| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 00:06 | |
| - mjw (QUIT: Killed (NickServ (GHOST command used by mark_!~mark@gnu.wildebeest.org))) (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 00:29 | |
| * mark_ -> mjw | 00:29 | |
| + wielaard (~mjw_@2001:1c06:2488:1400:9e5c:8eff:fe8f:a440) | 00:29 | |
| - ajr (QUIT: Quit: WeeChat 3.7) (~ajr@user/ajr) | 00:56 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 02:50 | |
| flowy | frick yea, i finally got moonlight working perfectly smooth! 1080p 60fps | 03:07 |
|---|---|---|
| flowy | its' EGL output renderer was failing to initialise, as it wanted > GL ES3. the fallback renderer is really slow and introduced around 25ms just for painting decoded frames. after 5ms for decoding, 60fps was impossible out of the box | 03:12 |
| flowy | but the only things that actually required GL ES3 were some very simple shaders | 03:13 |
| flowy | after backporting the shaders to ES2, the EGL renderer works and now frames are painted in 1ms | 03:14 |
| flowy | i might find a clean way to upstream this. however, the other ingredient is patched ffmpeg for v4l2-request support, which is out of our control. | 03:19 |
| bkeys | Boostisbetter I am a happy Moonlander mk1 keyboard, got mine and haven't looked back since | 05:44 |
| - Boostisbetter (QUIT: Ping timeout: 252 seconds) (4a410829d7@irc.cheogram.com) | 05:55 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 06:32 | |
| Boostisbetter | bkeys: Yeah, I know what you mean. I am prefer using mine as well. However when it comes time to write some software, I find that I have to hunt too much on it, and so I tend to switch up to a UHK which has clear key legends and has been optimized for productivity. | 06:35 |
| - bkeys (QUIT: Ping timeout: 268 seconds) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 06:36 | |
| + bgs (~bgs@212-85-160-171.dynamic.telemach.net) | 07:46 | |
| - bgs (QUIT: Remote host closed the connection) (~bgs@212-85-160-171.dynamic.telemach.net) | 07:46 | |
| - S0rin (QUIT: Ping timeout: 252 seconds) (~S0rin@user/s0rin) | 09:39 | |
| + S0rin (~S0rin@user/s0rin) | 09:45 | |
| - S0rin (QUIT: Ping timeout: 256 seconds) (~S0rin@user/s0rin) | 09:50 | |
| + S0rin (~S0rin@user/s0rin) | 09:52 | |
| josch | flowy: i thought the ffmpeg v4l2-request patch doesn't apply anymore for current ffmpeg? | 09:59 |
| Boostisbetter | josch, what keyboards are preferable to? Just whatever one is in front of you? | 10:04 |
| Boostisbetter | minute, same question to you, although I suspect you are partial the Reform board. | 10:04 |
| josch | Boostisbetter: i prefer the keyboard that i got used to. Switching to the reform took a bit but now I'm as fluent on it as on my old keyboard. I don't know much about keyboards so i wouldn't be able to tell which ones are better or worse. I typed on normal laptop keyboards for most of my life so I don't have high standards. :D | 10:05 |
| minute | Boostisbetter: i found that low profile mechanical keyboards work the best for me, or very light full profile keyboards (like the one in the a500). i have a hhkb lite that i find absolutely horrible, and a WASD with mxes that i find meh nowadays | 10:06 |
| minute | i am looking forward to reform keyboard v3 | 10:06 |
| minute | i can't get used to split keyboards or ergo kbs | 10:07 |
| Boostisbetter | minute, I have a hhkb as well, and while I used to like it, now I find it to be no where near as good as the other boards I have. I think the Reform kb is very good as far as standard staggered layouts go, but I am also really looking forward to the ortho pocket reform kb. | 10:07 |
| minute | yeah, with refkb3 i will implement a more traditional stagger to accommodate people who are used to it | 10:07 |
| minute | and pocket is ortho as ortho gets yes :D | 10:08 |
| Boostisbetter | I used to think that I couldn't do split kbs either, but then I gave it a shot, and it worked. Now, of course, I find them so good that I really preferred using them only on desktops. But at work I use a standards rubber dome thing and have been for a long time. | 10:08 |
