| - stites (QUIT: Ping timeout: 272 seconds) (~stites@2603:3005:b69:4100:ffbe:3c4e:c9c0:ad22) | 00:00 | |
| + stites (~stites@2607:fb91:dcd:e247:3117:f970:e6:e12c) | 00:01 | |
| - jacobk (QUIT: Ping timeout: 255 seconds) (~quassel@utdpat241106.utdallas.edu) | 00:05 | |
| josch | abortretryfail: can i help you with your u-boot-menu problem? | 00:07 |
|---|---|---|
| abortretryfail | josch: maybe? I removed it and the problem went away. | 00:10 |
| + jacobk (~quassel@utdpat241106.utdallas.edu) | 00:11 | |
| vkoskiv | So I clone linux, checkout roughly the same commit as what's on rescue sd, copy in the config, then build it | 00:12 |
| minute | vkoskiv: yup, with cross compile env vars | 00:13 |
| josch | abortretryfail: can you post your /boot/extlinux/extlinux.conf? | 00:14 |
| josch | (maybe you already did and i missed it in the backlog) | 00:14 |
| vkoskiv | the build script in reform-debian-packages seems to just pull in master | 00:14 |
| minute | vkoskiv: hmm i don't think so... it uses debian's kernel version | 00:14 |
| vkoskiv | I guess tag 6.5 is close enough? | 00:15 |
| minute | for this purpose probably | 00:15 |
| abortretryfail | josch: yeah just a sec | 00:16 |
| - stites (QUIT: Read error: Connection reset by peer) (~stites@2607:fb91:dcd:e247:3117:f970:e6:e12c) | 00:18 | |
| abortretryfail | ttps://paste.debian.net/hidden/27c74584/ | 00:18 |
| + stites (~stites@130.44.147.204) | 00:18 | |
| abortretryfail | worth noting i didn't write that. something in the debian package makes it. | 00:19 |
| josch | abortretryfail: ah that should be an easy fix -- what is in your /etc/default/u-boot? | 00:19 |
| josch | abortretryfail: yes, u-boot-menu writes this | 00:19 |
| josch | and it reads /etc/default/u-boot for the cmdline | 00:19 |
| josch | abortretryfail: are you on imx8mq? | 00:19 |
| abortretryfail | right now, nothing. I uninstalled the package because I don't want it making my laptop randomly unbootable after updates. :) | 00:19 |
| abortretryfail | yeah, imx8mq | 00:20 |
| josch | abortretryfail: writing /etc/default/u-boot will prevent something from breaking after updates because reform-tools does not mess with /etc | 00:20 |
| josch | putting this in that file should make it work again: | 00:20 |
| josch | U_BOOT_PARAMETERS="ro no_console_suspend cma=512M@3G cryptomgr.notests console=ttymxc0,115200 pci=nomsi console=tty1" | 00:20 |
| josch | you could do that, install u-boot-menu again and then look into your /boot/extlinux/extlinux.conf to see if the right entries are set in the "append" lines | 00:21 |
| abortretryfail | i guess the part i'm confused about here is how all the different parts get applied for u-boot to figure out the kernel cmdline | 00:21 |
| abortretryfail | does extlinux.conf override boot.scr? | 00:21 |
| vkoskiv | Hmm, some patches don't apply. meson/g12a.{c,h} at least | 00:22 |
| josch | abortretryfail: no, boot.scr is created by flash-kernel | 00:23 |
| josch | abortretryfail: extlinux.conf is created by u-boot-menu | 00:23 |
| abortretryfail | but u-boot reads both of them | 00:23 |
| josch | you can create both manually but that sucks because you'd have to manually re-do that on every kernel upgrade | 00:23 |
| josch | abortretryfail: u-boot tries reading both of them | 00:24 |
| josch | it first triest extlinux.conf and if that does not exist tries boot.scr | 00:24 |
| abortretryfail | so yes, extlinux.conf overrides boot.scr | 00:24 |
| vkoskiv | I'm guessing I probably need to clone a debian kernel, instead of the mainline one? | 00:24 |
| vkoskiv | Mainline doesn't even have a 6.5.3 tag | 00:24 |
| abortretryfail | so whatever script logic in boot.scr is trying to create the right kernel cmdline never happens, explains what i was seeing earlier. | 00:25 |
| - aliosablack (QUIT: Ping timeout: 240 seconds) (~chomwitt@2a02:587:7a24:b000:1ac0:4dff:fedb:a3f1) | 00:25 | |
| josch | vkoskiv: mainline doesn't have 6.5.3 because that's a tag from the stable kernel | 00:25 |
| vkoskiv | Ah, got it. And is 6.5.3-1 == 6.5.31? | 00:27 |
| vkoskiv | I misread, there's just 6.5.3 here. | 00:29 |
| josch | ye | 00:29 |
| vkoskiv | I've never actually applied patches in a plain tarball, the obvious for patch in $(find patches); do patch < $patch; done didn't work | 00:32 |
| josch | it has to be patch -p0 | 00:33 |
| josch | err... -p1 | 00:34 |
| josch | you are in an extracted tarball and not git? | 00:34 |
| vkoskiv | Yeah | 00:34 |
