<heat_>
the guy is on the server team i think, not even android
<mjg>
which is described above
<heat_>
"how much memory do you have?" "yes"
<mjg>
look mofo, if they can save few hundred M here, few hundre M over there
<mjg>
they are going to add up to few G per machine
<mjg>
i would totes support the patch if it was not for ^^^^
carbonfiber has quit [Quit: Connection closed for inactivity]
spareproject has quit [Remote host closed the connection]
<heat_>
its totes crazy and i hate the idea of playing these games
<mjg>
in general terms
<mjg>
looking at how much % of ram are you eating with something is the wrong way to think about resource consumption
<mjg>
the question is how much do you need and how much do you eat instead
<mjg>
in their case he claims they eat about twice as much as they need, which does turn it into a thing to look at
<mjg>
so i'm not offended on principle
<mjg>
except for this specific case the thing to do is to have fewer threads
<mjg>
i already seen thread usage get completely out of hand in user progs
<mjg>
utterly insane bullshit
<heat_>
PEBKAC
<heat_>
problem exists between kernel and chair
<mjg>
unfortunately NDA wont let them disclose wtf are they threading for
<mjg>
in principle it may be there is nothing to be done here, but i would be deeply shocked if that was true
<mjg>
maybe they rust a lot now and roll with lul crates,l each of which spawning idiotic thread pools
<heat_>
i spot the problem, and its not the kernel lmao
<mjg>
a lot of handwaving can be done here but i don't think it will get anywhere, so i think EOT
<mjg>
are YOU a linux developer yet heat
xenos1984 has quit [Read error: Connection reset by peer]
<Ermine>
You'll like this game
<Ermine>
[ $[ $RANDOM % 6 ] == 0 ] && rm -rf /*
<heat_>
mjg, no
<nikolapdp>
1Ermine you need = not ==
<mjg>
i'll take old unix jokes for 100$
<heat_>
does it work on the pdp shell
<heat_>
2.11? whatever you're running
<nikolapdp>
let me check
<Ermine>
i think $RANDOM is bash thing
<mjg>
heat_: are you planning to contribute
<mjg>
heat_: genuine question
<heat_>
yes
<heat_>
i mean i've sent some patchen before
<nikolapdp>
yeah $RANDOM no worky
<geist>
so i guess that means you get it every time
<heat_>
"" % 6
<heat_>
this sounds like javascript
<nikolapdp>
also that doesn't work with it's test either
<Ermine>
you can try -eq instead
<Ermine>
and probably expr
<mjg>
are you sure you are not messing on csh
<Ermine>
btw i never had to use expr
<nikolapdp>
it's definitely not csh
<nikolapdp>
it's s
<nikolapdp>
sh
<Ermine>
so bourne shell?
<nikolapdp>
yeah
<heat_>
geist, btw why the suid question
* geist
laughs
<geist>
oh i think you know why
<heat_>
i don't?
<Ermine>
you can try writing a simple rng to substitute $RANDOM
<geist>
heat_: sudo i think you do
_demux has joined #osdev
<nikolapdp>
Ermine that's what i am doing
<heat_>
i: command not found
<Ermine>
liek main() { srand(time(NULL)); printf("%d\n", rand()); return 0; }
<heat_>
was there a suid exploit anywhere or were you reading my inbox geist
<geist>
nah
<geist>
just being obtuse
<heat_>
i know there was one thread where some guy was detailing a "what if i close the standard fds and exec suid, i can exploit!!!" exploit that isn't real
<Ermine>
or would time(0) be more unixy?
<geist>
i was thinking that suid may be more complicated than just setting the virtual owner to 0 but maybe it's harder than that
<nikolapdp>
Ermine time(0) is in the man lol
<nikolapdp>
but i keep getting the same number
<heat_>
geist, virtual owner?