| minute | typing on pocket is kind of a fun experience i think... might need a little getting used to at first, especially for people who have never used ortholinear | 10:09 |
| minute | most people cannot touch type anyway and just hunt for the legends | 10:09 |
| minute | i touch type but probably not 100% correctly | 10:10 |
| Boostisbetter | yeah, I'm used to typing on small keyboards, AND the Penkesu I built uses a Plunck style board. I'm totally used to them. | 10:11 |
| Boostisbetter | I'm really looking forward to the Pocket. | 10:11 |
| - S0rin (QUIT: Ping timeout: 260 seconds) (~S0rin@user/s0rin) | 10:11 | |
| Boostisbetter | Hope things are progressing well on that front. | 10:11 |
| Boostisbetter | but to make one final comment on keyboards, I find the Moonlander to be the most comfortable I have used, BUT the UHK with the thumb modules is probably the most practical keyboard I have used. | 10:12 |
| Boostisbetter | There is just something about being able to mouse and type all without moving your arms. | 10:12 |
| minute | Boostisbetter: yes, case parts are in first test production run | 10:13 |
| Boostisbetter | that is great! Do you already have a line on all of the components your'll need for production? Specifically the SoC? | 10:13 |
| Boostisbetter | Really exciting stuff! | 10:13 |
| minute | we are doing something slightly unusual, so we have 2 aluminum bodies and each has a top and bottom plate/bezel made from PCB material | 10:13 |
| minute | yesterday i chatted with pcbbuy about doing counterbores in these 1mm pcbs | 10:14 |
| c-keen[m] | Is there an update on the battery boards? | 10:14 |
| minute | Boostisbetter: the soc via boundary has 24wk lead time | 10:15 |
| + S0rin (~S0rin@user/s0rin) | 10:15 | |
| minute | c-keen[m]: yes, they're done and on the way to us | 10:15 |
| Boostisbetter | holy cow, that is quite the lead but I guess to be expected. | 10:16 |
| c-keen[m] | We have orders here with a delivery estimate of 72 weeks 😕 | 10:20 |
| c-keen[m] | (here == at work) | 10:20 |
| c-keen[m] | minute: cool! | 10:20 |
| minute | Boostisbetter: yeah, i hope it will get better but as c-keen[m] says, it's not even bad tbh | 10:21 |
| minute | our first batch of cameras is almost done too btw | 10:22 |
| minute | batch of 100x | 10:22 |
| c-keen[m] | \o/ | 10:23 |
| minute | and for cm4, i think i'm around 1 day (fulltime) work away from finishing the layout for the hdmi->edp adapter | 10:24 |
| JC[m] | Looking forward to paying more attention to Reform soon. I lent mine to a friend a couple of months ago (he's super stoked about it but understandably can't afford one right now) but unfortunately he was tinkering with the battery board while the batteries were connected and shorted out a trace. Oof. | 10:36 |
| c-keen[m] | on the battery board? | 10:36 |
| JC[m] | Yep. The mainboard was fine and the only casualty was an exposed trace. Lucked out. | 10:37 |
| c-keen[m] | That's a good reason to order the newer battery boards then | 10:37 |
| JC[m] | I was previously tinkering with the original battery board PCB to add socketed fuses (because someone's unit potentially catching fire seems like an intense liability) but haven't really looked into the updated version. I don't know much about BMS circuitry. | 10:39 |
| c-keen[m] | the redesign added undervoltage protection for each cell | 10:39 |
| c-keen[m] | amongst other things | 10:40 |
| JC[m] | There is someone who attended a past open hardware meetup where I live and he creates BMS systems professionally. I was hoping to get his info and see if I can coerce or pay him for some peer review. | 10:40 |
| Boostisbetter | minute, cm4 is just a rpi4 right? RAM amount is 8 gb? If so that is great because it sounds like it will be more powerful than the current Reform SoC. | 10:40 |
| JC[m] | I'd want to ensure it not only has undervoltage protection but overvoltage/short-circuit protection as well | 10:41 |
| minute | Boostisbetter: yes, it has a lot of pluses and some drawbacks | 10:41 |
| minute | Boostisbetter: cpu is faster (4x A72), gpu can do vulkan, but it can't sleep and it can't do accelerated disk encryption | 10:41 |
| Boostisbetter | can't sleep is the deal break for me. | 10:49 |
