klange changed the topic of #osdev to: Operating System Development || Don't ask to ask---just ask! || For 3+ LoC, use a pastebin (for example https://gist.github.com/) || Stats + Old logs: http://osdev-logs.qzx.com New Logs: https://libera.irclog.whitequark.org/osdev || Visit https://wiki.osdev.org and https://forum.osdev.org || Books: https://wiki.osdev.org/Books
X-Scale` has joined #osdev
X-Scale has quit [Ping timeout: 240 seconds]
X-Scale` is now known as X-Scale
nyah has quit [Ping timeout: 245 seconds]
<Jari--> hi all dudes
<emartinez> ]]Sup?
<bslsk05> ​www.army-technology.com: General Dynamics wins $883m US Army CPM Plus contract
<Jari--> just for the laughter of it
<Jari--> Anyone still remembers CP/M Plus?
<Jari--> Pretty neat on Commodore 128, everyone told my Commodore 128 looked like a "PC" with it!
<Clockface> ive thought about getting one of those 8051 derivative microcontrollers and hooking it up to some peripherals with CP/M
<Clockface> maybe it could be a calculator
<Clockface> i know it is designed to do everything through pretty simple BIOS, meaning you could put it on a lot of weird devices
ElectronApps has joined #osdev
<mrvn> you can buy a Z80 and run CP/M
<emartinez> //buffer 1
[itchyjunk] has quit [Ping timeout: 240 seconds]
[itchyjunk] has joined #osdev
<geist> recommend a rc2014 z80 machine. lots of fun
<geist> and you get to build it!
vdamewood has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
pretty_dumm_guy has quit [Quit: WeeChat 3.4]
<gog> i kinda want some little project like that
<gog> get one of those usb-powered solderin pens
dude12312414 has joined #osdev
dude12312414 has quit [Remote host closed the connection]
sprock has quit [Quit: brb]
sprock has joined #osdev
gog has quit [Quit: byee]
sprock has quit [Client Quit]
sprock has joined #osdev
sprock has quit [Client Quit]
nitrix has left #osdev [Leaving]
<bslsk05> ​thenewstack.io: SerenityOS: Building an Operating System from Scratch – The New Stack
sprock has joined #osdev
sonny has joined #osdev
sonny has quit [Remote host closed the connection]
sonny has joined #osdev
sonny has quit [Remote host closed the connection]
sprock has quit [Client Quit]
sprock has joined #osdev
JerryXia1 has quit [Quit: Bye]
JerryXiao has joined #osdev
sonny has joined #osdev
<energizer> idk how many people make their own OSes but i was surprised to hear the one getting attention is just rebuilding something that existed in the 90s
<energizer> i guess it's nostalgic for a lot of people
emartinez has quit [Quit: WeeChat 3.3]
<energizer> well, i suppose that's not quite right, it'll be better than the 90s version
<sonny> energizer what new ideas do you have?
<energizer> everybody's full of a million ideas, the hard part is implementation
<sonny> not really, I've been thinking about OS design but I didn't come up with much
<energizer> oh sorry
<sonny> nothing to apologize for
<sonny> I was just saying the ideas aren't a dime a dozen, at least in my experience
<sonny> I suspect the innovation is going to be at a lower level of abstraction, from overall design principles I think OS is already a well defined class or programs
<sonny> s/or/of/
<sonny> what I'd be interested in seeing is how to handle concurrency and parallelism in an OS, seems like theirs lots of room there
<sonny> s/theirs/there's/
<sonny> security and distributed computing seems to have arrived by trial and error :P
<sonny> also I'm not sure if the jury's still out on database filesystems
<moon-child> 'how to handle concurrency and parallelism in an OS' erlang?
<moon-child> memory protection is tyranny
<sonny> yes
<sonny> I was thinking of the actor model
<sonny> but I haven't explored much yet
<sonny> "memory protection is tyranny" -- yeah but the alternatives don't work with existing software
<energizer> not all OSes need to run all software
<moon-child> 1) too bad for existing software
<sonny> moon-child: what are your thoughts if it doesn't matter about legacy?
<moon-child> 2) don't they? I don't particularly like wasm, but it is becoming an increasingly common target
<moon-child> llvm, go, I think somebody also did js->wasm
<moon-child> that accounts for quite a bit
<sonny> I'm a bit confused, how did we get to wasm?
<moon-child> wasm is a way to run existing software without hardware memory protection
<sonny> oh
<sonny> the vm approach
<sonny> well, you got me there
<moon-child> everything is a vm
<sonny> iirc the ibm i has a vm in the kernel
<moon-child> c is a vm
<moon-child> x86 is a vm
<sonny> I think that was the original idea for .NET but it failed
<moon-child> read c spec. It refers to an 'abstract machine'. And x86 just happens to have an implementation in silicon
<bslsk05> ​en.wikipedia.org: Singularity (operating system) - Wikipedia
<sonny> curious, why don't they formalize the c abstract machine?