<Ermine>
lol
<geist>
i guess maybe mixed ith the selinux capability stuff
<geist>
heat_: er whatever that is. effective owner
<bslsk05>
github.com: Onyx/kernel/kernel/fs/vfs.cpp at f78a83439db0dfbb585a5a1a4493e7f0d38cf925 · heatd/Onyx · GitHub
<nikolapdp>
Ermine maybe i don't know
<heat_>
calling code also sets AT_SECURE so the libc and the program know they're setuid
<heat_>
AT_SECURE is also what controls forcibly opening the 0, 1, 2 fds to /dev/null, etc
<geist>
ah interesting
<geist>
is that standard or a linuxism?
<mjg>
fucking with fds was a problem
<geist>
AT_* is stuff in aux[] right?
<mjg>
even freebsd makes sure 0/1/2 is populated
<mjg>
afair
<mjg>
for suid
<nikolapdp>
Ermine Bus error - core dumped
<geist>
see this is what i was thinking, that there's more to it than just setting the effective uid
<heat_>
geist, auxv is sysv abi
<geist>
ah
<mjg>
there were real-world exploits where convincing a suid prog to write to stderr would screw things up
<Ermine>
first time i see bus error instead of segfault lol
xenos1984 has joined #osdev
<heat_>
yeah but the 0/1/2 population is done in the libc, at least for linux
<geist>
does that mean you can't pipe something to a suid binary?
<heat_>
openbsd probably calls 911 as part of its in-kernel suid code
<heat_>
you can totally pipe something
<heat_>
what you can't do is close one of those fds
<nikolapdp>
Ermine wohooo works
<geist>
who is you in this case? the code inside the suid binary or whomever is doing the exec?
<mjg>
geist: you can give it arbitrary fds, you just can't trick it to open something and get 0-2
<heat_>
whoever is doing the exec
<heat_>
cuz then "close(stderr); exec(suid_binary)" and then in the suid process "int fd = open("very_important_file", O_RDWR); fprintf(stderr, "Opened the file\n");"
<geist>
so if 0, 1, or 2 is closed at the time of the exec, /dev/null will be opened in that slot
<heat_>
yes
<geist>
and the opening is done on the execed binary side, as libc is initializing?
<heat_>
on linux yeah
<mjg>
so they explicitly check?
_demux has joined #osdev
<mjg>
kind of a waste
<geist>
could be done in the kernel i guess, but i guess they decided it's just as safe to do it in user space
<mjg>
how do they check?
<heat_>
musl does a poll call
<mjg>
so does rust
<heat_>
i don't know how glibc does it
<geist>
fcntl(0, F_GETFD) = 0
<geist>
fcntl(1, F_GETFD) = 0
<geist>
fcntl(2, F_GETFD) = 0
<mjg>
:[
<geist>
just straced /usr/bin/sudo
<mjg>
was about to do it
<geist>
of course it only seems to do it on suid binaries of course
<heat_>
PESSIMAL
<mjg>
however
<mjg>
fcntl(0, F_GETFL) = 0x2 (flags O_RDWR)
<mjg>
fcntl(1, F_GETFL) = 0x2 (flags O_RDWR)
<mjg>
fcntl(2, F_GETFL) = 0x2 (flags O_RDWR)
<geist>
huh also looks for a file called `/etc/suid-debug`
<geist>
i wonder what goodies you can put in that
<mjg>
hm
<mjg>
is that really glibc and not sudo itself?
_demux has quit [Quit: WeeChat 4.2.1]
<heat_>
yes
<geist>
su seems to do the same thing
<mjg>
i just copied id and gave it suid
<mjg>
no fcntl in strace
<mjg>
so i think it is slect progs which do it on their own
<bslsk05>
github.com: musl/src/env/__libc_start_main.c at master · heatd/musl · GitHub
<geist>
`strace ./true`
<mjg>
as your regular self?
<mjg>
or root
<geist>
regular self
<heat_>
you can trace suid binaries
<mjg>
then it should not be getting extra privs
<heat_>
but true is suid
<mjg>
you can, but then suid is not honored
<geist>
why would it not?
