<illiliti>
glibc ldd doesn't print it and i see no point why it should
<wael[m]>
yep that fixed it
<wael[m]>
maybe that should be removed in sh
<wael[m]>
-alternatives
<illiliti>
it should yeah
<wael[m]>
shall i open a PR?
<illiliti>
yeah why not
<wael[m]>
alright
<illiliti>
dilyn: tinyramfs doesn't use readelf
<illiliti>
i don't remember why i implemented only ldd
<wael[m]>
i messed up the commit name but it works
<illiliti>
thanks for the PR
rohan has quit [Ping timeout: 256 seconds]
<dilyn>
I just meant try readelf to see a better picture than what ldd provides
<dilyn>
but it doesn't matter in this case :v
Torr has quit [Quit: leaving]
rohan has joined #kisslinux
ella-0 has joined #kisslinux
ella-0_ has quit [Read error: Connection reset by peer]
<testuser[m]12>
Hi
<dilyn>
o/
<dilyn>
testuser I've been thinking lately
<dilyn>
should we provide a generic kernel and initramfs in the main repo?
<dilyn>
I think it's basically trivial to provide now
<testuser[m]12>
Sounds good, can save a lot of time for newcomers
<dilyn>
that's what I was thinking
<dilyn>
it basically trivializes a new setup and offers the *option* for doing the "hard part" of Linux
<wael[m]>
well if a generic kernel is going to be provided, would there be a way to provide a custom configuration?
<dilyn>
Theoretically it minimizes questions concerning kernel problems, but I have no idea what clusterfuck of a monsoon it might open up because we essentially end up removing a major bar to entry (at least, those who aren't knowledgeable enough to know "oh I can just copy this binary package from arch and create a handy initramfs thanks to illiliti")
<dilyn>
It's... semi-trivial to customize a kernel config in this case...
<dilyn>
make oldconfig reconciles the config and sanity checks it, so if you echo some new config option to the end of a .config file the Kconfig system will make sure everything works as it should
<dilyn>
users may be prompted to make decisions on new config options, but you can spoof selections on those in an automated way as well
<dilyn>
the way to provide a custom config would, I think, be best handled by forking a package, but that might just be due to lack of imagination on my part. Intervening in KISS package builds is difficult by design. You could always make it so that kernel builds prompt for config options, but that is the opposite of solving this problem unfortunately:\
<testuser[m]12>
wael[m]: kiss fork?
<dilyn>
you could use a pre-build hook...
<testuser[m]12>
What if u want extra patches
<dilyn>
in that case you just have the normal problem with any other package in KISS
<dilyn>
not really a blocker here
<dilyn>
i think we might have to create some new variables for KISS specifically for this venture; an install path for the kernel (because nobody ever seems to just want to use /boot, thanks arch...), potentially the location of a config file a user might want to supply (keep a simple pre-build hook...)...
<testuser[m]12>
Adding build variables is just use flags
dilyn has quit [Quit: Connection closed]
dilyn has joined #kisslinux
<dilyn>
I agree that it's USE_FLAGS but I think the kernel presents a special case due to the uniqueness of the kernel
<dilyn>
alternatively we could establish that all kernel-related artifacts are installed in a specific location to be handled by the user; /usr/src/ is the de facto
<wael[m]>
why would someone choose another kernel path different to /boot
<testuser[m]12>
Just have /boot/vmlinuz as default
<testuser[m]12>
Genpoo probably has a fixed path too
<wael[m]>
arch does too I believe
<dilyn>
have you ever seen documentation for other distros on where vmlinuz goes xD
<testuser[m]12>
Arch is not source based
<dilyn>
/efi, /boot/efi, /boot...