<moon-child> there have been attempts to formalize c semantics
<moon-child> I believe the k framework has one
<moon-child> and compcert of course
<sonny> moon-child yes, but also didn't M$ want apps to be done in C# at some point?
<moon-child> metamath 0 has a model of x86 machine code. And intel and amd have models, but those are not public
<Jari--> Any books on how to create a thread lock ?
<Jari--> I am using modules, functions, etc. as locking mechanism.
<sonny> moon-child ah ok
<moon-child> sonny: I don't know. Windows has been traditionally designed for modularity via 'subsystems'--nt subsystem, ce subsystem, posix subsystem, etc. I doubt they planned to give up on hardware memory protection
<moon-child> but could be wrong about that
<moon-child> Jari--: what exactly do you want to do? Or, rather--what exactly is it that you don't understand how to do?
<bslsk05> ​books.google.co.in: An Introduction to Parallel Programming - Peter S. Pacheco - Google Books
<Jari--> I am checking this out
<Jari--> sonny, thanks
<sonny> np
<energizer> are there any non-assembly languages that give direct control of caches?
<sonny> that a trick question?
<energizer> no
<moon-child> there are no assembly languages I know of that give direct control of caches
<sonny> the language isn't responsible for the control
<energizer> oh i see, i changed the question in the middle
<energizer> are there any languages that give direct control of caches?
<sonny> moon-child I'm curious, do you have another way to think about memory, or did you figure something out about VMs?
<sonny> It seems like most material is much in favour of memory protection and doesn't really discuss alternatives
<energizer> looks like gcc has __builtin_prefetch
<moon-child> sonny: I don't know of any literature on the topic, no
<moon-child> well, hmm, look up ocaps maybe
<moon-child> object capabilities
<sonny> thanks
<sonny> (though, capabilites seemed to have failed in general)
<moon-child> (also possibly http://metamodular.com/closos.pdf)
<moon-child> in what respect have capabilities failed?
<moon-child> an fd is a shitty capability, but it is nevertheless a capability
<sonny> til an fd is a capability
<energizer> http api keys are capabilities
<sonny> It seems that capabilities aren't pragmatic because they are too fine grain
<sonny> windows is built around object capabilities but somehow it doesn't seem to have helped
<energizer> you can choose how fine they are
<moon-child> (1: a file is a record; 2: a file is a byte steam; 3: wtf is ioctl())
<sonny> lol
<sonny> fair point
<sonny> http://bwl-website.s3-website.us-east-2.amazonaws.com/Slides/GoldAndFoolsGold.pdf is interesting if you have the time, just an overview of ideas for systems stuff by butler lampson
<moon-child> will take a look, thanks
<moon-child> https://0x0.st/oK0Y.png hahha
<sonny> yeah I've thought about that for a some time lol
<sonny> "it seems to have escaped the lab"
<moon-child> 'Reduce highway traffic deaths to zero' what if we just replace highways with trains? :P
<sonny> get rid of cars in NA? lmao
<moon-child> sure, but we're already dreaming
<moon-child> one of the dreams just seems a bit more realisable
<sonny> oh yeah lol
<moon-child> I feel like there are two distinct problems, which the slideshow is not quite acknowledging as distinct (maybe the presentation did, idk, but)
<moon-child> 1) how do we program the computers in airplanes and cars and spacecraft
<moon-child> 2) how do we program the computers we use to run texteditors and video games and irc clients
<energizer> 1) unikernel
<energizer> 2) in elisp
<energizer> [ok (2) is a joke]
<sonny> moon-child I hope their would be a general solution, but it almost seems like that's creating a taxonomy
<energizer> cache line management is very important
<energizer> the most important thing for some applications' performance.
<moon-child> sonny: I absolutely refuse to write automotive software without formally verifying it, and I absolutely refuse to formally verify a video game
<energizer> but i don't have great control of it
<moon-child> is that unreasonable?