| vkoskiv | Probably a good idea to clone, but I just want to see if I can even build the thing | 00:34 |
| josch | then it should work with patch -p1 | 00:34 |
| josch | the patches are all git format-patch format | 00:35 |
| vkoskiv | some patches in drivers/clk/meson/g12a.c didn't apply, but as long as the rejected ones aren't in the ti module I'm looking to modify, it should be fine? | 00:36 |
| josch | vkoskiv: then you should try to rebase the patches -- i'm doing that all the time when new kernels get uploaded to debian | 00:39 |
| josch | what are you trying to do? | 00:39 |
| josch | i only skimmed the backlog | 00:39 |
| vkoskiv | I want to set this up so I can modify the ti_sn65dsi86 kernel module | 00:40 |
| vkoskiv | https://mister-muffin.de/p/mIU-.txt | 00:40 |
| vkoskiv | There are those errors that happen when I bring my A311D from suspend | 00:40 |
| vkoskiv | Just want to see if I can build a .ko that does something more interesting than that | 00:41 |
| josch | ooh | 00:41 |
| vkoskiv | I've got UART hooked up and everything | 00:41 |
| josch | yes, building a custom kernel and not using reform-debian-package is definitely faster when you do builds locally | 00:41 |
| vkoskiv | I'm pretending to be a hardware person, you see :D | 00:41 |
| vkoskiv | I'm actually building on my desktop | 00:41 |
| vkoskiv | I intend to scp the .ko over and modprobe | 00:41 |
| abortretryfail | i wonder if there is special things like whats in /usr/sbin/reform-standby | 00:42 |
| vkoskiv | abortretryfail: I haven't modified it yet, but I know that script exits if it's not running on imx8 | 00:42 |
| abortretryfail | ok cool :) | 00:42 |
| vkoskiv | It kicked me into config, do I just spam return or should I pay attention here? | 00:43 |
| vkoskiv | I copied .config from the running reform from under /boot | 00:43 |
| vkoskiv | There are some new options I guess | 00:43 |
| vkoskiv | Hmm, I do wonder what a reasonable clocksource watchdog maximum allowable skew is :D | 00:43 |
| vkoskiv | Okay, this one I know. "Enable PC-Speaker support". I'll say no. | 00:44 |
| vkoskiv | (That's a NEW option? :D) | 00:44 |
| vkoskiv | Do some people actually sit through and answer all of these? There has to be over 100k switches in here | 00:46 |
| vkoskiv | It's compiling now, I just held down return, hoping for the best :D | 00:47 |
| vkoskiv | minute: forgot to comment earlier re: qwerty-se: It looks very reasonable. You can also sell it as QWERTY-FI (we use the same keyboard layout, but don't tell the Swedes) | 00:48 |
| abortretryfail | vkoskiv: i used to do that back in the Slackware days, 20+ years ago. | 00:50 |
| vkoskiv | Oof, why is it building x86 object files if I'm using a cross-compiler :( | 00:51 |
| vkoskiv | Did the compiler vendor sell me a fake compiler (0€) | 00:51 |
| minute | vkoskiv: export ARCH=arm64? | 00:54 |
| josch | vkoskiv: you should ask for your money back! | 00:55 |
| vkoskiv | I forgot to do that, it seems | 00:55 |
| josch | vkoskiv: much success with all that :) | 00:55 |
| vkoskiv | Hm, the toolchain I found from AUR isn't up to the task it seems. The build went right for a bit, but it blew up right away | 00:56 |
| vkoskiv | unrecognized '-mlittle-endian' | 00:56 |
| vkoskiv | But now it certainly was building arm-y things instead of a bunch of x86 stuff | 00:57 |
| vkoskiv | exporting CROSS_COMPILE too, let's see if that sticks | 00:59 |
| abortretryfail | maybe one of those situations where the slow little ARM box would've finished compiling by the time the cross-compiler gets sorted out. :) | 00:59 |
| vkoskiv | Yeah. Now it's saying there is no rule to make the target for our bananapi dtb | 00:59 |
| vkoskiv | So that's probably not a cross-compiler thing maybe | 00:59 |
| vkoskiv | Let me wipe the config once more | 01:00 |
| vkoskiv | Oh wait, this is probably those failed patches | 01:00 |
| vkoskiv | meson-g12b | 01:01 |
| minute | vkoskiv: did you copy the dts in? | 01:14 |
| vkoskiv | I did not, it seems | 01:14 |
| vkoskiv | I though the patches would add 'em | 01:14 |
| minute | needs to go to boot/arch/arm64/dts/amlogic | 01:15 |
| vkoskiv | Yeah, I see it now | 01:15 |
| minute | sorry, forgot to mention that earlier | 01:15 |
| vkoskiv | It's fine, I was just about to figure it out as well :D | 01:16 |
| vkoskiv | "Hmm, the file is missing..." | 01:16 |
| minute | btw https://community.mnt.re/t/mnt-reform-system-image-v4/1671/8 | 01:16 |
| vkoskiv | 'unknown type name u128' :D | 01:17 |
| minute | vkoskiv: what's going on there | 01:18 |