| josch | Boostisbetter: mine can't sleep either but since the battery runtime with display turned off is 6+ hours and since the reform barely produces heat when packed in my laptop, it has been working out alright so far for me :D | 10:56 |
| josch | JC[m]: a burned trace might mean that two of the adjacent battery clips touched and created a short circuit even while the unit is off and the batteries not connected to the mainboard: https://community.mnt.re/t/isolating-battery-clips/962 -- according to Lukas, the new board design is also different in that the adjacent clips are less prone to accidentally touching each other when changing the cells | 10:59 |
| JC[m] | This just occurred to me: what if we stagger the battery clips in the next rev? They wouldn't come into contact unless the clips are bent diagonally, which is very unlikely. | 11:01 |
| JC[m] | Like 1cm or so (don't have a unit in front of me to check the width) | 11:02 |
| josch | JC[m]: looking at my reform i don't think there is space for shifting them like that | 11:03 |
| JC[m] | Even if rotated 90 degrees? | 11:04 |
| JC[m] | Just brainstorming | 11:04 |
| josch | staggering the batteries like that will create some dead volumes above every even and below every odd battery. i don't see where that free volume is to come from | 11:05 |
| josch | it's not a bad idea but since there are very simple fixes against the problem i think those are easier than "wasting" more volume :) | 11:05 |
| JC[m] | True. I'm still a fan of having socketed fuses as a failsafe... just don't want any lawsuits to happen due to a fire. | 11:09 |
| minute | the lifepo batteries are highly unlikely to burn | 11:12 |
| minute | also there have been some nice 3d printed mods that go in between the clips? | 11:13 |
| minute | funfact: the pcbs have these slots cut into them exactly for this purpose | 11:14 |
| minute | to insert any kind of isolation "sheet" | 11:15 |
| JC[m] | Any plans to have these distributed as OEM in future shipments? | 11:15 |
| JC[m] | The isolators, I mean | 11:15 |
| minute | so far we weren't conviced that they are needed, but we can certainly come up with something | 11:16 |
| JC[m] | @minute did you happen to perform short-circuit tests on various boards between each of the 3 points of concern to demonstrate a consistent result? | 11:20 |
| JC[m] | with mobo connected, battery-only, mains connected, etc | 11:24 |
| minute | which 3 points of concern? | 11:26 |
| minute | the battery boards themselves were not protected against shorts. | 11:27 |
| JC[m] | 3 points of (previous) concern, shown in red in the linked post josch provided above | 11:28 |
| JC[m] | minute: are you referring to the original or revised board, or both? | 11:29 |
| minute | JC[m]: original! | 11:30 |
| minute | i haven't myself managed to have these clips touching when inserting the batteries. | 11:30 |
| JC[m] | Ok | 11:31 |
| minute | hmm, maybe if one inserts two batteries at the same time? | 11:31 |
| JC[m] | My friend shorted the original board by trying to unscrew the battery boards with a metal screwdriver while the batteries were still connected. | 11:31 |
| minute | JC[m]: yes, but this is forbidden in the manual and security sheet | 11:31 |
| minute | safety, not security | 11:32 |
| minute | if you short the batteries directly to each other with a tool, there's nothing a circuit can do | 11:32 |
| minute | the only option here would be to cover the wings of the clips with plastic | 11:33 |
| JC[m] | are you against the idea of socketed fuses as a failsafe? | 11:34 |
| minute | they would not help in this case at all | 11:34 |
| vkoskiv | I thought about dipping/coating the clips with some kind of lacquer, and leaving just the points of battery contact exposed. | 11:34 |
| minute | or, maybe they could | 11:34 |
| vkoskiv | But I haven't had issues with the clips coming into contact. | 11:34 |
| minute | actually they probably could help because the user only shorts one side of the circuit | 11:35 |
| vkoskiv | I could see them possibly touching very briefly if I dropped Reform from some height onto the left or right side, but I tend to avoid dropping my hardware. | 11:35 |