<sonny> no, that's a good point
<sonny> verifying a video game sounds like dev hell
<moon-child> it might be a continuum rather than a discrete categorization, but I think that there are definitely qualitatively different tradeoffs to be made for different applications
<sonny> yeah, I agree
<sonny> it's really interesting, I am going to think about this some more
<sonny> I think alan kay may have done some work in this area too because I recall when he made a comparison to biological systems and said we havne't really made much progress in systems programming
<sonny> (the comparison to how biological systems scale)
<bslsk05> ​'"We Really Don't Know How to Compute!" - Gerald Sussman (2011)' by Strange Loop Conference (01:04:18)
<kingoffrance> another channel, someone said "glass cockpit stuff is running X on many commercial airliners." "usually vxworks" "For aviation, you have a usual designed lifetime of 30 years." i pointed out, what is the game lifespan, 6 months? this was about whether to replace X or not, not arguing either way, just agree with moon-child i suppose; different people have different needs
<kingoffrance> it is not that i dont think a general solution is possible, rather that's not a "market" it will be one side steamrolling the other i think
<sonny> ah
<sonny> yeah, I see what you mean
<kingoffrance> some giant unlimited budget research project, maybe could pull it off :D
<sonny> "multics II"
* moon-child writes 'struct O *restruct RO'
* moon-child bonks compiler
<moon-child> hmm, I want some sort of 'inline' lookup table
<moon-child> static inline int lut[4] = {7, 3, 1, 4}
<moon-child> if I write lut[2], I want it to be constant-folded to 1
<sonny> macos?
<moon-child> but I also want it to be merged with identical instances of 'lut' from other TUs
<moon-child> can I do that?
<sonny> not sure, that was my first thought
<moon-child> macos?
<moon-child> oh typo for macros?
<sonny> macros
<sonny> yeah, my bad
<moon-child> like ICE_P(x) ? ((int[]){7,3,1,4}[x]) : real_lut[x]?
<moon-child> could work
<energizer> i dont think game lifespan is so short
<moon-child> though I wonder if there's other analysis it can do. Even if it doesn't know what x actually is, maybe it calculates bounds on x and hence bounds on lut[x]?
<energizer> the biggest games have a $500M budget and they go for quite a while
<kingoffrance> energizer, sure, but how old of a video card or hardware can you use to run them
<bslsk05> ​en.wikipedia.org: List of most expensive video games to develop - Wikipedia
<kingoffrance> or, which console lasts 30 years?
<energizer> i'm not really plugged into the game world so idk how long consoles last these days
<energizer> i assume it's still ~6 years?
<kingoffrance> im not either, it was somewhat exaggerated
<sonny> moon-child like sizeof(x) ?
<energizer> just checked, 6 is about right for playstation
<moon-child> I mean suppose it knows x < 2. Then ideally it should also know that lut[x] is either 7 or 3, but not 1 or 4
<kingoffrance> point is of course moot if you have to be online/subscribed to something
<moon-child> but with the macro it wouldn't be able to ascertain that
<kingoffrance> because then it is really just discussing the uptime of whoever runs the servers
<kingoffrance> i did not mean to single video games out either -- just seems like there is very little long-term anything anywhere, although maybe that is just how the web seems
<kingoffrance> or, long-term things are quiet without fanfare. short-term things make all the noise
<kingoffrance> i dont believe there is a real "split". but i also believe perception > reality
<kingoffrance> its more of a ...how you gonna convince ppl :/ a marketing problem, not technical
sonny has quit [Quit: Client closed]
<kingoffrance> sonny died, but the gold paper i think agrees with me "perfect is the enemy of good" its all about tradeoffs
[itchyjunk] has quit [Read error: Connection reset by peer]
hyenasky has joined #osdev
<hyenasky> hey anyone know anything about fundamentally asynchronous IO systems like NT's
<moon-child> io_uring! :3
<hyenasky> reading about it and idk if that counts as fundamental
elastic_dog has quit [Ping timeout: 256 seconds]
<moon-child> asynchrony is a fundamental aspect of io_uring
<moon-child> and io_uring is an io system
<moon-child> ergo, io_uring is a fundamentally asynchronous io system
<hyenasky> i dunno it looks like an asynchronous API on top of a synchronous IO system
<hyenasky> unless im misunderstanding
<moon-child> ¯\_(ツ)_/¯
elastic_dog has joined #osdev
srjek has quit [Ping timeout: 240 seconds]
<ThinkT510> check out managarm for async
hyenasky has quit [Quit: Client closed]
hyenasky has joined #osdev
hyenasky has quit [Ping timeout: 256 seconds]
_xor has joined #osdev
GeDaMo has joined #osdev
the_lanetly_052 has joined #osdev
heat has joined #osdev
<Clockface> mymacro mylable, myotherlabel
<Clockface> thing %1 thing %2
<Clockface> in the preprocessed ASM its going to ..@4.2
<Clockface> instead of the label i am intending
<Clockface> what?
<heat> what
<Clockface> i dont know either man
<heat> source code?
<GeDaMo> Is this gas?
<heat> i didn't understand a word of what you said
<GeDaMo> Clockface: show code
<Clockface> im giving a macro 2 labels, and i checked the preprocessed code
<Clockface> the labels in the preprocessed code are ..@4.2 instead of what i was intending
<GeDaMo> Also, there's an ##asm channel
<heat> what preprocessor?