| minute | vkoskiv: what compiler are you using? | 01:18 |
| vkoskiv | arm-none-linux-gnueabihf-gcc | 01:18 |
| vkoskiv | I found it on AUR | 01:18 |
| vkoskiv | It is of questionable origin. | 01:18 |
| minute | that's no good | 01:18 |
| minute | that's for 32 bit | 01:18 |
| vkoskiv | Ah, figures. | 01:19 |
| minute | ye olde | 01:19 |
| minute | how about gcc-aarch64-linux-gnu | 01:20 |
| minute | or probably just called aarch64-gcc in arch land? | 01:21 |
| vkoskiv | Ah, and I even had it installed this whole time, just couldn't find the tab complete | 01:22 |
| vkoskiv | *facepalm* | 01:22 |
| vkoskiv | Easy fix though | 01:22 |
| vkoskiv | It's going again | 01:23 |
| vkoskiv | Now it's happy, I think. device trees compiled, it's building objects now | 01:23 |
| minute | nice | 01:24 |
| minute | you mentioned some patches did not apply? which ones? | 01:24 |
| vkoskiv | > find . -name '*.rej' | 01:24 |
| vkoskiv | ./drivers/clk/meson/g12a.c.rej | 01:24 |
| vkoskiv | ./drivers/clk/meson/g12a.h.rej | 01:24 |
| minute | ok, probably not too bad | 01:24 |
| vkoskiv | I tried to look at those, didn' really grok what I was looking at, or how to resolve | 01:24 |
| minute | my guess is those have maybe been mainlined | 01:25 |
| vkoskiv | Meanwhile, I can take a closer look at the speaker cables to see if those need work | 01:25 |
| minute | vkoskiv: did you already take a look at all those bars in alsamixer? | 01:25 |
| vkoskiv | I did, yeah, nothing obvious popped out | 01:27 |
| vkoskiv | v3 kbd uses hyper+up/down for pgup/pgdown, right? | 01:31 |
| vkoskiv | I really grew to like the dedicated pgup/pgdn on v2 | 01:31 |
| minute | yep | 01:31 |
| minute | they've been sacrificed on the altar of tradition | 01:32 |
| minute | although you could potentially remap AGR and menu key to them | 01:32 |
| vkoskiv | Yeah, are the 1u caps same size as on V2? | 01:32 |
| minute | yep | 01:32 |
| vkoskiv | Could even replace caps then | 01:32 |
| minute | 1u and 1.5u are the same, there are just a few additional caps now | 01:33 |
| minute | 1.25, 1.75 and a 2u | 01:33 |
| vkoskiv | I kind of want a clicky one, so that'd be the time for a v3 too | 01:33 |
| minute | i've been typing on the clicky one for a few days at work now | 01:33 |
| minute | it's fun, except if i'm stressed, then the noise adds to the stress lol | 01:34 |
| vkoskiv | Heh, yeah. I think it'd be less than ideal in some settings | 01:34 |
| vkoskiv | I'd also be interested in swapping the keyswitches in my trackball to whites | 01:35 |
| minute | we always ship the TB with whites now | 01:35 |
| minute | sometimes i think that might be too aggressive but.... so far no complaints | 01:35 |
| vkoskiv | Ah, well that would explain it | 01:47 |
| vkoskiv | One of the wires on the left speaker just popped off when I pulled it back :D :D | 01:47 |
| vkoskiv | "Hmm yes, the wire here is made out of air..." | 01:47 |
| vkoskiv | Good thing I've got my soldering iron handy | 01:48 |
| minute | oh! | 01:50 |
| vkoskiv | Quick dab of solder, it's fixed! | 01:54 |
| vkoskiv | The left speaker is also the one I peeled back some months ago to take a closer look at the speaker unit out of curiosity. | 01:54 |
| vkoskiv | So probably just one too many twists. It's now solidly connected. | 01:55 |
| vkoskiv | minute: Hope you have a relaxing weekend! Someone on the forum mentioned they got furnace to work on their A311D reform | 01:57 |
| vkoskiv | I hadn't heard of it, but it looks like an interesting program | 01:57 |
| minute | thank you, same to you! n8n8 | 02:01 |
| - mtm (QUIT: Ping timeout: 255 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 02:04 | |
| - mjw (QUIT: Ping timeout: 272 seconds) (~mjw@gnu.wildebeest.org) | 03:35 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 04:10 | |
| - vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:20) | 05:12 | |
| - jacobk (QUIT: Ping timeout: 248 seconds) (~quassel@utdpat241106.utdallas.edu) | 05:57 | |
| - mtm (QUIT: Ping timeout: 255 seconds) (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 06:16 | |
| + jacobk (~quassel@64.189.201.150) | 06:27 | |
| josch | minute: about your "no complaints" from yesterday: i actually prefer the brown switches on my trackball v1 over the white ones of v2 :) | 06:56 |
| pandora | anyone knows whether there is a working matrix bridge to this room again? i kinda miss all the messages when i am not online | 07:35 |
