midfavila acheam Rio6 wttr.in has a few options for better output: oneliners (github.com/chubin/wttr.in#one-line-output), JSON output (github.com/chubin/wttr.in#json-output), and an experimental inline images format (github.com/chubin/wttr.in#map-view-v3)

I use that one-liner output formatting and shove it in my bar

Thanks ominous_anonymou!

midfavila golem required, at least on my system, libXpm, gettext, libX11, libXext, libXt, libtool, gcc, ncurses. However, I had to build golem itself from source (I used sourceforge.net/projects/golem/files/golem/0.0.5/golem-0.0.5.tar.gz) because there were a couple modifications that needed to be made to get it to compile.

shit i deleted the source that compiled and now I can't remember what i did

sigh

oof

ok i think these were the changes needed: ix.io/2Ga8

anybody have ideas on how to build cmake with a statically built curl/libressl?

get a bunch of symbol errors :S ix.io/2Gac

tried to pass OPENSSL_USE_STATIC_LIBS to cmake during the bootstrap step but it doesn't seem to work

set to TRUE?

I had ./bootstrap -- -DOPENSSL_USE_STATIC_LIBS:BOOL=ON

analogous to how they suggest building a Release build

what options does bootstrap --help give you?

nothing useful for passing options to cmake unfortunately

looking over the CMakeCache.txt I get, it seems to find libcurl.a and *ssl.a just fine, and adds the libs to ldflags... could it be that they're in the wrong order??

that would be irritating

haha wrong order has happened before: gitlab.kitware.com/cmake/cmake/-/issues/20082

yeah that's why I thought of it xD

would passing --system-curl work?

I also pass --system-curl yeah

hm, i'm all out of ideas! haha

xD yeah linker errors do my head in

i wonder if you're right about the .a ordering, do the ssl ones get added to ldflags after the curl ones?

yeah -lcurl is the first one and then the rest follow

dilyn libbytesize in dilyn-kde has a tpyo

:o

also congrats on having pam accept so many patches that must feel good

they weren't mine xD

what's the typo?

is it in the sed

checking my local fork now

--witk-gtk-doc=no => --with-gtk-doc=no

in the build

your fingers too fast you were already on the next work

*word xD

lol yeah

what a silly mistake smh

tyty

i really wanna create a github action that would build whatever package was updated post-commit that would at least catch build failures

isn't that what traverse is for

travis* w/e

the build never failed for me, hence not noticing haha

yeah i mean i dont care what CI

the logic should be 99% contained in regular old sh scripts and just configure the ci to call said script so its platform agnostic

imagine if you make travis build a package like firefox once its updated... travis will take hours just to compile rust and the other prereqs

i wonder if they have a limit on CPU time spent per CI job

different CIs have different limits

and many will not share the exact limits because a) they change as the infra gets better or worse and b) knowing the exact limits allow people to get around them

my CI is manual, i build all stuff on a build server before i commit

same here

where the server is my own laptop that i use daily

lmao

:>

same dilyn

ideally there would be three things coming out of a little CI automation

it's unclear to me that CI could properly test KDE lmao

1. build failures. 2. missing deps 3. automatic creation of binary packages

i mean 3 happens when building, but would require a bit of glue to make it easy for people or processes to use

and optimally 3 would produce reproducible binaries

sh4rm4^bnc yeah hopefully we can stand on the shoulders of other giants

but I bet it would require some auditing of each package and enabling

(i think there was previous discussion in here as an example, re: timestamps)

yeah, it's not trivial...

also when you have split debug info builds, the usual way to do it is with a build-id

if we were running something like urbit and not linux it *might* be easier

or if we were running an oooooold linux userspace it would be easier

just like, less lines of code xD

actually some packages in my distro are quite old

but they are stable and work well, so why update

i *think* it was void linux who started the trend that every package must be bleeding edge

usually the package in void repo is bumped already like 4 hours after a release came out...

yeah i got screwed by that once

bleeding edge qt5 is bae

reddit.com/r/voidlinux/comments/73caqx/sshfs_broken_after_update lol

on my home systems i rarely update packages unless there's a bug that hits or there's a security issue that comes up. i'm on 1.5Mbps DSL so I gotta be judicious in my downloads...

