself implement the pulse network protocol?
<ashkitten>
like, a library
cjpbirkbeck has quit [Quit: Quitting now.]
drakonis has quit [Quit: WeeChat 2.6]
endformationage has quit [Ping timeout: 240 seconds]
<etu>
Hmm, it appears that my work laptop has more than one webcam
<etu>
The default one is an IR camera
<etu>
which is totally garbage :D
<ar>
isn't this one of those "windows hello" things?
<ar>
that they also claim is good enough for face recognition login
<etu>
Maybe it has more data that mpv doesn't show
<eyJhb>
etu: almost looks like you! Or, something :D
<etu>
middle top is my hand
<etu>
So it shows *something*
<eyJhb>
Knew it!
<eyJhb>
etu: what is a alternative to Google Maps? (desktop)
NinjaTrappeur has quit [Quit: WeeChat 2.6]
NinjaTrappeur has joined #nixos-chat
NinjaTrappeur has quit [Client Quit]
NinjaTrappeur has joined #nixos-chat
NinjaTrappeur has quit [Client Quit]
NinjaTrappeur has joined #nixos-chat
NinjaTrappeur has quit [Client Quit]
NinjaTrappeur has joined #nixos-chat
vika_nezrimaya has quit [Remote host closed the connection]
<sphalerite>
etu: probably some settings would need adjusting on it, I'm guessing the normal auto-adjust wouldn't be suited for it
Jackneill has joined #nixos-chat
<elvishjerricco>
Upgrading to 19.09. Being kinda prompt in my major updates for once
<eyJhb>
elvishjerricco: any breakage?
<elvishjerricco>
eyJhb: So far so good
<eyJhb>
Considering changing, but I don't think it would be wise right up to a meeting
<elvishjerricco>
Maybe some weirdness with the graphics
<eyJhb>
Ahh
<eyJhb>
What GPU?
<elvishjerricco>
nvidia gtx 1050 (ti I think?)
<elvishjerricco>
graphics mostly seem fine. It's just my terminal
<elvishjerricco>
Suddenly the font's all weird
<elvishjerricco>
So probably the terminal's fault
<eyJhb>
:( I should get something better than the INtel HD Graphics 4000 (or 4400 I think)
<eyJhb>
Kitty?
<elvishjerricco>
Terminator, though I've been meaning to find an alternative. Don't like this one much
<eyJhb>
"trace: warning: types.string is deprecated because it quietly concatenates strings", wops.
<eyJhb>
Same, but I use urxvt
<elvishjerricco>
Yea I had to update home-manager to get rid of that warning
<eyJhb>
No clue what gives that warning, I don't use home-manager
<eyJhb>
But lets see
<eyJhb>
Should probably have plugged into ethernet beforehand, and not just WiFi
<etu>
eyJhb: osm.org?
<eyJhb>
ty
<eyJhb>
Forgot the name of it
__monty__ has joined #nixos-chat
psyanticy has joined #nixos-chat
<__monty__>
eyJhb: How is hmac-sha2-512 insecure?
<eyJhb>
__monty__: it isn't I assume, it was just not in the default allowed Macs in NixOS. So I just assumed insecure because there most be a reason :)
ctp has quit [Quit: Connection closed for inactivity]
<__monty__>
Ripemd160 is another generally considered secure hash btw. And HMAC is pretty standard, no sure about UMAC (I just don't know it, doesn't mean it's not secure.)
veske has joined #nixos-chat
nocent has quit [Read error: Connection reset by peer]
atopuzov[m] has quit [Remote host closed the connection]
sphalerit has quit [Read error: Connection reset by peer]
tokudan[m] has quit [Write error: Connection reset by peer]
colemickens has quit [Write error: Connection reset by peer]
worldofpeace has quit [Remote host closed the connection]
dtz has quit [Write error: Connection reset by peer]
thefloweringash has quit [Write error: Connection reset by peer]
colemickens has joined #nixos-chat
<__monty__>
Anyone have experience running xfs_repair -n on a running system? Like, is it even possible?
