ChanServ changed the topic of #kisslinux to: KISS Linux | https://kisslinux.org | https://kisscommunity.org | post logs or else
Uks2 has quit [Read error: Connection reset by peer]
Uks2 has joined #kisslinux
kyxor has joined #kisslinux
<kyxor> Well I hate those guys who say "you may not run this program as root", like dude I believe in security by good design, not security by obscurity. If one is able to hack into a regular user, it is equally as bad as hacking root, it does not save my ass in any way, cause if he hacked the user - then hacking your way in further is just is level 2 defence. Level 1 defence should not be hackable by design.
<kyxor> user privilages in unix is the worst nightmare one may ever experience, probably a big reason for not being user friendly because of dumb complications put in by incompetent programmers not completent in their code design
<kyxor> *** rant end ***
<acheam> yeah thats annoying
<acheam> although your scenario only applies to single user systems
<kyxor> yep
<acheam> the main thing is just that its annoying. Put that info in the usage dialog and manpage, but don't stop me from running as root
<acheam> if you really feel compelled, then write it to stderr
<acheam> but please not stdout
<kyxor> oh god. this dude's site is laggy as hell https://camille.sh/philosophy/infinity
<kyxor> chrome is flexing that %100 core usage when scrolling
<kyxor> i don't really understand what causes that, it's like a hidden js crypto miner somewhere LOL
dilyn has joined #kisslinux
dilyn has quit [Client Quit]
dilyn has joined #kisslinux
dilyn has quit [Client Quit]
<akspecs> what are you all trying to run as root (and why?)
GalaxyNova has joined #kisslinux
<kyxor> akspecs: everything
<kyxor> I don't even have user permissions built into my kernel
<kyxor> useless bloat, I don't run multi user server so it's pointless
<GalaxyNova> Here's an idea: Community maintained ncurses installer for kiss
<GalaxyNova> it would have an interface to add the core repositories, etc
<akspecs> GalaxyNova: who would want to maintain that?
<GalaxyNova> I don't see how it could be too difficult
<akspecs> kyxor: i've literally never heard that before (unix permissions are bloat)
<GalaxyNova> if making it in C is too much one could do what void linux does and just call dialog
<kyxor> why would you use ncurses though? just use write() systemcall into the kernel
<akspecs> kyxor: that sounds like something terry davis would say
<kyxor> Yep, Terry Davis is 100% right
<GalaxyNova> Eitherway I feel like an installer would make it much easier to just whip up a kiss install
<kyxor> unless you wanna do some crazy utf-8 block art, ncurses is basically an unecessary abstraction
<kyxor> or some crazy other wierd encodings
<GalaxyNova> then just make it a shell script and call dialog https://linux.die.net/man/1/dialog
<GalaxyNova> it's going to be run on a live usb anyway so depending on dialog woudn't be a problem
<akspecs> you all ever hear of notcurses?
<akspecs> bloat, but pretty awesome bloat imho
<GalaxyNova> woah
<GalaxyNova> binary's probably 30 megabytes
<GalaxyNova> lol
<kyxor> hacker culture generally dislikes tuis. refer: http://www.catb.org/jargon/html/hates.html
<GalaxyNova> what is a "tui" anyway
<acheam> kyxor: you know the creator of camille.sh is a kiss user, right?
<acheam> not that that makes the website not laggy
<acheam> just cool
<acheam> GalaxyNova: text user interface
<acheam> a user interfaced created with text
<GalaxyNova> yes ik
<kyxor> acheam: yep I am collecting all personal sites, for bookkeeping on my site, so most links I can find on github from kiss users is where I look first
<GalaxyNova> but can you consider something like vim a text user interface
<kyxor> it is
<acheam> kyxor: nice
<acheam> acheam/armaanb/armaanb.net is me
<kyxor> yep, already got your site, acheam
<acheam> im honored
<GalaxyNova> love
<GalaxyNova> the bumper stickers
<acheam> kyryl.tk refuses to connect :(
<kyxor> GalaxyNova: I don't hate tuis, btw. It's just I will hate tuis that are badly designed
<acheam> GalaxyNova: heh thanks
<acheam> GalaxyNova: instead of a curses installer, a shell installer works well
<kyxor> and sometimes they are, espicially if they block your terminal text selection, or have no means to redraw screen
<GalaxyNova> perhaps
<acheam> i do like the openbsd installer
<acheam> which is just a shell script with some prompts
<phoebos> ^
<kyxor> basically tuis are substantially more complex ususally than just relular cli tools with flags
<phoebos> the mailx(1) interface is nice too
<GalaxyNova> I'd like to see a text editor as a non TUI
<phoebos> ed
<kyxor> ed
<kyxor> ex
<GalaxyNova> ok
<GalaxyNova> but is that really a text editor
<acheam> yes?
<phoebos> man -f ed
<noocsharp> ed is the standard text editor
<acheam> what makes it not a text editor
<phoebos> ed - text editor
<acheam> lol
<kyxor> well if there was a measure like "is it turing complete?" but for text editor I would say yes.
<phoebos> "does it edit text?"
<GalaxyNova> just barely though
<GalaxyNova> it's not much better than echo "text" >> file
<acheam> you haven't used it then
<acheam> its rather powerful
<GalaxyNova> perhaps that's also true xD
<acheam> and ex is pretty much just headless vi
<acheam> on obsd, it is exactly that
<phoebos> it's more like an interactive sed session
<acheam> on other OSs, its a bit more complex
<acheam> sed is derived from ed
<phoebos> aye
<kyxor> also my neatvi has ex mode too, I don't normally have a need for it, but ocasionaly I use it to test things out. Probably has some ed commands missing though.