Press arrowkey up in the terminal, and see if you cant find it C:

adding --verbose to cmake bootstrap, it would appear that the linking step doesn't actually call the ssl libs. ix.io/2GaF

I think I'm being trolled

if libcurl was built statically, wouldn't it already include the links to libressl?

wouldn't that be the other way around?

jk

hm one would think

i have no idea how it works, i was just wondering

i've never identified so much with a sentence before

mood

well what do you know, libressl wasn't actually static

wtf

eat my shorts ld

:thonk:

rebuilt libressl, rebuilt curl. libcurl.a grew 80kb that day

presumably, my missing ssl symbols

welcome to ld where the linking is made up and the libraries don't matter

are you building an entire static KISS system, dilyn, or just libressl and friends?

static kiss

Very cool. I think I'm going to try that on my new laptop when it arrives

everything in core builds pretty fine with minimal changes

I fear I am far too dumb to be trying this, however

you'll never become slightly less dumb without doing something dumb and learning from it

:v

it's true tho

ld-musl-x86_64.so.1 is v important for a properly functioning system

don't not build it

yeah I would imagine so

something something c compiler doesn't work something something get rekt

god that's such a useless error

could be literally anything that causes it

like an 0 instead of an O

or a broken libc!

or a nonexistant $PATH!

or or or

or an awk failure, or sed failure, or cat failure in a configure script

or a malfuncting shell

or the dir isn't writeable lmao

or a broken implementation of shell testing functions

(thanks again suckless)

f

yeah

i spent all my morning meetings troubleshooting this problem instead of paying attention. ope

relatable

most of the time I'm messing around with KISS instead of actually paying attention to my classes

dylan has unleashed a monster none of us can control

this work is #valuable. suck it, school

basically

yeah

yesterday it took my class collectively an hour to install packet tracer on their winshit machines

so I was like...

"yeah fuck this. I'm leaving"

One time i set my CFLAGS with `-0s`, set ff to build, and went to attend my class (about 40 min), came back, "C compiler doesn't work"

rip

that's why I kinda want to set up a VPN so that I can ssh into my machine from pretty much anywhere. no need to worry about skids that way either

ack that is fucking triggering

mfw my smal brane can't debug a simple C program

I can't even say that I have morning brain or some shit, because it's 1PM

grah. still getting the same linker error

so via nm i have divined that MD5_Init is undefined. so where should it be defined...

md5.h certainly exists, so i can only assume...

it's defined in libressl, i wonder if you've graduated to the linking order problem now?

although that still wouldn't make sense to me since the curl libraries should already know about it

nm says that it's also undefined in (at least one of) the libressl libs (libcrypto.a)

something dastardly is happening

thanks to stackexchange i now have a rudimentary understanding of nm output. huzzah. for some reason, libressl is not including md5 (and a few other symbols relevant for libcurl->cmake). I wonder why...

stupid question but

did you check that there was an include statement?

included where?

in main.c or whatever, I guess. I'm not really a developer so it was just an idea

i'm building someone else's thing

but yes md5.h is include in at least the relevant crypto/*.c files in libressl

we'll see what ld is up to when it links libcrypto I guess

oh, also

totally unrelated

sometimes you'd think "how did they even get this thing to build on their own machine!?"

but hardware scalpers need to be themselves scalped

abolish currency

all the good PSUs were bought out and are being resold at 3x the price

at least on newegg

:/

maybe there's another supply chain issue like there was earlier this summer?

I suppose it's possible

that's one benefit to not having any current hardware! i don't have to deal with that

Hah, fair

I've been putting together a new rig for a couple months now though

PSU is the last component and I was waiting for Black Friday to get an RM850x

but then the RM850xes sold out

and the 750xes got their price hiked

oh damn you weren't kidding

No, I wasn't :v

850xes are like 300$ now

$288 cheapest for a $150 PSU

they were 180$ a week ago

i knew about the whole graphics cards hullabaloo lately but not PSUs

PSUs have been a problem for years believe it or not

and RAM

nah i had no clue, is it to power their cryptomining rigs?

