2026-05-01.log

+ kensanata (~alex@user/kensanata)00:01
- pomel0 (QUIT: Ping timeout: 244 seconds) (~pomel0@user/pomel0)00:09
+ pomel0 (~pomel0@user/pomel0)00:10
- Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149.210.0.44)00:33
+ Ar|stote|is (~linx@149.210.0.44)00:39
- mjw (QUIT: Killed (uranium.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2486:4600:5952:3a9:6e0e:555a)00:41
* wielaard -> mjw00:41
+ Guest2781 (~mjw@2001:1c06:2486:4600:5952:3a9:6e0e:555a)00:42
- mjw (QUIT: Ping timeout: 255 seconds) (~mjw@gnu.wildebeest.org)00:48
- Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149.210.0.44)01:09
+ mjw (~mjw@45.83.234.184)01:11
minutejosch: yeah, i noticed the second one after a while, so the first one is harmless?01:14
joschminute: i do not know yet why we get this. It is not happening for the normal Debian linux kernel builds. I think it's part of the try-run macro which is allowed to fail as run by cc-options in headers_install. I have pushed some changes into a custom branch to maybe get to the bottom of this and remove this confusing output.01:17
minutejosch: aw ok!01:17
+ Ar|stote|is (~linx@149.210.0.44)01:19
+ paul (~paul@c-68-43-179-233.hsd1.mi.comcast.net)01:25
* paul -> Guest495001:25
- Guest4950 (QUIT: Client Quit) (~paul@c-68-43-179-233.hsd1.mi.comcast.net)01:27
- Ar|stote|is (QUIT: Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) (~linx@149.210.0.44)01:28
+ Ar|stote|is (~linx@149.210.0.44)01:37
- pomel0 (QUIT: Read error: Connection reset by peer) (~pomel0@user/pomel0)01:49
+ pomel0 (~pomel0@user/pomel0)01:51
- pomel0 (QUIT: Ping timeout: 264 seconds) (~pomel0@user/pomel0)01:57
+ pomel0 (~pomel0@user/pomel0)01:58
- bkeys (QUIT: Ping timeout: 244 seconds) (~Thunderbi@98.19.131.29)02:07
- vagrantc (QUIT: Quit: leaving) (~vagrant@2600:3c01:e000:21:7:77:0:50)02:14
- mjw (QUIT: Ping timeout: 244 seconds) (~mjw@45.83.234.184)02:14
- a3f (QUIT: Read error: Connection reset by peer) (~a.fatoum@flummi.grey.stw.pengutronix.de)02:19
+ a3f (~a.fatoum@flummi.grey.stw.pengutronix.de)02:22
- pomel0 (QUIT: Ping timeout: 245 seconds) (~pomel0@user/pomel0)03:11
- paperManu (QUIT: Ping timeout: 252 seconds) (~paperManu@204.244.197.237)03:35
+ pomel0 (~pomel0@user/pomel0)04:15
- aelius (QUIT: Remote host closed the connection) (~aelius@user/aelius)04:32
+ aelius (~aelius@user/aelius)04:33
jfredmmm that's what I was afraid of, swapped modules on my pocket and it won't boot... glad I bought that faster serial cable or I'd really be dead in the water. time to hook up and see what's going on04:39
jfredhm. okay, it's trying to load the device tree from the reform2 still. I set fdtfile to the rk3588 pocket reform one and tried booting, and... the display came on that time, but just showing noise05:04
jfredreflashing an SD card with the debian image to see if I can boot off that05:18
jfredwoo okay that worked. (now to figure out how to fix my install XD)05:29
- pomel0 (QUIT: Ping timeout: 246 seconds) (~pomel0@user/pomel0)05:45
rick_@vagrantc oh interesting, i had the same problem yesterday. I thought the adapter may have have loose contact again. I keep my eyes open and look around more when this happens next time05:46
+ pomel0 (~pomel0@user/pomel0)07:22
joschminute: 7.0.3-1 is now in unstable -- i saw you were doing qcom things yesterday. Should I just merge the linux7.0 branch or do you first want to finish the quasar work or do you want more testing of the 7.0 kernel?07:51
- Guest2995 (QUIT: Ping timeout: 264 seconds) (~ndufresne@transit.collaboradmins.com)08:32
+ mjw (~mjw@gnu.wildebeest.org)09:37
- mjw (QUIT: Ping timeout: 245 seconds) (~mjw@gnu.wildebeest.org)10:47
* Guest2781 -> mjw10:51
- cli (QUIT: Ping timeout: 245 seconds) (~m-vsauiy@user/cli)11:36
+ cli (~m-vsauiy@user/cli)11:38
minutejosch: merge away (7 to main)! i will later merge/rebase into my qcom branch. last night i made that build successfully and today i'll make a sysimg from that branch and try to see if it works on next+649012:12
minuteopenbsd really using efi gop @ barebox https://fosstodon.org/@a3f/11649830752435156112:13
+ paperManu (~paperManu@204.244.197.237)12:24
joschminute: great, I'll also build another sysimage with linux 7.0 and test on a311d, rk3588, imx8mq and ls1028a and then merge if things are booting (i'm not intending to test much more than that today because holiday)12:36
minutejosch: awesome :3312:39
- lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah)12:40
+ lidstah (~lidstah@gateway/tor-sasl/lidstah)12:41
- lidstah (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah)13:08
+ lidstah2 (~lidstah@gateway/tor-sasl/lidstah)13:08
+ wielaard (~mjw@gnu.wildebeest.org)13:28
- wielaard (QUIT: Ping timeout: 244 seconds) (~mjw@gnu.wildebeest.org)14:05
- paperManu (QUIT: Ping timeout: 248 seconds) (~paperManu@204.244.197.237)14:33