<Clockface> the one for NASM
<heat> ah
<Clockface> im just doing a macro
<heat> *shrug*
<heat> why do you care?
<heat> i bet it anonymises the labels
<Clockface> because i dont want to write 256 functions manually
<heat> hard to know if what you're doing is right if you don't give us a snippet of any kind
<GeDaMo> "You can call this macro as many times as you want, and every time you call it NASM will make up a different `real' name to substitute for the label %%skip. The names NASM invents are of the form ..@2345.skip" https://nasm.us/doc/nasmdoc4.html#section-4.3.2
<bslsk05> ​nasm.us: NASM - The Netwide Assembler
<Clockface> im generating labels within the macro to call from a jump table
<Clockface> so thing_%%1 will be called later expecting %%1 to be something
ravan has quit [Remote host closed the connection]
<bslsk05> ​pastebin.com: %macro modrm_mod00 2BX_PLS_SI__%%2:mov WORD ax, [R9 + rsi]mov bx, %%1 - Pastebin.com
<GeDaMo> And how are you calling this macro?
<Clockface> what?
<Clockface> i know the program is incomprehensable garbage
<Clockface> like i said, part of the macro is having labels within getting named from the macros inputs
<GeDaMo> What sort of parameter are you passing to the macro?
<Clockface> modrm_mod00 r8W, AX
<GeDaMo> So you want BX_PLS_SI__AX ?
<Clockface> yea
ravan has joined #osdev
<Clockface> im assuming it was never expected to be used this way?
<GeDaMo> I'm getting things like mov bx, ..@1.1
<GeDaMo> Clockface: it should just be %1 not %%1
<Clockface> thank you
<Clockface> now it assembles
<Clockface> no
<Clockface> it only has 3 erros now
<Clockface> yay
<Clockface> yes now it works
<Clockface> thank you
<GeDaMo> :)
<Clockface> can you make good code or garbage code in assembly?
<Clockface> or is it implied that all assembly is job secure
<GeDaMo> Code is code
<heat> assembly will usually result in garbage code
<heat> you're not a better asm writer than the compiler
<Clockface> yes but i can certainly be weirder than the compiler, and weird programs are fun!
<Clockface> an explicit design goal of this emulator was to only use registers directly "mapped" to the emulated CPU's registers, and use the real CPU's stack instead of emulating one in software
<Clockface> its working so far actually
pretty_dumm_guy has joined #osdev
<bslsk05> ​en.wikipedia.org: Weird machine - Wikipedia
<heat> Clockface, how do you do that without using the stack?
<Clockface> the program simply does not make use of the stack unless the emulated program is doing something involving it
<heat> how do you emulate without using the registers and the stack
<Clockface> at most the stack is push'd and pop'd before it moves onto the next opcode
<Clockface> i reserved 14 of my 16 registers for the emulated CPU
<Clockface> so the emulator just operates on them directly
<Clockface> i was told it was impossible on this channel in the past so i decided to try going about it
<Clockface> it is suprisingly doable
<GeDaMo> That just sounds like an assembly interpreter
<Clockface> thats exactly what it is
<Clockface> isnt that what an emulator is?
<Clockface> or does an emulator imply more than just the CPu
<Clockface> in which case im just making a bad version of something common
<heat> an emulator emulates more than the cpu
<heat> sorry :(
<Clockface> oh well, still nice to have around
<Clockface> does V86 mode allow you to put the 1 MB in paged memory or does it only use the lower 1 MB of physical memory like real mode?
<Clockface> im relatively sure its just like real mode
<Clockface> but ive never gotten this clarified
<Clockface> did protected mode windows have to reserve the lower megabyte of memory just for NTVDM?
<bslsk05> ​www.businesswire.com: Wejo Makes Its Data Available in First-of-its-kind Operating System for Electric Vehicle Infrastructure, Available Through Palantir Foundry | Business Wire
<Jari--> ecars not running Windows 11?
