💾 Archived View for gemini.ctrl-c.club › ~phoebos › logs › kisslinux-2022-11-22.txt captured on 2024-03-21 at 15:51:43.
⬅️ Previous capture (2023-01-29)
-=-=-=-=-=-=-
[2022-11-22T00:30:53Z] <phoebos> looks like sbase and ubase dd aren't POSIX
[2022-11-22T00:31:19Z] <phoebos> sad_plan: try the dd branch
[2022-11-22T00:31:27Z] <sad_plan> well thats abit ironic..
[2022-11-22T00:31:48Z] <sad_plan> will do
[2022-11-22T00:36:24Z] <sad_plan> yeah, that seemed to work
[2022-11-22T00:39:44Z] <sad_plan> it seems the developers just satisfied with not including ibs at all. according to the readme, its flagged as finished
[2022-11-22T00:40:18Z] <sad_plan> s/just/was/
[2022-11-22T00:40:39Z] <phoebos> ubase doesn't include ibs, but sbase does. trying to track down the bug.
[2022-11-22T00:41:25Z] <sad_plan> hm
[2022-11-22T00:44:42Z] <sad_plan> hm, can I not specify branch and commit at the same time with kiss?
[2022-11-22T00:45:12Z] <phoebos> maybe just specifying a commit from any branch works?
[2022-11-22T00:45:19Z] <sad_plan> ill test
[2022-11-22T00:45:40Z] <sad_plan> it did
[2022-11-22T02:54:58Z] <midfavila> http://0x0.st/olGU.tar
[2022-11-22T02:55:11Z] <midfavila> ^alpha of my http client if anyone wants to look at it
[2022-11-22T02:55:58Z] <midfavila> super not-good but it *does* seem to work reliably with small files, at the very least. caveats: only implements super basic HTTP1.0 GET, doesn't validate URIs properly, doesn't have TLS support yet
[2022-11-22T03:10:07Z] <wael_> Hi
[2022-11-22T03:25:31Z] <phoebos> mid: looks like a nice start, you have better overall design than me; I tend to just dump everything in main and only make structs when absolutely necessary
[2022-11-22T03:25:51Z] <phoebos> btw recv(2) can return -1 on error, but your for loop will ignore that
[2022-11-22T03:26:25Z] <phoebos> so it enters an infinite loop for me
[2022-11-22T03:26:31Z] <midfavila> do you happen to use glibc?
[2022-11-22T03:26:35Z] <phoebos> no
[2022-11-22T03:26:39Z] <midfavila> hmmmmmmm
[2022-11-22T03:26:53Z] <phoebos> that's POSIX
[2022-11-22T03:26:58Z] <midfavila> yeah, I know
[2022-11-22T03:27:04Z] <midfavila> only asking 'cause someone else mentioned the same thing to me
[2022-11-22T03:27:29Z] <midfavila> anyway, main's layout is going to be overhauled once I implement TLS -- i threw it together in like five minutes
[2022-11-22T03:27:33Z] <midfavila> should be fine then
[2022-11-22T03:27:52Z] <phoebos> also, your Makefile has a bug: fetch is built because BIN is the default target, but `make all` will dereference a ${B} variable that doesn't exist
[2022-11-22T03:27:59Z] <midfavila> yep
[2022-11-22T03:28:05Z] <midfavila> also something i threw together without thinking
[2022-11-22T03:28:22Z] <phoebos> no worries
[2022-11-22T03:28:27Z] <phoebos> just mentioning it in case
[2022-11-22T03:28:50Z] * midfavila nod-nods
[2022-11-22T03:29:02Z] <midfavila> it also doesn't help that i don't know how to write makefiles lmao
[2022-11-22T03:29:23Z] <midfavila> but yeah, from the get go I wanted to make the design as generic as I could
[2022-11-22T03:29:29Z] <midfavila> it should be easy to extend it to handle other protocols
[2022-11-22T03:29:43Z] <phoebos> looks like what hurl wanted to do
[2022-11-22T03:29:58Z] <phoebos> in terms of code layout
[2022-11-22T03:30:14Z] <midfavila> that's high praise :p
[2022-11-22T03:30:41Z] <phoebos> :v hurl is a mess
[2022-11-22T03:30:56Z] <midfavila> i need to replace the calls to open and write with fopen and fwrite at some point too
[2022-11-22T03:31:01Z] <midfavila> and yeah hurl is a clusterfuck
[2022-11-22T03:31:11Z] <phoebos> one function per protocol, every one essentially the same
[2022-11-22T03:31:20Z] <midfavila> bleh
[2022-11-22T03:31:38Z] <wael_> macslash1:
[2022-11-22T03:32:31Z] <phoebos> i can't stand your syntax style, though :p
[2022-11-22T03:33:21Z] <midfavila> yeah it seems to be a controversial thing
[2022-11-22T03:33:28Z] <midfavila> whitesmith>knr though
[2022-11-22T03:33:35Z] <midfavila> i can't stand KnR style
[2022-11-22T03:33:43Z] <midfavila> makes it way too hard to follow the shape of the program
[2022-11-22T03:35:24Z] <phoebos> i'm accustomed to using K&R as a visual aid, but I can imagine it just depends on what you're used to
[2022-11-22T03:35:33Z] <midfavila> almost certainly
[2022-11-22T03:35:46Z] <midfavila> i'm used to programming in scheme, where indentation is like
[2022-11-22T03:35:58Z] <midfavila> the *only* reflection of the procedure's evolution
[2022-11-22T03:36:12Z] <midfavila> so if it's not a somewhat lispy style of indentation i get lost really fast
[2022-11-22T03:36:36Z] <phoebos> the first program i wrote in scheme i tried to indent like K&R C
[2022-11-22T03:36:39Z] <phoebos> idk if you saw it
[2022-11-22T03:36:41Z] <midfavila> lmao no
[2022-11-22T03:36:44Z] <midfavila> but that sounds amazing
[2022-11-22T03:37:19Z] <phoebos> https://tmp.bvnf.space/pearson.scm
[2022-11-22T03:37:32Z] <midfavila> oh, hm, you know what
[2022-11-22T03:37:55Z] <midfavila> checking for recv failing actually breaks the program
[2022-11-22T03:37:55Z] <midfavila> lmao
[2022-11-22T03:38:21Z] <midfavila> also wow that scheme program is hideous
[2022-11-22T03:38:25Z] <phoebos> hehe
[2022-11-22T03:38:33Z] <midfavila> jesus christ
[2022-11-22T03:38:36Z] <midfavila> it gets worse the longer i look at it
[2022-11-22T03:45:09Z] <midfavila> yeah i'm gonna need to debug this tomorrow
[2022-11-22T03:45:24Z] <midfavila> i'm running on seven cups of tea and dreams of not needing curl or axel
[2022-11-22T03:49:37Z] <phoebos> i also struggle to read the whole if (!foo())
[2022-11-22T03:49:46Z] <phoebos> but i think you're ignoring a failed dial()
[2022-11-22T03:50:12Z] <phoebos> need to loop through getaddrinfo's results until successfully connected
[2022-11-22T03:52:05Z] <phoebos> also why are using using XOR in open's flags? should be just OR
[2022-11-22T03:52:22Z] <phoebos> some things for you to do tomorrow :D
[2022-11-22T03:57:35Z] <midfavila> i'm ignoring and screwing up a lot of things in fetch rn
[2022-11-22T03:57:47Z] <midfavila> main() is super bad as is the uri validator
[2022-11-22T03:57:57Z] <midfavila> both need to be completely rewritten
[2022-11-22T03:58:16Z] <testuser[m]> Hi
[2022-11-22T03:58:20Z] <midfavila> goal today was mainly to get the networking and http functions semi-working
[2022-11-22T03:58:24Z] <midfavila> testuser[m] hi
[2022-11-22T04:00:54Z] <midfavila> as for getaddrinfo i actually have no clue what i'm doing when it comes to berkeley sockets
[2022-11-22T04:01:06Z] <midfavila> this was the first time i've used them... and fetch is really my first not-completely-trivial program
[2022-11-22T04:03:29Z] <phoebos> there's a manpage with a good example of using getaddrinfo, socket, connect correctly
[2022-11-22T04:03:40Z] <phoebos> freeaddrinfo(3p)
[2022-11-22T04:04:33Z] <phoebos> there's also a medium-depth tutorial about network programming at https://beej.us/guide/bgnet/html/index.html
[2022-11-22T04:05:04Z] <midfavila> will peek, tenku
[2022-11-22T04:05:31Z] * phoebos thumbs up
[2022-11-22T04:06:21Z] <phoebos> one thing that confused me was the whole stuff about shutdown and recv returning EOF, but skimming through a TCP textbook helped with that
[2022-11-22T04:07:04Z] <testuser[m]> > very minor potential for a segfault here
[2022-11-22T04:11:29Z] <midfavila> i never said it was gooooooooooooooooooooooooooooooooooooood-
[2022-11-22T04:16:41Z] <phoebos> doesn't have to be good, i hope you enjoy fiddling around learning and digging up bugs