| josch | pandora: I think Boostisbetter is using cheogram and might be able to tell you more | 07:37 |
| Boostisbetter | pandora: I am using cheogram to bridge xmpp to irc. There was a matrix to irc bridge, BUT that was taken down because of some issues. It has not, as of yet, been restored. | 07:44 |
| Boostisbetter | my own experience with the matrix bridge was that it was kind of spotty. Cheogram for bridging xmpp is pretty good, although sometimes libera chat will block it and i'll have to cycle the xmpp account being enabled to get it to try again. | 07:48 |
| - jacobk (QUIT: Ping timeout: 260 seconds) (~quassel@64.189.201.150) | 07:51 | |
| pandora | maybe i should just setup an irc bouncer :D | 08:01 |
| Boostisbetter | If you could extend that to me on my Matrix account that would be awesome. | 08:07 |
| Boostisbetter | By bouncer you just mean one way traffic right? | 08:08 |
| + aliosablack (~chomwitt@2a02:587:7a24:b000:1ac0:4dff:fedb:a3f1) | 08:54 | |
| - Boostisbetter (QUIT: Ping timeout: 240 seconds) (4a410829d7@irc.cheogram.com) | 09:59 | |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 10:03 | |
| Boostisbetter | minute: btw, regarding my left speaker, the cable came undone again. I have resoldered it, but I am thinking the length of the wire is the real issue. Is this something that happens for your own Reform? Any tips on preventing the constant need of resoldering the wire? (The white wire is the one that is kind of short.) Don't worry I see this as a normal thing and I don't consider it a design or qu | 10:53 |
| Boostisbetter | ality issue. | 10:53 |
| Boostisbetter | also it is Saturday so I have no expectation of a response today. | 10:53 |
| - aliosablack (QUIT: Ping timeout: 252 seconds) (~chomwitt@2a02:587:7a24:b000:1ac0:4dff:fedb:a3f1) | 11:12 | |
| ec0 | pandora: I use heisenbridge to connect IRC and my Matrix homeserver, it might be useful for you if you run your own | 12:16 |
| minute | pandora: we also have logs, there's a link in the topic. | 12:27 |
| minute | Boostisbetter: sorry to here. maybe it needs some strain relief? it doesn't happen to me. although i have a mod with bigger speakers living in the main body. | 12:28 |
| josch | the question about boot order on ls1028a was easily answered as minute already took care to instruct it to do the right thing: https://source.mnt.re/reform/reform-ls1028a-uboot/-/blob/main/patches/0001-tweak-boot-commands.patch :) | 12:35 |
| pandora | @Boostisbetter I was briefly looking into psyBNC as a bouncer. Maybe I set one up and make it somewhat public so other people could use it too | 12:46 |
| pandora | @minute yes that is how i stay up to date rn :D | 12:47 |
| josch | pandora: depending on the IRC client, prefixing a nick with another character might not create a highlight | 13:17 |
| + aliosablack (~chomwitt@2a02:587:7a24:b000:1ac0:4dff:fedb:a3f1) | 13:21 | |
| pandora | oh yeah ... using matrix too often | 13:26 |
| + mjw (~mjw@gnu.wildebeest.org) | 13:30 | |
| josch | minute: as you know, I'm maintaining a fork of reform-debian-packages and reform-system-image on reform.debian.net to provide the same content for stable releases. I'm now wondering and wanted to agree with you on what you think should happen when somebody who downloaded a system image from reform.debian.net boots that and then runs "reform-flash-rescue"? Should that download the rescue image from MNT | 13:33 |
| josch | or should it download the rescue image from reform.d.n? What do you think? | 13:33 |
| minute | josch: i think it should download the image from debian net | 13:34 |
| josch | that's also my gut feeling | 13:34 |
| josch | minute: with that in mind, would you object if reform-flash-rescue in the official MNT repos would check (for example by investigating /etc/apt/sources.list.d/reform_bookworm.sources) whether it's running a Debian stable system (the one from reform.d.n) or not and then choose where to download the rescue image from depending on that? | 13:37 |
| minute | josch: sounds good to me | 13:37 |
| minute | josch: i would say that it should print the download URL and let the user confirm it | 13:38 |
| minute | so they know what they're getting | 13:38 |
| minute | or, no extra confirmation, but show the url in any case | 13:38 |
| josch | i could make it do a confirmation only if it thinks that it's running stable | 13:39 |
| josch | that way there is just auto-downloading on the MNT images and confirmation for the stable reform.d.n images | 13:39 |