<wael[m]>
*efi is just for efi studd why would kernel be in there
<dilyn>
I mean I'm fine with instituting the policy of '/boot is the location deal with it', it's just a thing to keep in mind
<dilyn>
because if the kernel is built as an EFI binary (which it almost certainly would be) it theoretically goes there
<wael[m]>
I have never seen that happen
<testuser[m]12>
Also a portable kernel would need blobs so do we package 200mb of blobs too
<wael[m]>
the efi partition would be mounted at /boot if that's the case
<dilyn>
we would have to have a linux-firmware package or bundle firmware with the kernel (i vote a separate package)
<dilyn>
wael[m]: again, other distributions handle this differently
<wael[m]>
non separate firmware is a bad idea
ejjdhfjsu has joined #kisslinux
<dilyn>
agreed
<testuser[m]12>
Gnetoo just installs kernel source to /usr/src and u have to build manually
<dilyn>
yes
<dilyn>
that's the historical location for all kernel builds
<wael[m]>
gentoo kernel sources
<wael[m]>
then have to do some weird eselect stuff
<wael[m]>
btw, didn't dylan intentionally want go have the kernel managed by the user?
<dilyn>
correct
<wael[m]>
hmmmm
<noocsharp>
package splitting when
<wael[m]>
wdym
<dilyn>
-dev et al
<noocsharp>
multiple packages from the same build
<dilyn>
never not once
<wael[m]>
I think this is difficult to do
<noocsharp>
most major package managers have it
<wael[m]>
many distros do this by having a package entry function inside the build file of some sort
<dilyn>
and its the wrong choice for a sources distribution I think
<wael[m]>
does gen2 do it
<dilyn>
I'd hope not that would be insane
<noocsharp>
idk why major distros don't package split kernel modules
<wael[m]>
the idea could save some storage space but muh package count
<noocsharp>
[nihal@debussy ~]$ du -sh /usr/lib/modules
<noocsharp>
198M/usr/lib/modules
<dilyn>
it's a lot of labor to split modules
<dilyn>
a standard Ubuntu kernel generates about 1GB worth of modules
<noocsharp>
which i need 10MB of
<wael[m]>
wait hold on splitting kernel modules???
<wael[m]>
nahhhh that's a lot of work
<dilyn>
technically canonical *does* split some modules up
<dilyn>
between modules, modules-extra, others...
<dilyn>
it's certainly a lot of labor
<illiliti>
8.0M/usr/lib/modules
<noocsharp>
well you're presumably not running artix illiliti
<noocsharp>
i'm using their stock kernel
<noocsharp>
why do you have modules at all, why not build a static kernel?
<dilyn>
-rwxr-xr-x 1 root root 9.8M May 30 17:41 vmlinuz.efi
<dilyn>
fuck a module
<dilyn>
it'd be smaller but kvm adds a lot of shit:X
<wael[m]>
vmlinuz.efi?
<dilyn>
just a name i chose for no good reason
<wael[m]>
it made me get confused for a moment
<dilyn>
:P
<illiliti>
i need modules to test some things
<illiliti>
tinyramfs, mdevd and others
<noocsharp>
ah
ioraff has quit [Quit: ioraff]
ejjdhfjsu_ has joined #kisslinux
ejjdhfjsu has quit [Ping timeout: 248 seconds]
soliwilos has quit [Remote host closed the connection]
soliwilos has joined #kisslinux
<virutalmachineus>
what you guys think of harden kiss version?
<virutalmachineus>
like gentoo harden
rohan has quit [Ping timeout: 252 seconds]
rohan has joined #kisslinux
ecs has quit [Remote host closed the connection]
ecs has joined #kisslinux
ecs has quit [Remote host closed the connection]
ecs has joined #kisslinux
<Ogromny>
If you want a "good" default kernel .config, we can use voidlinux .config, that's what I did and it worked pretty well
<Torr>
Book recommendation: Gene Wars - Kristin Dawkins
midfavila has joined #kisslinux
ioraff has quit [Quit: ioraff]
ioraff has joined #kisslinux
midfavila has quit [Remote host closed the connection]
midfavila has joined #kisslinux
<noocsharp>
how's the math learning going mid?
<midfavila>
slowly
<midfavila>
i've been having a lot of trouble sleeping lately, and there's been a lot of stuff that needs to be addressed beyond my studies
<midfavila>
i haven't touched my textbooks in maybe two weeks
<midfavila>
feelsbadman
<midfavila>
i was thinking of trying to finish off my logic textbook this upcoming week since there's only a few chapters left
<midfavila>
atm it's covering the use of trees to reason about statements in monadic predicate logic
<midfavila>
i think the remaining chapters are on generalized predicate logic, set theory, and basic proofs
<midfavila>
oh, the discrete math textbook i got recently is pretty neat, though
<midfavila>
it focuses on applications, so there's chapters on stuff like digital logic
<noocsharp>
what's the "monadic" in monadic predicate logic refer to?
