<virutalmachineus>
is there a openbsd kiss edition?
Torr has quit [Quit: leaving]
<wael[m]>
i believe someone has tried it with bsd
<wael[m]>
not sure which one of the bsds
<wael[m]>
but i believe the openbsd package manager is sufficient
ella-0_ has joined #kisslinux
ella-0 has quit [Ping timeout: 268 seconds]
dilyn has joined #kisslinux
<dilyn>
rohan: gdk-pixbuf is in repo/extra
<dilyn>
virutalmachineus: as stated before, a BSD KISS has been tried several times, but they all ended unsuccessful (for a variety of reasons)
<dilyn>
I've been out of the loop for too long in this community and for that I apologize
<dilyn>
I can point the kisscommunity.org URL to phoebos' site I suppose; that begs the question of how we would want to maintain an ML
<dilyn>
I'd also love to just entirely drop github outside of *as a mirror* and do PRs strictly via format-patch/am...
<dilyn>
given dylan is gone for... who knows how long, I think we can take more liberties as a community. the Org has matured enough at this point (along with some fresh blood) that I think we can make something better and bigger of this
<dilyn>
i'd also like to see the server I'm paying for be used more...;)
<wael[m]>
the what
<wael[m]>
what is the server
<dilyn>
ah
<dilyn>
I have a linode server which previously hosted the community's mailing list
<dilyn>
it also hosts the git and fossil mirrors for the community repository
<ioraff>
sourcehut would be nice if it had orgs. that would also handle the mailing list
<dilyn>
:)
<wael[m]>
how would someone even upload a package
<dilyn>
patches are pretty easy; you make a commit and do a git format-patch
<dilyn>
you can use git send-mail as well, and maintainers would use git am to apply the patch
<wael[m]>
hmm interesting
<dilyn>
yeah
<wael[m]>
well the only way you can get most kiss users to go to it is to mirror the current kiss community repo
<dilyn>
basically github is gross; we have to rebase/squash every PR because github injects gross merge commits for every PR if you don't do a rebase :|
<dilyn>
git am avoids that
<dilyn>
afaik even when I was maintaining kiss fewer than ten people were using this mirror of community or the mirror of the main repo
<wael[m]>
well it's last update was in 5-31
<wael[m]>
that's pretty old
<dilyn>
that's about in line with the time I stopped maintaining community and ioraff stepped in:)
<dilyn>
but they don't have access to the server
<wael[m]>
how does someone gain permission to commit
<dilyn>
currently we don't have any rules or infra for that
<dilyn>
I never implemented it, though it was on the roadmap
<dilyn>
then dylan returned and I swiftly abandoned my post... and now here we are
<wael[m]>
well since Dylan uses github why can't we stay at thay
<wael[m]>
s/that/that
<dilyn>
that was the argument I made before
<dilyn>
but given the volatility of dylan's presence, I think that now gives us license to do whatever we want
<dilyn>
the overwhelming support last time was to completely abandon github (shithub, as they called it...) in favor of a self-hosted thing
<dilyn>
I'd favor simply not supporting GitHub PRs and only supporting ML patches this time around
<wael[m]>
why not something like codeberg? Its similar to how github functions
<dilyn>
i liken codeberg to gitlab
<dilyn>
in that I don't like that
<dilyn>
them*
<wael[m]>
why not
<dilyn>
well, codeberg has always been slow for me
<dilyn>
neither are as intuitive or as simple as I would like (I hate navigating freedesktop projects, for instance)
<dilyn>
this is a personal disdain, but this is my position
<wael[m]>
hmm
<wael[m]>
well switching to old style git is gonna take time to get used to
<dilyn>
maybe, but I think a large contingent of users have refused to contribute because the primary/only way has almost entirely been through github
<illiliti>
dilyn: you can merge PRs in cli to avoid merge commits. it works without rebasing if PR branch isn't diverged from master(i.e fast-forward merge)
<dilyn>
that's true
<dilyn>
but nobody is doing that and git am is easier far easier for our use-case
<wael[m]>
I have only contributed via github and world-write suckless repositories
<dilyn>
ah yes those...
<dilyn>
hm
<dilyn>
I suppose I would need ioraff: phoebos: and others to chime in
<wael[m]>
since KISS is so small I believe that world read write repositories could work
<dilyn>
yeah. at this point I am far more interested in community feedback than I was before.
<dilyn>
Before I wanted to maintain continuity in the hopes that it wouldn't be too jarring when Dylan returned. But now, I'm not in charge. and I think we can have far more liberty in our direction this time around
<dilyn>
essentially we've become a steering committee
<wael[m]>
what would the kiss community do once Dylan returns?
