Conversation

Notices

1.  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 09:11:52 JST 
   The GNU Hurd experience.
   Screenshot_vm1_2024-09-03_03:11…
   • † top dog :pedomustdie: likes this.
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 09:14:01 JST 

     in reply to
     2023 snapshot worked fine, guess I have to reinstall it.
   • :ihavenomouth: (inginsub@clubcyberia.co)'s status on Tuesday, 03-Sep-2024 09:15:20 JST :ihavenomouth:

     in reply to
     @mint didn't know they had wayland
      likes this.
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 09:20:25 JST 

     in reply to

     • Pawlicker
     @Pawlicker Debian's port apparently had an amd64 release recently, but I couldn't get it to install since it doesn't detect the most basic emulated IDE controller. Tried i386 snapshot, worked fine until I decided to upgrade it. Thought that it's just the framebuffer that's corrupted, but couldn't SSH either after reboot.
   • Pawlicker (pawlicker@makai.chaotic.ninja)'s status on Tuesday, 03-Sep-2024 09:20:26 JST Pawlicker

     in reply to

     @mint@ryona.agency decades in development and nothing to show

      likes this.
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 09:31:49 JST 

     in reply to
     No SMP support, it seems. Got Erlang in repos but not Elixir, thinking I should build both instead in case it was built with no SSL support.
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 10:06:19 JST 

     in reply to
     Bitchin'. OTP 25 in snapshot's repos is fairly fresh, however (unlike the one in haikuports stuck at 20), so if that fails might as well use the prepackaged one, but that's less fun.
     Screenshot_20240903_035904.png
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 10:18:16 JST 

     in reply to
     Too much IPv6-related shit to either gut or redefine in prim_socket_nif.c, too lazy to do so at 4 AM. Will try building the rest first.
     † top dog :pedomustdie: likes this.
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 10:28:27 JST 

     in reply to
     Yeah, as expected.
     Screenshot_20240903_042812.png
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 10:32:15 JST 

     in reply to
     Bitch.
     Screenshot_20240903_043055.png
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 10:43:43 JST 

     in reply to
     Apparently there are no source packages in snapshot repos. Anyway, found a patch in source package, applied fine on newer version. Looks like Hurd also isn't Y2K38-compliant.
     Screenshot_20240903_044146.png
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 11:22:52 JST 

     in reply to
     Edited the whole file with #ifndef __GNU__, it compiled but when it came to erlc part assertion error happened. Apparently it was all unneccessary since this could be solved by setting --disable-esock during configure (which I found in the same .deb build rules). Slow as usual since no SMP.
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 12:18:20 JST 

     in reply to
     https://www.gnu.org/software/hurd/faq/smp.html
     >yeah, in theory we should support SMP, but it's not yet done
     How long it's been? Anyway, OTP 27 with SSL cert paths patched in built successfully after a bunch of fiddling.
     Screenshot_20240903_061735.png
     † top dog :pedomustdie: likes this.
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 20:01:31 JST 

     in reply to
     Built Elixir and went to sleep in meanwhile. Stuck at this now.
     Screenshot_20240903_140025.png
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 20:06:18 JST 

     in reply to
     Right, had to add nother -D_GNU_SOURCE in makefile. Also some other dep complains about the lack of cmake.
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 20:23:35 JST 

     in reply to
     https://stackoverflow.com/questions/19901934/libpthread-so-0-error-adding-symbols-dso-missing-from-command-line
     Finally, someone assuming that you're just compiling someone other's shit.
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 20:44:46 JST 

     in reply to
     Bitchin'.
     Screenshot_20240903_144438.png
     † top dog :pedomustdie: likes this.
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 20:50:27 JST 

     in reply to
     It runs, kinda. Can't post anything, can't fetch anything because of the smae fasthtml_worker.
     Screenshot_20240903_144832.png
     Screenshot_20240903_144852.png
     † top dog :pedomustdie: likes this.
   • † top dog :pedomustdie: (dcc@annihilation.social)'s status on Tuesday, 03-Sep-2024 20:51:05 JST † top dog :pedomustdie:

