<geist>
also between core and socket sometimes there's a cluster
<heat>
what's a core cluster?
<heat>
like a chiplet?
<heat>
fwiw i always thought zen chiplets were somewhat organized in NUMA domains but apparently not
Dead_Bush_Sanpa1 has joined #osdev
Dead_Bush_Sanpai has quit [Ping timeout: 265 seconds]
Dead_Bush_Sanpai has joined #osdev
<geist>
well not really in x86 (though maybe not) but would be a lot more in the ARM world
<geist>
especially with shared L3s
<geist>
actually Zen2+ yeah would work out each chiplet would be a cluster
<geist>
was only Zen 1 that had the NUMA style arch. Zen2 they moved the memory controller out of the cpu chiplet into a separate io chiplet
<geist>
but then each chiplet talks to each other via the io one, so there's a 2 tier architecture there
Dead_Bush_Sanpa1 has quit [Ping timeout: 248 seconds]
dysthesis has joined #osdev
<heat>
one of these days i need to figure out how to parse the x86 topology out of cpuid
<heat>
when i eventually add scheduling domains to my scheduler
<Matt|home>
o\
sbalmos has quit [Quit: WeeChat 4.5.0]
sbalmos has joined #osdev
<geist>
yeah i got the SMP x86 stuff working on LK enough that i forgot how much completele nonsense it is to learn the topology, find all the clocks, etc
<geist>
like 5 ways to get each things, ugh
<karenw>
5 ways to perform the same task? In my x86? Surely not.
<geist>
sometimes i'm in the mood to just type all of it in and it's kinda fun
<geist>
but sometimes it feels like a slog
sbalmos has quit [Ping timeout: 272 seconds]
Lucretia has quit [Remote host closed the connection]
sbalmos has joined #osdev
<heat>
the children yearn for x86 timer calibration
<heat>
i need HPET support since the PIT seems to be in complete bozo mode for a bunch of hardware
<heat>
quick children, get it done
<sbalmos>
time is such a relative, fluid construct
dysthesis has quit [Remote host closed the connection]
<sbalmos>
ugh, and then the concept of unified power management. just got bit by the fact that GDM has separate auto-sleep settings than the rest of Gnome, and it naturally doesn't respect not being logged in at the GUI console, but an SSH session open, since it has no concept / visibility to CLI-only connections
sbalmos has quit [Ping timeout: 246 seconds]
sbalmos has joined #osdev
goliath has quit [Quit: SIGSEGV]
dysthesis has joined #osdev
steelswords9436 has quit [Read error: Connection reset by peer]
<heat>
they finally clocked onto the amd fun mmu stuff
<geist>
yeah guess so
<karenw>
Oooh, fun stuff?
<heat>
also it seems that cores may get noticeably a little slower when handling remote invlpgb's
<heat>
which kinda makes sense? i guess it operates in a mailbox kind of way at the hardware level
<heat>
karenw, yeah newer amd has a fun invlpgb instruction that does the tlb shootdown for you, asynchronously, arm-style
<heat>
and these translation cache extensions bits are a little older but relax the TLB requirements for invlpg
<heat>
it's x86 tradition to invalidate page tables (all levels, all of them) when invalidating a single address
<karenw>
Oh I saw a link about that. 80% faster in synthetic benchmarks iirc. Nice win for some workloads I guess.
<heat>
this relaxes it into "invalidate all levels that map the given address"
<heat>
yeah i guess you saw phoronix's garbage articles
<heat>
we meme'd a 2000% speedup patch in #mm some months ago lol
<karenw>
You said phoronix twice. And no, I saw a mailing list message
<karenw>
*lkml
<karenw>
I have an infinity% speed up for all pcs. Just don't use them.
<karenw>
There, job complete.
<heat>
based
<heat>
fwiw this is good for certain kinds of workloads
<heat>
but ofc workloads dont spin in a for (;;) {allocate_and_free_memory()}
<zid>
across cores!