<heat> Clockface, paging
<heat> cars use linux
<heat> /generally/
<Clockface> linux is faster than windows
<Clockface> and people like fast cars
<Clockface> its only obvious
<Jari--> Clockface hehe
<Jari--> :-)
<GeDaMo> I'd be surprised if the engine management was running on Windows, maybe the entertainment systems
<Jari--> I actually seen an evehicle, like yesterday
<Jari--> heard it, nothing to be heard
<Jari--> most gasoline cars here
<clever> Jari--: https://www.youtube.com/watch?v=gidJopKtcnc this guy has been converting his hummer into an electric vehicle, so you can see every single component in the drive system
<bslsk05> ​'Want a Ride? - My Electric Hummer Battery is MOUNTED!!' by JerryRigEverything (00:31:10)
<GeDaMo> Petrol cars are much quieter than they used to be
<GeDaMo> Also, the automatic turning off of the engine at traffic lights is still odd to me
<GeDaMo> The lights turn green and all the engines start :P
<clever> hybrid cars can get away with that more easily
<clever> where its electric until the battery runs low
<Jari--> sometimes before I thought VM86 would not have a paging, but a flat linear memory instead in implementation, lol that you would have to share the first 1Mb of RAM with all VM86 tasks. How does Windows 95 do this? next to pool I guess :)
<clever> and the gas engine runs at a fixed rpm, to recharge things, usually
<clever> staying in its peak performance band
<clever> throwing in a dual-engine transmission and drive train would just be so much more complex
<Jari--> make your OS so stable everyone will be jealous
<Clockface> sometimes i miss how windows would be complete garbage sometimes
<Clockface> linux works way better, but windows is kind of fun
<heat> linux is not faster than windows
<moon-child> sometimes it feels like windows is better suited to servers and linux is better suited to desktop
<moon-child> like
<heat> linux also works horribly
<moon-child> sure
<heat> wtf do you use
<moon-child> but I absolutely hate using windows as a daily driver. I tolerate linux (and freebsd)
<heat> I enjoy using windows and I enjoy using linux
<Clockface> better at being a server?
<moon-child> but whereas linux has this beastly 'overcommit', windows actually distinguishes reserve and commit (at the user level)
<Clockface> how so?
<moon-child> (for instance)
<moon-child> solving the problem entirely
<moon-child> I think there are other things win32 gets right; I don't know it super well, but hear such things
<moon-child> e.g. I understand it had a completion-based io system years before io_uring
<Clockface> i like how DOS/windows works, linux is only really better in practice because microsoft doesnt make a good windows distrubition
<moon-child> and, well, pick your poison from among dll hell and so hell. Windows at least has a nominally backwards-compatible userspace
<Clockface> *distrobution
<Clockface> eh i cant spell
<heat> linux is better for server stuff and devving
<heat> but not anything else really
<heat> because few companies actually care about desktop linux
<heat> red hat, canonical, suse? collabora too
<heat> the rest is just "hurr durr look you can run linux on 512 CPUs instead of the old 384 cpu limit"
<heat> meanwhile you don't video gpu acceleration
<heat> don't have*
<moon-child> well, what else is there? Most of the things I do with a computer can be classified as: server stuff, compute stuff, dev stuff, communicate stuff, and play video games
<heat> general purpose usage
<moon-child> server/compute is definitely linux's ball game. dev ditto. Communication is fine (I have adequate email, irc, etc. clients, and windows doesn't have anything better)
<Clockface> linux has given me less weird problems than windows has
<moon-child> video games are ehhh admittedly
<heat> use the web, write documents on microsoft word, do a zoom call
<Clockface> as long as source engine games and dosbox work im satisfied for gaming
<moon-child> web and zoom can be classified as communication
<heat> i'm not even talking about video games
<moon-child> web browsing is fine. Zoom i will grant you--pulse can go die in a hell fire pit
<heat> those are a separate issue
<heat> zoom works amazingly until you notice the linux version is out of date (vs windows) and it's burning your cpu omg it's literally on fire
<moon-child> I literally stopped taking video calls on my linux box because I could not get my microphone to work reliably, and webcam usually took a couple tries too
<Clockface> really?
<Clockface> ive had more driver problems on windows oddly enough
<moon-child> really
<moon-child> I had to use windows exclusively at one point, because my laptop's builtin wifi card died, and the usb dongle I got didn't work with linux
<heat> using a web browser is great until you try to play a youtube video in the background and oh my god really why is my cpu on fire again I thought I had uninstalled zoom
<moon-child> there was a driver but it crashed
<moon-child> heat: never had that problem
<heat> congrats for the beefy CPU then
<moon-child> I have a nice cpu now, but worked on a shitty laptop with an i3 for some time
<moon-child> never had a hitch w/youtube
<heat> if I don't explicitly enable gpu video accel (read: hack the firefox and chromium options) everything gets so fucking slow
<moon-child> don't get me wrong firefox is abysmal and chrome is too. But they're that way on windows too sooo
<heat> like, the rest of the system
<moon-child> Clockface: I have also had full-system freezes with both windows and linux. Not sure which one had a higher incidence rate
<heat> then you realise linux has no UI, really
<Clockface> are you guys installing "experimental" versions of stuff or the stable ones for linux?
<Clockface> its been stable as a rock for me
<Clockface> a big one at that
<moon-child> heat: why does it need 'a' ui?
