[itchyjunk] has quit [Remote host closed the connection]
nuid64 has joined #osdev
nuid64 has left #osdev [#osdev]
agent314 has joined #osdev
agent314_ has joined #osdev
agent314_ has quit [Client Quit]
agent314 has quit [Ping timeout: 255 seconds]
craigo has joined #osdev
xenos1984 has quit [Read error: Connection reset by peer]
heat has quit [Ping timeout: 252 seconds]
xenos1984 has joined #osdev
netbsduser`` has joined #osdev
gbowne1 has quit [Remote host closed the connection]
gbowne1 has joined #osdev
netbsduser`` has quit [Ping timeout: 258 seconds]
netbsduser`` has joined #osdev
dude12312414 has quit [Quit: THE RAM IS TOO DAMN HIGH]
khrbtxyz has joined #osdev
netbsduser`` has quit [Ping timeout: 258 seconds]
pretty_dumm_guy has quit [Ping timeout: 258 seconds]
netbsduser`` has joined #osdev
vai has joined #osdev
gbowne1 has quit [Quit: Leaving]
GeDaMo has joined #osdev
netbsduser`` has quit [Ping timeout: 240 seconds]
vai has quit [Ping timeout: 260 seconds]
masoudd has quit [Quit: Leaving]
masoudd has joined #osdev
frkzoid has quit [Ping timeout: 260 seconds]
kalj has joined #osdev
<kalj>
I am writing a floppy driver, and I am trying to understand if a SEEK is necessary before reading data. It seems to me that READ_DATA would work fine on its own since takes `cylinder` as an argument?
<bslsk05>
wiki.osdev.org: Floppy Disk Controller - OSDev Wiki
duderonomy has quit [Remote host closed the connection]
<kazinsal>
based on the original IBM disk controller manual I don't see anything saying you need to do a seek before a read
duderonomy has joined #osdev
<clever>
kalj: one guess, less wasted cpu cycles, you can initiate a seek, and then go do other things, before you initiate a read and wait for data
<kalj>
That makes sense. It seems like a good idea to have one in a decent driver. But wording in various places (including the wiki) seems to indicate that it was necessary.
<clever>
i have seen controllers like the DSK-II, where it is nessesary
<kalj>
My driver seems to be able to read correctly in qemu, but I still have to try it on real hardware.
<clever>
the cylinder# in the read command, is purely something to check the sector-addr header against
<clever>
and if you dont seek first, then it reads the wrong track, and throws an addr error
<kalj>
Okay, so it could just ignore `cylinder` and read wherever it's at?
<clever>
more, that the DSK-II doesnt know what track the drive is on
<clever>
when you initiate a read, it just reads the given sector# within the track
<clever>
and then compares the track# in the header, against what you asked for
<kalj>
Forgive my ignorance, but what is the "sector-addr header"
<clever>
every magnetic drive ive heard of, has a header at the start of every sector
<clever>
and that header includes the CHS addr of the sector
<clever>
they typically also have a checksum at the end of every sector
<kalj>
oh, i see
bauen1 has quit [Ping timeout: 255 seconds]
<clever>
the drive controller checks the header/footer, and returns on the data in the sector
<clever>
so all of that is hidden from you
<clever>
depending on how the drive and controller work together, i can see how read may not seek, and will just read whatever the current track is
<clever>
in the old days, the hard-drive controller was also its own ISA card
<clever>
and that is where IDE comes from, integrated drive electronics, putting the controller on the hdd itself
<clever>
and the PATA ribbon was basically just an extension of the ISA bus at first
<kalj>
wow, so much stuff unknown to me. thanks for describing!
<clever>
so you basically had an ISA card glued to the hdd, and connected back to the mobo over a ribbon cable
<bslsk05>
'Not forgotten: MFM hard drives' by Adrian's Digital Basement (00:30:22)
<clever>
from back when the controller wasnt part of the drive
<clever>
those hdd's function a lot more like a floppy drive does
kalj has quit [Quit: Client closed]
danilogondolfo has joined #osdev
duderonomy has quit [Remote host closed the connection]
duderonomy has joined #osdev
KitsuWhooa has joined #osdev
<geist>
em eff em
<kazinsal>
back in mah day we had to bitbang frequency modulations onto disks
<kazinsal>
then along came emacs and it had a meta key combo to do it for us!
