<midfavila>
gfdi the repos for my openwrt build are gone now, too
<midfavila>
today is turning into a real mess
<zr>
hm, where did #oasislinux on freenode move to, libera or oftc?
<zr>
ah, #oasis
<zr>
kernelc: you may want to ask in that channel too ^^
midfavila has quit [Ping timeout: 264 seconds]
<acheam>
zr: mcf was very happy about getting #oasis lol
<zr>
A redirect should be setup in #oasislinux :v
<acheam>
yeah that would be wise
<acheam>
or at least put it in the topic
zenomat has quit [Ping timeout: 264 seconds]
zenomat has joined #kisslinux
zenomat has joined #kisslinux
midfavila has joined #kisslinux
Uks3 has joined #kisslinux
Uks2 has quit [Ping timeout: 268 seconds]
<m3g>
zr: yeah, figured I just had to specify 4 for mod4 (super/windows key)
necromansy has joined #kisslinux
Uks3 has quit [Ping timeout: 268 seconds]
Uks2 has joined #kisslinux
mint has joined #kisslinux
zola has joined #kisslinux
mint is now known as beans
zola has quit [Client Quit]
<beans>
hello, im just playing with setting up a kiss system, and while building the kernel i have encountered issues related to the use of non-posix flags in the kernel build scripts. in the kiss wiki it says to disable the config flag CONFIG_IKHEADERS, which i do, but upon running make, it prints that it is overriding that config, causing the build to fail in the middle of a script. is there a
<beans>
simple step i am missing here?
<acheam>
without full logs its hard to tell unless someone's experienced the exact same error
riteo has quit [Quit: epic sleepless night moment]
m3g has quit [Ping timeout: 268 seconds]
eyepatchOwl has quit [Ping timeout: 272 seconds]
strajder has joined #kisslinux
smartin has joined #kisslinux
gtms has joined #kisslinux
progenyx has joined #kisslinux
claudia has joined #kisslinux
<claudia>
o/
<testuser[m]_>
hi
<claudia>
testuser[m]_: Do you use lagrange in a wayland session? On my end it does start with a fresh config, but after the second start it does crash so hard, that it crashes the compositor and I have to delete the *fresh* config.
<testuser[m]_>
works fine here, just resizing is a bit buggy and thats a known issue
<claudia>
testuser[m]_: This lagrange thing seems to be compositor related. Sway works fine.
<testuser[m]_>
wayland moment
<testuser[m]_>
does wayfire ont use wlroots ?
<soliwilos>
lagrange seems to run fine here on hikari as well, though I tend to simply use gmni. I haven't yet browsed gemini much, but like it so far.
mahmutov has joined #kisslinux
m3g has quit [Ping timeout: 268 seconds]
<claudia>
It does, but What I imagine is some protocol implementation or something is this direction.
<claudia>
Altough dwl builds on wlroots, things like firefox menus open outside the screen and some games have problems to capture the mouse (shooter -> 360c view)
mahmutov has quit [Quit: WeeChat 3.1]
mahmutov has joined #kisslinux
<claudia>
I have not much preferences until now, but bombadillo, amfora and lagrange seems to be all greate programs.
<claudia>
soliwilos: Have you changed keybindings in hikari? Mind sharing your config?
<soliwilos>
claudia: I've so far kept my hikari config close to the default, only changed input layout and added some actions.
<soliwilos>
Still want to see it?
<claudia>
nah :D
<claudia>
I thought I could spy something about this grouped thing.
<soliwilos>
Figured as much. :p
<claudia>
I struggled with a simple "alt + tab" window switch
<claudia>
because some windows misteriously where signed to other groups or something?
<soliwilos>
Ah, yes, like terminals are grouped together. You can use meta+shift+p and +n to swap between individual windows within a group.
<soliwilos>
Oh, I meant to write super, not meta.
<soliwilos>
You could change L+Tab to use "view-cycle-next".
<soliwilos>
Then super+tab will cycle between all views (windows).
<midfavila>
damn, I missed getting my textbook by literally two minutes
<midfavila>
fucking phone didn't go off
mahmutov has quit [Read error: No route to host]
<illiliti>
midfavila: did you get a chance to test libudev-zero?
<midfavila>
...fuck, no, sorry
<midfavila>
the past few weeks have been really hectic for me
<illiliti>
ok! no problem
<zr>
acheam: what was the consensus regarding cloaks? I think we were giving kisslinux/* cloaks out to anyone who's been here a while?