[2022-11-22T04:17:21Z] <midfavila> well, hope is that it'll eventually become good
[2022-11-22T04:17:31Z] <midfavila> i kinda need it to :v
[2022-11-22T04:17:35Z] <phoebos> :v
[2022-11-22T04:17:56Z] <midfavila> can't build any decent http clients on my latest rootfs sooooooo
[2022-11-22T04:18:03Z] <phoebos> if you want a patch to fix things lmk
[2022-11-22T04:18:22Z] <midfavila> maybe once it's beta quality
[2022-11-22T04:18:26Z] <midfavila> will keep in mind
[2022-11-22T04:19:10Z] <phoebos> sure; if it were me, I'd rather be left to figure it out
[2022-11-22T04:19:43Z] <phoebos> anyway, goodnight \o
[2022-11-22T05:02:04Z] <testuser[m]> Personally, I wish they would step back and stop adding hacks on top of mmap. The API is seriously flawed because it conflates a load of things: allocating virtual address space, permissions on that address space, and backing store. It does all of these things with no permission model. OpenBSD would be in a good position to design and ship a deconflation of these. Brooks Davis proposed a ‘cookie mmap’ a few years ago that started
[2022-11-22T05:02:04Z] <testuser[m]> this, where the call to allocate address space returned a capability that then gave you the rights to change the mapping, you could only do mprotect or MAP_FIXED-like things if you had that capability. Once you have such a model, it becomes quite easy to add a richer set of permissions, including dropping the right to subsequently modify the mapping.
[2022-11-22T06:37:52Z] <testuser[m]> How should sandbox be handled for go/rust packages (network) or pgo packages (filesystem)
[2022-11-22T06:38:14Z] <testuser[m]> i mean foot pgo / firefox pgo
[2022-11-22T06:38:56Z] <testuser[m]> Also ccache
[2022-11-22T06:50:09Z] <illiliti> sandbox should not allow network access for go/rust packages. it is out of question
[2022-11-22T06:50:34Z] <illiliti> instead, packages should provide tarballs with vendored dependencies
[2022-11-22T06:51:02Z] <illiliti> filesystem sandboxing is harder yeah
[2022-11-22T06:51:58Z] <illiliti> maybe there should be some hook to dynamically add path exceptions
[2022-11-22T06:52:11Z] <illiliti> dunno
[2022-11-22T06:54:33Z] <testuser[m]> Ye that's wat i thoight
[2022-11-22T07:57:35Z] <testuser[m]> Same here, I would rather trust my data with our own government instead of the Chinese or US ones.
[2022-11-22T11:33:35Z] <wael[m]> testuser you scammed me
[2022-11-22T11:34:11Z] <testuser[m]> what
[2022-11-22T11:34:19Z] <wael[m]> wait i think it works
[2022-11-22T11:34:32Z] <testuser[m]> bruh u cant see history for irc rooms
[2022-11-22T11:34:34Z] <testuser[m]> cuz of settings
[2022-11-22T11:34:39Z] <wael[m]> oh yea ru
[2022-11-22T11:34:47Z] <wael[m]> well it DOES work fluffychat lied to me
[2022-11-22T11:34:53Z] <wael_> yeah u right
[2022-11-22T11:35:35Z] <testuser[m]> lol does this channel have no modertators
[2022-11-22T11:35:48Z] <wael[m]> ioraff is the irc moderator technically
[2022-11-22T11:36:11Z] <testuser[m]> only acheam is according to the bridge
[2022-11-22T11:36:20Z] <wael[m]> who dat
[2022-11-22T11:49:17Z] <illiliti> testuser[m]: speaking of pgo, do you think it is good idea to move foot-pgo to community?
[2022-11-22T11:49:35Z] <illiliti> because i don't believe it should be in main repo. we already have general-purpose foot package after all; i don't see why we should maintain foot-pgo
[2022-11-22T11:50:09Z] <wael[m]> hi illiliti
[2022-11-22T11:50:26Z] <illiliti> hi
[2022-11-22T11:53:25Z] <testuser[m]> all variations of a given pkg should probably be in the same repo as the original
[2022-11-22T12:01:20Z] <illiliti> the point is, does it really bring so much value to be in main repo?
[2022-11-22T12:01:54Z] <illiliti> or alternatively, who uses it?