<kazinsal>
(disclaimer: "back in mah day" was actually 1440 kB, we had some 5.25" stuff when I was a wee proto computer toucher but it was mostly just a curiosity)
GeDaMo has quit [Ping timeout: 260 seconds]
CaCode has joined #osdev
bauen1 has joined #osdev
nyah has joined #osdev
CaCode has quit [Ping timeout: 272 seconds]
CaCode has joined #osdev
<clever>
kazinsal: does a dedicated 6502 cpu in the floppy drive count as having to bit-bang the write heads? :D
Left_Turn has joined #osdev
Osmten has joined #osdev
Burgundy has joined #osdev
gog has joined #osdev
<totally-not-mjg>
where is heat when i have somethin' juicy
<mcrod`>
ffunction-sections generates a separate section for each function, then the linker removes unused functions at link time
<heat>
your job as a shitty C userspace programmer is not to stick all sorts of opinionated compile options unless you really need to
<heat>
spoiler: --gc-sections takes time
<mcrod`>
it's virtually imperceptible right now
<mcrod`>
wtf shitty
<mcrod`>
you're shitty :(
<heat>
you're going full google
<heat>
never go full google, at least without having millions and millions of lines of code first
<heat>
if you want so much control over a build, ship over a bunch of prebuilts and use gn (ignores CC and CFLAGS!)
<mcrod`>
why wouldn't I want this type of control whenever I pleased
<mcrod`>
anything else is awful
GeDaMo has quit [Ping timeout: 248 seconds]
<mcrod`>
and so far this is working okay, barring some CMake stupidity I have to figure out
<Ermine>
CMAKE
<heat>
the more control you have, the less the user does
<mcrod`>
that's crap
<heat>
it's not
<mcrod`>
define "the user" here
<Ermine>
just build the fucking binary and don't mess with flags
<mcrod`>
no
<heat>
$whoami
<mcrod`>
flags matter
<heat>
they do not
<mcrod`>
are you smoking something
<mcrod`>
do you not build with -Wall -Wextra
<heat>
those are warnings
<heat>
i do not build with -Werror
<mcrod`>
they are warning flags to be precise
<mcrod`>
so, sorry if I got that confused
<mcrod`>
yeah, I build with -Werror but I make sure that the really annoying ones remain passable warnings
<puck>
[updates compiler] [new warning appears, instantly breaks your build]
<heat>
my build is horrendous and huuuuuuuuuuuuuugely complex but it's what I have to do in order to build a full system
<puck>
at best, restrict -Werror to debug builds
<heat>
and in fact I do *need* the control here, and I still give you kernel config knobs to give you more control
<Ermine>
If someone needs get functions in different sections, they would do it
<mcrod`>
puck: i manually control the compiler used for the project
<Ermine>
Otherwise someone will definitely get inexplicable build failures
<heat>
<insert joke here>
<mcrod`>
so if that *would* happen, it wouldn't be instant death for someone who did pacman -Syu
<mcrod`>
there is no "someone" here
<mcrod`>
keep this in mind
<Ermine>
Your code is in public lol
<mcrod`>
yeah, and?
<heat>
if a random project made me build or download a whole other toolchain i would rm -rf it straight away
<mcrod`>
that's fine
<heat>
i do it for onyx because i have no other option
<heat>
you do it because you saw my build_toolchain.sh and thought "haha that's funny i want that"
<mcrod`>
no, i've been wanting to do something like that for a while
<mcrod`>
for what it's worth, you can certainly upgrade your compiler and shit breaks due to a compiler regression
vai has joined #osdev
<vai>
hi
<mcrod`>
hi
<vai>
anyone else noticed banks are sometimes pretty slow? I got slow, been waiting for some time.
<Ermine>
and stop drawing attention to your project if it's not intended for use by other people
<puck>
also it doesn't lolok llike your toolchain is actually portable
<vai>
someone stole me 2500 euros from paypal i.e. meant for software projects, etc.
<mcrod`>
puck: not really right now, no
<mcrod`>
that script fails straightaway on macOS
<puck>
even in linux distros
<totally-not-mjg>
heat: what if i told you openbsd does *NOT* fall over
<totally-not-mjg>
:]
<mcrod`>
ok that's crap
<mcrod`>
well, let me back up
<heat>
totally-not-mjg, should've mentioned that in the email to netbsd, maybe cc'd theo too
<mcrod`>
because i might be wrong
<mcrod`>
why do you think that?