<GalaxyNova> `i only know basic write and read commands for ed
<akspecs> GalaxyNova: ed is the beginning
<kyxor> acheam: did you try my editor on BSD, I just wanna know if it builds with no errors or not
<acheam> no but i'll build it for you
<acheam> oh boy you're not going to like this
<kyxor> Thanks, cause for me to test it id have to configure vm support for my kernel, then figure out how to get bsd working, lots of stuff
<kyxor> what did it say?
<acheam> oops
<acheam> thats only stdout
<kyxor> hmm, what's the compiler error?
<phoebos> SIGWINCH not defined
* acheam has to look up how to redirect stderr to stdout every damn time
<kyxor> phoebos: send log
<acheam> same as what I sent, kyxor
<phoebos> ^
<kyxor> ok i see
<phoebos> in fact you've got more warnings, acheam
<kyxor> -Wsign-compare is harmless there
<acheam> because I compiled with "-W -Wall -Wextra -g -O0"
<acheam> my default development CFLAGS
<kyxor> yeah it's from -Wextra
<kyxor> i'll see if I can make that SIGWINCH portable
<phoebos> ah i was comparing two copies of your output lmao
<phoebos> looks like SIGWINCH is #defined in <sys/signal.h> for obsd at least
<acheam> wait whats a dual-panel UI?
<kyxor> try to add it at the top of hund.c #include <sys/signal.h>
<acheam> nope
<acheam> doesnt help
<phoebos> no <signal.h> includes <sys/signal.h> on obsd
<acheam> ew headers including headers?
<phoebos> and <bits/signal.h> on linux
<phoebos> acheam: literally most headers lol
<phoebos> but i don't see why it's not getting defined
<acheam> yeah I don't know why thats bad
<acheam> but the kcgi guy doesn't like it
<acheam> kcgi/bchs/lots of other cool obsd stuff
<kyxor> try #include <signal.h> ?
<acheam> kyxor: doesnt help
<kyxor> damn where is it defined then? it should be supported, right?
<acheam> signal.h(3) doesn't mention it
<acheam> signal.h(3) nor signal.h(3p)
<phoebos> kyxor: it's in there, probably some preprocessor crap getting in the way
<phoebos> grep -r SIGWINCH /usr/include
dilyn has joined #kisslinux
<acheam> oh yeah it is
<acheam> weird that its not in the manpage though
<dilyn> curl's build system has appeared to change a bit and now curl is no longer static?
<dilyn> hmhmhm
<phoebos> oof
<kyxor> dilyn: hi!
<noocsharp> bruh are you serious
<dilyn> yeah needs LDFLAGS=--static now
<dilyn> o/ kyxor
<acheam> if you wanted to get really rough, you could #ifndef SIGWINCH \n #define SIGWINCH 28 \n #endif
<noocsharp> oh, well that's reasonable
<dilyn> it's better than what it was before lmfao
<acheam> dilyn: good to see you around here again
<kyxor> acheam: ok doen't sound too bad, if you do that any other errors you get? besides SIGTERM
<dilyn> :)
<acheam> manually defining SIGWINCH leads to a succesful compilation
<testuser[m]> Hi
<kyxor> acheam: nice
<phoebos> yeah but why doesn't it work? all the __BSD_VISIBLE _USER_SIGNAL_H _SYS_SIGNAL_H are correct
<acheam> hi testuser[m]
<kyxor> maybe -std=c99 -D_POSIX_C_SOURCE=200809L has some effect
<acheam> ah yeah it must be that
<acheam> as SIGWINCH isn't posix
<acheam> I should have noticed that earlier
<GalaxyNova> is there any reason why the git package explicitly statically links curl?
<kyxor> sadly, ehh
<testuser[m]> How do you detect resizes on pure posix systems then
<phoebos> whuttt
<kyxor> so removing -D_POSIX_C_SOURCE=200809L builds correctly without any C code mods
<kyxor> ?
<phoebos> SIGWINCH *is* in vi.c after the preprocessor
<acheam> kyxor: yes
<kyxor> testuser[m]: good question...
<akspecs> GalaxyNova: what have you against curl?
<phoebos> ah i wasn't `cc -E`ing with a _POSIX_C_SOURCE
<phoebos> nice catch.
<acheam> yeah I was trying to verify that it was included after preprocessing and couldn't find it
<acheam> I guess you can just poll tcgetattr?
<dilyn> galaxynova: so that git still works if you muckbang your libcurl.so
<kyxor> try replacing it with _DEFAULT_SOURCE
<kyxor> -D_DEFAULT_SOURCE
<dilyn> essentially, for the same reason you'd build any other thing statically
<acheam> _DEFAULT_SOURCE doesn't help
<GalaxyNova> akspecs: Nothing, it's just that statically linking it seems wasteful from a storage space perspective
<phoebos> <sys/cdefs.h> is responsible
<acheam> actualy now SIGTERM isn't defined
<GalaxyNova> since curl is expected anyway by kiss
<GalaxyNova> dilyn: oh
<acheam> not if you're an axel-using chad
<GalaxyNova> is axel smaller than curl?