<heat> want to do X? yeah why don't you use the cOmMaNd LiNe
<Clockface> yeah i have to give windows that
<heat> because regular people don't want to use a command line, and don't need to use a command line
<moon-child> sure, I use x
<moon-child> but I don't see why there should be a singular user interface associated with linux
<moon-child> the graphical configuration I have is probably markedly different from that most other people have, and I think that's fine
<heat> that's not the point
<heat> actually, that is the point
<heat> the unbelievable fragmentation of a really small user/dev base(linux desktop) makes it so every DE is crap instead of having 1 or 2 really good DEs
<Clockface> xfce works fine
<GeDaMo> Having a uniform separation between the UI (text or graphics) and the core code would be nice
<heat> ...
<heat> everything works fine but it's all crap compared to the rest
<heat> macOS and windows are leagues ahead in UX
<amazigh> really good DE (tm)
<Clockface> yeah you dont have to touch the command line in windows like you do in linux
<amazigh> I briefly used macos and windows recently, it does not match GNOME or my i3 / sway setup.
<Clockface> thats something pretty cool about windows
<heat> it's not only "pretty cool" but a requirement if you want to have a desktop linux
<amazigh> It is unclear what "really good DE" would look like?
<amazigh> imo GNOME is a good DE (idk about KDE tho)
<heat> my mom is not going to use apt-get to install chrome
<moon-child> I _have_ a desktop linux
<moon-child> the quality of the user interface is not my problem
<Clockface> im not advocating desktop linux for everyone, it just works well for a lot of people
<amazigh> ubuntu does not require to use apt directly (tho I do not use it much)
<Clockface> there isnt anything wrong with windows
<moon-child> it may be a different problem--for instance, it may be the problem if you want linux to see more users
<moon-child> however it is, I think, distinct from the _basic system stability_-related problems we have brought up thus far
<heat> desktop linux works great for what, 2-4% of the total user base?
<Clockface> yeah
<Clockface> it works fine for me
<Clockface> so i like it
<Clockface> i dont think my grandma would
<Clockface> and thats fine
<amazigh> Clockface: it started spitting error message 30 minutes after starting it for the first time
<Clockface> really?
<amazigh> yes
<Clockface> the dumb thing about some linux distributions is the devotion to muh freddom
<heat> moon-child: the problem is not stability, linux tends to be kinda stable
<Clockface> so they dont include non FOSS drivers
<Clockface> firmware-linux-nonfree on debian tends to fix most of that
<amazigh> Clockface: that is not the case of ALL distros, free ultras have lot of mindshare in the sense their position is very well-known, but in practice there is more "open" minded regarding licenses users
<amazigh> e.g. Arch and NixOS both allow blobs
<Clockface> arch tends to break itself randomly
<heat> no
<Clockface> debian never has
<heat> arch never broke itself
<heat> I've been using it since 2015
<Clockface> every time i destroyed debian it was my fault
<amazigh> ^^'
<Clockface> while arch seems to periodically do weird stuff
<klange> Outside of the one distro that very specifically bills itself as being purely free software, are any popular distros doing that? Even Debian, as noted, has its nonfree stuff.
<heat> no.
<heat> klange, fedora kinda IIRC
<j`ey> people have had different experiences :o
<heat> never heard someone complain about arch randomly breaking itself, and it never broke for me (and I've used lots of DEs and rando packages)
<heat> but that's not the point
<amazigh> gnu guix only document and support free or open software
<Clockface> its never major, and its usually involving pacman
<Clockface> idk
<amazigh> but it is possible to use blobs if you know your way
<Clockface> im off to sleep now, goodbye
<amazigh> +
Jari-- has quit [Ping timeout: 272 seconds]
heat has quit [Ping timeout: 256 seconds]
nyah has joined #osdev
xenos1984 has quit [Remote host closed the connection]
xenos1984 has joined #osdev
mahmutov has joined #osdev
gog has joined #osdev
the_lanetly_052 has quit [Ping timeout: 240 seconds]
<mrvn> moon-child: you can change overcommit in linux and it had async IO for decades.
<mrvn> Basically all problems I had with Linux in the last decade where because of the desktop. I whish linux didn't have one. It's a horrible mess of black box magic that nobody can understand.
<mrvn> Plain X was good enough.
<j`ey> w a y l a n d
<mrvn> is that still a thing?