[2022-11-22T12:03:24Z] <testuser[m]> me
[2022-11-22T12:03:37Z] <testuser[m]> its easier to keep in sync
[2022-11-22T12:05:48Z] <illiliti> ok
[2022-11-22T12:07:53Z] <illiliti> i will never understand these micro-optimizations, especially for small programs
[2022-11-22T12:08:08Z] <testuser[m]> ye i dont think it has any benefit lol
[2022-11-22T12:08:20Z] <testuser[m]> you're not catting /dev/urandom constantly
[2022-11-22T12:08:29Z] <testuser[m]> should probably be rmeoved
[2022-11-22T12:10:55Z] <wael[m]> should probably be meowed
[2022-11-22T12:10:59Z] <illiliti> i can undertand the need for pgo'd firefox or chromium. there it will have a use, but foot... i'll never understand that
[2022-11-22T15:01:32Z] <midfavila-mobile> holy shit you guys, i actually *slept* last night- feel like i'm gonna live forever
[2022-11-22T15:46:48Z] <wael_> do you not sleep
[2022-11-22T15:49:47Z] <midfavila> i do not
[2022-11-22T15:53:09Z] <wael_> well you gotta have some sleep
[2022-11-22T15:53:19Z] <wael_> its very important for you
[2022-11-22T15:53:32Z] <midfavila> yes I know :v
[2022-11-22T15:53:32Z] <midfavila> it's not for lack of trying
[2022-11-22T15:55:58Z] <wael_> can you re-phrase that
[2022-11-22T17:15:45Z] <aelspire> Hi, I've managed to make touchpad working in linux kernel. It was hellish ride. Much more problematic than my root drive missing due to lack od VMD. And now my kernel config is a mess. Anyone have idea how to nicely organize kernel config? I consider something like git repo containing empty config and in separate commits I'm adding more functionality for e.g commit: Make touchpad working. Anyone have better idea?
[2022-11-22T17:16:41Z] <aelspire> Main problem with git repo will be config changes due to new kernel versions
[2022-11-22T19:01:09Z] <van2022> How assemble from parts kisslinux-live.iso?
[2022-11-22T19:01:54Z] <van2022> I myself found
[2022-11-22T19:23:29Z] <van2022> I have kiss-chroot.tar.xz file. How can I make an iso? github.com/kiss-community/kiss-live
[2022-11-22T19:41:44Z] <aelspire> van2022: Why You want iso?
[2022-11-22T20:07:09Z] <van2022> to write it to a flash drive
[2022-11-22T20:16:26Z] <aelspire> van2022: I guess You want to install KISS and want Live USB?
[2022-11-22T20:17:45Z] <aelspire> kiss-chroot is just root without linux kernel it won't boot
[2022-11-22T20:18:07Z] <aelspire> https://kisslinux.org/install
[2022-11-22T20:18:29Z] <aelspire> You need liveCD from other distro to install KISS
[2022-11-22T20:19:28Z] <aelspire> it is possible to steal kernel with firmware and modules from other distro and repack kiss-chroot via cpio as initramfs and boot it
[2022-11-22T20:19:47Z] <aelspire> but it's rather cumbersome
[2022-11-22T20:21:35Z] <van2022> the iso file will weigh many megabytes?
[2022-11-22T20:22:00Z] <aelspire> you was offline so I repeat: "it is possible to steal kernel with firmware and modules from other distro and repack kiss-chroot via cpio as initramfs and boot it"
[2022-11-22T20:22:07Z] <aelspire> bare chroot will not boot
[2022-11-22T20:22:14Z] <aelspire> as it has no linux kernel
[2022-11-22T20:22:44Z] <aelspire> you should use other distro's LiveCD to unpack this kiss-chroot on freshly formatted disk
[2022-11-22T20:22:50Z] <aelspire> check: https://kisslinux.org/install
[2022-11-22T20:26:34Z] <van2022> then what to do?
[2022-11-22T20:27:26Z] <aelspire> follow official instalation guide and use other LiveCD, I suggest to use something with the same linux kernel version as You want
[2022-11-22T20:30:16Z] <van2022> I'll finish this job tomorrow. leaving the chat
[2022-11-22T21:23:58Z] <sad_plan> o/
[2022-11-22T21:25:12Z] <sad_plan> aelspire: what issue do you suppose is there when having your kernel config tracked with git? I do this, and tbh, its a great way to keep track of it. however I have in a kiss repo, and have kiss build my kernel for me instead of doing it manually