<puck>
host libc, plus dynamic builds of clang/etc
<totally-not-mjg>
heat: i did not know at the time :(
<mcrod`>
yes, you need the host libc to build the software for the host
<puck>
no?
<mcrod`>
why would I not want to use the host libc?
bauen1 has joined #osdev
<puck>
because it depends on your host's version of glibc/musl
<totally-not-mjg>
heat: also, here is the REAL winner
<mcrod`>
that's correct, and every bit of software on that machine already is probably dependent on that host version's glibc/musl already
<totally-not-mjg>
heat: *ILLUMOS* is doing the bench faster than linux
vai is now known as Jari--
<netbsduser``>
that's not surprising
<totally-not-mjg>
heat: linux openbsds itself on a global inode hash lock
<puck>
mcrod`: then what's the point of compiling your own toolchain and having a fetch script?
<totally-not-mjg>
netbsduser``: this is the only case i know of where it is not losing by a large margin
sbalmos has quit [Ping timeout: 240 seconds]
<mcrod`>
because I can control *exactly* how each piece of software is built
<netbsduser``>
at Sun, it was to be considered a bug if Solaris didn't outperform Linux
<mcrod`>
and that's something I care about
<puck>
you'll have to run a glibc at least as new as the one it's compiled on
<totally-not-mjg>
netbsduser``: linux is outperforming solaris big time on almost everything
<totally-not-mjg>
so
<totally-not-mjg>
it is surprising something is left
<netbsduser``>
and i wouldn't say that linux and openbsd *pride* themselves on global inode hash lock
<mcrod`>
well, on paper, this is supposed to be compiled on ubuntu 22.04 LTS
<totally-not-mjg>
on that note, illumos is doing the workload in 37 seconds while freebsd in 18
<puck>
which means that you get to deal with the fun of older glibcs on non-rolling-release distros; or distros where the dynamic linker is located elsewhere, or distros that don't thave glibc
<totally-not-mjg>
and linux in 70
sbalmos has joined #osdev
<netbsduser``>
though i'm not actually sure what you are referring to with that
<mcrod`>
are any of those distros of any real worth?
<puck>
mcrod`: then why not use the toolchain that ubuntu 22.04 lts has
<heat>
mcrod`, trying to control everything just makes everything less flexible
<heat>
i think that's the point puck is making
<mcrod`>
i get what is being said
<puck>
and less portable, mostly
<mcrod`>
but I don't care about porting to the 540000892034924789 linux distros
<mcrod`>
because only 4-7 of them are actually really relevant
<netbsduser``>
one of the big deviations that 4.4BSD made from SunOS when it implemented the VFS was that they have this big LRU queue of inactive vnodes, while on solaris, vnode inactivation is intended to kill the vnode (individual filesystems might deviate)
<puck>
nixos is growing massively, but a fetched toolchain will not run on it
<mcrod`>
if I *really* cared about extreme portability, I wouldn't be using CMake, or targeting C17
<mcrod`>
it would be shell scripts and ANSI C
<puck>
>
<puck>
However, it should work just fine using C99, too.
<mcrod`>
yes
<mcrod`>
embedded people still stick to C99
<puck>
yes but are you trying to target those or not
<mcrod`>
that's a matter of practicality
<mcrod`>
yes
<puck>
then why are you targeting C17?
<mcrod`>
why not?
<heat>
because you also want to be C99 compatible
<mcrod`>
yes
<heat>
so use C99
<mcrod`>
no
dude12312414 has joined #osdev
<heat>
then delete that line
<mcrod`>
no
<sham1>
> VLAs
<heat>
you are C17 compatible and compatibleness with older standards is by-chance
<puck>
also
<puck>
your code seems to be very dependent on CHAR_BIT == 8
<puck>
which isn't very portable to DSPs
<puck>
tbf the DSP i used had a C11 compiler
<mcrod`>
yes, it is dependent on CHAR_BIT == 8, i don't really deny that, fair point
<puck>
..it wasn't clang or gcc, though
<sham1>
Proprietary compilers, oh joy
<heat>
hello mcrod` sir do u support armcc
<puck>
it was surprisingly pleasing to use
<mcrod`>
heat: no
<mcrod`>
puck: let me guess
<mcrod`>
IAR?
<puck>
no?
<puck>
arm has bytes
<puck>
that are 8 b its
<mcrod`>
.
<puck>
or, hrmmm, ah
<puck>
google betrayed me
<mcrod`>
i wasn't talking about arm.