<sphalerite>
__monty__: sure it's possible. I'd expect it to complain and refuse if applied to a mounted filesystem.
<sphalerite>
"Regardless, the filesystem to be repaired must be unmounted, otherwise, the resulting filesystem may be inconsistent or corrupt."
<__monty__>
Yeah but that means I can't run it on /, right?
<sphalerite>
"-d Repair dangerously. Allow xfs_repair to repair an XFS filesystem mounted read only. This is typically done on a root filesystem from single user mode, immediately followed by a reboot."
<sphalerite>
__monty__: yes. You can probably run it from the initramfs.
<sphalerite>
__monty__: so add `boot.debug1devices` to the kernel command line, boot, enter a shell when it gives you the option, and run xfs_repair from there.
<__monty__>
sphalerite: Would -d combined with -n be harmless? This is a remote machine so I wouldn't be able to ssh in so early at boot.
<__monty__>
Afaik it should be harmless but dangerous does sound a bit dangerous.
<sphalerite>
I don't know
<__monty__>
It's still refusing : /
<sphalerite>
can you get a serial console or something? Or remote hands?
<__monty__>
Damn these fs developers and their (probably wholly sensible) design decisions!
<__monty__>
I'll have to go on-site.
<sphalerite>
hm, it _might_ be safe if you remount it read-only
<sphalerite>
mount -o remount,ro /
atopuzov[m] has joined #nixos-chat
thefloweringash has joined #nixos-chat
dtz has joined #nixos-chat
tokudan[m] has joined #nixos-chat
worldofpeace has joined #nixos-chat
sphalerit has joined #nixos-chat
nocent has joined #nixos-chat
<__monty__>
I think I'm just gonna not mess with it further. There hasn't been any observable damage. Just someone jostled the system and it made some noise.
<__monty__>
I'll visit tomorrow.
veske has quit [Quit: This computer has gone to sleep]
tilpner_ is now known as tilpner
<manveru>
hmm, my vpsfree node uses 8GB ram, but only has 4GB :P
<eyJhb>
Yay, 19.09, no DisplayLInk breakage. But still running my overlay ..
waleee-cl has joined #nixos-chat
<andi->
gchristensen: is that a hint that we might be seeing that in Nix soon (tm)?
drakonis has joined #nixos-chat
kcalvinalvin has quit [Ping timeout: 245 seconds]
<gchristensen>
andi-: hehe no
drakonis has quit [Quit: WeeChat 2.6]
<cransom>
if it's just an ld_preload, i'm tempted to explore that.
drakonis_ has joined #nixos-chat
<samueldr>
:/ anyone has experience with mirroring a git repo? looks like as usual git has all the plumbing to do a thing, but no convenient porcelain that works as expected
<gchristensen>
`git clone --mirror --bare`?
<samueldr>
I just figured that out after asking, that `clone --mirror` seems to be the magical invocation; if you already cloned it seems there's no clean way to do it
<gchristensen>
aye
<samueldr>
github's own import failed
<samueldr>
and it failed in a way that broke github, the "restart" option led to a 404 :|
<samueldr>
all because I want to optimize the fetching since google's own git services don't have a tarball download
<samueldr>
(new mobile nixos target kernel)
<samueldr>
1.3GiB download to checkout a single commit is a bit annoying for the end-user imo :)
<gchristensen>
yikes
<samueldr>
oh! I might be mistaken
<samueldr>
I think they now have a .tgz download link
<samueldr>
well, I see a tgz link
<samueldr>
or was it raw files it was lacking%
tilpner has quit [Ping timeout: 265 seconds]
tilpner_ has joined #nixos-chat
MichaelRaskin has joined #nixos-chat
cjpbirkbeck has joined #nixos-chat
<drakonis_>
github offers shallow clones on the interface
<drakonis_>
web UI
<drakonis_>
the full repository is 20mb top
<drakonis_>
ah wait, this is a source clone
<drakonis_>
android source clone, nevermind.