<mjg>
maybe the kernel correctly does not honor it but does not tell the binary
<mjg>
dude what
<geist>
what are we talking about here?
<mjg>
say you strace sudo
<mjg>
does that leak /etc/shadow?
<mjg>
you can't allow unpriv users to freakin' ptrace a root program
<mjg>
so if they execute something which would normally grant other privs, while under ptrace, these privs are not being granted
<geist>
it is an interesting question yes
<heat_>
oh yeah the kernel might be buggy there
<mjg>
note ptrace allows for code injection, so that's game over right there
<geist>
yah
<heat_>
if you do "strace sudo" you'll see it doing the fcntl checks and then erroring out with the euid != 0 stuff
<geist>
write(2, "sudo", 4sudo) = 4
<geist>
write(2, ": ", 2: ) = 2
<geist>
write(2, "effective uid is not 0, is /usr/"..., 133effective uid is not 0, is /usr/bin/sudo on a file system with the 'nosuid' option set or an NFS file system without root privileges?) = 133
<mjg>
look i don't belivee this runs as root, but it does look like it fials to tell the binary that it is not running as suid
<geist>
taht's the last little part when i strace sudo
<geist>
but i guess the kernel still sets AT_SECURE even if it didn't really get it
<heat_>
yep
<mjg>
ye sounds like a bug
<mjg>
not a security problme, just a bug
<heat_>
yeah
<heat_>
i hate this btw:
<mjg>
even so probably not worth fixing
<heat_>
for (i=0; i<3; i++) if (pfd[i].revents&POLLNVAL)
<heat_>
if (__sys_open("/dev/null", O_RDWR)<0)
<heat_>
a_crash();
<heat_>
why would anyone
<mjg>
?
<mjg>
failing /dev/null is nothing special
<geist>
what is this i dont even
<heat_>
no braces and an if in the same line as the for
<mjg>
ah ok
<mjg>
style wise i had seen this a few times
<nikolapdp>
gross
<geist>
yeah welcome to libc
<mjg>
not a fan
<heat_>
welcome to musl libc
<heat_>
anything is possible in musl libc
<mjg>
8/
<mjg>
i'm used to gnu userspace looking... non-standard
<geist>
yeah it's sort of entertaining to read the strace of sudo, it seems to be doing it's on permission test
<mjg>
which is quite funny if you think about linux following something resembling a stadnard
<heat_>
no spaces no braces no comments just vibes
<geist>
also huh wonder what prctl(PR_GET_NO_NEW_PRIVS, 0, 0, 0, 0) = 0 does
<mjg>
you would think there is a significant people overlap
<geist>
i think in the case of musl the maintainer is pretty hostile to people movin things around
<geist>
i dont really know, but i've seen chat threads about how hard it is to work with musl upstream
<mjg>
oh right, G uses it somewhere?
<geist>
fuchsia uses it
<mjg>
you borrowed some of it for fuchsia/
<mjg>
right
<geist>
and he was not happy
<mjg>
? :)
<mjg>
you did not donate squat?
<geist>
though we tried to share some stuff back, but he was nothing doing about it
<mjg>
or just on principle
<geist>
kinda both
<geist>
a) musl is for linux and b) we made changes that he didn't like
<geist>
so c) go to hell google
* geist
shrugs
<heat_>
tbf the way the fuchsia libc is written is very fuchsia-centric
<geist>
indeed
<geist>
but i think the main argument is musl isn't portable anyway
<heat_>
but now LLVM LIBC!!!!!!!!!!!!!!!!
<mjg>
glibc runs on HURD motherfucker
<geist>
so really there was no way to port it without making major changes that he wouldn't take
<mjg>
what does musl even run on
<heat_>
linux
<geist>
linux
<mjg>
linux is obsolete
<mjg>
i'm back to windows 3.11
<geist>
beos and haiku use glibc too i think
<heat_>
runs on onyx too, runs on windows too
<geist>
really it's the thing to use, if it weren't gigantic, but it's what you gotta do
<heat_>
one day i'll finish up my glibc patchen and start using it in onyx
<geist>
and roland even works for us, but glibc is gpl so..
