#
✕ nick:marcan
#asahi-dev · 2023-01-01 09:06
marcanjannau: is it possible "opaque" actually means "premultiplied alpha"?
#asahi-dev · 2022-12-28 13:47
marcanno idea...
#asahi-dev · 2022-12-28 13:15
marcangot it :)
#asahi-dev · 2022-12-28 13:09
marcanat that point there is nothing to do :)
#asahi-dev · 2022-12-28 13:09
marcanstable releases will make sense at the point where upstream already has all the drivers you actually want (and backporting actual bugfixes should be trivial and already taken care of by the kernel stable process)
#asahi-dev · 2022-12-28 13:05
marcanncopa: I don't think stable releases make much sense here, given asahi itself is very much not stable
#asahi-dev · 2022-12-28 13:04
marcanmaz: I'm going to do the rebase soonish, maybe this weekend
#asahi-dev · 2022-12-28 13:03
marcanDocumentation/ABI/testing/sysfs-class-power is the proper documentation
#asahi-dev · 2022-12-28 13:01
marcanthose are legacy nonstandard names
#asahi-dev · 2022-12-27 03:01
marcantpw_rules: not past ones, will tag going forward though
#asahi-dev · 2022-12-26 17:24
marcaneiln: vm-base/vm-size is usually about the ASC
#asahi-dev · 2022-12-26 03:16
marcanjannau: pushed dcp/sleep (blindly rebased, not compile tested)
#asahi-dev · 2022-12-26 03:11
marcanpractically: I think the only actual use case for non-native colorspaces in DCP is for video playback and such, where you're feeding in frames straight from a video decoder / scaler engine and bypassing the GPU (and normal ICC processing)
#asahi-dev · 2022-12-26 03:10
marcanIIRC macOS has the ICC profiles stashed somewhere in plain view and I imagine those are intended to output at the native colorspace
#asahi-dev · 2022-12-26 03:10
marcanI'm curious how good ICC support is in KDE/Wayland these days
#asahi-dev · 2022-12-26 03:10
marcanat some point we'll figure out the ICC profiles to do this all properly, feel free to look into that if you're interested
#asahi-dev · 2022-12-26 03:09
marcanbut also in general fewer conversion steps are preferred, so native is better and should result in less banding/quantization issues, and just let userspace deal with a single color correction step
#asahi-dev · 2022-12-26 03:08
marcanwe can't do 10-bit sRGB without either new colorspace/userspace support (ew) or matrix weirdness to compensate
#asahi-dev · 2022-12-26 03:08
marcanDCP has been doing sRGB, but that's a waste of a default; until we get both options exposed (which will require matrix gooeiness for 10-bit due to the wide gamut weirdness) it makes more sense to just expose native and let people who care do their own color management IMO
#asahi-dev · 2022-12-26 03:07
marcanthat's the native colorspace
#asahi-dev · 2022-12-26 03:07
marcannicolas17: yes, it only affected DCP, so you can expect oversaturated red again :p
#asahi-dev · 2022-12-26 03:07
marcanjannau: yeah, so we still need a completion or something at some point to block until DCP is properly up
#asahi-dev · 2022-12-25 17:34
marcanare you in a rush? :p
#asahi-dev · 2022-12-25 17:34
marcanmaybe tomorrow
#asahi-dev · 2022-12-25 17:31
marcanand it's also kind of useless since it doesn't persist after reboots: https://bugs.kde.org/show_bug.cgi?id=450551
#asahi-dev · 2022-12-25 17:31
marcanfunny enough the UI for that in KDE doesn't work: https://invent.kde.org/plasma/powerdevil/-/merge_requests/127
#asahi-dev · 2022-12-25 17:30
marcanand also battery charge threshold emulation in the macsmc-power kernel driver
#asahi-dev · 2022-12-25 17:30
marcanonly tested on t6k right now
#asahi-dev · 2022-12-25 17:30
marcanjannau: pushed the DCP 10-bit fix + switch to native colorspace (also for 8bit) to bits/200-dcp
#asahi-dev · 2022-12-24 11:46
marcanyou can add the other one as alt and just make sure to commit with the same email as the signoff
#asahi-dev · 2022-12-24 11:46
marcanI merged it anyway, but you might want to fix that
#asahi-dev · 2022-12-24 11:45
marcanas in the commit author
#asahi-dev · 2022-12-24 11:45
marcanChaosPrincess: oh, it's because it does not match your GitHub email
#asahi-dev · 2022-12-24 11:44
marcanhuh, weird.
#asahi-dev · 2022-12-24 11:43
marcanChaosPrincess: don't forget the signoff
#asahi-dev · 2022-12-23 04:34
marcanI think for me it'd be ideal if I can just bump that at the same time I bump mesa-asahi-edge (or even have both be part of the same PKGBUILD) and push to our repo as usual
#asahi-dev · 2022-12-23 04:33
marcancan we wrap that in a PKGBUILD so we can offer it as a package for users (via the `host` gldriver stuff)?
#asahi-dev · 2022-12-22 08:36
marcanwithout that input won't work
#asahi-dev · 2022-12-22 08:36
marcanit's about *input* not *output* there, because it gets called as curl|sh
#asahi-dev · 2022-12-21 12:15
marcan:)
#asahi-dev · 2022-12-21 11:38
marcanmaz: niiice!
#asahi-dev · 2022-12-19 11:44
marcanmaz: :<
#asahi-dev · 2022-12-17 03:46
marcanzzywysm: Needs a signoff, yeah, please fix that and I can merge it :)
#asahi-dev · 2022-12-16 16:43
marcansven: someone said that turning bluetooth off before suspend and back on after fixes the fail, so maybe that's all we need to do (though in that case, I wonder if it needs to be sequenced with wifi in some way...)
#asahi-dev · 2022-12-16 16:38
marcanwhich already has a disclaimer that you won't file bugs
#asahi-dev · 2022-12-16 16:37
marcanit only shows in expert mode
#asahi-dev · 2022-12-16 16:35
marcanI assume the folks here with legitimate reasons to do it won't mind :p
#asahi-dev · 2022-12-16 16:34
marcanI'm usually nicer than that but I'm considering explicitly choosing expert mode as permission to be shouted at.
#asahi-dev · 2022-12-16 16:34
marcanI give up: https://github.com/AsahiLinux/asahi-installer/commit/0ac64c9ce1c460f4576162a82d239d7e8688a79e
#asahi-dev · 2022-12-15 13:07
marcanI imagine some discussion happened offlist after the NACKs he got, or it just got stuck in limbo

plus ›