<acheam> I never compared, but I wouldn't be surprised
<phoebos> kyxor: https://envs.sh/ME.h is obsd <sys/cdefs.h>, you need __BSD_VISIBLE to be 1
<kyxor> so -D_BSD_SOURCE should set __BSD_VISIBLE to 1 at some point?
<phoebos> looks like you could define _BSD_SOURCE
<phoebos> yeah i reckon
<phoebos> yes! it worked
<kyxor> I wonder if I define _BSD_SOURCE, it won't break anything?
<testuser[m]> GalaxyNova: curl performs much more than just downloads so axel is smaller
<phoebos> wrap it in #ifdef OpenBSD or whatever
<kyxor> I mean it seems like defining -D_BSD_SOURCE on linux does not break build
<acheam> phoebos: nah
<phoebos> no?
<acheam> then it breaks on netbsd, freebsd, etc
<kyxor> wow why are they so different in terms of header files omg
<phoebos> #ifdef OpenBSD || NetBSD || FreeBSD
<acheam> what about dragonflybsd, or sunos, or darwin
<testuser[m]> ifndef linux
<phoebos> lmao
<phoebos> linux or *BSD
<phoebos> kyxor: even musl and glibc drastically differ in header layout
<acheam> don't forget hurd, minix, solaris, aix, and hp-ux
<kyxor> yeah I know about glibc vs musl, but other bsds use different C libraries?
<acheam> yes and no
<acheam> they're all forks of the same codebase, 25 years ago
<kyxor> forgot plan9
<acheam> well thats not UNIX-like
<testuser[m]> Windows
<acheam> lmao IBM is supporting AIX through 2035
<acheam> i mean, its nothing close to sqlite
<acheam> but still, can't enterprises move a /little/ bit faster
gtms has joined #kisslinux
<kyxor> hmm I don't really like #ifdef mess I might just make a separate makefile target for BSD* then
<phoebos> it would be simplest to add -D_BSD_SOURCE for *BSDs
<kyxor> ok wtf, probably not even worth bothering to try to do any kind of "automatic" os detection
<kyxor> like I could just do uname and then try to compare strings, but it's so awkward, not gonna do it
Uks3 has joined #kisslinux
Uks2 has quit [Ping timeout: 250 seconds]
<phoebos> hund.c uses SIGTERM but doesn't include <signal.h>
<kyxor> ok I'll add it in too
<phoebos> but with CFLAGS+=-D_BSD_SOURCE and including signal.h in hund.c, it builds!
<phoebos> nice editor as well, imma use it now
<phoebos> been missing some colors too
<kyxor> fastest file opening times ever it has
<acheam> i dont think comparing strings is that akward
<acheam> see how june does it here
<acheam> phoebos: noooo syntax highlighting will make you a weak programmer, and the devil will sell your soul to richard m. stallman
<phoebos> oh god no
<phoebos> :P
<phoebos> i don't use colors when writing software, only reading
<kyxor> does it though? It help me. The only problem I have with syntax hl in %99 of text editors is how it's implemented sucks
<kyxor> neatvi does it all under 700 lines of C
<phoebos> neatvi's way is really nice tbf
<kyxor> also read the readme, there are some cool stuff like if you do :se hww
<kyxor> that thing is a must have if you are reading source code, cause you can immediately see varibles and how they used in function of example
<phoebos> oh that's cool it updates when you move the cursor
<dilyn> so I *did* complain about ccmake a few weeks back, but ngl it's pretty fucking neat
<dilyn> for instance, after about twenty seconds I got llvm to build statically
<dilyn> unfortunately it does not like netbsd-curses...
<acheam> hmm
<acheam> i should look into it more
<testuser[m]> What's ccmake
<testuser[m]> Cmake in C ? Lol
<testuser[m]> Oh
<dilyn> it's the curses viewer for cmake
<testuser[m]> Why does a build system have a ui
<dilyn> so you can flip options in a nice little TUI instead of reading CMakeLists.txt for options or trying to remember what the command is to print all the options
<dilyn> have you every tried to configure something as big as LLVM? lmao
<dilyn> it can be helpful
<dilyn> oh look dylan fixed curl not building statically XD
<dilyn> hi dylan o/
jedavies has quit [*.net *.split]
jedavies_ has joined #kisslinux
kernelc has quit [*.net *.split]
kernelc has joined #kisslinux
mahmutov has quit [Ping timeout: 250 seconds]
<kyxor> wow din't know rofl0r was working on sabotage, cool stuff
<acheam> dilyn: lol
<acheam> kyxor: they used to hang out in here too
<dilyn> it's theirs!
<acheam> ?
<dilyn> didn't rofl0r make sabotage?
<dilyn> butch is their baby
GalaxyNova has quit [Ping timeout: 252 seconds]
<acheam> No
illiliti has quit [Quit: leaving]
<testuser[m]> what
<acheam> It was made by leah2 originally
<dilyn> f
<dilyn> sry leah2
<dilyn> didn't mean to rob you of your glory
<kyxor> phoebos: I pushed bsd build fix on master, check if everything works, thanks ;)
<testuser[m]> the changes make your makefile non portable outside gmak
<testuser[m]> gmake
<kyxor> what isn't portable ?
<testuser[m]> ifeq
<kyxor> ok that's suprising
<testuser[m]> Yeah it's too limited
<kyxor> so how do I portably do if statement ?