     in reply to
     @mint >fasthtml_worker O boy :alex_lol:
   • :blank: (i@declin.eu)'s status on Tuesday, 03-Sep-2024 20:53:28 JST :blank:

     in reply to

     • † top dog :pedomustdie:
     @dcc @mint who needs html sanitization anyway, rip it out
     † top dog :pedomustdie: likes this.
   • Lina Inver:z:e :mur_close: :nuts: (lina@eientei.org)'s status on Tuesday, 03-Sep-2024 22:00:49 JST Lina Inver:z:e :mur_close: :nuts:

     in reply to
     @mint so many sixes, you'd think this is a 7th grader's literature notebook
      likes this.
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 22:04:11 JST 

     in reply to
     Back at it. All lexbor tests passed, so it's the issue with fast_html.
     † top dog :pedomustdie: likes this.
   • Iska (iska@catposter.club)'s status on Tuesday, 03-Sep-2024 22:43:03 JST Iska

     in reply to

     @mint@ryona.agency how long it's been?I thought they finished it 2 years ago when zammit worked on it. You should check latest gnumach or build from git with some --flag

      likes this.
   •  (mint@ryona.agency)'s status on Tuesday, 03-Sep-2024 23:40:42 JST 

     in reply to
     Thanks to some debugging help from a friend, looks like the issue is broken pipe between fasthtml_worker and whatever process pleromer spawns for it. No idea what to do with it.
     "Hurd of interfaces representing depth"
     The depth in question -> :marioatse:
   •  (mint@ryona.agency)'s status on Wednesday, 04-Sep-2024 00:29:52 JST 

     in reply to
     https://git.pleroma.social/pleroma/elixir-libraries/fast_html/-/blob/master/CHANGELOG.md?ref_type=heads
     >[2.0.0]
     >The worker process now communicates with the node via stdio, instead of TCP, which was known to cause issues on BSD systems
     Yeah, and now it causes issues on Hurd systems. Either way, downgrading it takes a whole lot less friction than trying to debug the issue somewhere within Erlang on obscure unsupported platform.
   •  (mint@ryona.agency)'s status on Wednesday, 04-Sep-2024 00:32:02 JST 

     in reply to
     And yes, it now works.
     Screenshot_20240903_183029.png
   • feld (feld@friedcheese.us)'s status on Wednesday, 04-Sep-2024 00:32:39 JST feld

     in reply to
     @mint this is insane haha
      likes this.
   •  (mint@ryona.agency)'s status on Wednesday, 04-Sep-2024 00:33:32 JST 

     in reply to

     • feld
     @feld The idea of Hurd is infinitely funnier than Hurd itself. Running Pleroma on it ought to be the ultimate shitpost.
   • this ad CAN be blocked, bitch (kirby@lab.nyanide.com)'s status on Wednesday, 04-Sep-2024 00:50:05 JST this ad CAN be blocked, bitch

     in reply to

     • feld
     @mint @feld put it on the internet. i need to join an instance running on hurd
      likes this.
   •  (mint@ryona.agency)'s status on Wednesday, 04-Sep-2024 00:50:30 JST 

     in reply to

     • this ad CAN be blocked, bitch
     • feld
     @kirby @feld I might one day move cum.salon to a Hurd VM without telling anyone.
   •  (mint@ryona.agency)'s status on Thursday, 05-Sep-2024 21:11:48 JST 

     in reply to

     • Iska
     @iska Decided to play around with it more. Apparently there is gnumach-smp in non-snapshot port repos, but it immediately crashes with some APIC assertion error unless I change disk type from IDE to SATA (apparently Hurd supports later, that's surprising). But then for whatever reason they has different disk names, sd0 on up kernel and wd0 on smp so it looks like I have to change fstab every time I need to switch between kernels (adding both with errors=continue results in hang on boot). Also framebuffer corruption hasn't gone anywhere, but I at least can manually start ssh by blindly logging in. htop on SMP kernel still shows only 1 CPU, so not sure if it even works.
   •  (mint@ryona.agency)'s status on Thursday, 05-Sep-2024 21:18:17 JST 