<midfavila>
. 0.
<midfavila>
gib
<acheam>
zr: yeah sounds good to me
_whitelogger has joined #kisslinux
<sad_plan>
cloaks?
<midfavila>
used to show affiliation with a project or channel
<midfavila>
alas, you can't have more than one, as far as I'm aware.
<acheam>
covers your ip when you /whois someone
<midfavila>
also that
<midfavila>
(but if you're that concerned about IP leaks then you should really be using a proxy or VPN)
<midfavila>
i block all social sites' primary interface, and if I find one that works in Links that isn't cancerous, I'll set my plumbing script up to convert URLs automatically.
<testuser[m]_>
nice
<midfavila>
mhm.
<midfavila>
I need to convert plumber to a C program, eventually
<midfavila>
as it stands I just abuse sed and case statements
<midfavila>
it's kind of crusty with the invidious redirect
<midfavila>
i need to take a look at it again
<midfavila>
but it does pass shellcheck, save for the latest addition of libredd.it (which I've just fixed)
<acheam>
midfavila: can links do a user script?
<acheam>
beacuse I have a privacy redirect one I wrote for lariza
<midfavila>
links can associate MIMEs with arbitrary programs
<testuser[m]_>
what program sends stuff to this script
<acheam>
also, check out teddit.net as well, you might like it more than libreddit (its my favorite)
<midfavila>
testuser[m]_ I have Pidgin set up to send links to plumber
<testuser[m]_>
oh
<testuser[m]_>
nice
<midfavila>
so that I can open content in local programs, instead of relying on an actual browser
<midfavila>
but you can use it with any program
<midfavila>
all plumber does is determine what kind of content a given URI refers to and then handles it appropriately. basically,
<midfavila>
it's xdg-open
<midfavila>
can POSIX case handle ||?
<midfavila>
so can I have a single rule for youtu.be||youtube.com
<midfavila>
?
<testuser[m]_>
you do "youtu.be | youtube.com)"
<midfavila>
very cool, thanks
<midfavila>
that'll clean up my plumbing script a lot
<testuser[m]_>
is case in C supposed to handle || ? i think you're just supposed to fall-through
<midfavila>
I don't know about in C
<testuser[m]_>
wat case were you talking about then, that handles ||
<midfavila>
POSIX shell
<testuser[m]_>
huh
<midfavila>
when I convert plumber to C, I'll probably have it determine preferred programs using env vars
<midfavila>
like $W3B for your web browser, $XPG for your "X pager" (so you can view plaintext), etc
<acheam>
ew env var configuration is so annoyin
<midfavila>
it makes more sense for plumber than using a dotfile
<midfavila>
ew, there are these jackasses forcing their APs to run with 40mhz wide broadcasts in my apartment
<midfavila>
absolutely d i s g u s t i n g
demo- has joined #kisslinux
demo- is now known as eyepatchOwl
sad_plan has joined #kisslinux
<sad_plan>
Im trying to wrap my head around testuser[m]_: 's LTO stuff. but Im a bit confused about the flags. it stated to modify the buildflags to those mentioned, which is fine. do I rechange them, to those other mentioned ones when building gcc and binutils or...? gcc even has other build flags in the build script aswell.
<sad_plan>
I dont get if I have to change buildflags each time gcc/binutils is built/updated anyhow. if thats the case here anyway..
<eyepatchOwl>
Pretty much any GUI application except for ones which require rendering to pixels (e.g. 3D) can be shoehorned into a web layout description without an escape to canvas. I think the Electron apps, editors, etc. establish this. However, I think complexity of the HTML spec forces browsers that want to render pages which exploit that complexity to be complex. If we assume that we want to interact with the world of rich user applications,
<eyepatchOwl>
e.g. managing electricity bill or banking, what would a KISS layout stack look like? The separation of layout / view / elements (HTML), styling (CSS), and interactivity (JS) make sense to me, but each one of those particular solutions seems horribly broken to me, from specification of HTML as a state machine, to the box model of CSS, to the massive download size + latency + lag introduced by JS apps.
schillingklaus has quit [Quit: schillingklaus]
sad_plan has quit [Ping timeout: 264 seconds]
<testuser[m]_>
https://termbin.com/9ggb maybe this clears it a little sad_plan ? ill update the readme
<testuser[m]_>
you wont need to change anything once the 3 steps are done
kqz has joined #kisslinux
eyepatchOwl has quit [Ping timeout: 268 seconds]
vbt has joined #kisslinux
sad_plan has joined #kisslinux
<sad_plan>
testuser[m]_: ah, yeah know I get it. you just add them along the way basiclly
<sad_plan>
now*
<testuser[m]_>
yeah
<noocsharp>
does anyone know is there's a tool that will show you how many lines of code will be included for a given kernel config?