<kyxor> or I can't
<testuser[m]> not sure, need to ask illiliti
<testuser[m]> I guess you would use shell syntax
mps has quit [Ping timeout: 268 seconds]
<kyxor> damn, I have to write extra targets to get portable conditions
<dilyn> portable makefiles are bae :v
<acheam> Damn your bae ugly af
<testuser[m]> += isn't portable either right
<kyxor> Yeah I sought so
<kyxor> I should stop looking at that GNU manual page
<testuser[m]> What good build system even exists then
<kyxor> shell script
<kyxor> you can even do parallel compile, using the &
<testuser[m]> ninja seems good but the only generators for it are cmake and poogle gn both are meh
<kyxor> hm you know what, maybe I should just write posix shell script for build and call it a day, remove make. But I won't have the based on date/modification time compile
<kyxor> unless I can somehow emmitate that in script, using date, then checking and filter what needs to compile
<testuser[m]> Implement some make-like functionality in sh
<testuser[m]> Call it "shake"
<kyxor> I'll give it a try
<dilyn> nss generates headers for some god forsaken reason http://ix.io/3wIk
<dilyn> that whole build system is fuckery
<dilyn> http://ix.io/3wIl nss should be illegal
<dilyn> it bundles its own version of install that it COMPILES ITSELF, and then it uses that to put headers in the "right locations" during the build
<dilyn> except it just creates symlinks to those headers
<dilyn> it's like... just move the headers to include/ or something smh
<testuser[m]> Pozilla moment
<testuser[m]> Why does everything suck so much
Uks3 has quit [Ping timeout: 250 seconds]
Uks2 has joined #kisslinux
Uks2 has quit [Ping timeout: 240 seconds]
Uks2 has joined #kisslinux
GalaxyNova has joined #kisslinux
GalaxyNova has quit [Client Quit]
gtms has quit [Remote host closed the connection]
gtms has joined #kisslinux
<testuser[m]> Oh the new site is up
<akspecs> kisslinux.org ? or are you talking about something else?
<testuser[m]> Yeah
<testuser[m]> https://kisslinux.org/wiki/ bruh he wrote pages for every package, huge effort
<necromansy> yeah mental
<dilyn> love the new website
<dilyn> what a cute little github icon
<akspecs> nice!!
mobinmob has joined #kisslinux
<travankor> github icon isn't kiss
mps has joined #kisslinux
<testuser[m]> clang-tidy
<testuser[m]> It'll always be atleast 1 tho, since lbuf_len wont return negative
<travankor> jedavies_: btw how is xkiss going to work?
chaihron_ has joined #kisslinux
chaihron has quit [Ping timeout: 252 seconds]
mobinmob has quit [Quit: Connection closed for inactivity]
<kyxor> testuser[m]: impossible
<kyxor> I mean that's why MAX is there, clang is dumb
<testuser[m]> Doesn't it make max useless if the other arg is always going to be > 0?
<kyxor> it can't analyze the program 100% completely to know that I guess, because of translation units i suppose ?
<kyxor> it's this: #define MAX(a, b) ((a) < (b) ? (b) : (a))
<travankor> macros are handled by the preprocessor so that's why it's not detected?
<testuser[m]> I'm saying your code on that line would be equivalent to MAX(0,1) atleast so why not remove the macro
<testuser[m]> Clang warns since it doesn't know that lbuf_len() will always be >= 0
<kyxor> yeah not a bad idea
<kyxor> cause it doesn't really do anything there anyway
<kyxor> actually, no it does
<kyxor> that macro is necessary for some reason
<kyxor> cause if I just leave lbuf_len()+1 vi compiled with gcc thows arithmetic exception on empty file
<testuser[m]> It's hiding a bug due to double evaluation ?
<kyxor> nope
<testuser[m]> Weird
<kyxor> OHH, it's because of order of operations
<testuser[m]> Lol
<kyxor> division always comes first
<kyxor> so that +1 is not evaluated before division
<kyxor> yep, just gotta add parethesis
<acheam> Oof
<acheam> We need mid to come teach us math
<testuser[m]> I always use explicit brackets even if the result would be the same without em
<kyxor> I guess that max macro acted like parenthesis there, because if statement will happen first
<kyxor> so technically it was still non bugged code, i pushed the change though
<kyxor> and 1 sec forgot to remove debug statement, dammin
<kyxor> time to revert it (eh )
<testuser[m]> Just hard reset and force push if the commit is like 5 seconds old
<kyxor> yeah i do that sometimes
<testuser[m]> Do you use sanitizers?
<testuser[m]> Or valgrind
<kyxor> I do use valgrind
<kyxor> push done
<kyxor> testuser[m]: btw do you use valgrind? it does not work correctly on latest musl btw
<kyxor> well unless you patch it like I did
<testuser[m]> no i just use clang sanitizers since im on glibc, they dont work on musl yet i think
<travankor> i think sanitizers work on musl now
<travankor> clang 12
<testuser[m]> nice
<jedavies_> travankor: Started that it since I had trouble building GCC with the clang toolchain. Sorted that now so will bring that into glasnost and remove xkiss.
<jedavies_> Needed gcc since clang can't build arm firmware / uboot.
<travankor> you patched uboot for clang?
<travankor> or sorted compiling gcc with clang?