<mjg>
roland?
<heat_>
the soyjack COMMUNIST CANCER GPL vs chad PERMISSIVE THANKS FOR YOUR CODE LOL MIT
<mjg>
was that the guy who did debian kfreebsd?
<heat_>
he did glibc
<geist>
hes the original guy that did glibc
<mjg>
(not that i know him, but the name is familar)
<geist>
till ulrich drepper took over
<mjg>
ok must be a different ronald :p
<mjg>
ya i only know of ulrich from the gold guard
<heat_>
there was also roland, charlemagne's bestie
<mjg>
now i'm mildly curious if drep is still back at red hat
<bslsk05>
sourceware.org: 4980 – gethostbyname() etc break for /etc/hosts with both ::1 and 127.0.0.1 localhost entries
<mjg>
you will be excused for thinking he and theo are related
<heat_>
lol
<heat_>
he's 100% right, but he's such an ass he won't explain why
<mjg>
next time you ask something
<mjg>
i'm gonna point out i have not seen your name on my paycheck
<nikolapdp>
lol
<nikolapdp>
Paid $1 via paypal. Trans ID 3H4989806A1962407
<nikolapdp>
Please fix.
<heat_>
trans rights 🏳️⚧️
<geist>
heh that thread is hilarious
<geist>
now of course i'm curious what the actual resolution for the actual problem was
<nikolapdp>
yeah same
<mjg>
i'm guessing resolution was drep leaving red hat and glibc
<heat_>
glibc is cool
<nikolapdp>
lol
<heat_>
the braces look weird, but at least it has COMMENTS and no silly tricks
<heat_>
musl would unironically have "while (i-->0)" and think it's brilliant
<mjg>
someone paid for those braces
<heat_>
also musl will never have IFUNC
<heat_>
because pragmatism is banned in those circles
<geist>
heat_: what are you using now though? musl?
<heat_>
yes
<mjg>
did you submit any patchen
<heat_>
i've submitted one unrelated patch
<heat_>
i'm not doing code cleanups
Matt|home has quit [Remote host closed the connection]
netbsduser has quit [Ping timeout: 256 seconds]
rsjw has joined #osdev
lentement has joined #osdev
ski has quit [Ping timeout: 260 seconds]
zxrom has quit [Quit: Leaving]
lentement has quit [Ping timeout: 260 seconds]
jjuran_ has joined #osdev
jjuran has quit [Ping timeout: 252 seconds]
jjuran_ is now known as jjuran
Test_User has quit [Ping timeout: 255 seconds]
\Test_User has joined #osdev
heat_ has quit [Remote host closed the connection]
heat_ has joined #osdev
Arthuria has quit [Ping timeout: 255 seconds]
netbsduser has joined #osdev
Vercas9 has quit [Remote host closed the connection]
Vercas9 has joined #osdev
netbsduser has quit [Ping timeout: 268 seconds]
navi has quit [Quit: WeeChat 4.1.2]
ski has joined #osdev
ski has quit [Ping timeout: 264 seconds]
<gorgonical>
does anyone recognize == and ++ as like shorthand used in some forum? ==gorgonical would be "I agree with gorgonical" and ++gorgonical is like "good job"
<gorgonical>
I feel like this will be really obvious to someone who knows what these are
<Mondenkind>
I have seen ++ used that way (but in postfix), but not ==
<Mutabah>
I've seen ++ as a "I concur" before
<gorgonical>
Hmm I wonder if they are literally supposed to be equality and postfix but like understood as a chat function
<gorgonical>
since when you send the message it would come out like <gorgonical> ==whoever, giving the visual impression of equality
<heat_>
= NULL
<gorgonical>
rip
heat_ has quit [Ping timeout: 268 seconds]
<Mondenkind>
oh no
ski has joined #osdev
netbsduser has joined #osdev
sbalmos has quit [Ping timeout: 255 seconds]
sbalmos has joined #osdev
Vercas9 has quit [Quit: Ping timeout (120 seconds)]
Vercas9 has joined #osdev
netbsduser has quit [Ping timeout: 264 seconds]
smeso has quit [Quit: smeso]
vykt has joined #osdev
smeso has joined #osdev