| minute | but, the new battery board should protect too in this case. we'll test it | 11:35 |
| JC[m] | that's what I was wondering | 11:35 |
| JC[m] | Ok sounds great. Thanks for hearing me out. | 11:35 |
| minute | vkoskiv: this could work, if the lacquer does not break up from (re)inserting cells | 11:36 |
| JC[m] | I wonder about plastidipping them | 11:36 |
| vkoskiv | Possibly some kind of flexible coating then | 11:36 |
| minute | yep | 11:36 |
| vkoskiv | lacquers harden, so they would crack/flake off | 11:37 |
| vkoskiv | The surface of the clips also might not be the best for adhesion | 11:37 |
| JC[m] | probably would have to coat it first | 11:37 |
| minute | yeah. ideally the clips would be of a different design | 11:37 |
| sknebel | I've considered putting kapton-tape on the side bits | 11:37 |
| vkoskiv | Just pour in and fill the whole main box full of resin. Problem: solved | 11:38 |
| minute | there are full plastic 4x holders but they were too bulky to fit in the device | 11:38 |
| minute | vkoskiv: lol | 11:38 |
| vkoskiv | Mil-spec Reform | 11:38 |
| vkoskiv | Just need some EMP protection as well, job done | 11:38 |
| JC[m] | the heatsink has things to say about that | 11:38 |
| minute | just mill the case from lead instead of aluminum | 11:38 |
| JC[m] | that'll be fun in customs lol | 11:39 |
| vkoskiv | And then just wash hands after each use? :D | 11:39 |
| vkoskiv | Wonder how much it would cost to mill the case parts out of titanium | 11:39 |
| JC[m] | this laptop may cause cancer in the state of California | 11:40 |
| vkoskiv | They're all solid blocks of aluminium that get milled out, right? | 11:40 |
| vkoskiv | That would be a large volume of titanium :D | 11:40 |
| vkoskiv | I'm happy that I own a titanium PowerBook | 11:40 |
| minute | titanium is possible, just expensive | 11:40 |
| vkoskiv | (The titanium on it is non-structural, just basically decoration for the most part) | 11:40 |
| JC[m] | I had an old Dell XPS M1710 that had an aluminum/magnesium body | 11:41 |
| minute | some NeXT parts were cast magnesium iirc | 11:41 |
| minute | there was this article about setting one on fire | 11:41 |
| JC[m] | but it was also like USD$2400 in ~2005 | 11:41 |
| minute | https://simson.net/ref/1993/cubefire.html | 11:42 |
| JC[m] | Simson is quite the storyteller | 11:43 |
| + oomono (uid328183@id-328183.tinside.irccloud.com) | 11:44 | |
| josch | vkoskiv: haha nice :D There was a point in my life where i filled everything with epoxy to make it last forever. It's such a cool technique to make your electronices super durable. :) | 11:58 |
| flowy | josch: i've been using https://github.com/jernejsk/FFmpeg, merging branches v4l2-request-hwaccel-4.4.1-Nexus-Alpha1 + v4l2-drmprime-v6-4.4.1-Nexus-Alpha1 . i also see there's a 5.1.2 branch there... i would offer to maintain reform packages but perhaps the situation is a little too unstable | 13:04 |
| josch | aha, so there is finally a 5.1.2 branch, nice! | 13:05 |
| flowy | i'd mirror whatever libreelec are doing with their package | 13:06 |
| flowy | https://github.com/LibreELEC/LibreELEC.tv/tree/master/packages/multimedia/ffmpeg | 13:06 |
| josch | at the point where Debian upgraded to ffmpeg 5, that repo only had patches for ffmpeg 4 still | 13:06 |
| flowy | ah | 13:06 |
| josch | flowy: kodi is also still using ffmpeg 4 | 13:07 |
| flowy | well i'll try the 5.1.2 branch | 13:08 |
| josch | flowy: please come back to me with the results of your tests because we can certainly ship a patched ffmpeg in the reform Debian repo! | 13:08 |
| flowy | will do | 13:08 |
| josch | nice :) | 13:08 |
| flowy | interestingly, i also compiled mpv from git yesterday against patched ffmpeg, and the colors are all wrong when using hwaccel and mpv's 'gpu' driver | 13:09 |
| josch | oh no :( | 13:10 |
| flowy | however the 'dmabuf-wayland' vo driver works very well | 13:10 |
| flowy | w/ hwaccel | 13:10 |
| josch | ah okay :) | 13:10 |
| josch | on the hwaccel front it's also noteworthy that linux 6.1-rc7 is already packaged in Debian experimental and together with gstreamer 1.22 this will give hardware accelerated h265 :) | 13:18 |
