#
#asahi · 2025-11-15 15:56
campfirechaos_princess: i have some experience in reverse engineering and find it really interesting
#asahi · 2025-11-15 15:55
twbAh, sorry, I defer to your better understanding
#asahi · 2025-11-15 15:55
chaos_princesscampfire: do you have any skill with reverse engineering, embedded programming, or anything of that nature
#asahi · 2025-11-15 15:54
chaos_princessit is not _that_ bad, we are getting things upstreamed
#asahi · 2025-11-15 15:54
chaos_princesswhat
#asahi · 2025-11-15 15:54
twbBasically a waiting game for enough of the Old Guard to die, retire, or be found out
#asahi · 2025-11-15 15:52
twbcampfire: the next step is to get existing drivers into mainline linux, which is a "human problem" not a "tech problem"
#asahi · 2025-11-15 15:40
campfirei don't really know whether this is the place to ask this in, but what's missing from the mac mini m4 support right now? how can i help maybe?
#asahi · 2025-11-14 20:04
jbenot entirely sure why just building it myself worked and not e.g. unpacking the rpm, but I guess I didn't exactly try putting that stuff outside my homedir either perhaps. counting it as a success anyhow
#asahi · 2025-11-14 20:03
jbeok, so I got wpunix working by checking out the project and building it myself (on an x64 fedora machine though, rather than cross compiling), putting the opt/ part of the output under my /opt on asahi and starting using a sh script containing muvm -- /opt/wp80/wpbin/wp -c ${COLUMNS:-$(tput cols)} -r ${LINES:-$(tput lines)} -t ${WPTERM:-xterm} "${@}"
#asahi · 2025-11-13 13:17
the_eyethat seems relatively painless.. gonna give that a try later on, thanks!
#asahi · 2025-11-13 10:03
jbethe_eye: I installed muvm, unpacked the 123elf rpm in a dir under ~/opt, then I did muvm -e TERM=xterm -- ~/opt/123/bin/123
#asahi · 2025-11-13 09:24
the_eyejbe: can you give some pointers what steps you used to get 123elf running? would love to do that too..
#asahi · 2025-11-12 22:10
jbehmm, ok, I did get 123elf running (though that had x64 binaries), so my setup isn't at least completely broken. will have to get back to this at some point...
#asahi · 2025-11-12 18:59
jbeseems to be the same. might try some more later, but thanks for all the help so far :)
#asahi · 2025-11-12 18:53
chaos_princessif you are directly running the binaries - you are using binfmt_misc, if you are passing them as an argument to FEXInterpreter - you are bypassing it
#asahi · 2025-11-12 18:53
jbeor rather, how do I bypass binfmt_misc?
#asahi · 2025-11-12 18:50
jbesame thing it seems
#asahi · 2025-11-12 18:47
chaos_princessmaybe directly run the main executable with fex then and don't rely on binfmt_misc?
#asahi · 2025-11-12 18:43
jbeah, true, didn't think to see that it was a sh file. getting "Invalid or Unsupported elf file." instead now, which I guess is a bit of progress :)
#asahi · 2025-11-12 18:38
chaos_princessi am assuming you are using wp8.sh, so maybe try removing setarch at all and hope fex finds the right architecture
#asahi · 2025-11-12 18:35
chaos_princessenter FEXBash, then run `file $(which setarch)`
#asahi · 2025-11-12 18:35
chaos_princessprobably calling out to native version of setarch instead of emulated one
#asahi · 2025-11-12 18:27
jbehmm, I'm trying to run https://github.com/taviso/wpunix from within a muvm (unpacked the rpm, basically). I'm getting "setarch: i686: Unrecognized architecture" when trying to run it though. haven't really touched x86 emulation before so unsure on where to start troubleshooting
#asahi · 2025-11-11 20:03
JamesB192If only someone could sell the Bursar's dried frog pills to OpenAI
#asahi · 2025-11-11 20:02
JamesB192is a shroomy Grok.
#asahi · 2025-11-11 19:44
jbestill twitching I'm afraid
#asahi · 2025-11-11 19:41
chaos_princessdead internet theory is true, everyone was replaced with a markov chain while you weren't looking
#asahi · 2025-11-11 19:40
lmEveryone in this chat is dead actually, you are talking to ghosts
#asahi · 2025-11-11 19:40
svenobviously I don’t want to do X, but X 🙄
#asahi · 2025-11-11 19:32
chaos_princessdead as a doornail. sell your macs, buy thinkpads
#asahi · 2025-11-11 19:32
tsujpobviously i don't intend to discount the work of others and those who remain but hector martin being gone, asahi lina who did all the GPU work etc those are very significant contributors no?
#asahi · 2025-11-11 19:31
tsujpi'm not trying to fear monger but with so many important public people leaving the asahi project is it effectively dead now? what's the future look like?
#asahi · 2025-11-11 09:01
oddline(I wonder what non-asahi platforms use that hook? if it were asahi-only it could just be moved to asahi-scripts, but I don't think it is)
#asahi · 2025-11-11 09:00
oddlineI don't normally install sdubby, but there's no reason I couldn't, and iirc the "supported" systemd-boot setup on recent (non-asahi) fedora versions does install it
#asahi · 2025-11-11 08:59
oddlinejannau: as a short-term solution, what about adding a copy of that install script to sdubby, and changing the dependency to (grubby OR sdubby)?
#asahi · 2025-11-11 08:57
oddlinechadmed: on fedora the directory itself is provided by systemd-udev, and various hooks come from various packages - it just happens that /that/ hook is in the grubby package
#asahi · 2025-11-11 08:44
chadmedsince the dtbs end up in /lib/modules/${kv_full}/dtbs we can just do that same thing under the expectation that 99% of people running fedora are not going to have weird and wonderful kernel configurations
#asahi · 2025-11-11 08:43
chadmediirc it looked wherever alarm installed the dts to, then got back the latest kernel by doing sort -rV on that dir listing
#asahi · 2025-11-11 08:41
jannauI don't think that's old behaviour. as far as I remember update-m1n1 always looked at a fixed path for the dtbs, certainly for the archlinux arm variant since that allowed only a single installed kernel version
#asahi · 2025-11-11 08:39
chadmedwe dont have eselect kernel or other nonsense like that to really worry about
#asahi · 2025-11-11 08:38
chadmedfor a distro like fedora would it make sense to just go back to the old behaviour of looking for the most recently installed kernel in /lib/modules
#asahi · 2025-11-11 08:37
jannauupdate-m1n1 should either grab the dtbs from the installed kernel or maintain its own symlink
#asahi · 2025-11-11 08:36
chadmedyeah of course, but i think the days of kernel point releases rendering machines completely unbootable due to dtb mismatches are probably over, or close to being over
#asahi · 2025-11-11 08:35
jannauwe are not there yet until the upstream DT contains the full display setup
#asahi · 2025-11-11 08:34
chadmedit is an issue of diminishing importance as dt schemas find their way upstream and all of that stuff stabilises anyway
#asahi · 2025-11-11 08:34
jannauso until that is changed it's a hard depencency from my point of view
#asahi · 2025-11-11 08:33
jannauoddline: yes, the goal is to become independent of that installkernel script provided by grub or grubby but the fedora update-m1n1 config depends on that. There were broken installs due to that
#asahi · 2025-11-11 08:33
chadmedits a hard problem to solve in a universal way at the asahi-scripts level because every distro will want to handle this completely differently
#asahi · 2025-11-11 08:32
chadmedmaybe because grubby symlinks the installed dtbs to /boot, but we should probably just make update-m1n1 look elsewhere for those

plus ›