#asahi · 2022-03-15 17:12
marcanPSA: pushed a m1n1 bugfix that affects stage 1. I *could* work around it in stage 2, but this doesn't affect normal u-boot chainload cases yet. however, it will break SEP once we get there.#asahi · 2022-03-15 16:12
marcanbut not merged yet#asahi · 2022-03-15 16:12
marcanwell, the spi controller was#asahi · 2022-03-15 16:11
marcannope#asahi · 2022-03-15 16:08
marcanand then I need to hack up calamares to preconfigure the keyboard model properly on the macbooks... which is also going to require some extra devicetree shenanigans to forward the layout from the ADT... (as we just confirmed in #asahi-dev)#asahi · 2022-03-15 16:07
marcanhaha, sure :)#asahi · 2022-03-15 16:05
marcanI mean, I think I got close to option #2 in this conversation? :)#asahi · 2022-03-15 16:05
marcan*sure#asahi · 2022-03-15 16:05
marcansurde#asahi · 2022-03-15 16:04
marcanheh#asahi · 2022-03-15 16:03
marcanthe hid-apple driver does the same, defaulting to mapping to media keys unshifted and Fx with Fn, but you can swap it via module arg, or disable the conversion entirely#asahi · 2022-03-15 16:03
marcanso Fn, the HID codes the top row reports are indeed Fx, but the keycaps say they are actually media keys unshifted and F-keys with Fn, though on macOS you can swap them if you want#asahi · 2022-03-15 16:00
marcan(I actually don't know and have never met any japanese computer users that use kana mode)#asahi · 2022-03-15 15:59
marcan("you start typing like this", no, that is not legible in japanese either)#asahi · 2022-03-15 15:59
marcanbecause right now if you select it ンラナ トカチスカ カンセニミキ リニノイ カクニト#asahi · 2022-03-15 15:58
marcanit should probably be called "Japanese (Macintosh, Kana)" or something#asahi · 2022-03-15 15:58
marcanthe only remaining confusing thing is jp(mac) has, if nothing else, a terrible name#asahi · 2022-03-15 15:58
marcanyeah, makes sense#asahi · 2022-03-15 15:57
marcanbut it's in a git repo that is in a machine that's not booted right now#asahi · 2022-03-15 15:57
marcanfwiw I had already done the ~same thing except as layout variants instead of options :-)#asahi · 2022-03-15 15:56
marcanpc106 should be AE13#asahi · 2022-03-15 15:54
marcanproblem is evdev Fn is >255 so Xorg can't see it#asahi · 2022-03-15 15:54
marcanbut that still leaves Fn+everything else unused#asahi · 2022-03-15 15:54
marcanthe latter is probably a decent idea given that having those keys missing kind of sucks, and most people wouldn't want to do anything else there#asahi · 2022-03-15 15:54
marcanthe driver properly reports it in evdev, and also synthesizes Fn+top bar = Fxx keys, and Fn+arrow keys = paging/home/end keys in-kernel (although that can be disabled)#asahi · 2022-03-15 15:53
marcandaniels: there's also the Fn thing, any ideas about what to do with that?#asahi · 2022-03-15 15:49
marcanwhere the two options for "somewhere sane" are replacing ¥ or putting it where it would be on every other PC keyboard#asahi · 2022-03-15 15:49
marcanbut given we can't do that in xkb, I argue the reasonable thing to offer by default is "provide a layout that matches the keycaps without trying to do fancy third level shifts that don't fit into the standard linux modifier model, and stick backslash somewhere sane"#asahi · 2022-03-15 15:48
marcanand if you want to fix *that* you need userspace daemons (which is how people actually do it)#asahi · 2022-03-15 15:48
marcanif you want to do what macos does then we need the option thing, but then that kind of clashes with the whole "macos doesn't do keybindings like linux does"#asahi · 2022-03-15 15:47
marcanI mean, I'm talking about people coming from the PC world#asahi · 2022-03-15 15:47
marcanespecially considering the _ key having no shifted variant is completely silly#asahi · 2022-03-15 15:47
marcanand then apple kind of went with JIS plus that option hack, but most Linux users probably would rather have either of the above two modes#asahi · 2022-03-15 15:46
marcanhowever, jp(106) is the default because microsoft#asahi · 2022-03-15 15:46
marcanso if you want what's printed on the keycaps on the vast majority of machines today, you want that#asahi · 2022-03-15 15:46
marcanmy normal PC keyboard is OADG109A, most are these days#asahi · 2022-03-15 15:46
marcan(which is actually the default)#asahi · 2022-03-15 15:45
marcanwhich is what jp(106) does#asahi · 2022-03-15 15:45
marcanhowever, some people might prefer the AE13 -> backslash,bar version#asahi · 2022-03-15 15:45
marcanwhich is the OADG109A version#asahi · 2022-03-15 15:45
marcanthat was originally key <AB11> { [ backslash, underscore] };#asahi · 2022-03-15 15:45
marcandaniels: the problem is replace key <AB11> { [ underscore ] };#asahi · 2022-03-15 15:40
marcanapple tries to do the right thing in the OS with font rendering, but the keyboard is still weird because of this#asahi · 2022-03-15 15:39
marcanand I get emails with currency denominated in backslash#asahi · 2022-03-15 15:39
marcanso most japanese people don't *know* what a backslash is#asahi · 2022-03-15 15:39
marcanwhere even in unicode mode, to this day, \ renders as ¥ if you have a japanese locale#asahi · 2022-03-15 15:39
marcanyou know this is all microsoft's fault, right?#asahi · 2022-03-15 15:38
marcan:)#asahi · 2022-03-15 15:38
marcanmake it do a different thing in terminal apps?#asahi · 2022-03-15 15:37
marcanwell, considering xkb can't magically do the nonsense macOS does, the next best thing is make the special keys do the right thing and then let users pick between 106 and OADG109A modes, which is what they'll be used to from other machines anyway