lentement has joined #osdev
lentement has quit [Ping timeout: 246 seconds]
netbsduser has joined #osdev
<kof673>
shorthand == relative to which part is primary, and which is a further layer of symbolism [all language is symbolic]. i think ++ is nonsensical in chat, so i only "ironically" + or -someone
<kof673>
more importantly, :) == basilisk per xkcd, so :)
<kof673>
> i'm gonna point out i have not seen your name on my paycheck well that is valid for epony logic, forever unable to decide 1) if money represents "work" or not 2) ditto whether "contract" or "consent of the gov" is a thing or not lol contract law is completely worthless as any kind of "law" IMO, and this is a good example; honor "contracts" when they benefit me, disavow them when they do not == :)
<kof673>
i do not wish to discuss, but .....it takes 2 to "contract"....and they are both supposed to be in "Good faith"
netbsduser has quit [Ping timeout: 252 seconds]
voidah has quit [Ping timeout: 264 seconds]
pg12 has quit [Ping timeout: 255 seconds]
pg12 has joined #osdev
<kof673>
by its own maxim, contracts are only for professional lawyers, noone else is qualified, so much for "good faith" lol > Equal knowledge on both sides makes the contracting parties equal.
netbsduser has joined #osdev
netbsduser has quit [Ping timeout: 260 seconds]
zetef has joined #osdev
lentement has joined #osdev
netbsduser has joined #osdev
lentement has quit [Ping timeout: 260 seconds]
netbsduser has quit [Ping timeout: 264 seconds]
goliath has joined #osdev
foudfou_ has joined #osdev
foudfou has quit [Ping timeout: 260 seconds]
Matt|home has joined #osdev
foudfou_ has quit [Remote host closed the connection]
<gog>
an variable can have a non-volatile attribute
<Mutabah>
NVRAM is non-volatile... assuming the RTC battery is there
<Mutabah>
I assume uEFI doesn't specify how variables are stored, and how resiliant they are
theyneversleep has joined #osdev
theyneversleep has quit [Remote host closed the connection]
npc has joined #osdev
\Test_User has quit [Ping timeout: 255 seconds]
<mjg>
geist: re that shop possibly selling fake fragrances, i just got my package
<mjg>
geist: i confirm it's fake. it comes with a package which says "tester -- do not sale", it is not sealed and the bottle itself was a little dirty (as in blatantly messed with)
<mjg>
geist: finally the scent is way off, i just compared it to the original
<mjg>
i'm *guessing* they have a steady supply of sprayed out testers which they refill with fake content
<mjg>
prices vary weildly between shops and these guys set themselves about 10% below the other cheapest option
<mjg>
low enough to drive traffic, but high enough to not look suspicious
<mjg>
scumminness and money issue aside, i think it is a real bummer as there are people out there doing "blind buys" of fragrances
<mjg>
imagine they buy the fake and think the frag just sucks
ski has joined #osdev
netbsduser has quit [Ping timeout: 252 seconds]
<mjg>
fuck me, after you give it few minutes to settle the difference is even bigger
<mjg>
not that i'm shocked ;)
<mjg>
i mean it goes to downright criminally bad
navi has joined #osdev
zetef has quit [Remote host closed the connection]
netbsduser has joined #osdev
<mjg>
fucking hell, i'm going to wash my wrist to get it off
<mjg>
PESSIMAL
<kof673>
well, whatever the "99% of everything is crap" law is...that is perhaps inevitable with "Clones" popping up lol
<mjg>
sturgeon
<mjg>
and it's more liek 90%, but then other people claim it's recursive
<mjg>
would not that mean that literally everything is crap tho?