<jedavies_> Sorted building gcc with clang. Wanted to avoid having GCC but building the low-level stuff with clang is not easy right now.
jedavies_ is now known as jedavies
<travankor> I see, thanks
<travankor> btw will glasnost follow upstreams/dylan's repo changes (wayland, mold, etc.)
mps has quit [Ping timeout: 252 seconds]
progenyx has quit [Quit: progenyx]
<jedavies> Yeah, working on merging the package manager changes. Will push them soon, been busy with work.
<jedavies> Mold? The linker?
<testuser[m]> yea
akira01 has joined #kisslinux
<jedavies> Interesting... can it do lto with clang/llvm?
<akira01> acheam: did you use wayland?
<testuser[m]> soon
<testuser[m]> doesnt do lto at all rn
<acheam> akira01: not really
<acheam> I toyed with it
gtms has quit [Remote host closed the connection]
<akira01> Maybe is the problem
<akira01> When i used xorg
<testuser[m]> what
<akira01> Webtorrent works great
<akira01> Now with sway i get segmentation fault
<akira01> Other apps like joplin had some bugs to
<acheam> Seems like an unrelated thing, probably something else in the build file
<testuser[m]> and the segfault is in lib what
<testuser[m]> webtorrent itself /
<testuser[m]> gtk ?
<acheam> Time to git bisect run an 8 hour browser build
<testuser[m]> gdb /usr/bin/firefox then type 'run' then crash the browser then type 'backtrace'
<kyxor> what do you guys think about using httpd from busybox as a webserver?
<kyxor> it seems to have ssl/tls support out of the box unlike quark
<akira01> testuser[m]: dont know well
<akira01> is not gtk because is the cli version
<akira01> webtorrent-cli
<testuser[m]> then do the gdb thing
<akira01> Okay
<testuser[m]> replace /usr/bin/firefox with /usr/bin/webtorrent-ci
<testuser[m]> cli
<testuser[m]> Lol that name sounds contradictory
<testuser[m]> webapp - cli
<akira01> lol
<testuser[m]> the segfault is in nodejs i guess
<testuser[m]> that's what you deserve for running such heretic software on kiss
<kyxor> ok never mind I got busybox httpd confused with apache version. busybox does not have tls
<akira01> Pardon my man
<kyxor> I guess no choice, have to use another tunnel prog for https
<kyxor> there is also this: http://www.acme.com/software/mini_httpd/
<noocsharp> openbsd httpd
<konimex> jedavies: I don't think we can even build mold with an LLVM-only toolchain
<travankor> tbf lld is quite fast, so it's gcc setups that benefit the most
<travankor> and mold doesn't have ppc64le support
<akira01> testuser[m]: i had some errors with gdb
<akira01> First do gdb /usr/bin/webtorrent
<testuser[m]> what happened
mps has joined #kisslinux
<akira01> "0x7ffdce07eb00s": not in executable format: file format not recognized
<akira01> when i type run
<akira01> It start program and said no executable file specified
<testuser[m]> what
<testuser[m]> show ss
<akira01> one sec
<acheam> noocsharp: if your system has busybox, i would just use an ssl proxy
<acheam> oops sorry meant to tag kyxor
<acheam> was thinking of you beacuse of your ssl tunnel software :)
<acheam> but yeah, I second the openbsd httpd reccomendation
<acheam> its the perfect balance of minimalism and features IMO
<testuser[m]> dilyn: This is the issue with chromium i asked you few weeks back, is it the same ? https://0x0.st/-tKr.jpg
<testuser[m]> it doesn't seem to be a chromium issue in specific, since it fixes itself whenever something prompts a redraw, eg typing `grimshot save area` from ssh, or just launching bemenu
<testuser[m]> dont mind the potato
<akira01> testuser[m]: had some solution?
<testuser[m]> oh sorry i didnt notice you sent the link
<testuser[m]> wait
<testuser[m]> oh ofcourse
<testuser[m]> its nodejs crapware so wont be a binary
<testuser[m]> binary binary
<testuser[m]> `ulimit -c unlimited`
<testuser[m]> launch it
<testuser[m]> it should print a dump
<dilyn> konimex: yeah hopefully the llvm devs finish up their support for c+20
<akira01> I do on terminal and show nothing
<dilyn> testuser: mine is just black boxes that generate over the tab/status bar/bookmarks bar area
<testuser[m]> idk akira01 just get a better torrent downloader
<testuser[m]> aria2c
<akira01> i mean
<akira01> i even used webtorrent to download
<akira01> Just to stream
<akira01> can aria2c do that?
<testuser[m]> dont think so
<akira01> yeah so i pretty fucked up
<testuser[m]> what's the contents of /usr/bin/webtorrent-cli
<testuser[m]> you can launch it via node directly
<testuser[m]> and get the backtrace
<akira01> i can launch it in terminal
<akira01> it launch okay
<akira01> but when i do some
<akira01> webtorrent 'magnet link' --mpv
<testuser[m]> oh just do `gdb /usr/bin/node` and `run /usr/bin/webtorrent-cli`
<akira01> node?
<testuser[m]> its a nodejs program bruh
<akira01> okay
<testuser[m]> oh you dont need to do it
<testuser[m]> i forgot
<testuser[m]> just check in dmesg
<testuser[m]> segfault at ffffaddress bla bla in libc.so.6
<testuser[m]> what does it say
<akira01> cant do
<testuser[m]> what cant do
<akira01> dmesg
<testuser[m]> why cant do dmesg
<akira01> it recognizes differents commanda
<testuser[m]> wat
<akira01> i mean
<akira01> webtorrent itself run okay
<testuser[m]> I know
<akira01> the problem is when i run webtorrent "magnet" --mpv
<testuser[m]> yes
<testuser[m]> have you got the segfault atleast once on your current boot?