<ioraff>
I'm in favor of ML-only development. it'd disentangle the distro from any particular hosting platform
<wael[m]>
what is ML-only?
<ioraff>
mailing-list-only
<dilyn>
I think at this point we should accept that our direction for the main repositories is incongruous to Dylan's work, and KISS-Community is now it's own thing with its own rules, obligations, and decisions
<dilyn>
^
<dilyn>
Constantly trying to emulate Dylan was a virtuous goal in the past but I think it's only going to limit us -- ideally, I think, KISS-Community will follow the general intention of the community. at least broadly speaking
<dilyn>
ofc, given that the maintainers are opinionated, the Org will be opinionated, and I think ML-only is an excellent example of that stance
<wael[m]>
I have no idea how a mailing list works tbh, it feels a bit old
<dilyn>
it's certainly old, but it's tried and true
<dilyn>
and well-supported and established
<dilyn>
our ML worked quite well after I worked out most of the kinks
<dilyn>
basically it's a quote that encourages a BDFL, someone who has veto authority
<ioraff>
I get what it's trying to say, but the analogy is just wrong. a horse is prettier, sure, but a camel is stronger and can survive in much harsher conditions.
<dilyn>
the point is that the committee tried to design a horse and arrived at a camel
<dilyn>
Ultimately I would move that we actually rigorously define the KISS-Community org, much like I wanted to do 20 months ago:)
<dilyn>
I would like some actual infrastructure and rules surrounding our org to make things easier. that way it isn't just me, or git-bruh, or ioraff, or whomever, reviewing and accepting PRs and making changes
<dilyn>
rules for who's in charge of what, who access to what, what needs review and what doesn't, bla bla bla
<wael[m]>
why not have multiple people in charge
<dilyn>
that's the goal
<wael[m]>
this is community after all so
<dilyn>
I don't want to be in charge lmfao
<wael[m]>
bro you are literally a Dylan typo
<dilyn>
:P
<dilyn>
there was speculation that I was secretly dylan
<wael[m]>
the British looking guy in your github profile is your right
<dilyn>
do I look british?!
<dilyn>
that's the nicest, meanest thing anyone has ever said about me
<wael[m]>
you look like you operate for a British radio station who makes funny jokes
<dilyn>
man that would be the dream
<dilyn>
instead i'm just some bloke who lives on that strange border between new england and the midwest:'(
<wael[m]>
so you are British haha
<dilyn>
if only
<ioraff>
pennsylvania?
<dilyn>
pittsburgh PA!
<dilyn>
you're a genius
<wael[m]>
wowie
<ioraff>
I mean the only two options were PA and OH. or ontario I guess
<ioraff>
I'd like to visit pittsburgh sooner than later
<dilyn>
ohio is absolutely in the midwest
<dilyn>
I'll die on this hill
<dilyn>
pgh is chill. if you find yourself here lmk and we can get a drink:P
<dilyn>
i'm trapped here for a year or two yet
<ioraff>
fair enough. and will do
<wael[m]>
kiss linux convention - at a random bar
<ioraff>
sounds about right
<dilyn>
goals
<wael[m]>
~~also holds 80% of the total kiss linux userbase~~
ella-0 has joined #kisslinux
ella-0_ has quit [Ping timeout: 256 seconds]
<virutalmachineus>
is kiss going to be for the niche user or there's going to be iso disks?
<wael[m]>
there are isos available
<wael[m]>
not officially however
<virutalmachineus>
yeah i saw, i mean a full installer
<wael[m]>
well in kiss I don't think you need an installer
<wael[m]>
look at mkrootfs
<wael[m]>
would be fun to try to make tho not sure how
<dilyn>
there was an initiative to create a KISS installer ISO a few years back from eudaldgr
<dilyn>
I tried to make a KISS-kde installer iso leveraging their project but I didn't quite have the time or energy for it
<wael[m]>
kiss-live, I didn't know it had an installer
<virutalmachineus>
tails-kiss-live edition with browser only would be cool
<illiliti>
we don't have binary repositories. we would have to either rebuild deps for each package on each update or store binaries somewhere
<illiliti>
it's also impossible to build bloated packages with github actions, such as rust, llvm, firefox...
<illiliti>
ci would simply die
<phoebos>
at our scale it would be more efficient to build binary packages on an actual computer, so that not everything has to be downloaded and set up for each package
<phoebos>
but there could still be a gh mirror for testuser[m]12 to play with
ioraff has quit [Ping timeout: 268 seconds]
ehawkvu has joined #kisslinux
<testuser[m]12>
I've seen people build chromium on it lol
<testuser[m]12>
Ill store bins on my server, sha256 will be from ci
<illiliti>
GitHub will remove any cache entries that have not been accessed in over 7 days. There is no limit on the number of caches you can store, but the total size of all caches in a repository is limited to 10 GB.