Afaik it's something to do with a shortage of rare earth metals

i could see that -- China has kind of cornered the market on a lot of them: defensenews.com/opinion/commentary/…-china-cuts-off-rare-earth-supplies

that's called foreshadowing

sometimes it can be subtle

yeah it's certainly something to worry about. oil isn't the only nonrenewable resource being abused right now

I wouldn't be surprised if China replaces the USA in terms of global power

just based on their economic position

I feel there's little recycling of electronic devices currently

oh, absolutely

it's cheaper to just throw it away and replace it than actually fix or recycle shit

at least for me, I see lots of people tossing them right into trash

Same.

heh, even for those with recycle bins like cans and paper, people still throw everything in the trash can right beside

you guys'll like this, it's in the spirit of KISS -- John Deere locked down their tractor computer systems so farmers couldn't repair anything themselves

yeah, I heard about that a while ago

it's bullshit

even if it's just a mechanical problem you still need a tech to check it out

wideopencountry.com/john-deere-tractor-hack so some Ukrainians started making firmware the farmers could use instead

re PSUs: get a seasonic

aren't seasonics supposed to be kind of mediocre?

they'll always be cheaper than most other brands because most other brands are just rebranded seasonics

oh, really?

they're the best psu manufacturer for the last like, fifteen years

yarly

also, not only do we have to worry about the scarcity of rare earth metals; the amount of hydrogen in the atmosphere is nonrenewable (naturally, at least) and we're slowly running out :v

what do you guys use to buy computer stuff nowadays? are newegg and tigerdirect still good?

I just use ebay

go on newegg, look for what I want, look up prices on ebay, order whichever is cheaper

one of my last purchases through newegg was an EVGA 500W PSU and a "Replace Power" slimline 420W PSU

$50 and $40 respectively lol

hydrogen can still be extracted from water, though in a higher cost. Helium is getting delpleted more

pcpartpicker will show you what the prices are for most things at many major retailers and outlets

in addition to ebay, you can source parts from places like overclock.net and other enthusiast forums

merci

Hi there!

I'mjust getting started with this OS, but I've ran into a roadblock. I can't build GCC -- it errors out.

With what error?

I'll upload the log somewhere if anyone would like me to (just a bit awkward getting it off from my testing vm).

Oh sorry I didn't see your message

It says Error 2, which is obviously undescriptive

It's best for me to upload the log to pastebin. I'm doing that rn

you can upload the logs to ix.io by cat'ing the log into this curl command

| curl -F '1:1=<-' ix.io

Oh thank you! That'll speed things up a lot for me

er, replace that first 1 with a lowercase f

you can 'curl ix.io' to get instructions on how to use it

also i think ix.io has a cli you can download to use it

yeah `curl ix.io/client > ix`

The URL is ix.io/2Gbq

Thank you both for your help so far :)

I think it may be to do with this line:

checking whether objcopy supports debuglink... objcopy: /tmp/ls195000: cannot fill debug link section `x': No such file or directory

Oddly, I have managed to set up this OS before -- same VM settings, same setup proceedure, and it worked fine, so something must be going on.

yash, that's weird

I've never run into that error before

like, ever.

my only suggestion would be to rebuild binutils with sane cflags and then try gcc again with sane cflags, if you're setting any

I'll try that now!

no promises it'll work

If it doesn't solve it, I'll try installing Kiss again from nothing, and see if it reoccurs.

Ofc

just a guess, maybe there's problem writing file to /tmp

that's the other thing

if it's mounted with ro or just like

not mounted

then that'd do it

It is chrooted.

yeah, that's what I figured.

what does df -h say about /tmp's size?

and ls -l /, tmp's perms should probably be 1777

give me one sec, i'm just waiting for it to build binutils :)

ofc

tmp perms are drwxrwxrwt -- that's 1777 iirc?

mmhm

df -h gives me nothing helpful -- just the size of my root

probably because I'm chrooted?

Anyway, I'm trying to build GCC again - wish me luck...

Same error, will just reinstall kiss...

Thank you all for trying!

hopefully you can get it running

i had weird problems with KISS in a VM too, fwiw

I did earlier on a different VM with exactly the same setup and proceedure. Perhaps it is just VMs being unreliable. I did have to rebuild GRUB multiple times to get it to boot then

sometimes it just be like that

:(

sooo...a bit late to the convo, but one thing that confused the hell out of me with linux and now i just take as granted

if you try and do something, and it says 'file not found', but you can see the file... it could me a bunch of different things. the most common i see is the executable file format is 'wrong' for your system - either a different architecture or missing libraries sometimes even.

kinda like 'the printer says paper jam but there is no paper jam'

Yeah igy. It's odd though, since it seems that it is writing to that file in tmp, and also that the same install procedure worked fine before

this is a strange line too: configure: error: C++ compiler missing or inoperational

yeah, file not found can be just the file doesn't exist, missing shared library, symbolic link pointing to non existing file, typo in file...