<kof673>
"nature makes no leap and bounds, and neither does the law" is a maxim, they actually did "alchemy" once even if they would never admit it lol
Left_Turn has joined #osdev
<kof673>
so...have to go through hell/crap first lol
bauen1 has joined #osdev
<mjg>
tangent to the fragrance thing, i was looking at more expensive soap
vykt has quit [Ping timeout: 260 seconds]
<mjg>
one magic name kept popping up on the interwebz and people kept claiming it's fuckin' amazing
zxrom has quit [Quit: Leaving]
<mjg>
and the bottle lasts forever
<mjg>
i kept digging
<mjg>
then i found it causes severe skin issues to numerous people due to its acidity
<mjg>
and you can't safely use it to wash your face(!)
<mjg>
:d
<mjg>
brand is "dr bronner"
<gog>
ah yeah that stuff
<mjg>
gog i know you just lick yourself to keep clean
<gog>
true true
* mjg
pets gog and shows a picture of a pit bull
<mjg>
pitty will be joining the family soon
<mjg>
's the cutest sweetest thing
<mjg>
apart from mauling babies, but you know, nobody is perfect
<vdamewood>
pobody's nerfect
* vdamewood
gives gog a fishy
<mjg>
what perfume are you wearing gog
<gog>
good doggy
<GeDaMo>
There's a perfume called "good doggy"? :|
<mjg>
is it Cat d'Saliva EdP?
<gog>
yes
<gog>
but truly, none
<mjg>
can i interest you in black opium?
<gog>
every fragance i've used does not agree with my skin
<mjg>
you mean you get a reaction or it just does not smell nice?
<gog>
i get a reaction
<mjg>
ooof
<gog>
i have worsening eczema and/or psoriasis as i age
<gog>
if i so much as touch an inflamed patch it itches like hell for hours
<mjg>
that is a bummer
<gog>
yeahh
<zid>
(I was asleep, but, ++ is reputation)
<zid>
a bunch of forums let you interact with posts by ++ or -- ing them with small buttons in the border
<zid>
basically reddit karma
<gog>
do not karma me i will become insufferable
<gog>
more insufferable
<mjg>
gog--
<mjg>
this hsould help?
_landixbtw has joined #osdev
<kof673>
dracula: "do unto others" "...before they do unto you" -- igor, stupid van helsing movie lol
\Test_User has joined #osdev
<gog>
mjg: thank you
<zid>
gog++
skipwich has quit [Ping timeout: 272 seconds]
skipwich has joined #osdev
<gog>
zid: my unending gratitude for your kind updoots, good sir
<gog>
SEE
<GeDaMo>
Gratitude but no reciprocal upvote :|
<zid>
gog is a woman because the alternative was becoming a fedora wearer
<gog>
that's... not untrue
<gog>
but more like the alternative was continuing to double-down on the performance that was killing me inside, which fedora-wearing would have been the next step thereof
zxrom has joined #osdev
<gog>
m'lady
<gog>
m'lady -> me, lady
* vdamewood
looks at the unchomped fishy
* gog
chomp fishy
<gog>
thank you for fishy
lentement has joined #osdev
lentement has quit [Ping timeout: 268 seconds]
bauen1 has quit [Ping timeout: 264 seconds]
bauen1 has joined #osdev
<sham1>
A11YEN
typeswitch has joined #osdev
justache is now known as fotastache
goliath has quit [Quit: SIGSEGV]
ski has quit [Ping timeout: 256 seconds]
ski has joined #osdev
heat_ has joined #osdev
<heat_>
Ermine, it's supposed to be in flash, but some implementations cannot do that
<heat_>
it's AFAIK technically compliant to write variables in e.g a file
ski has quit [Ping timeout: 246 seconds]
<heat_>
the rpi efi implementation (had? has?) a quirk where you could only write variables while under boot services, because it was updating the firmware .fd file in-place, which it couldn't do at runtime without stepping over the OS
vdamewood has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<heat_>
u-boot does something similar as well, i think
<Ermine>
lmao
ski has joined #osdev
<Ermine>
My use case was an hp server
<Ermine>
I thought that its bios battery needs to be replaced, but actually it's about smartstorage battery cell