<illiliti>
that have not been accessed
<illiliti>
we could setup an action to access them
<illiliti>
iirc it's possible to setup cron job
<testuser[m]12>
but i dont think we can expose the actions cache to public
<illiliti>
why would we need that
<testuser[m]12>
hmm yea it would be useless outside large packages
<illiliti>
you can expose artifacts if you really need that though
<illiliti>
dilyn: if you're going to reinstate ML, consider enabling ipv6
fitrh has quit [Remote host closed the connection]
<illiliti>
nice
soliwilos has quit [Remote host closed the connection]
soliwilos has joined #kisslinux
<wael[m]>
ehawkvu: the kiss-live github link leads to kisslinux.xyz
<wael[m]>
also very nice its rewritten
<wael[m]>
but, why not have the entire iso be built without needing a tarball?
<ehawkvu>
wael[m]: We already have a tool that generates tarballs, and it would be a duplication of effort
<wael[m]>
interesting
<ehawkvu>
any tarball made with mkrootfs should work with it
<wael[m]>
just a off-topic question, isn't it possible to bootstrap KISS?
<wael[m]>
i mean this as in to make KISS make a system by itself
<wael[m]>
i made this a reality using XBPS and having it install all the system packages straight from nothing
<wael[m]>
well, using binary packages of course
<wael[m]>
my question is already answered due to how the previous kiss-live did it, it packaged all the system packages and installs them at runtime, via KISS_ROOT i assume.
<wael[m]>
i really liked that method of isos because of how simple and small the iso was
<ehawkvu>
The previous kiss-live script would use the prebuilt packages, store them in the iso, and on boot, extract them to /
<ehawkvu>
the new one does essentially the same thing, except it just extracts one archive which has the entire system
<wael[m]>
squashfs?
<ehawkvu>
Long term I'd like to move to squashfs
<ehawkvu>
I was more concerned with just getting the iso to build
<ehawkvu>
I'd like to have it be similar to either void's mklive or cloveros's livecd build script
<wael[m]>
how big is the iso?
<ehawkvu>
for kiss-live? a little over 100MB
<wael[m]>
gooood
<ehawkvu>
Yeah it's pretty slim
<ehawkvu>
It could likely get below 100MB if some more programs were dynamically linked
<ehawkvu>
But 109MB is ok to me
<wael[m]>
why are they not dynamically linked?
<ehawkvu>
a lot of the software in core/ has been made to be statically linked (which I think is good btw), but in this one case it makes the isos a bit bigger
<ehawkvu>
not by much, maybe a few megs
<testuser[m]12>
Can there be a mix of generic and native code in a binary if you statically link a library that was buily with march=native ?
<ehawkvu>
testuser[m]12: I would think so
<testuser[m]12>
Then shouldn't the tarball rebuild itself atleast once
<virutalmachineus>
you think kiss will ever be as big as void?
<wael[m]>
oh god I hope not
<ehawkvu>
mkrootfs will build all packages with general cflags so that should be a non-issue
<virutalmachineus>
i mean as popularly
<testuser[m]12>
But it doesn't link every package to the packages built for the tarball right? It just builds them all individually ehawkvu
<wael[m]>
@virutalmachineuser : sure it means more packages or maintainers but the demand is too much
<testuser[m]12>
Like static curl will link to the system openssl and zlib it was built on, not the one built for the Tarball
soliwilos has quit [Remote host closed the connection]
<ehawkvu>
testuser[m]12: I see what you mean - I generally do rootfs builds in a chroot to avoid this problem
soliwilos has joined #kisslinux
<ehawkvu>
I'm not entirely sure how you mitigate it completely, and conviently
<virutalmachineus>
i build it two times
<ehawkvu>
Perhaps the CFLAGS could be modified to put the libraries that are available in the new rootfs before the host systems
<ehawkvu>
So that only the general ones are linked against
<testuser[m]12>
You'll have to bootstrap a new sysroot
<testuser[m]12>
To do it cleanly
<testuser[m]12>
I'll try
soliwilos has quit [Remote host closed the connection]
soliwilos has joined #kisslinux
chomwitt has joined #kisslinux
chomwitt has quit [Ping timeout: 268 seconds]
chomwitt has joined #kisslinux
chomwitt has quit [Ping timeout: 255 seconds]
soliwilos_ has joined #kisslinux
soliwilos has quit [Remote host closed the connection]
<wael[m]>
ehawkvu: in mkrootfs, if KISS is sourced as a 'library' why is die() and msg() still a function although KISS provides it? also there is no real check on if kiss exists in $kissloc, unless -e takes precedence and quits
<wael[m]>
i very much find the reliance on mkrootfs for kiss-live a bit weird
<ehawkvu>
wael[m]: I inherited the script from carbs, it was originally written by cem - I think it should be possible to remove the sourcing of kiss since it doesn't make much of a difference for the script in the end
soliwilos_ has quit [Remote host closed the connection]
soliwilos has joined #kisslinux
<ehawkvu>
testuser[m]12: looks like it might be possible to avoid the 2 step build, gcc at least supports the '--sysroot' flag
<ehawkvu>
which makes gcc act like 'dir' is root, and will look for libraries in the right places
<ehawkvu>
clang also supports it
<wael[m]>
ehawkvu: why not have the system similar to how void does it? IIRC it makes a rootfs but from there it either makes/converts it to a iso or as is rootfs
<ehawkvu>
That's what we do already
<ehawkvu>
If they make a rootfs, then use that rootfs to make an iso, then we are doing the same thing
<wael[m]>
its not really automatic plus both are in seperate repositories
<illiliti>
by that day i'll be on netbsd with wayland
<ioraff>
I'm doubtful that anything besides drivers will be written in rust so long as linus is around.