     in reply to

     • Iska
     @iska Actually, Erlang does show more cores than one, but the number appears to be partially wrong (I gave the maching 3 cores with 2 threads per core topology).
     Screenshot_20240905_151609.png
   • BallsackGroyper1488 (ballsackgyroper1488@varishangout.net)'s status on Thursday, 05-Sep-2024 21:22:33 JST BallsackGroyper1488

     in reply to

     • Iska
     @mint @iska GNU/Turd looks like a clusterfuck. Tho as a layman, I do like the idea of a microkernal.
      likes this.
   • Iska (iska@catposter.club)'s status on Thursday, 05-Sep-2024 21:22:39 JST Iska

     in reply to

     • Iska

     @mint@ryona.agency @iska@catposter.club Also framebuffer corruption hasn't gone anywhereVery classic. Happens on any recent system or qemu with -cpu host. X11 may still work though iirchtop on SMP kernel still shows only 1 CPU, so not sure if it even works.Just checked the IRC logs. Apparently you need a program called smp to use smp. logs.guix.gnu.org/hurd/2024-03-07.log

      likes this.
   •  (mint@ryona.agency)'s status on Thursday, 05-Sep-2024 21:48:27 JST 

     in reply to

     • Iska
     >Happens on any recent system or qemu with -cpu host
     Changed CPU type to kvm32, seemed to work now. Well, for a moment.
     >Apparently you need a program called smp to use smp.
     Why the fuck. Pinned cores to the physical ones, it appears when running stress only one of them gets used. With that smp.c thingy, all but the first one. htop still shows only one core.
     Screenshot_20240905_154438.png
   • Iska (iska@catposter.club)'s status on Thursday, 05-Sep-2024 21:55:05 JST Iska

     in reply to

     • BallsackGroyper1488

     @BallsackGyroper1488@varishangout.net @mint@ryona.agency I haven't checked up on GNU/GNU in a while. Mezzano got virtio, sound and GL upstream before Hurd.
     
     It's ironic when fungus man wrote things in C thinking CL would be too big or slow or whatever.
     Nowadays, emacs is slow, hurd is less complete than Yandere Simulator, and both have younger CL counterparts better in every way, except that Lem won't do email and Mezzano isn't a POSix.

      likes this.
   •  (mint@ryona.agency)'s status on Thursday, 05-Sep-2024 21:56:10 JST 

     in reply to

     • Iska
     @iska Also hangs up with some device timeouts on shutdown.
     Screenshot_hurd_2024-09-05_15:5…
   • Phantasm (phnt@fluffytail.org)'s status on Thursday, 05-Sep-2024 22:02:53 JST Phantasm

     in reply to

     • BallsackGroyper1488
     • Iska
     @iska @mint @BallsackGyroper1488
     >Nowadays, emacs is slow, hurd is less complete than Yandere Simulator,
     The main issue with Hurd are the maintainers and mainly Stallman. I was somewhat involved in Hurd a long long time ago and almost every time some discussion about implementing features or optimization occurred, Stallman stepped in and gist of the conversation was almost always "The way I think it should be done is the only way it should be done." Every discussion got stonewalled by stupid nitpicks and Stallman's sometimes obviously bad opinions about the code.
      likes this.
   • Iska (iska@catposter.club)'s status on Thursday, 05-Sep-2024 22:18:18 JST Iska

     in reply to

     • BallsackGroyper1488
     • Phantasm

     @phnt@fluffytail.org @mint@ryona.agency @BallsackGyroper1488@varishangout.net I thought stallman had long abandoned hurd and left it to the 3 autists still caring about it

      likes this.