<midfavila>
the fact that primitive propositions have two components, and are always expressed as something of the form Ax
<midfavila>
as opposed to propositional logic, where primitive propositions are expressed simply as A
<midfavila>
capital letter is the predicate, and the lowercase is the name. predicates tend to ascribe traits to names
<midfavila>
so to use a cliche, M might be "x is a man", and s might be "Socrates", so to express the claim "Socrates is a man" in MPL would require the prop Ms
<noocsharp>
ah
<midfavila>
truthfulness of a prop is established by determining whether the name in a prop is a member of the set of things referred to by the predicate
<midfavila>
so on and so forth
<midfavila>
MPL also contains quantifiers, unlike PL
<noocsharp>
ah, ok, i've used mpl but didn't call it mpl
<midfavila>
what was it referred to in your work?
<midfavila>
s/to/to as/
<midfavila>
christ I can't grammar today
<midfavila>
didn't get any sleep last night xwx
<noocsharp>
actually wait, MPL is just a subset of first-order logic
<midfavila>
yeah
<midfavila>
GPL is first-order logic proper, I think
<midfavila>
i think it's also called polyadic predicate calculus..?
<midfavila>
or is that also a subset of first-order logic?
<noocsharp>
it must be
<noocsharp>
subset or the same
<noocsharp>
non-strict subset
<midfavila>
looks like it
<midfavila>
i'm looking at the BNF grammar for first-order logic on wikipedia right now
<noocsharp>
but no sane perso nwould call it polyadic predicate calculus
* midfavila
shrugs?
<midfavila>
idk dude i just like
<midfavila>
read books and act like i know things
<midfavila>
sometimes people even believe me
<noocsharp>
when in doubt, use the less fancy expression of the idea
<midfavila>
yeah, no, fair enough
<noocsharp>
lest you become a philosophy professor
<midfavila>
if i can't even attend university i don't think i need to worry about becoming a professor of anything lmao
<midfavila>
brb checking on dinner
<midfavila>
and we're back
<noocsharp>
what's for dinner?
<midfavila>
housemates are making nachos tonight
<midfavila>
i don't have to do any of the cooking since i made dinner yesterday
<midfavila>
(that was butter chicken curry)
<noocsharp>
nice
<midfavila>
yeah, 'twas okay
<noocsharp>
i've been on a bit of a lua binge recently
<midfavila>
oh yeah?
<midfavila>
doing what?
<noocsharp>
i made a ledger program, and started poking at ideas for a more minimal lilypond that doesn't use guile
<midfavila>
neat
<midfavila>
i've not heard of lilypond -- what is it?
<midfavila>
i'm assuming it's a GNU thing, since, you know, guile
<midfavila>
is it also a financial tool?
<noocsharp>
music engraving engine
<noocsharp>
sheet music
<midfavila>
aaah
<midfavila>
gotcha
<noocsharp>
it uses guile as an extension language, which for me makes most of the code incomprehensible
<midfavila>
that's fair
<midfavila>
lisp can kind of be like that at times
<midfavila>
wait until you learn about collector functions
* midfavila
shudders
<noocsharp>
what?
<midfavila>
one of scheme's party tricks. basically it's a way to treat the state of a function as an immutable parameter
<midfavila>
and computation is expressed by passing that parameter to recursive calls of the same function
<midfavila>
it's... weird
<midfavila>
the end result is that the base case is evaluated first
<midfavila>
i don't fully understand the concept myself, and even GJS called it "black magic"
<midfavila>
oop, food
<noocsharp>
just sounds like recursion to me
<midfavila>
i wish i could explain it better
<midfavila>
but it's not the same as a regular recursive function, that's for sure
<noocsharp>
recursion with an extra parameter
<midfavila>
except that parameter is a function that's constantly being modified