<testuser[m]> dmesg | grep fault | nc termbin.com 9999
<akira01> Yeah
<testuser[m]> weird
<testuser[m]> do the gdb /usr/bin/node thing i said
<testuser[m]> segfault is probably in mpv
<akira01> cant to because the magnet uri is fucking big and buf the gdb
<akira01> bug*
<testuser[m]> `gdb /usr/bin/node` then `run /usr/bin/webtorrent "magnet" --mpv`
<testuser[m]> pipe it ?
<testuser[m]> paste ?
<akira01> i past
<akira01> and bug
<testuser[m]> gdb /usr/bin/node -ex "run /usr/bin/webtorrent ... args"
<testuser[m]> paste it in the shell's prompt
<testuser[m]> no need to use gdb tui
<akira01> okay
<akira01> had some info
<akira01> i will send
<testuser[m]> type backtrace
<testuser[m]> or "bt"
<testuser[m]> do you know the C library this program binds to
<akira01> dont know
<testuser[m]> how'd you install it
<akira01> npm install webtorrent
<akira01> is in webtorrent-cli github
<testuser[m]> find /usr -name 'utp'
<testuser[m]> `find /usr -name '*utp*'` glob, idk if the prev msg sent properly
<testuser[m]> gyp files seem like compiled bindings to me so you somehow need to build a debug build of libutp and build that as a nodejs module
<testuser[m]> is X and wayland the only thing that affects this ?
<akira01> I mean
<akira01> whem
<akira01> When i used X
<akira01> Works out of the box
<testuser[m]> what has changed since you used X
<akira01> no problems
<testuser[m]> when did you last use X
<akira01> yeah
<testuser[m]> yeah is not a timeframe
<akira01> the last time is like when dylan put sway in repo
<testuser[m]> Lot of stuff has changed since then
<testuser[m]> wait
<testuser[m]> let me check the source
<testuser[m]> idk man
<testuser[m]> strace it once
<testuser[m]> strace webtorrent-cli bla bla > log 2>&1
akira01 has quit [Ping timeout: 240 seconds]
akira01 has joined #kisslinux
<akira01> testuser[m]: some solution?
<testuser[m]> Check logs
<testuser[m]> I asked you to strace it
<akira01> when i get see logs for node stuff?
<testuser[m]> what
<akira01> u mean i need to check logs for webtorrent
<akira01> i dont know where the logs are
<akira01> i tried to look yesterday
<testuser[m]> No
<testuser[m]> I said, just strace it with the argumsnts that cause it to segfault
<testuser[m]> strace webtorrent-cli bla bla > stracelog 2>&1
<akira01> okay
illiliti has joined #kisslinux
<acheam> hi illiliti
<mps> illiliti: I've built eiwd on alpine, works fine.
<acheam> any reason why it wouldn't? The alpine toolchain is very similar to kiss
<mps> only not sure should I patch it to use different paths (dirs) for /etc/iwd and /var/lib/iwd, to be co-installable with 'real' iwd
<mps> acheam: yes, didn't expected there will be any serious problem
<mps> my dilemma now is should I make iwd adn eiwd installable together or simply make them conflicting pkgs so only one can be installed
<mps> there are pros and cons for both idea
<omanom> my first instinct is to mark them conflicting. is there any reason someone would have them installed at the same time?
<mps> omanom: well. strange but yes. sometimes I travel around and having iwctl working is useful thing
<testuser[m]> why would someone use eiwd though, if they don't have an issue with dbus
<mps> but mostly I use my boxes at home so eiwd is quite enough
<testuser[m]> if you have regular iwd installed anyway, then i dont see an advantage
<mps> testuser[m]: I don't like dbus and I don't like it to run always on my boxes
<mps> I'm 'fighting' for minimum software on machines
akira01 has quit [Read error: Connection reset by peer]
<mps> anyway, thought to write simple shell script to use 'iw scan' and get list of available APs in range
<omanom> how hard would implementing a iwctl for eiwd be? does iwctl give you anything beyond an "easier" way to manipulate configuration files?
<acheam> iwctl heavily uses dbus IIRC
<mps> acheam: right
<omanom> sure, but the actual functionality of the tool... like what user actions does it expose and do those actions have a non-dbus equivalent
<acheam> idk, never used it
<mps> omanom: iwd 'exports' all data over dbus
<mps> I asked upstream to add option to control iwd over socket or something else but that idea is strongly denied by them
<omanom> is there another way to see the data needed to perform the actions that iwctl exposes?
<mps> afaik dbus only
<mps> and I don't want to fork iwd to have such option
<omanom> so like iwctl allows you to list wireless devices, scan for networks (and display the results), and connect/disconnect to networks.
<omanom> the connect/disconnect can be done by manipulating configuration files, there's no need for dbus
<mps> omanom: right
<omanom> list wireless devices can be done
<mps> iw with some shell scripts can do these things
<omanom> what about scanning for networks -- is there any way on a machine using eiwd to see available networks?