<zid>
importanterly
<zid>
It's one of those "ofc it's slow, pls avoid" things to begin with
<zid>
making the slow path faster kind of patch
<heat>
i mean, it is plausible it's important
<karenw>
heat: Yeah, why I'm careful to always mention if a workload is synthetic or not.
duderonomy has quit [Ping timeout: 272 seconds]
<heat>
rik was saying tlb flushes were very noticeable in facebook prod
<karenw>
And 80% improvement on a synthetic test is good, but it's nothing like 80% real world
<heat>
if malloc is any good you probably wont see much improvement, but there's other stuff
<heat>
like thread spawning and destruction
<heat>
JITters
duderonomy_ has joined #osdev
<zid>
that's because facebook runs php on huge webserver cpus
<zid>
ofc tlb shit is slow
<zid>
When you've got 96 cores all running the same process
hwpplayer1 has quit [Remote host closed the connection]
<karenw>
Don't they use their own PHP interpreter? Somehow that was less work than not using that god awful language.
<zid>
zend or whaever?
<heat>
yeah they have their own high performance stuff
<zid>
high-er :P
<heat>
surely it doesn't fucking suck that hard
<heat>
>96 cores all running the same process
hwpplayer1 has joined #osdev
edr has quit [Quit: Leaving]
<heat>
yeaahhhhhhhhhh that's what like most linux servers do
<zid>
webserver cpus can suck a dick
<zid>
gave us THE CLOUD and ruined desktop
<heat>
woah found a cloud hater
<heat>
4 choices if you work for a company that contributes to linux
<heat>
1) the embedded arm32 guy (you've never seen a multicore system in your life and LOVE mmc)
<heat>
2) the phone guy (you're also probably chinese)
<heat>
3) web servar on all corez
<heat>
4) qemu on all corez
<zid>
They should make them all buy 8 socket mobos again and limit them to 6 cores
<zid>
if you want 96 cores, buy two of those
<zid>
instead of a $15000 xeon, get sixteen $600 cpus, and maybe they get decent clockspeeds
<heat>
i think you're just jealous you dont have a 96 core system
<zid>
I barely used my 6 core
<heat>
that's because you're lame
<zid>
now I have 8 and I definitely don't
<heat>
get a cronjob that fires up cinebench every 30 minutes
<zid>
clockspeed is lower than my 6 though :(
qubasa has quit [Ping timeout: 246 seconds]
heat has quit [Ping timeout: 246 seconds]
hwpplayer1 has quit [Remote host closed the connection]
<ring0_starr>
hmm
<ring0_starr>
are there any hobbyist OSes written for PA-RISC?
<ring0_starr>
are there any OSes written for PA-RISC other than hp-ux and netbsd for that matter
hwpplayer1 has joined #osdev
dysthesis has quit [Ping timeout: 264 seconds]
<kof673>
there is older bsd i believe, nextstep (but think it was dropped for openstep) ....digital unix ...probably some obscure university projects
<kof673>
internet says "SPP/UX" "Many Unix and similar operating systems were available for PA-RISC from the 1980s on, first HP-UX, followed by many Mach and BSD research projects."
<ring0_starr>
problem is if i wanted to do something like this i wouldn't be able to make it smp
<kof673>
and there is/was linux
<ring0_starr>
the hardware i have is single core
<ring0_starr>
15ish years ago i could've probably splurged an extra $20 or something and gotten a really beefy machine
<ring0_starr>
now vintage everything is +$5000000
<ring0_starr>
so many regrets, so much hardware thrown out as garbage... man I threw out a vt220 that only needed a fuse because it'd cost too much to ship back home
dysthesis has joined #osdev
vdamewood has joined #osdev
dysthesis has quit [Remote host closed the connection]
hwpplayer1 has quit [Quit: ERC 5.5.0.29.1 (IRC client for GNU Emacs 29.4)]
karenw has quit [Ping timeout: 265 seconds]
<geist>
yeah that's true
<geist>
i did kinda a purge of machines with screens on them a while back because they were hard to ship
<geist>
though i didn't toss em as much as i gave em to someone else
craigo has joined #osdev
hwpplayer1 has joined #osdev
tjf has quit [Ping timeout: 252 seconds]
<ring0_starr>
i'm shocked that there's, relatively speaking, not too many people interested in vintage/historical stuff. i especially love reading old documents on the new thing that promises the world and ended up going nowhere, what it set out to solve, how they approached it, how/why it failed, etc.