<puck>
it wasn't IAR anyways
<mcrod`>
sucks
<mcrod`>
we could've been friends
<puck>
(first result googlle gave me was IAR Embedded Workbench for Arm)
<mcrod`>
we use IAR at work
<mcrod`>
yeah that'll happen
<puck>
also IAR does C17
<mcrod`>
heat: it's literally only clang and gcc
<netbsduser``>
i think there exists some outrageous and stupid implementation of "C" which is strictly conformant to the ISO language and deliberately defies all sense and sanity, the better to determine whose software really is written in ISO languaeg
<heat>
hello mcrod` do u support icc
<mcrod`>
no
<heat>
hello mcrod` do u support gcc 3.4
<puck>
this feels like a trick question
<puck>
(the ICC one)
<mcrod`>
oh
<mcrod`>
you might be right
<mcrod`>
cuz classic vs LLVM
<heat>
it is and is not
<puck>
eh, looks like icx is the oneapi one
<mcrod`>
anyway, the point of the toolchain script stuff was so you can git clone, `./toolchain_fetch.sh` and have everything you need, everything that I've tested and trust, and everyone is using the same shit
<sham1>
mcrod`: tcc support?
<mcrod`>
no I'm not going to support the 892374982378492939429347989 random compilers out there
<puck>
mcrod`: have you heard of our reproducible lord and saviour Nix
<mcrod`>
no
<sham1>
Guix > nix
<mcrod`>
for what it's worth i'm not trying to be arrogant
<sham1>
Damn proprietary apologists
<mcrod`>
i'm trying to understand any real problems that can emerge from this
<heat>
people can't use your shit
<puck>
mostly just people being really annoyed if they want to use your software
<heat>
that's the real problem
<mcrod`>
i don't see why they *can't*
<puck>
because now they have to deal with the fact that they have a more modern gcc locally that has a new warning that breaks your -Werror and they're not gonna go and either compile an entire LLVM or download one from an untrusted source
<puck>
they're just going to hack out the -Werror line
<heat>
because they may need to integrate it into another giant build and do not want to build toolchains for every piece of software
<heat>
like i'm fairly sure fedora would not package your software just because of that
<puck>
or NixOS, or arch, or .. really any distro
<sham1>
As long as it compiles with clang it's fine
<sham1>
We use gcc by default but clang is fine esp since it's ABI compatible
<netbsduser``>
i hate werror
<netbsduser``>
it is a beastly flag with beastly consequences for its use
<mcrod`>
being slightly a prick here, people could just take core.c/core.h and such and drop it in
<mcrod`>
i. understand. that's. not. the. point.
<sham1>
Unlike with c++ where ABI breaks can break *you*
<netbsduser``>
who among us has not had to use sed to patch out every refer to `-Werror' in a package to get a working build?
<Ermine>
netbsduser``: it's good when you're the developer and you want a clean build process
<heat>
sham1, C++ has no ABI breaks
<zid>
nearly everyone, who the fuck ships makefiles with Werror
<mcrod`>
the toolchain stuff was to also avoid someone doing `pacman -Syu` and something totally breaks that I can't control
<sham1>
Thankfully not needed that yet. But sed is a godsend in patching naughty builds in other ways
<mcrod`>
yeah I probably should make a SAMETHING_WARNINGS_ARE_ERRORS option instead
<mcrod`>
fuckers
<sham1>
heat: of course it doesn't. Not having an ABI makes that trivially true
<mcrod`>
also, re (un)trusted: then they shouldn't use the project to begin with if there are trust issues
<mcrod`>
and... that would be just fine
<heat>
sham1, the last time g++ had an ABI break was in gcc5 due to new C++ standard library requirements
<netbsduser``>
google i think tried to invent their own language to replace C++ after a failed attempt to win control over the committee and push for breaking changes
<heat>
and they didn't actually break it, just hacked around it
<netbsduser``>
no, not golang, dart, starklark, or any of the other google langs
<netbsduser``>
this one they called Carbon or something like that
<sham1>
Yes, carbon
<heat>
there's also cpp2 in msft's end
<sham1>
It's a weird thing
<netbsduser``>
to me a large part of the merit of C++ is that it is almost a C superset so there is very easy compatibility between the two
<pog>
restrict
<netbsduser``>
some people claim that you can get a compatibility by embedding an entire C frontend into your new language's compiler. well, show me a language that isn't C or C++ where i can use the C bsd sockets apis for manipulating ancillary data, the CMSG macros
<sham1>
Zig
<mcrod`>
restrict
<netbsduser``>
i don't think zig can handle C preprocessor macros (which in any case would expand to non-Zig code)
Turn_Left has joined #osdev
<sham1>
Zig does actually have a thing for translating c code
<heat>
the more i know about zig the less i want to
<netbsduser``>
i heard they were doing it before by embedding Clang (this is what i was alluding to)
<netbsduser``>
but now they want to invent their own back-end
<netbsduser``>
they have become alienated by LLVM
<netbsduser``>
so i'm curious as to whether the limited C compatibility till survive
Left_Turn has quit [Ping timeout: 240 seconds]
<sham1>
Probably. I actually don't know zig but that's just what I've heard some evangelists say
<sham1>
So grains of salt
GeDaMo has joined #osdev
CaCode has joined #osdev
Turn_Left has quit [Remote host closed the connection]
Turn_Left has joined #osdev
CaCode has quit [Client Quit]
Burgundy has joined #osdev
<heat>
rust
<zid>
RUST!?