<mps> no iiuc
<omanom> and that's probably the function of iwctl that's most useful when on travel
<mps> yes
<mps> omanom: fyi, there is also iwgtk tool, you guess it is gui for iwd
<illiliti> omanom: you need iw command
<illiliti> ssu iw dev wlan0 scan -u | grep SSID:
<omanom> can iw be used on a KISS Linux machine?
<illiliti> yes, why not
<omanom> so you can use iw in addition to eiwd and then you'd have wireless scanning capability right?
<omanom> so iwctl no longer needed
<illiliti> yes
<omanom> i like that "solution" better than installing iwd alongside eiwd just to get iwctl
<mps> i'm thinking to 'wrap' iw scan in perl script to print scan result in little nicely format
<mps> thank you all for 'brainstorming', I will make iwd and eiwd confliting pkgs on alpine
<illiliti> s/perl/shell/
<cotangent> <illiliti> yes
<illiliti> perl is overkill
<mps> I'm lot better in perl than in shell
<mps> but yes, shell with dialog looks 'atractive' to try
<illiliti> omanom: https://envs.sh/Md.tar
<omanom> is libnl the only depends?
<illiliti> +
<omanom> @mps got off my lazy butt and installed from that archive, can confirm it builds and works on my laptop at least :)
jess has joined #kisslinux
<acheam> hi jess
<jess> good morning
micro_O has joined #kisslinux
<micro_O> really appreciate all the package documentation / readme stuff
<micro_O> I am slightly meh on the 'draw an ascii box around terminal stuff' as it sometimes includes commands to run, sometimes includes text to ignore, etc
<micro_O> but 100x to having local-first (in-repo) documentation, and making that documentation accessible via http, or filesystem/cli
mahmutov has joined #kisslinux
<mps> omanom: didn't understood, you installed eiwd on kiss or you built it from the source directly?
<omanom> i already had eiwd installed, i downloaded the iw .tar you linked earlier and built iw from that and confirmed it allowed me to scan for wireless networks
<mps> omanom: I didn't linked this iw.tar, illiliti did it
<omanom> oof sorry, my fault
<omanom> thank you illiliti
<mps> np
akira01 has joined #kisslinux
<acheam> omfg I swear this headline isn't sarcasm
<acheam> "They’ve finally done it! Kali on WSL! Windows just keeps getting better."
<akira01> testuser[m]:
<acheam> this is what I get for deciding to browse medium
<akira01> i get the logs
<acheam> then share them?
<testuser[m]> Send
<akira01> yeap
<akira01> just sec
<testuser[m]> acheam: lol how is it windows getting better if you're just gonna run loonix on them
<testuser[m]> it
<acheam> tbf that headline sounds like something I would have said a few years ago
<akira01> Gosh
<akira01> some pastebin option to a 35K lines file?
<acheam> envs.sh
<acheam> ix.io
<acheam> termbin.com
<acheam> 0x0.s
<acheam> 0x0.st
<akira01> 3 of these not give me the link
<akira01> Or get segmentation fault
<akira01> yep now is all the links
<acheam> well how big is the file
<acheam> in *bytes, not lines
<acheam> cant hurt to gzip it
<akira01> 2.5M
<acheam> shouldnt be an issue then...
<akira01> yeah
<akira01> strange
<akira01> maybe is the number of lines
<omanom> how are you trying to post the files?
<omanom> what's the command that you're using? the sites all give examples of how to use them
<akira01> the command how is in the website
<akira01> just cp paste
<akira01> all sites
<akira01> works with other archives
<akira01> not this
<akira01> and i gzip now
<akira01> how can i send?
<dilyn> 0x0.st supports uploading arbitrary files
<akira01> okay
<acheam> if curl is too difficult for you, envs.sh has an upload button
<akira01> strange, with buttom works and without it segmentation fault
<acheam> wowee
<mps> illiliti: I'm finishing eiwd for alpine, want to add https://github.com/kisslinux/repo/blob/master/extra/eiwd/files/iwd_passphrase to pkg. what about license in it and do you agree we can ship it
<acheam> noocsharp: congrats on finishing the summer of pinephone
<acheam> ngl I got a shock realizing that the summer is almost over when seeing the conclusion article
<illiliti> but i guess it's better to ask Dylan to clarify license status
<mps> illiliti: ok
<mps> dilyn: ?
<illiliti> dilyn != dylan :)
<mps> ah, sorry
<mps> anyway, eiwd is LGPL already, so I think there is no problem
<dilyn> dylan almost always licenses MIT so you can always just rerelease LGPL anyways
<mps> ianal so I'm always feared of such things
<dilyn> i mean it's certainly kind to get his blessing :)
<mps> sure
<mps> (speaking someone whose works stolen some times attributed to someone else and never tried to force changes)
micro_O has quit [Quit: micro_O]
<akira01> forget webtorrent guys
<akira01> found new nodeshit
<akira01> peerflix do the work
<omanom> glad you found an alternative!
GalaxyNova has joined #kisslinux
progenyx has joined #kisslinux
Guest94 has joined #kisslinux
<Guest94> Anybody wanna tell me why [ -z "$string" ] and [ -n "$string" ] exist when you can just do [ "$string" ] and [ ! "$string" ]
<Guest94> :|
<GalaxyNova> posix is weird like that
<illiliti> -z and -n are more safer than just [ "$var" ] and [ ! "$var" ]
<Guest94> Why is it?