| josch | in any case, yes, having some confirmation sounds like a very good idea | 13:39 |
| minute | ok | 13:40 |
| josch | there will ne another larg-ish diff to review soon (i'm at 284 insertions(+), 136 deletions(-) right now) to make reform-tools be able to cope with a311d and ls1028a | 13:41 |
| josch | in that process i'm also updating reform.d.n to build stable images for ls1028a which should work fine even with the older kernel | 13:41 |
| josch | for a311d i'll have to wait for a backport of kernel 6.5 to stable | 13:41 |
| + jacobk (~quassel@64.189.201.150) | 14:14 | |
| Boostisbetter | excellent work! | 14:25 |
| + IchikaZou (~IchikaZou@36-231-76-229.dynamic-ip.hinet.net) | 14:58 | |
| - jacobk (QUIT: Ping timeout: 240 seconds) (~quassel@64.189.201.150) | 15:04 | |
| + mtm (~mtm@c-71-228-84-213.hsd1.fl.comcast.net) | 15:13 | |
| vkoskiv | minute: Is the pwmval (OCR0A) on the kbd limited to between 0-10 for hardware reasons? To keep things from breaking? | 15:45 |
| vkoskiv | I'm trying to debug why the keyboard brightness control tends to be a bit jumpy | 15:46 |
| vkoskiv | I think I under/overflowed pwmval with my changes, and the keyboard was much brighter than it usually gets briefly, just wondering if that has a chance of breaking things | 15:46 |
| Zaba | might be that the current consumption can get pretty high | 15:51 |
| vkoskiv | Yeah | 15:53 |
| vkoskiv | In the mcu, the pwmval does get adjusted in equal steps, but it feels like the ti backlight chip doesn't always get the new value right away, so there is a big jump in brightness | 15:53 |
| vkoskiv | And then stepping back down, the step isn't nearly as big | 15:53 |
| minute | vkoskiv: there was an issue in the beginning where the full brightness would cause it to flicker, maybe this is not the case anymore? | 16:15 |
| vkoskiv | I did observe it blink very harshly in that state | 16:16 |
| vkoskiv | Not flicker, blink | 16:16 |
| minute | vkoskiv: and you're right, there has always been some weirdness in the steps... would be cool if you figured sth out | 16:16 |
| minute | vkoskiv: yep, it blinks. the current is too high then | 16:16 |
| minute | vkoskiv: i think the blinking is a reset loop | 16:16 |
| vkoskiv | Yeah, investigating. I verified that the pwmval itself isn't the issue | 16:16 |
| vkoskiv | So basically I'm unlikely to break things if I set the brightness pwm to 0xff? | 16:17 |
| abortretryfail | Trip report: I got my friend's Wacom Cintiq 22HD to work on the Reform via HDMI. Everything just works out of the box. :D | 16:22 |
| abortretryfail | She pre-ordered a pocket and wanted to know if it'd be able to drive the tablet display, so maybe? | 16:22 |
| minute | vkoskiv: i don't think it'll break. | 16:37 |
| - stites (QUIT: Ping timeout: 255 seconds) (~stites@130.44.147.204) | 16:37 | |
| minute | abortretryfail: wow, that's cool! and yes, the pocket has good dual screen support | 16:37 |
| minute | with the defaut imx8mplus at least. the gpu is weaker than that of the imx8mq though | 16:38 |
| minute | the cpu speed makes mostly up for it | 16:38 |
| minute | i'm looking forward to the day when we have one chip ticking all the boxes | 16:38 |
| + stites (~stites@2607:fb91:dc9:51d1:1285:280d:59e4:7da2) | 16:38 | |
| abortretryfail | Tbh, it's probably enough to run something like inkscape or gimp. | 16:44 |
| Boostisbetter | gimp runs on the Reform, so I would imagine the Pocket having any problems with it. | 16:49 |
| abortretryfail | yeah, that's what we were testing with. Even the pressure sensitivity worked. I'm pretty impressed | 16:52 |
| - IchikaZou (QUIT: Remote host closed the connection) (~IchikaZou@36-231-76-229.dynamic-ip.hinet.net) | 16:55 | |
| - dodo (QUIT: Quit: dodo) (~dodo@user/dodo) | 17:06 | |
| + dodo (~dodo@user/dodo) | 17:07 | |
| Boostisbetter | that's awesome. | 17:24 |
| Boostisbetter | Ever since I got the Reform I have been impressed with just how capable the SoC is. I mean it feels like a Pentium class cpu. | 17:24 |
| Boostisbetter | ARM has come such a long way | 17:25 |
| vkoskiv | I think I did manage to bust the backlight partially :D | 17:39 |
| vkoskiv | Going any higher than 10 on the pwmval seems to do it. | 17:39 |
| vkoskiv | Nope, not busted. Just temporary, it seems. | 17:40 |
| vkoskiv | Power-cycling the keyboard resolves it | 17:40 |
| vkoskiv | Perhaps reading the docs on the TI chip properly is in order, before I venture further :D | 17:43 |
| abortretryfail | Boostisbetter: you never used a "Pentium class" CPU did you? This is so much faster. | 17:44 |