<pounce>
🦀
<pog>
ruuuuuuuuuuuuuuuuuuuuuuuuuuuust
<mcrod`>
pog
<mcrod`>
hi
<pog>
hi mcpog
<mcrod`>
i want mcdonalds now
<zid>
I never don't
<mcrod`>
puck: i am curious as to what that compiler was
<zid>
oh, heat, did you pick me out a birthday present yet
<bslsk05>
mail-index.netbsd.org: Failing installations for various NetBSD versions on MicroVAX 3100
<geist>
yeah that's kinda my experience. tis why i have netbsd 3 on mine
<heat_>
we're all doomed
<heat_>
totally-not-mjg, you should've ran your experiment on the VAX
<heat_>
or a 486DX
<geist>
but really the cpu is too slow to run much newer than that anyway
<geist>
even netbsd 3 is pretty bad.
<totally-not-mjg>
you will be pleased to learn they recently committed a patchset to improve vax toolchain
<geist>
the microwax KA42 in mine is a really slow cpu, even by vax standards. it's basically a late model cost reduced thing, which is why you see them on ebay
<totally-not-mjg>
ooff
<geist>
probably 486sx territory
<geist>
well, okay it'd have been contemporary to a 486 anyway, since we're talking like 1990
<totally-not-mjg>
now that you mention it though, i genuinely have no idea how it compares to early x86
<geist>
but 25Mhz
<totally-not-mjg>
heh
<geist>
4KB cache, etc
<totally-not-mjg>
thank you for answering before i asked
<geist>
there were much faster vaxen in the server world by then
<totally-not-mjg>
so ye, that would be pretty bad indeed
<geist>
well over 100mhz, etc
<heat_>
whats up with the -en plurals
<heat_>
pluralen
<pog>
german
<pog>
deutsche
<totally-not-mjg>
i fully admit i have no idea, i remember viro and probably some other people referring to "boxes" as "boxen"
<totally-not-mjg>
i thught it would be funny to perpetuate
<pog>
þýsku
<totally-not-mjg>
i have no idea about geist
<geist>
yah i think it shows up with things tht end with -x or other harsh vowels
<heat_>
yeah but I think vaxen is indeed THE plural for vax
<geist>
vaxes would work, but i've heard folks say vaxen
<totally-not-mjg>
vexing innit
<geist>
but yeah i think it's some archaic germanic influence
<mcrod`>
first off I don't want anything other than the GUI stuffs
<zid>
qt is a mess of a project to begin with
<mcrod`>
and not *only* do I have to provide the longest -skip ever
<mcrod`>
but I also need to provide maybe 60-90 -no-feature options
<mcrod`>
because otherwise, stupid shit like CUPS support will become part of qtbase
<heat_>
to build a C++ project you need: 1) a custom patched toolchain 2) 2-10 hours depending on compile options 3) luck 4) lots of memory 5) lube
<zid>
USE="-qt" emerge qt
<mcrod`>
it'd be nice if I had an EPYC or something
<zid>
heat_: you forgot 128GB of ram, and 14 attempts, to nail down the --with-
<mcrod`>
heat_: don't worry, I'm almost ready to nuke the toolchain idea entirely
<heat_>
zid, you're *lucky* if it uses autoconf
<mcrod`>
for the simple fact that I'm sitting here waiting for shit to go wrong, and then it does
<mcrod`>
except 2 hours later
<zid>
whatever the cshitmake equivalent is
<mcrod`>
then I try again
<mcrod`>
then it fails again
<zid>
I don't know, I refuse to touch it
<mcrod`>
by the time it's all over
<mcrod`>
it's dinner
<mcrod`>
sick of it
<heat_>
it probably uses cmake and bespoke non-documented custom options, and when you attempt to regen cmake and rebuild it fails and you have to do it all over
pog has quit [Quit: Konversation terminated!]
