💾 Archived View for gemini.ctrl-c.club › ~phoebos › logs › kisslinux-2022-09-04.txt captured on 2023-01-29 at 03:45:49.
-=-=-=-=-=-=-
[2022-09-04T02:18:53Z] <midfavila> http://0x0.st/opkY.xdump [2022-09-04T02:19:03Z] <midfavila> xwud or gimp or somesuch to view, it's a raw X dump [2022-09-04T02:19:09Z] <midfavila> first screenie from the new machine [2022-09-04T02:20:18Z] <midfavila> biggest thing holding it back now is limited TLS support tbqh - all of the major systems are in place and working [2022-09-04T02:20:45Z] <midfavila> working on getting ogg playback and encoding support running, since ffmpeg isn't really an option [2022-09-04T02:21:06Z] <midfavila> (sorry for blogposting, just figured some of you guys might have passing interest in this stuff) [2022-09-04T02:36:34Z] <Torr> Nice [2022-09-04T02:55:54Z] <midfavila> hey sad_plan [2022-09-04T02:55:59Z] <midfavila> check logs :p [2022-09-04T02:56:20Z] <sad_plan> on it :p [2022-09-04T02:59:11Z] <wael[m]> <testuser[m]12> "btw any idea how genpoo deals..." <- on void, they track all sonames with a file [2022-09-04T02:59:11Z] <wael[m]> if the soname does not belong to a package it will not build [2022-09-04T03:00:14Z] <sad_plan> nice work midfavila [2022-09-04T03:00:23Z] <midfavila> uwu~ [2022-09-04T03:02:55Z] <midfavila> i'm going to be learning athena and writing some tools for this system over the next week [2022-09-04T03:04:10Z] <midfavila> ultimate goal would be a terminal, text editor, text viewer, image editor, image viewer, basic browser (using links), file manager, window manager, IRC client, FTP, erm... that's probably the extent of it [2022-09-04T03:04:36Z] <midfavila> maybe a mailer and display manager [2022-09-04T03:08:04Z] <midfavila> ...oh, I need to write a multiplexer at some point too [2022-09-04T03:08:10Z] <midfavila> all the existing ones are kind of shit [2022-09-04T03:08:26Z] <midfavila> dvtm crashes randomly, mtm isn't particularly portable, and screen and tmux are both massive [2022-09-04T03:11:43Z] <illiliti> multiplexers suck in general [2022-09-04T03:12:08Z] <midfavila> well, clearly as a god among men, i should be able to succeed where everyone else has failed = w= [2022-09-04T03:12:29Z] <midfavila> very /s in case that wasn't uh, very obvious [2022-09-04T03:13:34Z] <midfavila> http://0x0.st/opk5.png [2022-09-04T03:13:38Z] <midfavila> ^png version of earlier screendump [2022-09-04T03:13:44Z] <midfavila> SDF has netpbm installed [2022-09-04T03:14:05Z] <sad_plan> yeah, if I werent running a artix live usb atm, I woulndt have been able to view the first one [2022-09-04T03:14:20Z] <midfavila> that's not surprising :P [2022-09-04T03:14:30Z] <midfavila> i don't imagine X dumps are very common any more [2022-09-04T03:14:58Z] <sad_plan> never seen one before tbh :p [2022-09-04T03:15:10Z] <midfavila> for good reason [2022-09-04T03:15:17Z] <sad_plan> I figured :p [2022-09-04T03:15:49Z] <midfavila> ...oh, actually, i lied earlier [2022-09-04T03:15:52Z] <sad_plan> im amazed by how quickly oasis is capable of compile the whole damn system though. [2022-09-04T03:15:55Z] <midfavila> XPM and XBM support is available [2022-09-04T03:16:08Z] <sad_plan> I dont even know what that is C: [2022-09-04T03:16:17Z] <midfavila> plaintext image formats [2022-09-04T03:16:21Z] <midfavila> standard with X11 [2022-09-04T03:16:28Z] <sad_plan> i see [2022-09-04T03:16:37Z] <midfavila> i'm quite fond of them [2022-09-04T03:16:47Z] <sad_plan> why? [2022-09-04T03:16:50Z] <midfavila> despite their relatively massive size they're easy to understand and work with [2022-09-04T03:17:00Z] <midfavila> since, you know, no binary data [2022-09-04T03:17:18Z] <midfavila> libXpm is also quite small and easy to get building [2022-09-04T03:17:33Z] <sad_plan> aah :p [2022-09-04T03:17:48Z] <midfavila> ye [2022-09-04T03:17:55Z] <midfavila> x11 has a lot of little nuggets hidden away [2022-09-04T03:20:21Z] <sad_plan> yeah, Ive figured after reading some of your websites, aswell as other