<dangole>
dwfreed: no, but auc on 19.07 is borked because of lacking knowledge of the ABI version of library packages of the package management system. this was added only in 21.02...
schwicht has joined #openwrt-devel
<dwfreed>
dangole: it also doesn't work at all because it still uses the old asu API that doesn't exist anymore
<PaulFertser>
[Pokey]: does your MUA not support threads?
GNUmoon has joined #openwrt-devel
<[Pokey]>
PaulFertser: my mail client? On mobile, it is an ass
<PaulFertser>
If you weren't subscribed you can usually download the whole mbox from archives and then reply as usual.
<PaulFertser>
[Pokey]: K9-Mail is no good?
<[Pokey]>
Samsung mail
<PaulFertser>
How is some proprietary shit relevant? ;)
<[Pokey]>
Because I've used it for 5 years and it works :P
<[Pokey]>
I rarely email
<robimarko>
Does it even support plain text emails only?
madwoota has joined #openwrt-devel
<[Pokey]>
robimarko: nah just chains are a pain
<stintel>
all Android mail clients I've tried are shit, I gave up and ended up using outlook
<PaulFertser>
Termux + mutt ;)
<stintel>
yeah no, terminal based stuff on smartphone is last resort
<rmilecki>
ynezz: oh, thanks for reminding me of that
<rmilecki>
i forgot about it of course
<rmilecki>
ynezz: still in this case I don't expect build failures, I think I thought of every target, it's runtime failures I'm afraid of
<PaulFertser>
stintel: I had openmoko freerunner with _resistive_ touchscreen, and I was able to precisely type using virtual keyboard with the tip of a fingernail, it was quite OK despite the small screen size.
<rmilecki>
still, would be safer to let github build targets ofc
Borromini has quit [Ping timeout: 480 seconds]
<stintel>
PaulFertser: I should have another attempt of installing a Linux distro on my Gemini
<stintel>
it types reasonably well and fits my pocket
<schmars[m]>
in recent days i've had sysupgrades on multiple devices (21.02-snapshot -> 22.03-snapshot) result in empty ssh host keys, which leads dropbear to fail to start with "Early exit: Bad buf_getptr" - bug or wrongdoing?
<PaulFertser>
stintel: plan to touch-type on it?
<stintel>
PaulFertser: no, it has a keyboard
<PaulFertser>
stintel: yes, but do you plan to teach yourself touch typing on it?
<stintel>
had to lookup what that means; most likely not, I can barely do that on a normal keyboard
<PaulFertser>
Oh, then you won't feel that much discomfort from having to look at the keys, good for you :)
cp-- has quit []
cp- has joined #openwrt-devel
<PaulFertser>
stintel: if you like clamshell designs you might fancy a Pinephone Pro with the Pinephone keyboard case.
<stintel>
nobody could confirm the Pinephone power issues are solved on the Pro, so I've decided not to get one for now (battery drain makes it no-go for daily)
<PaulFertser>
Yep, for now.
minimal has joined #openwrt-devel
<stintel>
I have a pinetab but it doesn't fit pocket, it's rather slow and also battery drain
johnf has quit [Quit: Lost terminal]
<ynezz>
rmilecki: well, there is now a check for dirty patches, so for example d80c91btarget/linux/mediatek/patches-5.15/501-auxadc-add-auxadc-32k-clk.patch
<ynezz>
argh
<ynezz>
rmilecki: well, there is now a check for dirty patches, so for example d80c91b made target/linux/mediatek/patches-5.15/501-auxadc-add-auxadc-32k-clk.patch dirty
<ynezz>
this is then causing false negative CI checks for PRs on GH
<ynezz>
and that pipelines seems to be now running on the push to master branch as well
gladiac has joined #openwrt-devel
floof58 has quit [Ping timeout: 480 seconds]
Borromini has joined #openwrt-devel
<f00b4r0>
would anyone know off the top of their heads what's the default way to get into failsafe on x86 (with squashfs)?
<robimarko>
Sending "f" just as procd takes over
<f00b4r0>
so mashing "f" during boot should work I guess. Thanks :)
<stintel>
f <enter>
<stintel>
it instructs you when to do so
<f00b4r0>
no display
<f00b4r0>
:)
<stintel>
it's x86, attach one? :P
srslypascal has quit [Ping timeout: 480 seconds]
<f00b4r0>
that would be my way of doing it, yes. I'm helping someone else remotely :)
<stintel>
gl
<f00b4r0>
meanwhile I notice that "uci import" implies "commit", so my hope to implement test-revert-or-commit that way goes to the drain ;P
valku has joined #openwrt-devel
<jow>
I actually tried doing just that yesterday (enter failsafe on x86/64 in a qemu vm), for me that just resulted in a hung console and inactive network
<jow>
but maybe it is qemu specific brokeness and works on actual hw
<stintel>
works for me, root@(none):/#
<jow>
hmm
<jow>
master nsapshot?