| vkoskiv | I own pentium machines. This is indeed far faster :D | 17:50 |
| vkoskiv | RPi1 in 2011 was roughly equivalent to a Pentium in the early 2000s | 17:50 |
| vkoskiv | Interesting fault mode on the TI chip, I managed to put pins 15 and 7, by the looks of it, stuck to low | 17:51 |
| vkoskiv | Rest of the keys still lit up, but most of the top row, and bottom row were dark | 17:51 |
| - xktr (QUIT: Quit: leaving) (~xktr@2602:fe3d:c01:10ca:1050:1ace:0:b) | 17:52 | |
| vkoskiv | Still takes up a lot of memory, but Discord is also quite usable on A311D | 17:55 |
| vkoskiv | The transition animations are a slideshow, but they are that on my desktop too. | 17:55 |
| vkoskiv | I'm running kicad and firefox with a bunch of tabs too | 17:56 |
| + xktr (~xktr@2602:fe3d:c01:10ca:1050:1ace:0:b) | 17:56 | |
| abortretryfail | Discord works on ARM? | 17:57 |
| vkoskiv | In the browser. | 17:57 |
| abortretryfail | ahh ok | 17:57 |
| vkoskiv | But the actual client is just an extra browser anyway | 17:57 |
| abortretryfail | Yeah the client is just Electron | 17:57 |
| Boostisbetter | exactly. You could make it an app though. I think it has a thing for that if you are on a chromium based browser. | 17:58 |
| Boostisbetter | Actually strike that. I don't think it will let you. | 17:58 |
| Boostisbetter | But it runs perfectly in the browser. | 17:58 |
| Boostisbetter | Although I really can't figure out what the appeal is with Discord | 17:58 |
| abortretryfail | lots of emojis everywhere. :) | 17:59 |
| sigrid | (not an ad) I used to run https://cancel.fm/ripcord/ when I needed a "lighter" slack on reform. though it's mostly for discord | 18:00 |
| sigrid | ran it via one of those box* emulators | 18:01 |
| vkoskiv | I love ripcord, I bought a license a year or two ago | 18:02 |
| sigrid | it was still much faster than whatever they come up with for a browser | 18:02 |
| vkoskiv | I'd just love to see a native aarch64 binary | 18:02 |
| sigrid | yeah, I asked the author about it but I think he's too busy with other stuff | 18:02 |
| vkoskiv | Yeah, I'd imagine it is a lot of work to maintain to begin with | 18:02 |
| abortretryfail | Didn't Discord start banning people using 3rd party clients? | 18:09 |
| vkoskiv | I heard that, yes, but I don't know how they go about it | 18:09 |
| vkoskiv | Since ripcord just uses the session token from the official client | 18:09 |
| vkoskiv | I guess they added some challenge response thing to their client, perhaps | 18:10 |
| sigrid | I vaguely recall discord being ok with ripcord specifically | 18:12 |
| sigrid | as in, the staff of discord being ok | 18:12 |
| sigrid | but also, like, I don't use discord. at all. so idk | 18:13 |
| vkoskiv | Yeah, if anyone wants to expell 3rd party client users, it's the people in suits with the money. | 18:13 |
| vkoskiv | "Oh, you people like our thing so much that you made a thing to use it? OUT!" | 18:13 |
| vkoskiv | Or I guess it's the other way, they disliked it enough to go write a better client :D | 18:14 |
| vkoskiv | But did like it enough for it to be worth the effort | 18:14 |
| vkoskiv | I'm happy I can just retreat to my corner and pretend Electron does not exist | 18:14 |
| - abortretryfail (QUIT: Ping timeout: 255 seconds) (~arf@146.ip-149-56-132.net) | 18:16 | |
| + jacobk (~quassel@2603-8080-b200-7b02-0000-0000-0000-0b71.res6.spectrum.com) | 18:36 | |
| Boostisbetter | I actually just fired Ripcord up, and I am very impressed with how fast and lightweight yet full featured it is. Kind of impressive that one person made something like this. | 18:40 |
| Boostisbetter | I use discord begrudingly. I really don't see the appeal at all. I mean XMPP and Matrix do everything I really need. | 18:40 |
| jfred | sigrid: How'd you get ripcord running on reform? Tried running it with box64 and it's failing to start for me due to missing qt wayland libraries, I wonder if I need the x86-64 version of whichever libs it needs | 18:48 |
| Boostisbetter | Yeah I didn't see it for ARM at all, so I was going to ask that myself. | 18:48 |
| sigrid | jfred: yes, iirc I went with installing x86-64 libs | 18:51 |
| sigrid | I used it with diy void-musl aarch64 installation too, so even in this case it was doable | 18:52 |
| - jacobk (QUIT: Ping timeout: 260 seconds) (~quassel@2603-8080-b200-7b02-0000-0000-0000-0b71.res6.spectrum.com) | 18:52 | |