<midfavila>
i don't think there is, but why are you interested in that?
<noocsharp>
*if
<noocsharp>
i think it would be interesting to know
<midfavila>
fair enough
<sad_plan>
I couldve sworn I seen that in the repos when browsing a while ago. but I dont recall what it was called.
<sad_plan>
it is however usefull to compare LOC. if thats of any consern to you that is
<midfavila>
as long as you're not trying to argue that fewer lines of code is necessarily better...
<sad_plan>
indeed
vbt has left #kisslinux [Leaving]
<midfavila>
meme minimalism
<sad_plan>
which is why imo suckless' software isnt inherently faster just because none of their tools is more than 2k LOC.
<midfavila>
besides, it's not about speed, necessarily
<sad_plan>
it does? coulvde sworn that was their *limit* atleast thats what Ive heard
<midfavila>
st alone exceeds that, afaik
<midfavila>
yeah, over 4k
<midfavila>
at least according to cloc
<sad_plan>
ah ok. I didnt know. but yeah, I dont have the impression of that speed is their primry goal here. just less lines of code really
<midfavila>
see,
<sad_plan>
^ that was the name of the tool that I found
<midfavila>
minimalism *happens* to cause a program to include fewer lines of code
<sad_plan>
Im gonna mark that just now, so I can compare some tools Ive been wanting to compare just for fun
<midfavila>
but that's merely a side effect, as far as I'm concerned
<midfavila>
i would argue that proper minimalism is concerned with getting the most out of the least. in short, it's optimizing for efficiency, combined with tight scope control
<sad_plan>
it should have less LOC, because less is more, you know?
<sad_plan>
this ^
<midfavila>
lines of code don't matter
<midfavila>
a good example of this is the fast inverse square root algorithm used by id
<midfavila>
that's waaaaaaaaaaaaaay more complicated than just having the computer calculate an inverse sqrt normally
<midfavila>
many more lines.
<midfavila>
but it's hilariously more efficient
<sad_plan>
exacly
<midfavila>
therefore, I would argue that it's more minimal, because even though it's a larger and more complex function, it's so much more effective that it outweighs its cost
<midfavila>
semi-related side note, this is why bloated code is unironically a risk to human life. i don't want soydevs programming traffic-stop microcontrollers or anything even remotely time-sensitive
<sad_plan>
there really no reason to choose a tool thats slower just because it has less lines of code. that is rather counter productive imo. Im all for having the best tool for the job. although I like experimenting though. trying out new tools is always fun for me.
<sad_plan>
hah, yeah indeed.
<sad_plan>
oh ffs, now gcc suddenly decides it no longer works after recompiling it with the new flags. fantastic
<testuser[m]_>
what does it say
<sad_plan>
its checking if gcc work, and the answer is no
<testuser[m]_>
check config.log
<testuser[m]_>
maybe you rebuilt the gcc from repo-main and added the graphite cflags ?
<sad_plan>
hm, i copied over from the tar archive, and rebuilding the original now. ill readd the flags in a bit.
<sad_plan>
I dont belive I did, I put LTO repo above the main one, but yeah, its possible. gonna rebuild straight from the repo in a bit
<testuser[m]_>
you can tell by the time, if it took 20-30 min then its the correct one
<testuser[m]_>
build time*
<midfavila>
we should really start using relative time.
<sad_plan>
it took way faster than that when I added the flags. so something is off :p
<sad_plan>
time is relative.
<midfavila>
= w=
<midfavila>
i mean in the sense of "relative to other things"
<midfavila>
like how LFS packages handle it
<midfavila>
where it uses standard build units.