<ring0_starr>
just one minor twist of fate and we might've been living in a world dominated by some obscure architecture instead of the arm we have today
<ring0_starr>
not saying i don't appreciate arm or even like it, it's just so ubiquitous that it becomes unexciting
<ring0_starr>
history feels more important now than ever because the younger generations especially don't seem to give a crap because they're so used of churn in things
<ring0_starr>
you see javascript frameworks reinventing the wheel every 2-3 years it seems
Dead_Bush_Sanpai has quit [Read error: Connection reset by peer]
Dead_Bush_Sanpa1 has joined #osdev
<ring0_starr>
they think they came up with something new and incredible when it had been done much better and had more adoption 40 or 50 years ago
<ring0_starr>
so i mean similar story with cpu architectures. just seems like civilization is going to forget what had been tried back then and it didn't work out for one reason or another, but due to modern circumstances boosting it up, the objectively crappier model of whatever could win out and we'd essentially devolve
Dead_Bush_Sanpa1 is now known as Dead_Bush_Sanpai
tjf has joined #osdev
<kof673>
there is a nice #virtuallyfun (neozeed, super hyper global mega corp blog, would virtualize many things) but the irc relay bot is currently down...goes to a discord i believe.
\Test_User has quit [Quit: \Test_User]
\Test_User has joined #osdev
\Test_User has quit [Client Quit]
\Test_User has joined #osdev
kfv has joined #osdev
vinleod has joined #osdev
vdamewood has quit [Ping timeout: 260 seconds]
vinleod is now known as vdamewood
emntn has quit [Ping timeout: 252 seconds]
housemate has joined #osdev
antranigv has quit [Ping timeout: 246 seconds]
housemate has quit [Remote host closed the connection]
ThinkT510 has quit [Quit: WeeChat 4.5.0]
kfv has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
ThinkT510 has joined #osdev
antranigv has joined #osdev
fedaykin has quit [Quit: leaving]
kfv has joined #osdev
kfv_ has joined #osdev
fedaykin has joined #osdev
kfv has quit [Ping timeout: 252 seconds]
larsm has joined #osdev
kfv_ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
kfv has joined #osdev
netbsduser has joined #osdev
GeDaMo has joined #osdev
hwpplayer1 has quit [Remote host closed the connection]
kfv has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Lucretia has joined #osdev
hwpplayer1 has joined #osdev
<nikolar>
ring0_starr: look into iapx 432
<nikolar>
It was nuts
emntn has joined #osdev
<zid>
nikolar is currently nuts
<zid>
I am will be being nuts from whence cometh beginning to have had going to be done has.
kfv has joined #osdev
<nikolar>
i think zid is having a stronk
kfv has quit [Remote host closed the connection]
kfv has joined #osdev
craigo has quit [Ping timeout: 252 seconds]
frytaped has joined #osdev
remexre has joined #osdev
housemate has joined #osdev
remexre has quit [Ping timeout: 265 seconds]
hwpplayer1 has quit [Remote host closed the connection]
housemate has quit [Quit: Nothing to see here. I wasn't there. I take IRC seriously.]
goliath has joined #osdev
TkTech has quit [Read error: Connection reset by peer]
TkTech has joined #osdev
hwpplayer1 has joined #osdev
kfv has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
hwpplayer1 has quit [Quit: brb]
netbsduser has quit [Ping timeout: 248 seconds]
hwpplayer1 has joined #osdev
jedesa has joined #osdev
larsm has quit [Ping timeout: 272 seconds]
kfv has joined #osdev
larsm has joined #osdev
aejsmith has quit [Quit: Lost terminal]
FreeFull has quit []
aejsmith has joined #osdev
kfv has quit [Client Quit]
netbsduser has joined #osdev
larsm has quit [Ping timeout: 260 seconds]
larsm has joined #osdev
Turn_Left has joined #osdev
Turn_Left has quit [Read error: Connection reset by peer]
Left_Turn has joined #osdev
FreeFull has joined #osdev
jedesa has quit [Remote host closed the connection]
Turn_Left has joined #osdev
Left_Turn has quit [Ping timeout: 276 seconds]
JerryXiao has quit [Ping timeout: 246 seconds]
JerryXiao has joined #osdev
duderonomy_ has quit [Ping timeout: 246 seconds]
dysthesis has joined #osdev
hwpplayer1 has quit [Quit: I'll be back later]
dequbed has quit [Quit: bye!]