| jfred | Hm, I'm on the stock reform debian image so might be trickier. (Then again, maybe x86-64/aarch64 multiarch works okay? I've admittedly never tried!) | 18:53 |
| + jacobk (~quassel@2603-8080-b200-7b02-0000-0000-0000-0b71.res6.spectrum.com) | 18:56 | |
| jfred | But yeah, I've got one friend who insists on using Discord and it doesn't run well on the Reform so Ripcord looks appealing | 18:59 |
| jfred | mostly otherwise I use Matrix via Nheko | 18:59 |
| Boostisbetter | Yeah Nheko is good, but one client that has really impressed me is Hydrogen. hydrogen.element.io | 19:04 |
| jfred | hm, got qt6-wayland:amd64 and qtwayland5:amd64 installed but no luck | 19:04 |
| Boostisbetter | Super light weight and really fast. | 19:04 |
| Boostisbetter | And you can make an app out of it. | 19:04 |
| jfred | Hydrogen is neat but I often rely on the tray icon for my chat apps, and it doesn't have that | 19:05 |
| vkoskiv | I've made some improvement on the keyboard backlight jumpiness when adjusting | 19:05 |
| vkoskiv | Now trying to understand why this works :D | 19:05 |
| jfred | oh it worked with QT_QPA_PLATFORM=xcb | 19:07 |
| vkoskiv | I think the issue has to do with clock prescaling in pwm setup. The comment says it's being set to clk_io/256, but the actual code disables prescaling entirely | 19:09 |
| vkoskiv | So it's at whatever the clk_io on this mcu is, I couldn't find it in the docs, I have to assum 16MHz? That's well beyond the spec of the backlight chip, that has a max of 2MHz | 19:09 |
| vkoskiv | Setting prescaling to /256 made the backlight blink again, but /8 seems to work nicely | 19:10 |
| vkoskiv | I get no jumpiness at all when I adjust backlight now. Feels good. | 19:10 |
| Boostisbetter | congrats, glad you got that sorted. | 19:20 |
| vkoskiv | I'm just bothered that there is now 9 brightness levels, instead of a nice round 10 :D | 19:21 |
| vkoskiv | But yeah, it feels much more responsive now that it doesn't jump arounnd | 19:21 |
| Boostisbetter | where I would really like to get Ripcord working is on 32 bit Raspbian OS. | 19:24 |
| Boostisbetter | fat chance of that, but that is where it would make a lot of sense. | 19:24 |
| vkoskiv | Ah, on an old pi? | 19:24 |
| Boostisbetter | Penkesu using a Pi Zero 2 W | 19:26 |
| Boostisbetter | It was my Pocket Reform before the Pocket Reform | 19:27 |
| Boostisbetter | still pretty useful, compact, and completely repairable. | 19:27 |
| Boostisbetter | https://github.com/penk/penkesu | 19:27 |
| sknebel | zero 2 should be able to run 64bit, but I guess 32bit makes some sense with the limited RAM? | 19:27 |
| Boostisbetter | yep. You end up loosing RAM with 64 bit due to larger name space requirements. | 19:28 |
| Boostisbetter | Not worth it at all. | 19:28 |
| Boostisbetter | If the Zero ever has a gig or more it would be great. | 19:28 |
| Boostisbetter | Although I don't think RPI will ever do that, because it would cut into people getting the RPI5 | 19:28 |
| Boostisbetter | Why buy a RPI5 with 1 or 2 gigs when you can just get a Zero for $15 | 19:29 |
| sknebel | I kinda doubt they are going to make smaller-RAM variants of the RPi5 | 19:30 |
| Boostisbetter | I watched a video from RPI where they said they were making 1 and 2 gb models for people who know what they are going to be using the RPI for. | 19:33 |
| Boostisbetter | and don't need much memory. So they could get it cheaper | 19:34 |
| Boostisbetter | in much of the same way that I have been impressed with the Reform, have I been impressed with the Penkesu and 32 bit Raspbian. With so many web based clients these days, you can still do pretty much everything on a 32 bit only system. | 19:36 |
| - Boostisbetter (QUIT: Ping timeout: 252 seconds) (4a410829d7@irc.cheogram.com) | 19:41 | |
| sknebel | huh, ok. I guess if you really need the CPU power of a 5 but not much RAM. but feels getting quite niche to me, esp given they still make RPi4 too afaik? | 19:41 |
| + Boostisbetter (4a410829d7@irc.cheogram.com) | 19:53 | |
| - stites (QUIT: Read error: Connection reset by peer) (~stites@2607:fb91:dc9:51d1:1285:280d:59e4:7da2) | 19:56 | |
| + stites (~stites@130.44.147.204) | 19:56 | |
| vkoskiv | Ooooh, I built something very neat now :D | 20:01 |
| vkoskiv | As the MNT logo appears and does the little brighten/dim effect, the keyboard does the same, nicely fading in the backlight | 20:02 |
| vkoskiv | Looks sweeeet | 20:02 |
| - jacobk (QUIT: Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) (~quassel@2603-8080-b200-7b02-0000-0000-0000-0b71.res6.spectrum.com) | 20:05 | |