<j`ey> ofc
[itchyjunk] has joined #osdev
vdamewood has joined #osdev
australopithecus has quit [Quit: ZNC 1.8.2 - https://znc.in]
ElectronApps has quit [Remote host closed the connection]
heat has joined #osdev
heat_ has joined #osdev
heat has quit [Read error: Connection reset by peer]
sonny52 has joined #osdev
sonny52 has left #osdev [#osdev]
heat_ has quit [Remote host closed the connection]
Burgundy has joined #osdev
dude12312414 has joined #osdev
xenos1984 has quit [Remote host closed the connection]
dude12312414 has quit [Remote host closed the connection]
xenos1984 has joined #osdev
<kingoffrance> "then you realise linux has no UI, really" x is "mechanism, not policy." compare unicode. one can easily say "windows" and "macos" are "fragmented" for not using x. its a bogus argument
<kingoffrance> linux desktop would be sooooooooooo much better, if macos and windows didnt fragment things
<kingoffrance> its a total rubbish argument
<mrvn> how many millions of handys run linux?
<kingoffrance> well, my question is why are UIs different than unicode?
<kingoffrance> in one area this generality is seen as a Good Thing
<mrvn> unicode makes more problems than it solves for me too
<kingoffrance> :D
<amazigh> how does macos and windows fragment things?
<amazigh> also emacs does a good job :p
<kingoffrance> it doesnt. my point is anyone can say that about anything, it means nothing but "work for me for freeeee"
<amazigh> understood
<amazigh> I am wondering how FLOPS compare between linux et al. and windows and macos
<amazigh> I mean, ok, most users have windows or macos / ios
<mrvn> trying to read in strings in python3 is the horror. You can't read in anything not utf-8.
<amazigh> excluding android that is a linux,and on the other side of the tube there is many many linux boxes
<mrvn> amazigh: handys and servers.
<amazigh> basically everyone use linux, they just do not know it.
<amazigh> web2 was made possible because of linux et al.
<amazigh> google is linux on steroids.
<amazigh> remove linux, bsd, x whatever oss. Try to imagine the how the world would be. It is not hard to imagine for me, I see the struggle daily.
<amazigh> also macos borrowed many components from foss
<amazigh> I am not sure what is the point of that conversation. anyday, engineering side: foss >> privateer;
<kingoffrance> well yeah, if everyone would just switch to openstep and gnustep, all this fragmentation would just vanish :D
<GeDaMo> Meh, let's just go back to 80x25 text displays :P
<mrvn> please, vt100
<gog> single-wire telegraph
pretty_dumm_guy has quit [Quit: WeeChat 3.4]
pretty_dumm_guy has joined #osdev
srjek has joined #osdev
<geist> 132 cols or gtfo
m3a has quit [Quit: leaving]
<geist> huh which got me thinking, why 132? and then there's a nice stackoverflow that answers it: https://retrocomputing.stackexchange.com/questions/7838/why-did-line-printers-have-132-columns
<bslsk05> ​retrocomputing.stackexchange.com: history - Why did line printers have 132 columns? - Retrocomputing Stack Exchange
<geist> so TIL why 132 cols was a thing
<geist> and 80 was because FORTRAN cards (and that probably has a reason that goes back further)
<bslsk05> ​www.eecis.udel.edu: Anatomy of a Hollerith Card
<geist> yah exactly, i assumed the size/shape of a punch card predates digital computers getting ahold of it
emartinez has joined #osdev
emartinez has quit [Quit: WeeChat 3.3]
<gog> looms?
<gog> ohh tabulating machines
<kingoffrance> you might even say, international business machines :D
<kingoffrance> for things such as a census
<gog> i....b.....m?
<gog> never heard of 'm
<gog> jk i used to live near endicott and the mess of toxic shit they left in the soil from their fab is an ongoing problem lol
<kingoffrance> they had an eye bee m screensaver. i forget what the m was
<gog> endicott,ny*
mahmutov has quit [Ping timeout: 240 seconds]
<kingoffrance> so they also do hieroglyphics apparently
<kingoffrance> big market lol
<gog> IBM? not before i've had a cup of coffee! XD
[itchyjunk] has quit [Quit: Leaving]
[_] has joined #osdev
[_] has quit [Remote host closed the connection]
[itchyjunk] has joined #osdev
mavhq has quit [Quit: http://quassel-irc.org - Chat comfortably. Anywhere.]
mavhq has joined #osdev
<bslsk05> ​www.righto.com: Simulating the IBM 360/50 mainframe from its microcode
<geist> ah yeah forgot he was working on that
<jimbzy> geist probably has 3 or 4 of those in his garage.
<geist> hah no. the 'biggest' old machine i have is the microvax, which is mostly big by mass, not by size
<geist> and really the old G5 powermac probably has that beat
<graphitemaster> Every IBM mainframe I see is always insane all around.
<jimbzy> You know you've considered it, though.
<geist> i dunno. mainframes i have a bit less affinity for
<graphitemaster> IBM is just overkill all the time.
<geist> minicomputers are more my jam
<jimbzy> "If I park my cars in the street and sleep in the livingroom I can make it fit!"