nadja has joined #osdev
dysthesis has quit [Remote host closed the connection]
kfv has joined #osdev
foudfou has quit [Quit: Bye]
nadja has quit [Ping timeout: 246 seconds]
seds has joined #osdev
nadja has joined #osdev
kfv has quit [Read error: Connection reset by peer]
nadja has quit [Client Quit]
housemate has joined #osdev
kfv has joined #osdev
nadja has joined #osdev
kfv has quit [Read error: Connection reset by peer]
kfv has joined #osdev
housemate_ has joined #osdev
housemate has quit [Ping timeout: 245 seconds]
bauen1 has quit [Ping timeout: 244 seconds]
netbsduser has quit [Ping timeout: 260 seconds]
kfv_ has joined #osdev
xenos1984 has quit [Ping timeout: 248 seconds]
kfv has quit [Ping timeout: 260 seconds]
xenos1984 has joined #osdev
netbsduser has joined #osdev
hwpplayer1 has joined #osdev
hwpplayer1 has quit [Remote host closed the connection]
frankenstein__ has joined #osdev
housemate_ has quit [Ping timeout: 252 seconds]
frankenstein__ has quit [Quit: Nothing to see here. I wasn't there. I take IRC seriously.]
chiselfuse has quit [Remote host closed the connection]
kfv has joined #osdev
chiselfuse has joined #osdev
kfv has quit [Remote host closed the connection]
xenos1984 has quit [Ping timeout: 272 seconds]
xenos1984 has joined #osdev
pie_ has quit []
vancz has quit []
vancz has joined #osdev
pie_ has joined #osdev
wereii has quit [Ping timeout: 252 seconds]
xenos1984 has quit [Ping timeout: 252 seconds]
wereii has joined #osdev
xenos1984 has joined #osdev
Dead_Bush_Sanpa1 has joined #osdev
Dead_Bush_Sanpai has quit [Ping timeout: 252 seconds]
Dead_Bush_Sanpa1 is now known as Dead_Bush_Sanpai
Turn_Left has quit [Ping timeout: 245 seconds]
Matt|home has quit [Quit: Matt|home]
Turn_Left has joined #osdev
Turn_Left has quit [Ping timeout: 265 seconds]
eluks has quit [Remote host closed the connection]
eluks has joined #osdev
eluks has quit [Remote host closed the connection]
eluks has joined #osdev
Left_Turn has joined #osdev
remexre has joined #osdev
cloudowind has quit [Ping timeout: 265 seconds]
<sortie>
okay so there's that level of self-hosting where you can run gcc on your OS, then there's the level where you can run nginx and the website, then there's a level where you're building everything natively and got your own irc network, now ... has anyone run their production email infrastructure on their own OS
remexre has quit [Ping timeout: 276 seconds]
<sortie>
The prospect of porting postfix both seems like surely it can't be that hard and also likely will lead to endless horrors
<sortie>
Self-hosting email is also extra terrifying cus if it does gown you're fucked
cloudowind has joined #osdev
<sortie>
2FA? Oh, sorry, my kernel paniced.
<nikolar>
lel
<the_oz>
I'd rather port openbsd maild, whatever it is
<the_oz>
seems a good starting point
thinkpol has quit [Remote host closed the connection]
thinkpol has joined #osdev
thinkpol has quit [Remote host closed the connection]
thinkpol has joined #osdev
duderonomy has joined #osdev
<heat>
is it really an email stack if it doesn't include copious amounts of perl?
frytaped has quit [Ping timeout: 252 seconds]
ponup has joined #osdev
GeDaMo has quit [Quit: 0wt 0f v0w3ls.]
monkeyPlus has joined #osdev
<monkeyPlus>
hey. Instead of gettings share libraries on localhost, why dont use a remote server to get them? All we need was .text,.data,.bss, and preferred addresses (in case were talking about elf). The linkage, would be done in localhost...what do you think of it?