<drakonis_>
samueldr, are you targetting android devices equipped with treble?
<samueldr>
that question is misleading, I'll explain why
<samueldr>
any phone with treble support can be targeted individually
<samueldr>
so "yes"
<drakonis_>
rather, target treble only, since it separates device drivers
<samueldr>
but if the question is "do you intend to make a GSI, that treble enables", the answer is "dunno yet"
<drakonis_>
aight.
<samueldr>
there's some research to be done still
<samueldr>
but any phone with treble can still run non-GSI roms
<samueldr>
so in the end it'll be a convenience thing if treble is supported AFAICT
<samueldr>
and "misleading" might not have been the proper word; what I mean is that treble is more of a concept than an actual thing that can be targeted
<drakonis_>
GSIs is so good and i cant believe google took so long to achieve this
Jackneill has quit [Remote host closed the connection]
<MichaelRaskin>
Manufacturers actually satisfying their GPL2 obligations would be even better, though…
<drakonis_>
of course.
<drakonis_>
never worrying about the device becoming useless would be grand tho
<MichaelRaskin>
Well, given that you still cannot upgrade the actual kernel…
<samueldr>
not only satisfying the GPL, but also letting user control the boot flow :/
<samueldr>
(e.g. bootloader unlock)
<samueldr>
and then there's still the issue of the actually useful bootloader being awfully easy to break accidentally
<MichaelRaskin>
I am not sure this is rarer than non-blob drivers
<samueldr>
just a wrong `dd` and there poof there's a brick
<samueldr>
and without the signed secret special programs (often leaked) there's no way a user can recover
<samueldr>
that is, if the phone boots in the special mode when aboot or abl has been removed
* samueldr
sighs
<drakonis_>
that's fixable if the vendor is worried about it
<drakonis_>
see purism
<samueldr>
that's the point, existing devices and current implementation don't
<drakonis_>
now that the tech has gotten sufficiently cheap, it should be possible to start a wave of phones focused on avoiding this
<samueldr>
it's kind of starting, but the word "cheap" cheapens the effort
<MichaelRaskin>
The problem is that the chips still have ridiculously high per-design costs
<samueldr>
many enthusiasts don'T want cheap phones
<samueldr>
and what "cheap phones" often means
<drakonis_>
to be honest here, cheap phone means i don't have to spend a month's worth of wages on it
<drakonis_>
it has much wider appeal when it does not cost a lot of money to have a phone that isn't awful to use
<drakonis_>
xiaomis are getting adopted because they're not expensive and they deliver something on par with far pricier options
<MichaelRaskin>
Well, most flagships are strictly worse ootB than LineageOS-flashed carefully-selected ≤€200 phone…
<samueldr>
I tend to agree here
<MichaelRaskin>
Non-flagship Huwaeis before they went dumb
<drakonis_>
that's a non issue nowadays with newer phones
<samueldr>
hm?
<samueldr>
what's a non-issue with newer phones?
<drakonis_>
see michael's message
<drakonis_>
its not exactly a huge issue anymore for newer phones
<samueldr>
I don't understand
<drakonis_>
you can flash lineageos on newer phones without needing to worry about anything ever
<drakonis_>
treble and whatnot
<samueldr>
flagships are still worse ootb than lineageos-driven phones
<drakonis_>
yes, but the flashing part is not a problem anymore
<MichaelRaskin>
You still need to check the manufacturer hasn't pulled a Huawei
<samueldr>
yeah
<drakonis_>
what is pulling a huawei?
<drakonis_>
i have not looked into phone news in a while
<samueldr>
north american phones are specially bad about bootloader unlocks
<MichaelRaskin>
They provided official bootloader unlock
<drakonis_>
i have a samsung and i fear if i flash twrp it'll ruin the fuse
<samueldr>
here
<samueldr>
that's horrible!