| flowy | yeah that will be interesting to play with for sure | 13:20 |
| minute | josch: exciting | 13:22 |
| flowy | it's pretty uncanny feeling to play modern FPS titles on the reform, streaming over LAN at butter smooth 1080p 60fps | 13:22 |
| minute | flowy: what do you use for that? | 13:22 |
| flowy | i'm pretty happy finally about moonlight. i was using it before to play slower games like civ6, when i travel | 13:23 |
| minute | ah, moonlight | 13:23 |
| minute | via internet? | 13:23 |
| flowy | but the performance was too garbage for anything >25fps and there was input latency | 13:23 |
| flowy | yeah | 13:23 |
| flowy | but with the EGL renderer working... on LAN... it actually feels local | 13:23 |
| minute | wow, nice | 13:23 |
| flowy | zero perceptible latency, full framerate | 13:23 |
| flowy | really uncanny | 13:23 |
| flowy | i might take a vid | 13:24 |
| minute | cool | 13:24 |
| josch | that is very impressive! | 13:24 |
| minute | please do | 13:24 |
| flowy | and i'll be travelling again this month so will test what it feels like over the internet again | 13:24 |
| flowy | at the very least will be playing rimworld a bunch over the holidays through moonlight heh | 13:25 |
| flowy | in this particular case h265 would be overkill/unnecessary, over LAN | 13:27 |
| flowy | however over the internet... that will be very interesting to try | 13:28 |
| josch | the last time i heard of a software called moonlight it was the foss silverlight implementation which is dead these days | 13:29 |
| flowy | haha yes i kept stumbling over references to that moonlight while working on this yesterday | 13:32 |
| flowy | moonlight is an open source client to nvidia's GFE shield streaming | 13:33 |
| - mlarkin (QUIT: Ping timeout: 260 seconds) (~mlarkin@047-036-074-225.res.spectrum.com) | 13:34 | |
| flowy | you get exceptionally low latency from this method of streaming | 13:34 |
| + mlarkin (~mlarkin@047-036-074-225.res.spectrum.com) | 13:35 | |
| flowy | you turn streaming on in nvidia's drivers, pair a device, and it's basically like VNC for gaming | 13:35 |
| flowy | if everything is tuned correctly it's quite possible to have total transparency in terms of input latency and graphics | 13:36 |
| flowy | i've finally gotten there with the reform | 13:36 |
| flowy | for various security and perhaps even ideological concerns heh i've fully committed a PC to gaming, i don't even use email on it or whatever. with moonlight it can even be headless | 13:40 |
| + bgs (~bgs@212-85-160-171.dynamic.telemach.net) | 14:29 | |
| - bgs (QUIT: Remote host closed the connection) (~bgs@212-85-160-171.dynamic.telemach.net) | 14:30 | |
| flowy | lol i just realised it's possible that using moonlight on my reform might offer better responsiveness than when i use my PC directly, with its current monitor attached | 14:53 |
| flowy | as its 4k monitor has pretty awful latency | 14:53 |
| flowy | don't really have the setup to measure. but it would be interesting | 14:54 |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 15:18 | |
| - oomono (QUIT: Quit: Connection closed for inactivity) (uid328183@id-328183.tinside.irccloud.com) | 15:57 | |
| - GNUmoon (QUIT: Ping timeout: 255 seconds) (~GNUmoon@gateway/tor-sasl/gnumoon) | 15:58 | |
| josch | minute: would you mind if I quote you publicly on that? "also there have been some nice 3d printed mods that go in between the clips? funfact: the pcbs have these slots cut into them exactly for this purpose to insert any kind of isolation "sheet"" | 16:01 |
| josch | (i know the irc records are public but i wanted to make sure anyways :)) | 16:01 |
| josch | i'm currently using heat shrink tubes but they don't look as nice as an isolation sheet | 16:02 |
| josch | unfortunately the 3d printed mods shown in the forum don't work for me because i have too many cables that use up all the space already | 16:02 |
| josch | and i didn't notice these slots in the pcb before but they are *perfect* to put a sheet in without having it move around | 16:03 |
| + GNUmoon (~GNUmoon@gateway/tor-sasl/gnumoon) | 16:09 | |