<heat>
what's the advantage?
<monkeyPlus>
you could access say, a shared lib, on a remote server...say instead of downloading a program, download some shared libs..
<monkeyPlus>
say, call a internet host library
<monkeyPlus>
say, zip. If you had some remote libraries, we could use like zip library
<heat>
yes but, again, what's the advantage?
<monkeyPlus>
you can call remote libraries
steelswords9436 has joined #osdev
<heat>
what's the difference between downloading? isn't this inherently unsafe?
<monkeyPlus>
it is unsafe..thats one problem
<heat>
to me it seems like it's a solution looking for a problem
<monkeyPlus>
ok..just wondering :))
<heat>
one thing i've thought about is lazy software installs
<heat>
like instead of having a /bin/gcc installed, you'd have a script that downloads and installs gcc itself
<bslsk05>
​learn.microsoft.com: HTTP.sys web server implementation in ASP.NET Core | Microsoft Learn
<heat>
oh nice it supports HTTP/2, TLS, even gRPC
<monkeyPlus>
one advantage, no need to store libraries on HD. On a lan (e.g 255 hosts), thats signficant. Another, say you only need one library, from a remote program...you could use only tat
<heat>
big disadvantage: network goes out and you ded
<monkeyPlus>
can happen to HD as well
<monkeyPlus>
we do stuff, supposing that something works
<monkeyPlus>
if youre going to consider, something that brokes, you wouldnt do anything
<heat>
no it doesnt, if my network goes out i can still use my computer
<heat>
my network is far more unreliable than my hard drive
<monkeyPlus>
ah i see the point
<heat>
zid has been using the same hard drive for like 20 years
<monkeyPlus>
whos zid
<heat>
a person, i think
<monkeyPlus>
ok
karenw has joined #osdev
<nikolar>
heat, you think?
<Ermine>
it's _the_ person
<nikolar>
it?
<heat>
i dont have concrete proof of zid's existance
<nikolar>
well something called zid exists
<heat>
existence
<heat>
yeah but is it a person
<heat>
or just a really smart cat
<nikolar>
good point
<heat>
like a proper glasses-wearing cat
<heat>
with a tie
<CompanionCube>
also networks can involve shenanigans.
<CompanionCube>
Imagine your shared libraries getting blocked by a captive portal, lmao
<nikolar>
networking sucks to be honest
seds has quit [Ping timeout: 265 seconds]
<heat>
networking is great
xenos1984 has quit [Read error: Connection reset by peer]
<nikolar>
except when it isn't
<nikolar>
because it's unreliable as heck
<heat>
communicating with people far away is really cool, easily humanity's biggest achievement
<karenw>
Internetworking has made a lot of people very angry and been widely regarded as a bad move.
<nikolar>
heat, what i meant is that working with network sucks
<nikolar>
not that the idea itself is bad kek
<heat>
oh. yeah
<heat>
i mean
<heat>
it depends doesn't it
<Ermine>
heat: but now more people need to touch grass
<nikolar>
kek
<heat>
simple, point-to-point communication is really not that complex, usually
<nikolar>
sure
<nikolar>
except you always need to deal with drop packets and whatnot
<heat>
ofc you can make networking as complex as you'd like with lans and lans on top of lans and nat and natnat and natnatnat and tap and tuns and bridges and vlans and shit
<nikolar>
ye
<heat>
tcp solved dropping packets!
<karenw>
Real life sucks. Have you seen how few polygons the grass has? And don't get me started on water.
<heat>
tcp really aged well
<nikolar>
tcp is great
<nikolar>
and now google wants to replace it with quic
<nikolar>
you're welcome
<heat>
it's not just google
<nikolar>
it's very much primarly google
<heat>
not really
<karenw>
Google is the main player pushing for your entire OS to be web based.
<heat>
all big network players stan quic
<nikolar>
screw quic
<nikolar>
all my friends like tcp
<karenw>
And TCP has some downfalls, for some use cases. So why not re-invent TCP instead of just using something more appropriate!
