IRC channel logs

<bigmike-sit>etno: good to know!

<bigmike-sit>Alicia: LLM psychosis perhaps? i saw that abd

<bigmike-sit>and it didnt even look like they wrote it*

<bigmike-sit>just copy and pasted a chatGPT output into email

<damo22>i believe the person driving the LLM for that project speaks french, and he uses the LLM as a translation tool to communicate as well as write code. But it certainly is confusing and misleading to call it GNU-AI

<Alicia>bigmike-sit: I suspect so

<bigmike-sit>damo22: That makes sense

<bigmike-sit>does seem to be a bit of vibecoded ai hypism though... oh well.

<gcannata23>hello, any interest of picking up aarch64 port initiated by sergey bugaev for gnu mach ?

<gcannata23>I don't to add noise to the mailing list, so I tought it would have been better to write here

<gcannata23>*don't want to add noise

<damo22>sneek: later tell gcannata23: hurd related development ideas are always welcome on the mailing list

<sneek>Will do.

<gcannata23>My idea is to continue the port to aarch64 but I need a mentor that it will review my patches before submitting them to the mailing list

<sneek>gcannata23, you have 1 message!

<sneek>gcannata23, damo22 says: hurd related development ideas are always welcome on the mailing list

<gcannata23>I have also proposed to a friend of mine that he has a beer pub to let me do an event to gather and talk people about the Hurd

<gcannata23>My crazy idea is to boot GNU Mach idea on real hardware, maybe AWS Graviton processors.

<GNU_Hurd_Rocks>Arm is indeed everywhere, unfortunately not an area I'm familiar with though

<gcannata23>The beer pub is also relatively near an importat university in Italy - the Politecnino di Milano

<gcannata23>Any Italian guys or girls here that like the Hurd ?

<gcannata23>Maybe I will create a post on Reddit to test the waters

<Alicia>good luck!

<GNU_Hurd_Rocks>"why are you using hurd, just use linux" (I could see that happening, unfortunately)

<GNU_Hurd_Rocks>I swear the average iq on some users is so bad though, and I hate that search engines even provide such results

<hwpplayer1>I believe that I can contribute to Hurd/GNU

<hwpplayer1>But I want to make something inhouse

<GNU_Hurd_Rocks>POWER ISA EULA is a funny thing, specifically 1.4, anti bad hardware focus?

<hwpplayer1>does it count ? before giving the code back ?

<GNU_Hurd_Rocks>I still want to experiment with making a terrible OS and boot loader for BIOS :3

<GNU_Hurd_Rocks>only thing that stops me is not having the time, too much goals and not enough time

<hwpplayer1>you don't know algorithms then

<GNU_Hurd_Rocks>I'm bad at multi-tasking :3

<hwpplayer1>but computer can do that

<hwpplayer1>:)

<hwpplayer1>this is called operating system

<GNU_Hurd_Rocks>and I want to experiment with neural networks and some more advanced things

<GNU_Hurd_Rocks>nah I have a bad idea, Hurd for POWER, RPC ;3

<GNU_Hurd_Rocks>RCS* agha

<hwpplayer1>wdym ?

<GNU_Hurd_Rocks> https://www.raptorcs.com/

<hwpplayer1>I see

<hwpplayer1>thanks

<GNU_Hurd_Rocks>SMT4, crazy, https://www.raptorcs.com/content/CP9M38/intro.html

<GNU_Hurd_Rocks>apparently SMT8 is a thing though?

<Pellesco1rs>a

<Pellescours>acl 2.4.0 cannot compile because it uses O_PATH in libmisc/fchmodat_compat.c. attr 2.6.0 has the same compilation issue (O_PATH usage)

<GNU_Hurd_Rocks>acl?

<Pellescours>yes gnu acl

<GNU_Hurd_Rocks>this? https://savannah.nongnu.org/projects/acl

<Pellescours>yup

<hwpplayer1>thanks for the link acl

<hwpplayer1>GNU_Hurd_Rocks: ^

<GNU_Hurd_Rocks>looks like O_PATH is for opening a fd for performing some operations, rather interesting

<GNU_Hurd_Rocks>based on open(2) it suggests fchmod(2) and fchown(2) would fail though

<GNU_Hurd_Rocks>I found chown(2) then :3

<sam_>the context is always worth looking at

<sam_>use was introduced because of toctou issues

<GNU_Hurd_Rocks>I would have to look at where it fails in acl though

<GNU_Hurd_Rocks>with grep I found part of a comment, libmisc/proc-self-fd.c:71: * least kernel version 7.2 fail with EBADF when the dirfd is an O_PATH file

IRC channel logs

2026-10-10.log