<mcrod`>
it already failed
<mcrod`>
with default options
<mcrod`>
I'm not even using my toolchain anything
<mcrod`>
just -static -release
<mcrod`>
and host GNU
<heat_>
i bet 40 quadrillion USD on how it's -static fucking things up
Jari-- has quit [Ping timeout: 258 seconds]
<mcrod`>
that's what I'm going to test
<mcrod`>
I know that you *can* build a static Qt
<mcrod`>
but it probably fucks up something with qtwebengine, which it was compiling, despite it explicitly telling me it won't compile it because -static
<mcrod`>
this time I'm going to provide nothing but --prefix
<mcrod`>
and then cry
<heat_>
yes given enough memory luck toolchain patches time and lube anything can build
<mcrod`>
[QtWebEngine] -- Support check for QtWebEngine failed: node.js version 12 or later is required.
<mcrod`>
failed instantly
<mcrod`>
not even with qtwebengine, with qtgrcp
<mcrod`>
qtgrpc*
<mcrod`>
i'm done. done. i'm done with this.
<heat_>
if this were google software it'd bundle its own nodejs
<heat_>
problem solved
<mcrod`>
yeah but when I want to do the same type of shit it all falls apart
<mcrod`>
fuck me
<mcrod`>
just unimaginable to me that this shit is more complicated than it has to be
deflated8837 is now known as cow321
<heat_>
this is C++, it's implied
<mcrod`>
you guys win with the toolchain stuff by the way
cow321 has quit [Remote host closed the connection]
cow321 has joined #osdev
<mcrod`>
it would've been nice to not include 56000 things that a library provides that is simply not necessary or needed
<mcrod`>
but that pain isn't worth it.
<mcrod`>
toolchain people are geniuses
<Ermine>
Happy bday heat !
dude12312414 has quit [Quit: THE RAM IS TOO DAMN HIGH]
heat__ has joined #osdev
<zid>
mcrod`: how do you know when heat's birthday is?
<mcrod`>
he told me
<zid>
omg
<zid>
are you two secretly dating
heat_ has quit [Read error: Connection reset by peer]
<heat__>
Ermine, thanks
<heat__>
you're a bit far off but i appreciate it anyway :p
heat__ is now known as heat
<heat>
zid, no i don't date C++ programmers
<zid>
that's why it'd be a secret
<zid>
a dirty secret
<mcrod`>
ugh
<heat>
i want a girl who can i can show flamegraphs to
<heat>
and she'll be like "oh that's cool" and i'll say "yeah"
<Ermine>
heat: I'll market and be on time next time
masoudd has quit [Read error: Connection reset by peer]
masoudd has joined #osdev
masoudd has quit [Client Quit]
craigo has quit [Remote host closed the connection]
heat has quit [Read error: Connection reset by peer]
heat has joined #osdev
friedy has joined #osdev
gildasio has quit [Remote host closed the connection]
gildasio has joined #osdev
GeDaMo has quit [Quit: That's it, you people have stood in my way long enough! I'm going to clown college!]
goliath has joined #osdev
heat has quit [Remote host closed the connection]
heat has joined #osdev
<SophiaNya>
hey osdev, vsauce here
<SophiaNya>
but what _is_ a shinji?