<nikolar>
karenw: but hey, you save a roundtrip or two on initial connection
<nikolar>
that must be worth it, right
<heat>
yes actually
<nikolar>
highly doubt it
<heat>
a modern web page makes A LOT of http requests
<nikolar>
sure, i've never thought my web page loaded slowly because of the network
<heat>
and they're usually limited concurrency-wise
<nikolar>
modern web just sucks
<nikolar>
it's not the network's fault
<karenw>
See also: TLS 1.3's quick handshake. It turns out that sending data in the opening packet before establishing session keys is potentially a security risk, who'd have thunk it.
<nikolar>
kek
<heat>
and the web people will tell you that the network sucks
<heat>
it turns out everyone's a little correct
<nikolar>
except that the web people's opinions don't matter because they are the ones who made web suck :P
<karenw>
I feel like that "Stop using math" meme is appropriate here. But it's "stop using the web for non-web things"
<karenw>
The Internet is not just The World Wide Web
<heat>
what's a non-web thing?
<karenw>
IRC
<karenw>
Email
<heat>
then what is irc?
<Ermine>
MPLS MPLS MPLS MPLS
<nikolar>
Ermine: what was mpls again
<heat>
irc is pretty much by definition "tcp over two hosts"
<heat>
sometimes with tls
<Ermine>
nikolar: multiprotocol label switching
<nikolar>
oh yeah
<nikolar>
heat: does that make it web?
<the_oz>
back when I was young, the laguages like php4->5 was a trasitio away from "we've reimplemented the c stdlib" to "HAVE YOU HEARD OF OOP?"
<nikolar>
i am pretty sure web primarly refers to http+html+tcp+whatever other junk you have in your browser
<nikolar>
not just "shove bytes over the network"
<karenw>
The Web originally had a pretty narrow definition of hyperlinked pages served over HTTP.
<heat>
if you're using that definition then i dont know how irc can be web
<karenw>
IRC is not web
<nikolar>
heat: i assume he meant irc web clients
<heat>
you can use irc thru the web
<heat>
and that's alright
<heat>
it is actually useful
<the_oz>
How will Moolenar ev er live? :( buy mirc plx
<nikolar>
either way, i don't think irc on the web is the biggest issue with the web currently heat
<nikolar>
kek
<heat>
totally
<heat>
the web has some problems but a lot of those aren't really... consequential
<nikolar>
well, every webpage downloading megabytes of javascript junk is certainly a consequential problem
<karenw>
Having web pages that are just stubs for a javascript program which then gets and sends all it's data over HTTP connections instead of a classical program that uses appropriate protocols is the problem. Firefox and Chrome have become the target operating systems for applications now.
<heat>
i disagree
<nikolar>
you're wrong :P
<the_oz>
guys you know this hotness called ECMAscript? well now you can be retarded on the server too!
<heat>
im not
<heat>
two choices here:
<heat>
1) discord is a web page and they just distribute chrome
<heat>
2) discord is a proper program and you dont get any of it on non-windows and non-macos
<heat>
i vastly prefer 1)
<nikolar>
discord as a web page sucks ass
<the_oz>
the standalone is a reimplementation of the web version
<heat>
and, you know, for "the web" stuff: show someone a static html web page and a dynamic javascript web page
<heat>
90% of people will pick the dynamic js web page
<karenw>
Discord sucks ass. And if they really want to make it a web application, just leave it as a web application don't download your own security-bug ridden fork of chroe.
<karenw>
*chrome
<heat>
discord is good
<nikolar>
the_oz: reimplementation is the wrong word, standalone *IS* the the web version
<nikolar>
heat: it sucks
<heat>
you're on irc
<heat>
you can't get worse than this
<nikolar>
i am not talking about features
<the_oz>
the only thing discord excels at imho is their server invitation system
<heat>
i am
<nikolar>
i am talking about how it takes forever to do
<nikolar>
anything
<nikolar>
it's dog slow
<heat>
like what?