<illiliti> ... Unexpected results would occur if such a string expression were used and $1 expanded to '!', '(', or a known unary primary. ...
<Guest94> ok
Uks2 has quit [Quit: Byee]
Uks2 has joined #kisslinux
GalaxyNova has quit [Quit: Whoooooshh]
Guest94 has quit [Quit: Client closed]
vulpine has quit [Ping timeout: 272 seconds]
vulpine has joined #kisslinux
kyxor has quit [Quit: kyxor]
kyxor has joined #kisslinux
<kyxor> Hi, I finally figured out how to make the stupid ssl tunnel work, god damn it and my mistake was trying to route it though the localhost, but for some reason that crap does not work
<kyxor> this is very much under construction, but let me know if it loads fine for you
<kyxor> cause the weird part is that I can't open my own website from my server using https, but that's no big deal
<kyxor> I am running it on quark + hitch, very minimal in my opinion
<phoebos> lgtm
<acheam> looks good
Guest3 has joined #kisslinux
micro_O has joined #kisslinux
diegs has quit [Ping timeout: 240 seconds]
GalaxyNova has joined #kisslinux
<GalaxyNova> Oh cool
<GalaxyNova> the new website is out
<micro_O> kyxor I think you can remove all the 'ws' classes by taking advantage of the word-spacing css property
<micro_O> something like @media (max-width: 1024px) { .hex { word-spacing: '-100%' } }
<kyxor> micro_O: thanks I'll try it
<micro_O> yeah it will take a bit of work since everything is inline
<Guest3> Need to redirect standard output to file which needs root permission. Tried sudo printf '%s\n' "text" > file buit didn't work. I know I can just run script as root but a lot depends on the user's environment so wondering if there's an easier way???
<Guest3> Err and tried doas too
<micro_O> Guest3 tee is your friend
<micro_O> it would be something like printf '%s\n' "text" | sudo tee -a file
<micro_O> if you want it to print to stdout as well
<micro_O> or `echo text | sudo tee file`
<Guest3> ok
<Guest3> micro_O do you use sudo or doas
<Guest3> as a personal question, just wondering
<micro_O> Guest3 been using ssu
<micro_O> which had a better name back in the day but i cannot remember now
<micro_O> at least its easier to type now
<micro_O> i think it was sls before
<Guest3> Wondering if the tee command could be invoked without needing pass. Just for that specific file
<micro_O> with sudo, definitely
<micro_O> with doas, maybe
<micro_O> with ssu, definitely not
<Guest3> I tried doas but didn't get it working. The syntax is permit nopass USER cmd tee args [args]
<Guest3> The file from the script is shell variable so I am not sure...
<micro_O> if you need a user to be able to create a file in a particular directory, maybe its worth making that directory owned by a group that user is part of?
<micro_O> Guest3 have you tried making the tee command in doas.conf the full path to tee? something like /usr/bin/tee or whatever
<phoebos> ssu doesn't need a pass anyway :p
<Guest3> Ahh yeah in doas.conf I was putting the full path, but I guess it doesn't like the shell variable in the script so in the actual script I put the full path
<Guest3> Err the full path to the file not the command
<micro_O> sometimes i do a search like this to see what other folks do
<micro_O> looks like unbound has some scripts with comments on how to do it, thats nie
<micro_O> permit nopass Guest3 as root cmd tee args /path/to/file
<micro_O> but if the script has argv[1] set to "relative/path/to/file" i am not sure if it will match
<Guest3> The command in the script in doas tee "${path}/file" but if I put the full path "/path/to/file" it works.
<Guest3> Ok i got it working with the var too. It has to be exact with no extra slashes
<micro_O> cute youtube client https://github.com/pystardust/ytfzf
<Guest3> That's what I am using currently
<Guest3> Only issue I found was that thumbnail doesn't work with st if you have the swallow patch
<micro_O> #patchlife
<acheam> yeah images in the terminal are always shaky
<acheam> actually they're using ueberzug so not as shaky as w3m
Guest3 has quit [Quit: Client closed]
<acheam> the one thing I really dont like about the new website is the use of dashes instead of spaces
<acheam> it makes the page feel much too busy IMO
<acheam> I also miss the testimonials :(
<acheam> oh nvm its in the wiki now
<micro_O> I was wondering why i had so much trouble reading the new site
<acheam> i do love the inline GH logo
<micro_O> But-Now-that-You-mention-it,--it-seems-much-harder-to-scan-with-dashes-in-between-words------------
<acheam> indeed
<micro_O> Actually-Its-Even-Worse-With-Capitalization
<acheam> actually the GH logo is missing an alt
<acheam> so if your browser doesn't load inlined images, you wont see a link
<micro_O> its pretty dope when the browser does though
<acheam> indeed
<micro_O> Also, should we be writing the distribution over two lines, like
<micro_O> |/
<micro_O> |\ISS
<acheam> i think it looks cool
<acheam> is it accesible?
<acheam> probably not
<micro_O> not at all
<micro_O> petscii would be cooler
<GalaxyNova> the community section was also removed
<acheam> well now we have kisscommunity.org
<GalaxyNova> i think there should at least be a link to kisscommunity.org
<GalaxyNova> somewhere in the wiki
<acheam> (although to be completely frank, the design of is repulsive to me)
<GalaxyNova> lol
<GalaxyNova> i was worried the gh logo was making a request to github at first
<acheam> dylan would never do such a thing