<stintel>
d744fea2b9
<jow>
I tried it with the official 22.03 images
<stintel>
ah, that's a local commit
<jow>
and qemu-system-x86_64 -nographic -nic tap -hda ...
<stintel>
that's a5265497a4f6da158e95d6a450cb2cb6dc085cab + 2 local commits
<stintel>
so fairly recent
<stintel>
this is vis libvirt thought
<stintel>
via*
<stintel>
ugh seriously openssh
* stintel
aliases scp scp -O
<f00b4r0>
-O?
<Habbie>
something something scp deprecated
<f00b4r0>
unknown option -- O
<stintel>
OpenSSH 9 or so uses SFTP by default
<stintel>
which dropbear doesn't support
<stintel>
you can only use -O, there is no .ssh/config option
<f00b4r0>
-o
<stintel>
that uses the same options
<stintel>
there exists no option for scp fallback
<stintel>
no such* option
<stintel>
only -O
<f00b4r0>
i see. /me makes mental bookmark for when he'll be running more bleeding edge openssh
<stintel>
oh you'll notice
<stintel>
ash: /usr/libexec/sftp-server: not found
<stintel>
;)
<stintel>
should probably contribute an ssh_config option for that
<stintel>
although I suspect they did it on purpose to force people to stop using it
<stintel>
i.e. fix the other end
<robimarko>
I also think they are trying hard to make people change to SFTP
slate has quit [Quit: quit]
<zarzarzar_>
Let's say I have two services A and B, and I want to make A restart if B crashes. Is there any mechanism in procd to define such "dependence" between services?
<jow>
zarzarzar_: no
<jow>
you could wrap B's executable in a script that unconditionally restarts A before it exec's into the actual B executable
<jow>
then whenever procd relaunches the B process, the wrapper script will launch, restart A and exec into the B binary
<zarzarzar_>
jow: okay, i see, thanks
<zarzarzar_>
however, i did some research and found that procd emits 'instance.respawn' when services crash theoretically, i can ask procd to watch 'service' ubus object, set trigger on 'instance.respawn' to execute '/etc/init.d/A restart'
<rmilecki>
in both cases "top" show 50% idle and 50% sirq
srslypascal has quit []
Tapper has quit [Ping timeout: 480 seconds]
Gaspare has quit [Quit: Gaspare]
guerby has quit [Read error: Connection reset by peer]
guerby has joined #openwrt-devel
SamantazFox has joined #openwrt-devel
T-Bone has joined #openwrt-devel
f00b4r0 has quit [Ping timeout: 480 seconds]
robje has quit [Quit: "BRB"]
robje has joined #openwrt-devel
srslypascal has joined #openwrt-devel
dangole_ has joined #openwrt-devel
dangole has quit [Remote host closed the connection]
Tapper has joined #openwrt-devel
goliath has quit [Quit: SIGSEGV]
Gaspare has joined #openwrt-devel
Tapper1 has joined #openwrt-devel
Tapper has quit [Read error: Connection reset by peer]
dangole_ is now known as dangole
<dangole>
nbd: in your clean-up of mt7622 kernel config you may have overlooked that CONFIG_IIO is needed for CONFIG_MEDIATEK_MT6577_AUXADC. we could also provide that one as a module, as afaik it's only needed to calibrate thermal values then stored in efuse for production (ie. normal users never need to read the raw ADC values).
gladiac has quit [Quit: k thx bye]
winternull has quit [Quit: Leaving]
Gaspare has quit [Quit: Gaspare]
T-Bone is now known as f00b4r0
Gaspare has joined #openwrt-devel
floof58 has joined #openwrt-devel
bbezak has quit [Quit: Ping timeout (120 seconds)]
bbezak has joined #openwrt-devel
<Mangix>
stintel: you use outlook on mobile?