| + jacobk (~quassel@2603-8080-b200-7b02-0000-0000-0000-0b71.res6.spectrum.com) | 20:07 | |
| Boostisbetter | Ha! That is awesome, hopefully that can get merged so we can all get something like that! | 20:11 |
| - jacobk (QUIT: Ping timeout: 248 seconds) (~quassel@2603-8080-b200-7b02-0000-0000-0000-0b71.res6.spectrum.com) | 20:14 | |
| jfred | So I just grabbed the system image v4 version of the waybar config in /etc/skel and it looks super nice, but the battery widget seems stuck at 0%. Is there a dependency I'm missing since I started on an older system image? | 20:17 |
| Boostisbetter | what kernel are you on? | 20:18 |
| jfred | 6.5.0-1-reform2-arm64 | 20:21 |
| Boostisbetter | ok, yeah that is fishy. It should work. I wonder if your LPC version matters here. | 20:23 |
| Boostisbetter | Shouldn't mine is still bone stock from 21 and it works. | 20:24 |
| jfred | could be an LPC upgrade needed, I also see my reform-tools package is on version 1.29 with 1.31 available so maybe that's it | 20:25 |
| - Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149.210.32.204) | 20:25 | |
| + Ar|stote|is (~linx@149.210.32.204) | 20:27 | |
| jfred | yeah wasn't reform-tools, I probably need an LPC upgrade | 20:39 |
| + klardotsh (~klardotsh@c-67-170-115-80.hsd1.wa.comcast.net) | 20:45 | |
| vkoskiv | Oof, I just realised that the keyboard backlight fade-in on powerup would require changes to the system power on sequencing, since the backlight is on the 5V rail (it only turns on later during boot) | 21:48 |
| vkoskiv | I'll leave this out until I've gauged with mnt if that's an acceptable thing to do *purely* for a cosmetic thing :D | 21:49 |
| vkoskiv | It'd require a new lpc command to just turn on the 5v rail when the keyboard requests it, but that SOM/system startup code in the LPC looks like something I don't want to experiment with too much | 21:54 |
| vkoskiv | For the adventurous: https://source.mnt.re/vkoskiv/reform/-/commits/kbd-backlight-no-jumpy | 21:58 |
| vkoskiv | Just be aware, I'm a software developer, and I adjusted hardware parameters I only think I understand from reading the docs | 21:59 |
| vkoskiv | So, buyer beware. Works nicely on my keyboard, though. | 21:59 |
| Boostisbetter | Although I am curious by what you mean by backlight being jumpy. I never felt that way with mine. It was always just a quick adjustment. | 22:02 |
| vkoskiv | Basically, as you go up or down, there are keypresses that do nothing, and then a big jump | 22:03 |
| vkoskiv | And if you go back the other way right after a big jump, it's a much smaller increment | 22:03 |
| vkoskiv | With my fix, every keypress to adjust brightness brings it up or down the same increment, consistently. | 22:04 |
| vkoskiv | It just feels much more precise now. It's kind of a subtle thing. | 22:06 |
| sigrid | nice. i needed that | 22:07 |
| vkoskiv | Maybe the /8 prescaling works nicely, because 16MHz/8 = 2MHz, and that's the upper bound for the backlight chip | 22:22 |
| vkoskiv | Not sure why lower values make it go blinky blinky, though. | 22:22 |
| vkoskiv | Hmm, TI says the range is 0.1->20kHz. Some math is needed to figure out how that range jives with our pwm generation circuitry | 22:27 |
| vkoskiv | Seemingly nowhere in the doc does it actually tell me the clock io frequency, so I have to assume 16MHz | 22:31 |
| + jacobk (~quassel@129.110.242.224) | 22:35 | |
| - stites (QUIT: Read error: Connection reset by peer) (~stites@130.44.147.204) | 22:35 | |
| + stites (~stites@130.44.147.204) | 22:37 | |
| - stites (QUIT: Ping timeout: 255 seconds) (~stites@130.44.147.204) | 22:46 | |
| + stites (~stites@2607:fb90:ad63:e013:9e88:5faf:89ef:30) | 22:47 | |
| vkoskiv | Yeah, evidently it's just equal to the system clock. | 22:55 |
| vkoskiv | (section 12.1) | 22:55 |
| vkoskiv | Yeah, I think I get it now. Without my fix, the PWM frequency is 31.25kHz, which is over the max limit of the backlight chip spec. | 23:06 |
| vkoskiv | My change to the clock divide turned it down to 3.9kHz, if my math is right. | 23:07 |
| vkoskiv | Which is well within the 0.1-20kHz range of the backlight chip. | 23:07 |
| vkoskiv | But the old frequency was probably also not high enough to make it totally not work, hence the jumpy weirdness. | 23:08 |
| + abortretryfail (~arf@146.ip-149-56-132.net) | 23:23 | |
| - S0rin (QUIT: Ping timeout: 255 seconds) (~S0rin@user/s0rin) | 23:32 | |
| + S0rin (~S0rin@user/s0rin) | 23:41 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!