sites. theres alot of stuff thats since been abondend, but still works, or are still included somewhere :p [2022-09-04T03:20:33Z] <midfavila> indeed [2022-09-04T03:25:42Z] <illiliti> midfavila: job control is better than multiplexing [2022-09-04T03:25:52Z] <illiliti> unpopular opinion maybe [2022-09-04T03:26:18Z] <midfavila> job control is incomparable to a multiplexer [2022-09-04T03:26:27Z] <midfavila> they serve similar but distinct purposes [2022-09-04T03:26:41Z] <midfavila> if i need to read two or more terminals side by side, a multiplexer is necessary [2022-09-04T03:26:53Z] <midfavila> if i need to run multiple programs simultaneously, job control is [2022-09-04T03:27:25Z] <illiliti> then spawn another terminal [2022-09-04T03:27:37Z] <illiliti> multiplexer is a hack [2022-09-04T03:28:12Z] <midfavila> that's not always an option [2022-09-04T03:28:17Z] <midfavila> anyway, hacks are fine [2022-09-04T03:28:36Z] <midfavila> unless they're inelegant, but I've not heard a good reason for multiplexers to be considered inelegant intrinsically [2022-09-04T03:28:36Z] <illiliti> implementing terminal on top of another terminal is fine? [2022-09-04T03:29:01Z] <midfavila> yes, if the additional implementation adds sufficient capability [2022-09-04T03:29:24Z] <illiliti> that's bloat and crap design [2022-09-04T03:29:41Z] <midfavila> by your logic, every widget toolkit on X is "bloat" and "crap design" [2022-09-04T03:29:49Z] <midfavila> your opinion isn't unpopular, it's just stupid [2022-09-04T03:30:02Z] <midfavila> systems engineering is about making a series of tradeoffs [2022-09-04T03:30:07Z] <illiliti> no it's by your logic [2022-09-04T03:30:23Z] <illiliti> i didn't say that x is bloat [2022-09-04T03:30:32Z] <midfavila> i'm making a direct comparison [2022-09-04T03:30:40Z] <illiliti> don't [2022-09-04T03:31:06Z] <midfavila> as I understand it, your claim is that "implementing a thing B of type x on top of another thing of type x, A, is bloat and bad design" [2022-09-04T03:31:14Z] <midfavila> is that a misunderstanding? [2022-09-04T03:31:37Z] <illiliti> look, terminal is already kind of hack [2022-09-04T03:31:53Z] <midfavila> literally everything in modern computers is "a hack" [2022-09-04T03:32:04Z] <illiliti> multiplexer is another layer of hacks, needless hack [2022-09-04T03:32:11Z] <midfavila> the BIOS was a hack to allow OS designers to ignore some elements of hardware design [2022-09-04T03:32:24Z] <midfavila> compilers are a hack to allow programmers to write code more efficiently [2022-09-04T03:32:29Z] <midfavila> macro assemblers were the same [2022-09-04T03:32:52Z] <midfavila> toolkits are a hack to prevent programmers from making mistakes and bad choices fiddling with super low-level graphics APIs [2022-09-04T03:33:19Z] <midfavila> *everything* is "a hack". the question is whether it's a good or bad hack, and that comes down to whether adding that hack to the system adds more value than complexity [2022-09-04T03:33:37Z] <illiliti> you don't get my point [2022-09-04T03:33:50Z] <midfavila> no, I get your point. I just think it's invalid [2022-09-04T03:33:53Z] <testuser[m]12> Hi [2022-09-04T03:34:12Z] <sad_plan> hi testuser[m]12 [2022-09-04T03:35:11Z] <sad_plan> illiliti I couldnt help but laugh at your comment on multiplexers being a terminal on top of a termal. made me think of the yo dawg meme [2022-09-04T03:36:11Z] <illiliti> then look at the tmux code and rethink what you thinking [2022-09-04T03:36:56Z] <midfavila> tmux is shit because it's tmux, not because terminal multiplexers are shit [2022-09-04T03:36:57Z] <sad_plan> Im not gonna argue against you, I just thought it was abit funny :p [2022-09-04T03:37:29Z] <midfavila> hell, you haven't even defined what you mean by "hack" in this instance [2022-09-04T03:37:43Z] <illiliti> i did [2022-09-04T03:38:06Z] <midfavila> let alone actually explained why multiplexers are "bad designs" for reimplementing terminal protocols on top of a terminal [2022-09-04T03:44:19Z] <illiliti> do