#asahi-alt · 2024-09-02 10:20
leioError creating the Kvm object: Error(13)#asahi-alt · 2024-09-02 10:20
leiokrun is just panicking immediately#asahi-alt · 2024-09-02 10:17
leionot sure what the next steps were for steam here#asahi-alt · 2024-09-02 10:14
leioafter libkrun upgrade (had some old version of that too), it was fine with old virglrenderer too, as far as compiling is concerned#asahi-alt · 2024-09-02 10:07
leioif older doesn't work#asahi-alt · 2024-09-02 10:07
leiofine, but minimum versions should still be listed, e.g. for virglrenderer#asahi-alt · 2024-09-02 10:05
leiowell, not matching, but probably too old for it#asahi-alt · 2024-09-02 10:05
leioso that wasn't upgraded either#asahi-alt · 2024-09-02 10:05
leiolibkrun dep doesn't have a matching minimum version either#asahi-alt · 2024-09-02 10:04
leiothough I don't see that dep at all in krun ebuild#asahi-alt · 2024-09-02 10:04
leiofeels like too low virglrenderer dep or some such, as I had that installed at a lower version from long ago tries of this stuff#asahi-alt · 2024-09-02 10:03
leiook, https://bpa.st/QYU44 might be more readable#asahi-alt · 2024-09-02 10:03
leiowhat is this ansi stuff, grr#asahi-alt · 2024-09-02 10:02
leioyup, that was it. Now I'm on a real rust compile error like https://bpa.st/JDQ7E#asahi-alt · 2024-09-02 10:01
leionvm, I missed one of the env overrides I need for some bindgen stuff, it looks like#asahi-alt · 2024-09-02 10:00
leionow stuck with krun on https://bpa.st/7ZMA4#asahi-alt · 2024-09-02 09:59
leioFEX went fine with https://bpa.st/GXOSY#asahi-alt · 2024-09-02 09:58
leiobuilt without any -march/-mcpu/-mtune via the env override, with https://bpa.st/WB3IO#asahi-alt · 2024-09-02 09:56
leioI set -march=native -mcpu=native in my gcc flags override, which I need for this due to the glibc build#asahi-alt · 2024-09-02 09:51
leio(and default on arm64)#asahi-alt · 2024-09-02 09:51
leio-fomit-frame-pointer is the default on x86, unfortunately; I'm having to override it everywhere :(#asahi-alt · 2024-09-02 09:51
leioif I set some x86 -march there, it'll fail in binutils build (and probably gcc stage1 and more), as that's done with native compiler#asahi-alt · 2024-09-02 09:50
leiowhich is user intervention#asahi-alt · 2024-09-02 09:50
leioyeah, right now the only way to build this is to do a package.env that forces -O2 -pipe or similar#asahi-alt · 2024-09-02 09:50
leioit's my default go-to when doing manually because bpa.st doesn't like longer stuff, albeit that wasn't long#asahi-alt · 2024-09-02 09:49
leiothat's just reporting gcc can't compile in configure.log, the interesting stuff is in bpa.st#asahi-alt · 2024-09-02 09:38
leiochaos_princess: x86_64-multilib-toolchain build fails for me with https://dpaste.com/F7EYRRCSP (config.log https://bpa.st/UE6DS ), as somehow strip-unsupported-flags isn't getting rid of -march=native#asahi-alt · 2024-09-01 11:10
leiogtk core devs say it being enabled has otherwise much less problems#asahi-alt · 2024-09-01 11:09
leiobut I think in 46 or 47 it might be non-fuzzy regardless#asahi-alt · 2024-09-01 11:09
leioin GNOME it's about experimental feature "“scale-monitor-framebuffer” — makes mutter default to layout logical monitors in a logical pixel coordinate space, while scaling monitor framebuffers instead of window content, to manage HiDPI monitors. Does not require a restart." is enabled or not, if it's fuzzy or not, it's not exactly a "support fractional scaling" thing, but it enables it#asahi-alt · 2024-09-01 11:03
leiochaos_princess: fwiw, jetbrains EAPs have wayland builds now, no clue of the arm64 variants, haven't investigated until I can have them non-fuzzy by not using fractional in gnome#asahi-alt · 2024-08-22 08:23
leiooh, need to continue reading, it was explained to sam after :)#asahi-alt · 2024-08-22 08:22
leiochadmed: I don't know what fex thunks are. We still owe you some PR merges too, I think?#asahi-alt · 2024-08-19 17:52
leiodon't see anything for that either, but I don't trust it when I can't narrow it down with a case-sensitive search of M1#asahi-alt · 2024-08-19 17:50
leio`#ifdef M1` for example, iirc (been doing rust instead of C too long)#asahi-alt · 2024-08-19 17:50
leioit can be a different syntax for the M1 line#asahi-alt · 2024-08-19 17:02
leiobut doesn't seem to cover arm64#asahi-alt · 2024-08-19 17:01
leiothe existing ::guru version does something like this per-arch#asahi-alt · 2024-08-19 17:00
leiobut on line 190 they just add a M1 preprocessor macro definition from it - maybe that does something needed for asahi#asahi-alt · 2024-08-19 17:00
leioI think most of this is defaults based on arch, that you can sort of replicate based on arch implicit USE flags or something if need be#asahi-alt · 2024-08-19 17:00
leiorest, well, need to understand what it does in the project#asahi-alt · 2024-08-19 16:59
leiofor this to be suitable for ::gentoo#asahi-alt · 2024-08-19 16:59
leioif you do end up with having to pass M1=1, you need to patch out the CFLAGS modifications (if they override user one, cmake is confusing to me), etc#asahi-alt · 2024-08-19 16:56
leioit's a bit of a mess for downstreams#asahi-alt · 2024-08-19 16:54
leiodon't have a clone to quickly check#asahi-alt · 2024-08-19 16:54
leiothough add_definitions(-DM1) might be doing something inside the code :(#asahi-alt · 2024-08-19 16:53
leioand those feel like they are just for all arm64, hence `if use arm64; then ...`#asahi-alt · 2024-08-19 16:53
leiolike https://github.com/ptitSeb/box64/blob/main/CMakeLists.txt#L67-L72#asahi-alt · 2024-08-19 16:53
leioso just don't, pass what's needed manually#asahi-alt · 2024-08-19 16:52
leioyou mustn't end up having the ebuild let these be set in place of user CFLAGS, so you can't pass M1=1 in an ebuild without patching that out