<drakonis_>
can't have magisk
<MichaelRaskin>
via a moderately annoying procedure (which is kind of good, you cannot spend that much effort by mistake, and it was still less effort than the first reflashing)
<samueldr>
why marr a perfectly good device with a "mark of shame" for exercising a basic right?
<drakonis_>
yes
<MichaelRaskin>
Then Huawei stopped providing means for doing the unlock
<samueldr>
the best "bootloader unlock" procedure there is imo is the chromeos family of devices imo
<drakonis_>
there was a samsung galaxy that had a leaked official rom you could flash without tripping it
<samueldr>
the firmware is freely accessible, both read and write
<samueldr>
open source even
<samueldr>
every operation can be reversed
<drakonis_>
also snapdragon devices, for some reason, have crippled battery life if you flash a custom rom
<drakonis_>
this specific device with the official rom at least, did
<samueldr>
meaning that you can fully open it for development purposes, and then fully close it back to be secure
<drakonis_>
its kind of terrible
<drakonis_>
i'd like to have a google device at least but they're so expensive
<drakonis_>
its so insanely expensive to own one
<MichaelRaskin>
I actually ordered a >500€ smartphone
<emily>
samueldr: can you boot your own signed depthcharge kernels on e.g. a pixelbook without opening the thing up?
<emily>
it was never quite clear to me
<MichaelRaskin>
(Haven't received it yet)
<emily>
I'd like to get NixOS running on mine
<samueldr>
emily: without opening the device physically yes
<samueldr>
emily: just a clarification, do you mean with or without flashing a custom firmware?
<emily>
ah ok, I thought the signing keys were hardcoded in and you had to get at the flash
<samueldr>
(but yes, for both, without physically opening)
<drakonis_>
i should probably try to root my phone without flashing anything
<drakonis_>
exploit kernel flaws
<samueldr>
since about end 2017 beginning 2018
<emily>
I don't think you could do a signed depthcharge boot without flashing your own firmware
<emily>
as opposed to like SeaBIOS or whatever that doesn't have signing keys to worry about
<drakonis_>
seems like the only way to do it without tripping knox
<samueldr>
emily: you can boot signed with a non-verified key though
<samueldr>
emily: or do you want signed and validated? :)
<emily>
samueldr: at the expense of the 10 second WARNING DEVELOPER MODE PRESS ANY KEY TO DELETE EVERYTHING you mean?
<emily>
I'd rather avoid that
<samueldr>
yeah
<samueldr>
right, then yeah
<samueldr>
starting with those that have CCD (closed case debugging) you can flash the firmware without opening the device
<samueldr>
it needs an annoying and long process by design
<emily>
nice
<samueldr>
about 5 minutes
<samueldr>
to "open" the firmware
<emily>
I think mine should be new enough for that, it's a 2018 Pixelbook
<samueldr>
yes it should
<samueldr>
you can also go with tianocore
<emily>
will have to play about with that some time :)
<samueldr>
some of the terminology assumes you know everything there is to know about cr50, ccd and their infra
<samueldr>
... but the info is all there
<emily>
cool
<emily>
yeah the chromiumos docs are... mixed
<samueldr>
really, I have to give it to the chromeos hardware team: they know how to do it right
<emily>
in that most of them seem to assume you're already a GOOG employee :P
<samueldr>
yeah, I agree with that :)
<samueldr>
though most of the time you can end up finding the right info, and the source availability is great
<samueldr>
the cr50 chip is also open source
<samueldr>
though IIRC it is strictly veified signatures from the mothership [citation needed]
<emily>
hm, I guess you need to wipe all the data to do any kind of flashing first-time, right?