| - bkeys (QUIT: Ping timeout: 248 seconds) (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 16:24 | |
| minute | josch: sure you can! we never developed those sheets though. i don't think it would be a lot of work though | 16:34 |
| minute | they could also be lasercut from mylar | 16:34 |
| josch | yup, but i don't have a laser cutter so i'm just gonna 3d print some and publish the stl in that forum thread :) | 16:38 |
| minute | nice | 16:49 |
| josch | thank you for the tip with the ridges -- apparently nobody has realized they exist so far XD | 16:49 |
| minute | it appears so :3 | 16:51 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:40) | 17:37 | |
| josch | huh... is that an openscad bug or a sway/reform bug? i'm in openscad and press "enter" in the editor but no newline appears. I can copypaste newlines but nothing happens when i press enter. Can somebody else confirm this? | 18:00 |
| minute | but other letters work? | 18:04 |
| minute | which gui framework is openscad again? qt? | 18:05 |
| minute | yeah, maybe try `export QT_QPA_PLATFORM=xcb` before launching it | 18:05 |
| minute | wayland+qt sometimes have strange bugs | 18:05 |
| + bkeys (~Thunderbi@static-198-54-135-69.cust.tzulo.com) | 18:09 | |
| josch | yes, openscad is qt | 19:41 |
| josch | and yes, other characters work | 19:41 |
| josch | one can also just copy-paste a linebreak and that works too | 19:41 |
| josch | QT_QPA_PLATFORM=xcb doesn't make a difference though | 19:42 |
| josch | can i somehow force an application to use xwayland instead? | 20:02 |
| c-keen[m] | josch yeah that's annoying | 20:15 |
| josch | c-keen[m]: can you confirm the problem? | 20:26 |
| josch | (i'm currently building from git...) | 20:26 |
| c-keen[m] | huh, it's gone. I am pretty sure that I have seen that behaviour before | 20:33 |
| c-keen[m] | josch: atm works as expected for me | 20:33 |
| c-keen[m] | that's 2021.10 from packages no modifications | 20:33 |
| josch | you mean 2021.01-5+b2, right? | 20:34 |
| josch | drats... if it works for you then something else must be broken on my system. :( | 20:34 |
| c-keen[m] | ii openscad 2021.01-5+b2 | 20:38 |
| c-keen[m] | yep | 20:38 |
| c-keen[m] | the only thing I see on the console: GLEW Error: Unknown error | 20:38 |
| c-keen[m] | upon loading a file | 20:38 |
| josch | thank you for checking! | 20:39 |
| c-keen[m] | but rendering works | 20:39 |
| josch | yes, rendering works fine | 20:39 |
| josch | everything seems to work other than pressing enter... | 20:39 |
| c-keen[m] | when I close openscad it coredumps though | 20:44 |
| josch | when i open it it says: | 20:47 |
| josch | qt.qpa.plugin: Could not find the Qt platform plugin "wayland" in "" | 20:47 |
| josch | and yes, segfault when closing | 20:47 |
| minute | josch: btw you had a script ready for installing a rescue system on emmc, yes? | 20:53 |
| vkoskiv | I copied my Reform sway config over to my work linux VM today :3 | 20:54 |
| vkoskiv | It's just so... nice. | 20:54 |
| vkoskiv | Don't know why, but the gnome environment ubuntu comes with is unusably slow under VMWare | 20:54 |
| vkoskiv | Sway works perfectly fine, so I just use that. | 20:54 |
| c-keen[m] | ioctl(21, DRM_IOCTL_ETNAVIV_GEM_SUBMIT or DRM_IOCTL_MSM_GEM_SUBMIT, 0xffffeb7c83d0) = 0 | 20:55 |
| c-keen[m] | ioctl(21, DRM_IOCTL_ETNAVIV_GEM_SUBMIT or DRM_IOCTL_MSM_GEM_SUBMIT, 0xffffeb7c8300) = 0 | 20:55 |
| c-keen[m] | --- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=0x80} --- | 20:55 |
| c-keen[m] | +++ killed by SIGSEGV (core dumped) +++ | 20:55 |
| minute | josch: nevermind, found it (reform-flash-rescue) | 20:55 |
| minute | btw apparently there are some etnaviv fixes on the drm level | 20:56 |
| minute | pending on drm-next | 20:56 |
| c-keen[m] | josh: would need debugging symbols for openscad to further debug this, running it as it is in gdb does not yield anything useful after segfaulting, looks like the stack is corrupted | 20:59 |
| minute | https://cgit.freedesktop.org/drm/drm/commit/?id=a3b4c2f9b03917d5d640bc6e3f42c24695967552 | 21:01 |
| minute | https://cgit.freedesktop.org/drm/drm/commit/?id=819683a1fc2f7e64017d50caf539e7bafcb37b81 | 21:01 |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:40) | 21:02 | |