<definitely-mjg>
funny story mate, i'm liek 2 levels down in youtube debunk
<definitely-mjg>
person posts a video, there is a debunk
<definitely-mjg>
and there is a debunk of said debunk
<definitely-mjg>
at this point i trust noone
<definitely-mjg>
albeit the truth is out there
[itchyjunk] has joined #osdev
freakazoid332 has joined #osdev
heat_ has joined #osdev
heat has quit [Read error: Connection reset by peer]
<Ermine>
It's definitely not on youtube
<definitely-mjg>
it was total youtube mon
<zid>
SophiaNya: A miserable little pile of secrets
[itchyjunk] has quit [Remote host closed the connection]
[itchyjunk] has joined #osdev
sortiecat has joined #osdev
<Ermine>
youtube is much more useful for memes than for any content which requires any processing power?
sortie has quit [Ping timeout: 255 seconds]
<definitely-mjg>
at this point youtube is mainstream media
<definitely-mjg>
that is to say people intentionally misrepresent realities for money
gildasio has quit [Remote host closed the connection]
gildasio has joined #osdev
<Ermine>
As they do on TV, in newspapers, etc
<mcrod`>
buildjet is cool
<Ermine>
Well, on some TVs they go as far as representing alternative reality
<Ermine>
But, *very* strictly speaking, the actual scientific truth is so complex that you have to simplify it (and inherently misrepresent) so common folk could comprehend something
<gog>
nya
<mcrod`>
goggies
<mcrod`>
gogmania
<gog>
mcroddies
<mcrod`>
hi
<gog>
this forbes article i read claimed that youtube has "sophisticated technology" to detect adblocking
<gog>
it's probably some javascript bullshit, how sophisticated can it really be
<mcrod`>
i'm confident the ublock origin people will kill it
<gog>
i updated ublock lists and it sort of works
<mcrod`>
for fun
<mcrod`>
i'm about to see how long the toolchain takes to build with buildjet
<mcrod`>
32 vCPU/64 GB
<mcrod`>
what concerns me is
<mcrod`>
i'm on chrome
<mcrod`>
so I wonder why they haven't just polled what extensions I'm using and see "ublock origin" and say "no"
<gog>
¯\_(ツ)_/¯
<gog>
can they do that?
<gog>
seems like a good way to add a another datapoint to analytics tracking
<mcrod`>
i don't know if they can, or do
<mcrod`>
but if not, I wouldn't be surprised if that's next on their radar
<mcrod`>
the nuclear option is google demanding ublock origin respect google advertisements or be pulled from the extensions store
qubasa has quit [Ping timeout: 240 seconds]
<heat_>
my ublock still works on youtube
<mcrod`>
i get the banner
<nortti>
are you logged in?
<mcrod`>
me?
<nortti>
both you and heat
<mcrod`>
yes
<heat_>
yes
<nortti>
oh interesting
* nortti
had a hypothesis that is was targetting only people who're logged in at this point
<heat_>
i'm in the EU though
<heat_>
maybe they can't/haven't rolled it out in the EU
<nortti>
a friend #elsewhere is getting it while being in finland
<heat_>
is finland EU?
<nortti>
yes, since the 90s
<heat_>
ah ok. yeah it's hard to track which one of you NORDIC BASTARDS are part of the EU
<mcrod`>
we have friends?
<heat_>
since like half of the nordic countries aren't but still have close ties
<mcrod`>
holy
<mcrod`>
fucking
<mcrod`>
shit
<mcrod`>
buildjet build LLVM in 20 minutes
<mcrod`>
*with* LTO
<heat_>
you sure?
<heat_>
that seems unrealistic
<mcrod`>
i don't believe it
<mcrod`>
let me check the logs
<heat_>
thin lto or lto?
<mcrod`>
thin
<mcrod`>
nope, it definitely did
<mcrod`>
cmake/llvm/sdl etc, 20 minutes
<heat_>
oh ok thin is a lot easier to get done though
<heat_>
true LTO is still a bitch
<mcrod`>
oh I sent you the logs
<mcrod`>
for what it's worth the buildjet runner i'm using is $0.048/min
<mcrod`>
so I only spent a dollar
<mcrod`>
and they give you $5 to play with
heat_ has quit [Remote host closed the connection]
heat_ has joined #osdev
gbowne1 has joined #osdev
sbalmos has quit [Quit: WeeChat 4.0.5]
heat_ has quit [Remote host closed the connection]
heat_ has joined #osdev
tixlegeek has joined #osdev
heat_ has quit [Remote host closed the connection]
heat_ has joined #osdev
danilogondolfo has quit [Remote host closed the connection]
gabi-250 has quit [Remote host closed the connection]
gabi-250 has joined #osdev
tixlegeek has quit [Quit: tixlegeek]
pretty_dumm_guy has joined #osdev
Burgundy has quit [Ping timeout: 258 seconds]
Arthuria has joined #osdev
gog has quit [Ping timeout: 240 seconds]
Arthuria has quit [Killed (NickServ (GHOST command used by Guest684531))]