<samueldr>
as usual with chromeos devices
<samueldr>
so yes
<samueldr>
it's security thing
<emily>
was wondering whether this means leaving a pixelbook with someone untrusted for 10 minutes was risky but I guess it's just the usual "it might be wiped"
<samueldr>
TPM gets trippe
<samueldr>
tripped*
<emily>
right
<emily>
as long as I can set up similar security measures for what I flash after the fact I'm fine with that :)
<samueldr>
and when it gets tripped, now they *also* secure erase over the data partition
<emily>
thanks for all the info, this is really cool!
<samueldr>
how I wish for a chromeos hardware team developed phone target :/
drakonis_ has quit [Read error: Connection reset by peer]
drakonis_ has joined #nixos-chat
<samueldr>
so, yeah, to close the loop on GSI and mobile-nixos; I have two devices that technically support GSIs, with different treble and A/B scheme supports, so I'll be able to check it out... but I don't expect GSIs to work simply because the android kernels lack some configuration options needed for nixos, and they may also need custom patches
<samueldr>
in my experience all qualcomm phones to date need a patch to add right RGB subpixel order
<samueldr>
(or software would need to be enhanced to know how to use the BGR subpixel order)
<emily>
samueldr: wouldn't mind chromeos software team phone either tbh
<emily>
I'm a lot more fond of it than android
<emily>
and the android app compability would be more useful :p
<samueldr>
I mean, at that point they can ship android if they want, as long as its firmware and stuff is the usual coreboot+payload and accessible through a debug cable :3
<samueldr>
oh, and another great thing with the chromeos hardware team: base board designs shared across OEMs...
<samueldr>
no surprise there, 99% of the time when one works, all the series works
<emily>
I mean, I'm not quite dedicated enough for NixOS-on-phone, so understandably our phone OS preferences might differ :P
<__monty__>
Kinda hoping samueldr figures out NixOS on phones before I have to get a smartphone.
<__monty__>
So chop, chop.
<samueldr>
well
<samueldr>
NixOS on phones is mostly figured
<gchristensen>
sounds like a grand opportunity to help!
<samueldr>
oh
<samueldr>
you wanted as *useful* system?
<samueldr>
:)
<__monty__>
gchristensen: Does it? Kinda hard to contribute when you don't have any hardware to test.
<joepie91>
__monty__: based on what samueldr has said in the past, I suspect that much of the work is "there's presumably some information about how to do X somewhere on the web, but I have no idea where, please help me figure out X"
<samueldr>
some is more vague, but some is more about putting some time in with the information available
<samueldr>
"doing the thing" is often the only thing missing... but doing it takes time
<__monty__>
joepie91: Yeah ok. But not having a smartphone has also meant I don't follow any of the developments re bootloader unlocking and whatnot. I'm no help here. Just sharing my support for the cause.
<samueldr>
though there are tasks that are independent from having hardware devices
<samueldr>
anything regarding porting the PE (phone environments, read it like DE, desktop environments)
<__monty__>
If you need any haskell development or formal verification, I'm your guy. C or graphics? Not your guy.
<samueldr>
I need C and graphics :)
<samueldr>
(half joking)
<samueldr>
there sure is going to be some need for that, but it's not an "actual" need right now
<samueldr>
about packaging PEs it's mostly a nixpkgs integration thing
<samueldr>
just like packaging gnome, kde or xfce
psyanticy has quit [Quit: Connection closed for inactivity]
<samueldr>
this discussion makes me realize the talk will be about the plan and not details of how it works :)
<gchristensen>
does that cover radios and stuff too?
<samueldr>
"that" meaning "packaging PEs" no
<samueldr>
though some of relevant apps will require relevant libs, which AFAIU will interface with the relevant systems to make it work
* gchristensen
is pretty excited
<samueldr>
right now I'm focusing on "ports" only to get a palette of different vintage devices working to, hopefully, figure out hardships in porting for the next porter :)
<samueldr>
oh shoot, forgot I wanted to make a small update post announcing that there is a new device which boots to parity
<drakonis_>
qubes on nixos is now a thing
<drakonis_>
or at least becoming one, seems like an interesting and overdue thing