<testuser[m]_>
you can confirm via "kiss s gcc" too
<testuser[m]_>
which lists the repos irs available from in order
<testuser[m]_>
its
<sad_plan>
Ive never built lfs *yet*, so thats no something I can compare to anyway
<sad_plan>
ah yea, ill check that for sure
<midfavila>
i mean
<midfavila>
it literally is
<sad_plan>
indeed LTO repo is above and before main repo
<midfavila>
the SBU is the time it takes to build... fuck, I can't recall off the top of my head
<midfavila>
but it's the first package you build in LFS
<midfavila>
and then everything else is said to be $X SBU
<midfavila>
so GCC might be 10SBU, meaning it takes ten times longer to build than the reference package.
<sad_plan>
I never got that far, I was messing with trying to get a cross compiler to work, but I dont belive I ever got to download all the files, before I ran out of time ofr whatever reason
<sad_plan>
aah, yeah that stuff. I recall now
<midfavila>
yeah... LFS takes a few days
<sad_plan>
a few DAYS? I thought it could be done in an aftermoon or something :p
<midfavila>
well, it obviously depends on your machine
<testuser[m]_>
lfs is mostly just copy pasting the configure make insyall
<testuser[m]_>
install
<midfavila>
also that
<sad_plan>
well yeah, if its as slow a a turd, its gonna take forever
<midfavila>
the real benefit of LFS is that it provides a springboard to learning about what makes up a system
<sad_plan>
yeah, mostly. still a fun thing to do though. has the potential for alot of experimenting
<midfavila>
"huh, what's this zlib thingie?"
<testuser[m]_>
yeah
<midfavila>
"Expect? I've never heard of that program, let's check it out..."
<sad_plan>
yeah, thats why I wanted to do it. so I could take a deepdive into what makes linux work
<sad_plan>
^that how we learn, aint it?
<cem>
LFS is bloat
<midfavila>
i mean
<midfavila>
unironically, yes
<midfavila>
but that's because the standard LFS install uses GNU tools
<testuser[m]_>
the function of core libs and some stuff about fhs is mainly what i "learnt" from it
<sad_plan>
does it matter though? lfs is mainly for learning purposes anyway
<testuser[m]_>
yeah it has dbus and crap too
<midfavila>
there are some people who actually daily LFS
<noocsharp>
i mean nothing forces you to install the bloat, right?
<sad_plan>
I know, but still. its still made for learning
<cem>
I'm saying that just as a joke but it's true
<midfavila>
noocsharp yeah, but then you have to know what you're doing
<cem>
It's a good resource, I learned a lot from it
<midfavila>
and LFS is sort of built to teach you to know what you're doing
<midfavila>
so like...
<noocsharp>
i mean if you use kiss you probably have a reasonable idea what you're doing
<midfavila>
i like to think I'm reasonably competent, sometimes
<midfavila>
...then I do something stupid
<cem>
Don't we all
<sad_plan>
oh ffs. gcc errors out again. where is this config.log file you mentioned testuser[m]_: ?
<testuser[m]_>
KISS_DEBUG=1 kiss b pkg then cd to tmpdir and find it
<testuser[m]_>
ill go now
<sad_plan>
I belive I found it. it exits out on code 77. Ill see if ill find what that means
<sad_plan>
¯\_(ツ)_/¯
demo- has joined #kisslinux
demo- is now known as eyepatchOwl
<acheam>
OK something really weird is happening
<sad_plan>
what?
<acheam>
I'm kiss i'ing a package, but it doesnt actually install
<acheam>
it doesnt go to /var/db/kiss/installed
<acheam>
or install anything to the actual system
<acheam>
only this one package, a fork of pinentry with 1 configure flag changed
<cem>
Probably a circular dependency
<cem>
That's what happened to me recently
<cem>
It just exits
<cem>
Although to be fair, it doesn't just exit, it actually crashes the shell
<acheam>
ahhh yeah thats the issue
<acheam>
no pinentry-gnome3 I suppose
<cem>
Most of the time the case is that a package is wrongly marked as a dependency
<cem>
You can also package two things together
<cem>
As is the case for freetype-harfbuzz
<acheam>
ah got it
<cem>
I have been thinking of writing a circular dependency checker
<cem>
So the shell warns the user and exits gracefully, instead of just crashing
<cem>
Because I spent so much time to understand what the issue was
<acheam>
that'd be nice
<illiliti>
i still don't know how to properly implement circular dependency checker in king. i would like to see your idea
<illiliti>
i know how to handle package1>package2>package1, but what about deep circular dependencies?