| minute | josch: is this duplication here on purpose? https://source.mnt.re/reform/reform-tools/-/blob/main/sbin/reform-boot-config#L73 | 21:24 |
| minute | haha i finally migrated my reform to a emmc+nvme setup | 21:29 |
| minute | nice nice | 21:29 |
| c-keen[m] | yes it is! | 21:30 |
| c-keen[m] | godspeed! | 21:31 |
| * tadeus_brick -> TadeusTaD | 21:32 | |
| josch | minute: what is line 73 a duplicate of? | 21:50 |
| minute | josch: nevermind, R vs B | 21:51 |
| minute | josch: in any case, i had to comment out line 75 to make the script work for my case | 21:52 |
| minute | (i was already booted from nvme via sd card and wanted to run it with --emmc nvme, or /dev/mapper/reform_crypt in my case) | 21:53 |
| vkoskiv | For carrying the thing around, it's nice not to need the SD card in there to boot | 21:53 |
| minute | yep. i experienced this the hard way yesterday. | 21:53 |
| minute | (left sd card in the office) | 21:53 |
| minute | (didn't have a working rescue flashed) | 21:53 |
| c-keen[m] | I just wished that the i.mx boot sequence would be more flexible | 21:54 |
| c-keen[m] | but that's not anyone's fault in here | 21:54 |
| josch | i have uboot on sdcard but /boot on emmc | 21:55 |
| josch | that way i can try out sdcard images without needing to flip the dip switch | 21:55 |
| c-keen[m] | but not boot without one | 21:56 |
| josch | correct | 21:56 |
| josch | but i can remove the sd-card and use the sd-slot after i have booted | 21:56 |
| josch | the only other sd-card i use is for my digital camera | 21:57 |
| josch | by also having uboot on that one, i even can boot the reform if i forgot to swap sd-cards back :) | 21:57 |
| - lastebil (QUIT: Ping timeout: 248 seconds) (truck@shell.suomiscene.fi) | 22:01 | |
| c-keen[m] | heh | 22:03 |
| josch | c-keen[m]: you can run gdb with DEBUGINFOD_URLS="https://debuginfod.debian.net" to avoid having to install all the dbgsym packages | 22:06 |
| josch | Enable debuginfod for this session? (y or [n]) y | 22:06 |
| josch | Debuginfod has been enabled. | 22:06 |
| josch | that way, gdb will download the correct debug info from debuginfod.d.n | 22:06 |
| josch | minute: my openscad problem is fixed by installing the qtwayland5 package | 22:09 |
| josch | c-keen[m]: since you did not see the problem i guess you have this package installed? | 22:10 |
| vkoskiv | It's been ~6 months since I received my Reform, and I *still* haven't put in the serial # sticker | 22:35 |
| vkoskiv | I'm worried I'll place it in crooked and I'll get annoyed by that forever :D | 22:35 |
| sigrid | I lost mine | 22:35 |
| vkoskiv | Mine is safely stuck between the pages of my Reform handbook | 22:36 |
| josch | ACTION hasn't put the reform sticker either XD | 22:43 |
| vkoskiv | I feel like I need a 3D printed guide to get it in the "official" spot | 22:45 |
| josch | should you make one, please share your stl XD | 22:50 |
| minute | josch: ha, weird, so the problem was that it was actually falling back to x11/xcb and ~not~ using wayland | 22:56 |
| sknebel | Did the ship of Theseus have a serial number? | 22:57 |
| josch | minute: possibly, is there an easy way to find out whether a window uses one or the other? | 22:58 |
| josch | vkoskiv: i feel like attaching a LEGO sticker -- you only have one chance and it feels like you'll botch it on the only try you have :D | 22:59 |
| josch | though the sources of the sticker are probably in a git somewhere so that i can print myself a new one, right? | 22:59 |
| flowy | if anyone is curious to see Stray being played on the reform: https://static.ckut.ca/~warp/output_1704.mkv | 22:59 |
| flowy | ^^ (60fps h264) | 23:00 |
| josch | this is nuts XD | 23:02 |
| josch | it's awesome that you filmed your mouse and keyboard | 23:02 |
| josch | the delay seems very tiny to non-existant | 23:02 |
| vkoskiv | Seeing Windows on the Reform display just feels.. Wrong :D | 23:02 |
| minute | josch: pkill Xwayland will take the window down ;) | 23:02 |
| vkoskiv | I wonder if this setup would work with a Linux host with Proton | 23:02 |