<nikolar>
on my normal ass laptop
<nikolar>
to load for a star
<nikolar>
*start
<heat>
i have 0 of those problems
<nikolar>
well i have 0 of those problems on my blood pc
<nikolar>
that's not the isse
<heat>
my only big grievance is that it doesn't really understand what "idle" means
<nikolar>
"this thing is slow" "just get a faster computer lol"
<heat>
im on a shitty 2018 laptop
<ponup>
I connected to this IRC channel today after many years without using IRC because I missed using software that has low memory/cpu footprint.
xenos1984 has joined #osdev
<the_oz>
nikolar: I call it reimplementation because it's not using a system web client, so like ... it reimplements EVERYTHING about the browser itself
<nikolar>
the_oz: discord isn't implemnting (almost) anything though
<ponup>
It's crazy to think that a chat app like slack consumes hundreds of RAM to run!
<nikolar>
they take chrome and normal discord web app and put files next to each other and zip them
<nikolar>
that's like 99% of what the "standalone" version is
<nikolar>
ponup: absolutely
<the_oz>
they might have stuff upstream
<heat>
yeah and the alternative is not having fucking discord
<heat>
or having to use wine and pray to god that it works
<the_oz>
but they're still doing it...
<heat>
like, youtube, believe it or not, i don't want to have to use ytdl + mpv to watch a fucking video
<heat>
even though this is a dreamy situation for a solid chunk of le UNIX community
<nikolar>
youtube sucks too
<the_oz>
your idea of alternative is weird
<nikolar>
and it keeps sucking noticably more and more
<the_oz>
alternative would be like ... matrix
<nikolar>
unlike discord, which has remained at the consistent level of suck as far as i can remember
<ponup>
what I dislike about youtube is that is centralized. I would rather use a p2p alternative
<heat>
no one uses matrix
<ponup>
(that and the ads!)
<the_oz>
no one uses it but it is an alternative to discord regardless
<heat>
i would not be able to speak with my windows-using friends if i did not have discord
<nikolar>
heat: well *some* people do
<heat>
not enough to matter
<ring0_starr>
generally speaking, nobody is willing to compromise for each other, partially because society has grown more narcissitic, but i think moreso because the chasm between options has become so much greater
<the_oz>
ad populism is only to say someone captures crowds better
<heat>
i tend to be really fucking practical. like, yeah, discord uses a lot of ram, totally, but if i did not have a discord client i would not be able to use it, and i do want to use it, so i prefer the status quo
<ring0_starr>
there is no happy middle ground between irc and discord, because they're too different at this point, and they have different ideals, different goals, and the ppl who refuse to try out irc are IMPOSSIBLE to change
<ring0_starr>
like even though it's easier to go from discord to irc than vice versa
<heat>
i can use a perfectly cromulent office suite (microsoft or google's) thru my web browser, which i'm not able to do using fucking libreoffice, so im happy with these things existing
<nikolar>
heat: well what if discord didn't take hundreds of megabytes of ram to display text in my browser
<nikolar>
that's also an option
<nikolar>
(not the one they would consider apparently)
<heat>
it is not, and you know it
<nikolar>
i mean it's the option for discord the company
<nikolar>
not me
<heat>
virtually no one writes desktop programs for linux except like... google and steam
<the_oz>
take a bot and turn it into a TUI
<nikolar>
heat: i am not even talking about desktop software
<the_oz>
or there are clients out there I'm sure
<nikolar>
why do they need hundreds of megabytes of ram to display mostly *text*
<heat>
because web browsers are really complicated
<nikolar>
heat it's text
<heat>
like, if you have a specific idea i'm sure the chrome or firefox people would like to hear it
<nikolar>
and images
<nikolar>
web browsers could do those 30 years ggo
<nikolar>
*ago
<ring0_starr>
it's also 400 nested <div> tags
<ring0_starr>
have you opened up the developer tools window in discord? it should become quickly apparent
<heat>
either the whole web browser industry has been completely asleep at the wheel, or the problem domain is way more complex than you think it is
<ring0_starr>
web developers are the problem, heat
<ring0_starr>
web DEVELOPERES
<heat>
i can't pinpoint you exact reasons why they use so much memory, because im not a web browser guy
<ring0_starr>
not the browsers themselves
<nikolar>
heat: the problem isn't chrome/firefox
<the_oz>
10,000 cycle per second on embrace extend extinguish superloops
<nikolar>
the problem is with the discord the webapp
<nikolar>
it's not that hard to display text and images in a web browser