you why people still write TUIs? [2022-09-04T03:45:27Z] <midfavila> because they're relatively simple compared to GUIs and thus require fewer resources and less support from the environment whilst still providing a relatively easy to learn and use interface for endusers? [2022-09-04T03:45:39Z] <sad_plan> because they want an UI, but still wanna stick to the terminal? [2022-09-04T03:45:50Z] <illiliti> no [2022-09-04T03:45:59Z] <sad_plan> why? [2022-09-04T03:46:22Z] <illiliti> because there's no portable way to access os ui [2022-09-04T03:46:31Z] <midfavila> one, that's a lie [2022-09-04T03:46:48Z] <midfavila> two, if you're talking UNIX, that's because there's no standard UI for *unix* [2022-09-04T03:46:59Z] <midfavila> not even TUIs are portable [2022-09-04T03:47:19Z] <testuser[m]12> wat [2022-09-04T03:47:25Z] <illiliti> tuis are quite portable unlike guis [2022-09-04T03:47:54Z] <midfavila> uh-huh? write a TUI using only the C89 standard library and a minimal subset of the ANSI 79 spec or whatever [2022-09-04T03:48:18Z] <testuser[m]12> Why c89? [2022-09-04T03:48:27Z] <midfavila> make it work on CP/M, MS-DOS, System V, BSD, OS/9, and ancient Mac [2022-09-04T03:48:37Z] <midfavila> and c89 because it's the earliest standardized C [2022-09-04T03:49:09Z] <testuser[m]12> Bruh [2022-09-04T03:49:57Z] <midfavila> anyway, there's a reason we target specs instead of OSes [2022-09-04T03:50:30Z] <midfavila> OSes are constantly changing and are radically different from each other: some have unique GUIs, some have no GUI at all, some have a standardized system [2022-09-04T03:50:47Z] <midfavila> by applying "hacks" and increasing the level of abstraction used, programs can be made much more portable [2022-09-04T03:51:01Z] <midfavila> that applies to GUI and non-GUI programs alike [2022-09-04T04:04:13Z] <illiliti> you seem to believe that everything can be done properly, but no. some things are broken. they work yeah, but that isn't the point. if you want to implement usable multiplexer, you will end up with tmux-level amount of nonsense because that's inevitable [2022-09-04T04:04:32Z] <illiliti> escape sequences, terminfo, tty, ... just to note things you will be fucking with [2022-09-04T04:05:00Z] <midfavila> mtm and dvtm are both quite small and do their jobs relatively well, ignoring the problems I mentioned earlier [2022-09-04T04:05:15Z] <midfavila> manipulating escape sequences via terminfo and such is literally a non-starter [2022-09-04T04:05:21Z] <midfavila> like, for your argument [2022-09-04T04:05:42Z] <midfavila> any time you write a program targetting a terminal specifically, you need to futz about with terminal escapes [2022-09-04T04:05:54Z] <midfavila> what do you think your precious TUIs are built on in part? [2022-09-04T04:06:50Z] <illiliti> that's why i hate tuis [2022-09-04T04:07:11Z] <illiliti> because they built upon hacks [2022-09-04T04:07:17Z] <midfavila> then what isn't a hack? [2022-09-04T04:07:24Z] <midfavila> what's "the right thing:tm:" [2022-09-04T04:09:30Z] <illiliti> the right thing is to abolish everything and start from scratch addressing current problems [2022-09-04T04:09:37Z] <illiliti> like plan9 [2022-09-04T04:09:54Z] <illiliti> terminal needs redesign [2022-09-04T04:10:34Z] <midfavila> conceptually the terminal is fine [2022-09-04T04:11:02Z] <midfavila> 8+1/2 is great, but it's not a terminal. it serves many purposes, and multiplexing textual interfaces is but one [2022-09-04T04:17:54Z] <illiliti> https://nvd.nist.gov/vuln/detail/cve-2022-28391 [2022-09-04T04:18:00Z] <illiliti> just an example [2022-09-04T04:18:09Z] <illiliti> there are many of them [2022-09-04T04:18:18Z] <illiliti> you think it's good? [2022-09-04T04:18:45Z] <midfavila> i think that's a very specific implementation error [2022-09-04T04:18:55Z] <illiliti> not at all [2022-09-04T04:19:32Z] <midfavila> "if a specific implementation of a specific tool is used to print specific bytes to a specific non-standard terminal emulator