#
#asahi-alt · 2025-06-16 11:46
chadmedyeah but can we actually do anything about flatpak?
#asahi-alt · 2025-06-16 11:44
chaos_princessAnd i think i've found the odd proprietary /opt thing that assumed i had ffmpeg installed
#asahi-alt · 2025-06-16 11:44
chaos_princessfor b - there are two sharp edges: 1. AppImages do not work unless extracted, not an issue, but a bit annoying. 2. Flatpak. Arguably solves the problem perfectly as one could say "do not have the right libraries - use flatpak", except it does not work.
#asahi-alt · 2025-06-16 11:42
chadmedso i dont think the approach of "here is glibc and some multimedia libraries steam/most games use" is a bad approach
#asahi-alt · 2025-06-16 11:41
chadmedb) proprietary /opt shit that is all statically linked anyway
#asahi-alt · 2025-06-16 11:41
chadmeda) steam
#asahi-alt · 2025-06-16 11:41
chadmedi maintain that given the rapidly maturing state of aarch64 desktop right now, there are basically only two uses for linux fex anyway:
#asahi-alt · 2025-06-16 11:40
chaos_princessit is arguably a -bin package and may or may not be too single-purpose. We only care about steam and maybe resolve, others may have different ideas
#asahi-alt · 2025-06-16 11:39
chadmedwhats wrong with how we do the rootfs now?
#asahi-alt · 2025-06-16 11:39
chadmedlmfao yeah fair enough
#asahi-alt · 2025-06-16 11:39
chaos_princessand no, i would rather split linux and xta, mostly because i do not want to vendor half the cmake.eclass again
#asahi-alt · 2025-06-16 11:38
chadmedah
#asahi-alt · 2025-06-16 11:38
chadmedthe benefit would be that all the little bits like FEXConfig and AppConfigs etc can get installed irrespective of the user's choice between linux and xta
#asahi-alt · 2025-06-16 11:38
chaos_princessnah, it is not about the fex package itself, it is the rootfs that is the annoying one
#asahi-alt · 2025-06-16 11:38
chadmedchaos_princess: following on re packaging, my thinking was a unified package with use flags to control what actually gets built/installed but that ebuild would be quite nasty
#asahi-alt · 2025-06-13 08:14
kode54so basically, it's a no-win scenario :D
#asahi-alt · 2025-06-13 08:13
chadmedi have some sympathy for users in this regard. if they file useless/misplaced bug reports we get upset at them. we also get upset at them when they dont report bugs :P
#asahi-alt · 2025-06-13 08:12
sam_it's really hard to communicate it
#asahi-alt · 2025-06-13 08:12
sam_this is really The Problem with testing
#asahi-alt · 2025-06-13 08:12
sam_yep
#asahi-alt · 2025-06-13 08:12
kode54my favorite is people committed to the bit of macOS, but not committed enough to test the betas at all, under the impression that any bugs they encounter now are just Beta Crud, and therefore won't be a problem if they just Wait For Stable
#asahi-alt · 2025-06-13 08:11
sam_what I really want is for the previous branch to overlap a little bit with the new one
#asahi-alt · 2025-06-13 08:10
sam_I have made comments about this upstream and they are getting there (which is why the releases now last longer)
#asahi-alt · 2025-06-13 08:10
sam_on the other hand, if we wait until towards the end, if we find any bugs, they won't get fixed because the branch is EOL now
#asahi-alt · 2025-06-13 08:10
sam_*releases
#asahi-alt · 2025-06-13 08:10
sam_like, if we stable 20 immediately, we end up making stable users build many point releaes
#asahi-alt · 2025-06-13 08:10
sam_it's always a tricky thing because of how LLVM's release process is
#asahi-alt · 2025-06-13 08:06
chadmedik it broke a bunch of stuff when it dropped
#asahi-alt · 2025-06-13 08:06
chadmedit hasnt really been an issue, more of a curiosity
#asahi-alt · 2025-06-13 07:57
sam_you're free to ping and ask about it as well..
#asahi-alt · 2025-06-13 07:57
sam_plan is to do that shortly
#asahi-alt · 2025-06-13 07:10
chadmedgentoo hasnt even stabilised 20 yet :/
#asahi-alt · 2025-06-13 06:51
kode54Arch just adopted 20 within the last month :/
#asahi-alt · 2025-06-13 06:50
kode54LLVM 21 branching is happening that soon?
#asahi-alt · 2025-06-13 01:51
chadmedmost likely fex-related
#asahi-alt · 2025-06-13 01:51
chadmedi am concerned about SEH being largely untested by the sounds of it but there are probably going to be bigger issues before that matters
#asahi-alt · 2025-06-13 01:50
chadmedhm we will probably need to hard depend on llvm-libunwind then too
#asahi-alt · 2025-06-12 22:28
vimprovedand also i wouldn't really consider one patch a scuffed fork, just a backport but i get the point
#asahi-alt · 2025-06-12 22:27
vimprovedyeah i'm aware, i just had it my head that the LLVM 21 branching was still like 3 months away
#asahi-alt · 2025-06-12 22:24
chadmed_if we wanted to do that we couldve had something cobbled together ages ago
#asahi-alt · 2025-06-12 22:23
chadmed_the whole idea is to not package a bunch of scuffed forks and insane shit
#asahi-alt · 2025-06-12 21:58
vimprovedi should probably get on testing it
#asahi-alt · 2025-06-12 21:57
vimprovednevermind
#asahi-alt · 2025-06-12 21:57
vimprovedoh wait i did not realize llvm 21 was so close
#asahi-alt · 2025-06-12 21:57
chaos_princess-g
#asahi-alt · 2025-06-12 21:57
chaos_princessvery much do not want to make some franken-toolchaing
#asahi-alt · 2025-06-12 21:56
chaos_princessor we can wait like a month or so
#asahi-alt · 2025-06-12 21:55
vimprovedchaos_princess: couldn't this patch just be carried locally in the llvm-mingw64 package for now? it's not too large and any effects of it would be localized to only the mingw libunwind build
#asahi-alt · 2025-06-12 15:12
chaos_princessarm64ec update - mingw 13 got released and seems to work, now we need llvm 21 i guess? whatever will end up containing this: https://github.com/llvm/llvm-project/pull/138583
#asahi-alt · 2025-06-08 02:40
chadmedthe other annoying thing is that it doesnt seem like any of our little hacks to make steamwebhelper not eat shit work in wine

plus ›