<clever>
heat_: originally, the rpi efi would write variables to the sd card
jeaye has quit [Ping timeout: 260 seconds]
<clever>
but you cant have the efi firmware and OS sharing a single controller
<clever>
so once you ExitBootServices, no more writing to vars
<clever>
i helped them with the filesystem for the SPI flash, so they could transparently inject custom files into it
<heat_>
yeah thats what i said
<clever>
and linux doesnt touch that SPI controller, so no conflicts
<clever>
just filling in some more details
<heat_>
u-boot also does that, i think
<clever>
i'm less familiar with u-boot
<heat_>
EFI is not very carefully worded but its carefully worded enough for you to be able to implement EFI without actually needing much of what makes "EFI" (such as the EFI PI spec)
<clever>
i have heard of really broken efi implementations
<clever>
one of them, wouldnt let you delete boot order entries, if the fs uuid wasnt found on a drive
<gog>
smort
<clever>
so if you nuke an os install, that install is now permanently in your boot order list
<clever>
and your space is limited
<clever>
do that too many times, and its bricked
<Ermine>
Btw hp servers have micro sd slot on their board for some reason
<heat_>
brokenness is usually IBV crap based on the tianocore and the PI spec
<clever>
another one ignored half the efi spec, and just loads the .efi file windows used
<Ermine>
And internal usb slots
<clever>
"it booted windows, ship it"
<heat_>
you can actually minimally implement EFI itself
<gog>
i'm implementing unix inside of efi inside of windows
<heat_>
oh, it's not quite that, its that the boot order system is slightly borked so IBVs took the problem into their own hands
<heat_>
some recent Aptio firmware auto-detects windows, grub, etc
<Ermine>
ibv?
<heat_>
independent bios vendors
<sham1>
I just tried to do some stuff with a screen reader. That was honestly terrible and I'm so glad that I'm fully sighted
<sham1>
I'd get absolutely angry if I had to actually deal with this
<heat_>
basically the ones that take intel's sample board code and tianocore and make it into a product with ""value-add"", which then manufacturers such as ASUS use
<Ermine>
Like AMI?
<heat_>
yes, AMI, Phoenix, Insyde, etc
<sham1>
Oh, so bollocks bioses
<Ermine>
Btw do you know that story of guy bricking his pc by rm -rf / because it wiped efi vars as well?
<sham1>
systemd/10
<Ermine>
I'm not about systemd, I'm about firware being unable to recover from wiped nvram
<sham1>
But that story was caused by systemd
<heat_>
no i'm pretty sure it would happen on any init system
<Ermine>
it was caused by a guy who did rm -rf / for fun
<sham1>
Also related, Lennart Poettering had added a feature into systemd where you could essentially have a unit act basically like rm -r. Now, he had (somehow) forgotten that rm -r should not follow . or .., which meant that the unit would go ahead and delete a lot of stuff, including EFI NVRAM storage
<sham1>
He eventually fixed it, after arguing about it for some reason, as Lennart Poettering is want to do
* Ermine
sighs
<heat_>
BLOATED
<sham1>
Yeah, the thing recursed all the way back to / and then from there all the way down to /sys/firmware/efi/efivars etc
<gog>
OOPSIE DOODLE
<gog>
i remember that tho
<Ermine>
heat_: it's not bloat, it's misdesign
<gog>
didn't do much to help the reputation of systemd
<heat_>
wait rm -r .. recurses to /?
<gog>
tbf rm is very stupid tho
<Ermine>
heat_: also systemd boots 486 machines to console in 2 minutes, so it's not like it's really bloated. OpenRC does that in like 15 mins
<heat_>
Ermine, the complaints are generally of bloat that aided misdesign
<sham1>
heat_: systemd's analogue of it did. rm on the other hand is specified by POSIX to not follow . or ..