we can do RCE" [2022-09-04T04:20:04Z] <midfavila> not using busybox, or not using busybox netstat, or not using a VT100 terminal in any combination completely negates that [2022-09-04T04:21:52Z] <illiliti> you think it's good that arbitrary program can manipulate terminal via escape sequences? [2022-09-04T04:22:09Z] <illiliti> also do you know about ctrl+v exploits? when you copy and paste something from internet into terminal... [2022-09-04T04:22:09Z] <midfavila> you think strawmanning me is a good debate tactic? [2022-09-04T04:22:29Z] <midfavila> and "ctrl+v exploits" are just called "exploiting stupidity" [2022-09-04T04:22:37Z] <midfavila> that's PEBKAC, not a tech issue [2022-09-04T04:23:19Z] <midfavila> anyway, allowing programs to manipulate the state of the display hardware in a well-defined manner is important, yes [2022-09-04T04:23:30Z] <midfavila> but it should be treated with care [2022-09-04T05:03:08Z] <midfavila> oh shidd [2022-09-04T05:03:10Z] <midfavila> netsplit time [2022-09-04T06:01:03Z] <testuser[m]12> What [2022-09-04T07:44:15Z] <testuser[m]12> konimex: what have u been upto [2022-09-04T08:13:06Z] <testuser[m]12> illiliti: riteo: https://gitlab.alpinelinux.org/alpine/mdev-conf/-/blob/master/mdev.conf#L113 [2022-09-04T12:36:22Z] <sad_plan> hi [2022-09-04T13:35:38Z] <Ogromny> hi [2022-09-04T13:35:53Z] <Ogromny> Is here any repo yet for the C version of Kiss ? [2022-09-04T13:36:35Z] <sad_plan> there isnt one yet really, but dylan started on long ago [2022-09-04T13:37:07Z] <sad_plan> https://github.com/dylanaraps/k [2022-09-04T13:38:07Z] <Ogromny> :o thanks [2022-09-04T13:38:15Z] <sad_plan> wait, kiss-community actually forked it. just noticed https://github.com/kiss-community/k [2022-09-04T13:38:23Z] <sad_plan> havent been any commits there yet though [2022-09-04T13:38:32Z] <sad_plan> theres still talks of going for a c replacement anyway [2022-09-04T13:40:35Z] <sad_plan> Ogromny https://github.com/kiss-community/repo/issues/91 [2022-09-04T14:50:02Z] <testuser[m]12> illiliti: can we add pkgconfig file to lolvm and merge #51 or is there something else missing? pgo support is also merged [2022-09-04T17:46:15Z] <illiliti> testuser[m]12: muon has some bugs i need to fix before we can consider merging #51 [2022-09-04T17:46:47Z] <illiliti> i plan to resume work on muon soon, so [2022-09-04T17:47:02Z] <illiliti> we will merge it, don't worry [2022-09-04T17:47:08Z] <wael[m]> Woooooo [2022-09-04T17:50:47Z] <testuser[m]12> Ok [2022-09-04T18:47:47Z] <midfavila> ...wait, shit [2022-09-04T18:47:50Z] <midfavila> i'd still need a linker [2022-09-04T18:48:04Z] <midfavila> ...and an archiver [2022-09-04T18:48:07Z] <midfavila> gah [2022-09-04T18:48:16Z] <midfavila> still! the potential is there [2022-09-04T18:48:35Z] <sad_plan> doesnt tcc's linker work? [2022-09-04T18:48:55Z] <midfavila> oop, fuck, didn't realize I unquery'd myself- [2022-09-04T18:49:05Z] <midfavila> i was PM'ing a friend about replacing tcc with cproc as an experiment [2022-09-04T18:49:37Z] <midfavila> depending on qbe's output i might be able to substitute yasm and mcpp with qbe/cproc for a dev toolchain, sans linker and archiver [2022-09-04T18:49:51Z] <midfavila> i'm only considering it because i ran into a non-trivial bug in tcc today [2022-09-04T18:50:13Z] <midfavila> for whatever reason X programs that rely on the XtInherit family of functions/macros/whatever fail when built with tcc [2022-09-04T18:50:32Z] <midfavila> pcc, cproc and gcc handle it fine though [2022-09-04T18:55:58Z] <sad_plan> sure. d try cproc. I was thinking about doing a run for that aswell. but Ive got other issues to solve first tbh :p [2022-09-04T18:56:31Z] <midfavila> looks like scc provides some tools... maybe they could supplement cproc idk [2022-09-04T18:57:43Z] <sad_plan> i dunno. [2022-09-04T18:59:36Z] <sad_plan> in any case, theres a list of stuff that does/doesnt work with cproc on oasis anyway. https://github.com/oasislinux/oasis/issues/13 if its of any interest