<Torr>
After he let somebody write an email impersonating him about his behavior, feel things would surprise me.
<ioraff>
he certainly cares more about the kernel than his behavior or what people think of him
<virutalmachineus>
is rust really that bad?
<Torr>
virutalmachineus: Yes
<ioraff>
better than c++ imo
<Torr>
Actually it's worse than that.
<Torr>
(Responding to virutalmachineus)
<virutalmachineus>
don't you still rust for firefox? are all you running a text browser?
<illiliti>
worse ioraff much worse
<illiliti>
chromium
<virutalmachineus>
* all you guys running a
<ioraff>
how?
<virutalmachineus>
isn't chromium more bloated
<illiliti>
it is, but firefox at the same level tbh
<illiliti>
ioraff: let's start with the fact that rust has only one mainstream implementation
<konimex>
some people are insane enough to wait for ~24 hours to compile chromium (ymmv, ofc)
<illiliti>
i did wait 2 days and at 99% build failed due to OOM lol
<Torr>
Ouch
<ioraff>
I agree that's a problem. presumably to be solved in due time.
<ioraff>
konimex? been quite a while
<konimex>
aye, I've been busy for a while
<dilyn>
konimex :O :D <3
<dilyn>
chromium only takes forty minutes to build what are you talking about :eyes:
<konimex>
not in my thinkpad at least, 5 hours in I gave up after problems with the llvm stack
<Torr>
konimex: Went through something similar when trying to compile Librewolf.
<Torr>
I was like "ok, I don't have enough lifetime to spend on this"
<illiliti>
ioraff: it will never be solved if rust will keep breaking everything on each release
<illiliti>
see firefox for example
<illiliti>
it breaks on each rust release due to its dumbness
<illiliti>
how one in the sane mind would rely on rust if they simply can't provide stable releases
<illiliti>
there are many other problems actually
<illiliti>
for example it's impossible to use rust in ipv6-only environments due to its high relience on github
<illiliti>
so github is a single point of failure for rust
<illiliti>
if github is down, rust is down too
<virutalmachineus>
so mircosoft owns rust now
<virutalmachineus>
mircosoft hated open source, now they are at the top xd
<Torr>
They donated to Gnome ~.~
<Torr>
That's like promoting disgenics.
<Torr>
Dysgenics*
<konimex>
hey, Poettering's employed by them now
<Torr>
True
<illiliti>
i forgot in which stage are we? extend?
<illiliti>
or extinguish?
<Torr>
At the vicinity of extinguish.
<virutalmachineus>
how do we stop them!! is this the end to linux?
<Torr>
Depends on what u mean by end. It's certainly not gonna be Linux as we know today.
<Ogromny>
Linux is on a down path, rust everywhere, systemd, dbus, bloat, bash, glibc, flatpak/appimage/snap, etc
<illiliti>
yeah it's inevitable
<illiliti>
cuz it's being pushed by corpos
<Ogromny>
BSD is still solid tho
<Torr>
I wouldn't blame Systemd and its tentacles, Flatpak, Appimage and the likes, on Linux the kernel. Have not seen any core maintainer push that stuff.
<Torr>
These came mostly from some distro maintainers.
<illiliti>
look at the recent example
<illiliti>
they want to hardcode bash in linux
<Ogromny>
Honestly I think we should do a KissBSD
<Torr>
I'm out of the loop on the whole Bash thing, could somebody link me to the patch mail?