<jimbzy> ;)
<sham1> I wish I had a VAX
<geist> but properly maintaining a minicomputer is a responsibility i dont want to take on really
<sham1> Well maybe a VAX clone from modern parts
<geist> kazinsal: did your vax come in yet?
<jimbzy> Before I die I'm going to build an Altair clone.
<geist> well, yuo can do that pretty well now
<geist> just a matter of money
<geist> in that there are folks that sell parts or schematics for parts (like say the S100 backplane, etc)
<jimbzy> Yeah I've seen a couple of kits over the years. I just think it would be a fun project.
<geist> whatyou can do is start assimilating parts for it. like finding cheapo 8080s on ebay before they're unobtanium
<jimbzy> Good idea
<geist> though it's not a strict altair i like to fiddle with z80 based things, and z80s are still pretty easy to come by
<geist> or even z180 stuff which i think they still make
<GeDaMo> An eZ80 might be interesting
<GeDaMo> "the performance is comparable to a Z80 clocked at 150 MHz if fast memory is used" "The eZ80 also supports direct continuous addressing of 16 MB of memory without a memory management unit" https://en.wikipedia.org/wiki/Zilog_eZ80
<bslsk05> ​en.wikipedia.org: Zilog eZ80 - Wikipedia
<jimbzy> I'm not super picky.
<jimbzy> A lot of my professors have said, "Why bother?" when it comes to everything other than x86-64. I can't remember how many times I had to point out that a computer doesn't necessarily mean a PC.
<sham1> The fact that a professor would say that is concerning to me
<jimbzy> Well, I am attending a university that specializes in liberal arts, so that could have something to do with it. Mostly, though, I think their attitudes just reflect the attitudes of the students. A lot of them have as much interest in technology as I do in basket weaving.
<jimbzy> Actually, I take that last sentence back. I'd love to learn how to weave a basket. :p
<energizer> cs professor not interested in computers?
<GeDaMo> Basket weaving professor not interested in computers? :|
<jimbzy> A few. I've had a couple who were sick with it like the rest of us, but for the most part it's been pretty weak.
<mrvn> Make a basket weaving robot.
<mrvn> or a programmable loom
Teukka has quit [Read error: Connection reset by peer]
Teukka has joined #osdev
xenos1984 has quit [Remote host closed the connection]
xenos1984 has joined #osdev
<sham1> GeDaMo: of course not. The basket weaving professor might see the writing on the wall about robots automating basket weaving
<sham1> Which IIRC is already a thing
FreeFull has joined #osdev
GeDaMo has quit [Remote host closed the connection]
<geist> ah yeah the ez80 looks neat
<jimbzy> Yeah, it does look neat.
<mrvn> Not sure how run code on an ez80?
<bslsk05> ​www.conrad.de: %product-title% kaufen
<mrvn> cZEALw_wcB%3AG%3As
<bslsk05> ​www.conrad.de: %product-title% kaufen
<kazinsal> geist: the SCSI2SD came in, the console cable is somewhere between Berlin and here and the VAX itself is in ebay global shipping purgatory (known to the rest of the world as Kentucky)
<geist> noice
<eryjus> Hey!! I'm from Kentucky!
<geist> then i guess you gotta go get his vax
<eryjus> 'from' != 'in'
<eryjus> but I got family
<kazinsal> someone breaking into the ebay global shipping facility and then driving the damn thing to vancouver might actually be faster than pitney-bowes' logistics chain
<geist> well that was fun. upgraded my firewall (pfsense) and futzed around with adding another rule to pass multicast ipv6 to keep it from filling up the firewall logs with records about mdns broadcasts
<geist> ie, a constant stream of stuff to ff02::fb port 5353
<geist> doesn't route it, but at least keeps it from being a hard block at the interface level
<geist> the more i work with ipv6 the more it makes a lot of sense. dunno how well it works in the global internet routing and whatnot, but for local bits its pretty clear
Burgundy has quit [Ping timeout: 252 seconds]
gog has quit [Ping timeout: 240 seconds]
pretty_dumm_guy has quit [Quit: WeeChat 3.4]
<moon-child> mrvn: posix aio is crap
<moon-child> mrvn: you can turn off overcommit, but I don't think you have a good way to do separate reserve/commit in userspace, right?
Vercas2 has joined #osdev
gxt has quit [Ping timeout: 240 seconds]
Vercas has quit [Quit: Ping timeout (120 seconds)]
Vercas2 is now known as Vercas
gxt has joined #osdev
Jari-- has joined #osdev
vdamewood has quit [Remote host closed the connection]
vdamewood has joined #osdev
vdamewood has quit [Quit: Life beckons]
vdamewood has joined #osdev
epony has quit [Quit: QUIT]