Nothing to concern ourselves about though, since I'm reinstalling (so the data is gone, sorry).

I'm going to try setup a kiss chroot without having root myself

so I can hopefully compile some packages on a remote server than on my laptop

Not to jinx it, but it seems like GCC is compiling this time :D

vms are so unreliable lmfao

ikr

If I can get this working in a VM though, then I'll be happy to install this normally on my system (probably in dual boot for now though)

<dilyn> [19:05:37] also, not only do we have to worry about the scarcity of rare earth metals

maybe they aren't that scarce, but scarcity is advertised to keep prices high

just like the oh-so-scarce "fossil" oil

The amount present in the earth is inferrable if our cosmological models are sufficiently accurate, but i would certainly call those a rough-lower-bound rather than a defacto truth

oil scarcity is a good point, i'm unsure of exactly what data went into that. but the generation of oil is vastly different from that of rare-earth metals, so this evaluation is a bit more concrete and obvious

i understand how manufactured scarcity works

debiers has been doing it for forever

s/debiers/de beers/

must've been disappointing for openbsd when they forked openssl and couldn't use their usual "open" prefix cuz it already had one lol

i feel like *bsd gets frustrated by many things lmao

openopenssl

mmmmm perfect

so are any of you guys XLFD gurus

0x0.st/i7LQ.png can't get the fonts to display correctly. I've verified that the fonts are selectable by XLFD (using st -f "-*-blahh") but when golem tries to use them they're not found

well, to be more specific, the title bar font

is the title bar meant to be less tall than the buttons?

yeah. the font that the theme uses is small enough that it should appear correctly in the titlebar.

you can see the intended font here: golem.sourceforge.net/themes/AlbertoDark.png

Hi there! It's me again.

I'm trying to get Kiss working with an LVM root partition.

But I get a kernel panic

Which says: not syncing: VFS: Unable to mount root fs on unknown-block(0,0)

Which evidently shows that the kernel isn't able to recognise the LVM partition

dylanaraps: the `invalid KISS_SU` issue is still present in kiss 5.2.3?

I've installed the lvm2 package, but I don't know what else I could do to get it woking

Oh... so it seems I need an initrd?

Guess I'll have to work that out

tinyramfs is the most 'kiss' option

although last i checked the documentation was relatively sparse, mostly a trial-and-error thing

Okay. I'll look into that now

It's getting late though, so I'll have a proper stab at it tomorrow.

Once I have a working method this should get easy.

Is there a community package for it by the way?

i don't think so

Okay -- that's fine!

Thank you again for your help :3

wouldn't be hard to make one tho!

kiss-new tinyramfs git git+https://github.com/illiliti/tinyramfs; echo 'make DESTDIR="$1" install" >> tinyramfs/build; kiss c tinyramfs

anyone have weechat packaged?

iirc it used to be in community

fda6ddba98878891a5a84d3ed4d6b19ccefb3aeb

or just checkout 0573b7c1dd4ed2eb933ac4d7501c5d1f2e5dfc6b

I am back

Still getting kernel panics, but the initramfs did build, so I think I am getting there.

I'll probably make and push packages for things I use and can maintain at some point.

thanks dilyn. i stole it now: cd /var/db/kiss/repos/jedahan && patch -p2 < $(git -C ../community format-patch -1 0573b7c1dd4ed2eb933ac4d7501c5d1f2e5dfc6b)

Hmm so I managed to get the kernel to panic differently. It wants me to pass an init= option to the kernel. Is there a wiki page or something about the kernel options?

renami i just searched kiss help, kiss help install, and kiss wiki/kernel/config and didnt see anything specific

maybe it would be good to update install or wiki/kernel/config with whatever information you find

Yeah I'll look into writing a wiki page for LVM configuration

Since it involves setting up an initramfs, which may be difficult for new users to set up (including me)

so it sounds to me like the kernel finds the disk but either can't read it or it's not a disk containing /bin/init

but I *think* that this work comes down to configuring tinyramfs properly

Oh that was me being an idiot and having my efi partition as the root XD

Sorry i forgot to update on that

I'm still having trouble getting Kiss to recognise the LV on boot...

Glad this is in a VM