Borromini has joined #openwrt-devel
<stintel>
yes
<stintel>
it's a less bad experience than other Android clients I've tried
<Mangix>
Interesting
<mrnuke_>
svanheule: What I want to do for testing your LED controller patch, is to edit /etc/board.d/01_leds to automagically configure things. If we can do that, then I am willing to live with the quirkiness of the implementation
xutaxkamay has joined #openwrt-devel
<stintel>
I remember having tried one client, they'd connect to my mail server from their cloud shizzle, and the client would fetch things from there
<stintel>
seriously
<stintel>
how do you do that without being able to get my password
<xutaxkamay>
hi there, i've been interested into contributing to openwrt recently and i will surely read what ever i need to do before to do so, but i just wanted some feedback on my idea if this wasn't already implemented somehow, is it possible currently to have an extroot encrypted with luks? currently i don't see the fstools/block.c having any such options, and i thought it could be nice to somehow enable some ssh network
<xutaxkamay>
connection to unlock the device where as the original file system is used to unlock the device at reboot, similar to what gentoo has, what do you think?
Gaspare has quit [Quit: Gaspare]
indy has quit [Ping timeout: 480 seconds]
<Mangix>
stintel: does it do plaintext?
<[Pokey]>
xutaxkamay: I am very surious what your usecase for this is. Unless someone more experienced than me wants to pitch in and say no, I don't see why that wouldn't work so long as you can compile the driver to work with the target platform, fit all the required binaries into the device's memory and the other thing that may pose an issue is the performance as it would have to use the weak processor of the device to do cryptography on the fly
<[Pokey]>
Could potentially impact networking performance
<minimal>
[Pokey]: using Adiantum rather than AES (assuming no hard support) would reduce any CPU impact
<[Pokey]>
minimal: You wanna aim that at xutaxkamay not me :P
<minimal>
[Pokey] was trying to deflect/bounce it off you ;-)
* [Pokey]
deflects minimal's projectile ideas
<[Pokey]>
Hey, watch where you aim!
<[Pokey]>
nearly lost a marble
<xutaxkamay>
well, just operation security in general, i know it can run on my small 400mhz mips device with only 50mb of ram, except that well, formatting the extroot needs pbkdf2 instead of argon with a longer password
<xutaxkamay>
but yeah of course, some devices runs with lower memory and speed
<xutaxkamay>
i just wanted this as an option currently
Gaspare has joined #openwrt-devel
shibboleth has joined #openwrt-devel
<xutaxkamay>
for pbkdf2 i meant for cryptsetup
<xutaxkamay>
for the type for metadata
<xutaxkamay>
there's already kernel modules ready for crypto and such so yeah
<xutaxkamay>
just crafting the stuff together
<xutaxkamay>
i could see how this would be not wanted though
<nbd>
dangole: if it's not needed for production, i guess we should package it separately
csrf has joined #openwrt-devel
<minimal>
xutaxkamay: the point wasn't about whether there are kernel modules for crypto but rather when the device has hardware acceleration for any of those crypto algos
<minimal>
s/when /whether /
<xutaxkamay>
minimal: well not every cpu have such features i believe
<xutaxkamay>
i wouldn't see how it is a problem though
bluew has joined #openwrt-devel
<xutaxkamay>
since it is an option
<xutaxkamay>
(and can be software emulated)
<minimal>
xutaxkamay: indeed, which which is why I mentioned Adiantum which has less CPU impact than many other crypto algos
<xutaxkamay>
the plant?
<xutaxkamay>
xD
<xutaxkamay>
im joking i never heard of this one yet
<xutaxkamay>
oh chacha
<minimal>
Google use Adiantum on some Android handsets with lower spec CPUs
<xutaxkamay>
okay interesting
<xutaxkamay>
yeah i see it now
<xutaxkamay>
well, it would certainly have an impact in performance anyway but i guess 'we' could let decide just the user if he wants a slower system or not, encryption isn't without performance loss anyway at some point
<xutaxkamay>
hence why i say as an option >.<
MaxSoniX has quit [Quit: Konversation terminated!]
Gaspare has quit [Quit: Gaspare]
Gaspare has joined #openwrt-devel
Gaspare has quit [Quit: Gaspare]
Gaspare has joined #openwrt-devel
<[Pokey]>
Mangix: I take it you received the mail list patch?
<Mangix>
what about it?
cbeznea has quit [Quit: Leaving.]
<[Pokey]>
Mangix: Nothing in particular, just wanted to make sure youi're Aware
<[Pokey]>
Don't particularly agree with it being adjusted so soon but I guess thats that
shibboleth has quit [Remote host closed the connection]