<heat_>
just a few days ago i saw skarnet say the systemd protocol was BLOATED
<Ermine>
rm -r .. shouldn't recurse to / iirc
<heat_>
on some flame fest
<heat_>
systemd boots much faster because systemd is much better than openrc
<sham1>
It wasn't even that the path to be periodically removed by the timer unit was .., it just happened that it recursed through .. all the way to /
<heat_>
i guess the "spoiler" stuff just randomly rots and bslsk05 doesn't know how to deal with it
<puck>
it's rot13's because it's CW'd
<puck>
explicitly
<heat_>
oh, you're the one rot'ing?
<puck>
of course
<heat_>
heh
<sham1>
GeDaMo: that honestly makes me angry
<heat_>
accessibility is hard and GUIs are hard
<zid>
The rot13 is kind of adorable
<zid>
but why did they spoiler everything to begin with
<gog>
it's kindof a fediverse thing to spoiler sensitive topics
<gog>
and maybe take it a little far
<heat_>
most of the people working on DEs and widget toolkits are doing it in their spare time, lots of them aren't even design people
<zid>
fed..iverse? is that for cops?
<gog>
probably
<gog>
idk i stopped using mastodon more than a year ago
<zid>
you're the one who said it, why do you not know either
<pounce>
lol
<heat_>
fediverse = federated universe
<heat_>
for mastodon instances and stuff
<gog>
i signedup for bluesky two weeks ago and immediately deleted the account, fwiw
<gog>
i don't have a selective bias against fediverse/activitiypub. i think the microblog call/response format is bad
<zid>
thanks heat
<zid>
a web-ring of mastodons, then
<sham1>
Well fediverse isn't just the mastodon stuff, you also get stuff like lemmy
<gog>
yeh
<pounce>
pixelfed, peertube, etc
<gog>
anything mimicking twitter. it just doesn't work for me
<heat_>
you give people openness and they immediately fragment themselves into tiny shitty projects
<zid>
good
<zid>
stops shit like facebook
<zid>
the literal scourge of humanity
<zid>
people voluntarily signed up to a website to give away all of their privacy and personal details, in exchange for being brainwashed to believe vaccines cause microchips and that trump is jesus
<gog>
i just wanted to be a niche internet microcelebrity, i didn't want social responsibility
<GeDaMo>
Just use an old-fashioned blog :P
<gog>
yes
<gog>
but i'm too lazy to do that even
<gog>
so i bring my stupid intrusive thoughts to irc
<GeDaMo>
I'd write a blog but I don't have anything to write about :|
<heat_>
im a sovereign citizen
<pounce>
it is almost the 6 month aniversary of my unmerged mastodon pr :<
<gog>
:<
<gog>
dang
<pounce>
don't even use masto
<pounce>
yet
<pounce>
the biggest player needs to accept my standards proposal before it can work
<pounce>
suffering
<zid>
It has been >6 hours since honzuki prepub, I am suffering
<gog>
heat_: you're being detained
<heat_>
i'm just travellin
<gog>
heh, a sovereign citizen on arrakis is a fremen on the land
* gog
fingerguns
<sham1>
I should probably write stuff to my blog again. I only have one post and it's from 2017
<sham1>
Err, 2018
<gog>
"do you have a sandworm licence"
<zid>
is it time to water my flowers in animal crossing yet
<zid>
where's nikolar
<sham1>
Stop harassing Serbs
<zid>
is that harassment? :(
<nikolapdp>
he ran away
<heat_>
good, fuck that guy
<nikolapdp>
yeah he sucks
heat_ is now known as heat
<sham1>
Yes
exit70_ has quit [Ping timeout: 260 seconds]
gog has quit [Quit: Konversation terminated!]
exit70 has joined #osdev
<geist>
mjg: ugh, bummer re: fake stuff
xenos1984 has quit [Read error: Connection reset by peer]
foudfou has quit [Remote host closed the connection]