| josch | minute: thanks, works for me :) | 23:03 |
| vkoskiv | I rarely play games, but I do sometimes play Minecraft, and recently My Summer Car | 23:03 |
| flowy | yea i am pretty sensitive to latency and i don't notice any | 23:03 |
| + vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:20) | 23:03 | |
| vkoskiv | Reminds me, I really want a clicky keyboard on my Reform now :D | 23:03 |
| flowy | no, it wouldn't work with a linux host... i don't think geforce experience runs on linux | 23:04 |
| vkoskiv | Ah, gotcha | 23:04 |
| flowy | i don't think there's an open source solution for the host unfortunately | 23:04 |
| vkoskiv | I don't have Windows on any of my systems, so I guess I'll have to skip on this for now | 23:04 |
| vkoskiv | This is over the local network, right? | 23:05 |
| flowy | oh wait what's sunshine | 23:05 |
| flowy | uh nevermind maybe it is possible | 23:05 |
| flowy | yeah | 23:05 |
| flowy | check out sunshine | 23:05 |
| vkoskiv | I spot the OG iFixit screwdriver in the background :D | 23:06 |
| vkoskiv | I still have my set, I love that thing. | 23:06 |
| flowy | heh yeah i had to clean up my desk to record this and figured i'd leave that for you :p | 23:06 |
| bkeys | minute: I was looking at your mastodon and it seems that the HDMI-eDP adapter had some progress? | 23:10 |
| minute | bkeys: yes, the design is almost done | 23:14 |
| bkeys | That is great to hear, the cm4 adapter is gonna be big for me | 23:14 |
| bkeys | Especially since Fedora has official support for the cm4 | 23:14 |
| minute | cool | 23:14 |
| minute | yeah i'm pretty interested in the 4x CA72, 8GB RAM and vulkan | 23:15 |
| bkeys | Yeah it opens a lot of possibilities for sure; I wonder if you'll be able to use the UEFI implementation they have for the rpi | 23:16 |
| minute | bkeys: possibly | 23:16 |
| bkeys | https://github.com/pftf/RPi4 | 23:16 |
| bkeys | Do you forsee any other hardware challenges? I know the HDMI->eDP adapter wasn't expected at first | 23:17 |
| minute | i haven't brought up sound yet but that shouldn't be a big problem | 23:17 |
| minute | as there are pi hats with the same sound chip | 23:18 |
| minute | also, you should probably use a wifi version of cm4 as there is only one pcie lane and it goes to the nvme slot, so mpcie slot will be nonfunctional | 23:18 |
| bkeys | That's a good point; does the wifi on the cm4 use nonfree software? I am not too familiar | 23:19 |
| minute | no idea | 23:22 |
| minute | as it is broadcom i would think a fw blob is needed | 23:23 |
| minute | as far as i can see it is bcm4345 | 23:25 |
| JC[m] | <flowy> "if anyone is curious to see..." <- Nice! This may be good material to link in the gaming thread :) | 23:27 |
| minute | strange, vlc has a lot of issues for me playing this video | 23:29 |
| minute | clapper can do it though | 23:30 |
| flowy | JC[m]: yeah! i'll do a community post for the recipe | 23:32 |
| minute | flowy: fantastic video btw | 23:33 |
| flowy | minute: yea i haven't found a player that can decode 60fps in software on reform yet | 23:33 |
| minute | flowy: yeah, i didn't have a perf problem, vlc just had some issue with the data in the file and was showing some garbage | 23:33 |
| flowy | tx | 23:34 |
| flowy | oh weird | 23:34 |
| minute | i see that it was recorded with iPhone, maybe they do something special | 23:34 |
| minute | or you probably converted it? | 23:34 |
| flowy | i transcoded it yeah, ffmpeg passed the metadata through | 23:34 |
| minute | in vlc i got tons of: | 23:34 |
| minute | [h264 @ 0xffff20db9e90] mmco: unref short failure | 23:34 |
| minute | [h264 @ 0xffff20dbc120] co located POCs unavailable | 23:34 |
| flowy | it might just not like being starved | 23:36 |
| flowy | dunno | 23:36 |
| minute | ah hm! | 23:43 |
| - cwebber (QUIT: Ping timeout: 246 seconds) (~user@user/cwebber) | 23:45 | |
| minute | TIL that foot support hyperlinks https://codeberg.org/dnkl/foot#urls | 23:53 |
| minute | btw if you don't use foot, you miss out on an extremely fast terminal on reform | 23:53 |
| minute | (much faster than the default xfce4-terminal) | 23:54 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!