- lidstah2 (QUIT: Remote host closed the connection) (~lidstah@gateway/tor-sasl/lidstah)14:42
+ lidstah (~lidstah@gateway/tor-sasl/lidstah)14:43
+ paperManu (~paperManu@modemcable141.205-200-24.mc.videotron.ca)14:45
minutejosch: remind me, to build reform-system-image with that variable for another reform-debian-packages branch, can i just do that as a pipeline for main on reform-system-image or do i need to branch reform-system-image first?15:08
minuteah, well the CI answered the question > Refusing to run pipeline in the 'reform' namespace with REFORM_DEBIAN_PACKAGES_BRANCH: not set to ''15:18
minutei thought i had a fork set up somewhere...15:18
minuteah well i do, just 404'ed on "minute" vs legacy "mntmn" https://source.mnt.re/mntmn/reform-system-image :D15:19
joschminute: yes, the reason for that is that otherwise custom builds will show up as the "latest" artifact when following https://mntre.com/system-image/15:47
joschminute: I do these custom builds in my fork of reform-system-image15:47
josch(also slight change of my plans: the whole family is playing luanti in networked multiplay with all the reforms we have here at home right now -- even a311d is fast enough for it :D15:48
amospallaWrote this to send text/image to either keyboard leds or oled display https://github.com/amospalla/reform/blob/main/reform_leds/15:54
amospallaActually, sending things to the Pocket oled works only on my pocket, not on current firmware hehe.15:55
minutejosch: haha oh that's awesome @ luanti party15:57
kfxminute: that barebox stuff is only for pocket so far?16:14
minutekfx: for pocket it is known working. for classic, it should work if you have hdmi adapter. but to run on classic with dsi adapter, there is code/testing missing to activate the sn65dsi86. which should be trivial, but i didn't have time yet16:17
kfxok16:17
minuteanyone with some patience and C+I2C experience could do it16:17
kfxI have one classic with the hdmi adapter and one with dsi, so I can have fun making a fool of myself for a while!16:17
+ vyvv (~vyvv@user/vyvv)16:23
minutewat > quasar dtb exists -- remove this hack!16:32
+ reform7466 (~jfred@2603:301b:2df:8300::75ff)16:48
* reform7466 -> jfred-pocket16:48
- buckket (QUIT: Quit: buckket) (~buckket@vps.buckket.org)17:09
+ buckket (~buckket@vps.buckket.org)17:10
jfred-pocketTrying to basically rebuild my /boot since it's on emmc on both modules... I'm chrooted into my rootfs from the rescue image, and the reform-tools postinst script wants to install dtbs for a kernel that isn't installed on this disk17:11
jfred-pockethttps://paste.debian.net/hidden/5170c1c517:11
jfred-pocketI'm having a bit of trouble finding out where the list of kernels to install dtbs for gets built, because the zz-reform-tools script takes a version as an argument17:12
joschminute: in the past there were sometimes problems with the system images because for some reason the dtb for a machine.conf was missing from the system image and then the system would not boot. The code you saw there in mkimage.sh makes sure to error out if the dtb is missing and has a helpful comment so that you know to remove this exception for quasar once the dtb exists.17:20
- amospalla (QUIT: Ping timeout: 248 seconds) (~jordi@user/amospalla)17:24
jfred-pocketaha, figured it out - I just had to remove the kernel image from /boot because the postinst script uses `linux-version list` to discover them17:26
minutetss, i can't install blender on greta's debian computer (amd64), because17:27
minute> libopenexr-3-4-33:amd64 is selected for install because: ... libopenimageio2.5:amd64 Depends libopenexr-3-4-33 (>= 3.4.6+ds)17:27
minute> libopenexr-3-4-33:amd64 is selected for removal because: ... blender:amd64 Depends libopenexr-3-1-30 (>= 3.1.5)17:27
- jfred-pocket (QUIT: Remote host closed the connection) (~jfred@2603:301b:2df:8300::75ff)17:28
minuteso blender's dependencies were not updated, but one blender dependency was and made it incompatible with blender17:28
minute(?)17:28
joschminute: did you install blender from experimental or from my repo? Because that requires openexr from experimental. Do you have experimental configured on your machine?17:28
minutejosch: normal amd64 unstable debian17:29
minuteon a pc17:29
minutei can try blender/openexr from experimental there...17:29
minuteaha, that seems to work17:30
joschaha this is a known issue: https://qa.debian.org/dose/debcheck/unstable_main/latest/packages/blender.html#1a580111377584810a6212de3234d5df17:30
minutejosch: thanks, this trick worked!!17:30
joschyou are now using experimental and that worked?17:31
+ amospalla (~jordi@user/amospalla)17:38
- amospalla (QUIT: Client Quit) (~jordi@user/amospalla)17:39
+ amospalla (~jordi@user/amospalla)17:45
minutejosch: yeah i did apt install -t experimental blender libopenexr-3-4-3317:45
minutejosch: not sure if libopenexr-3-4-33 was needed, probably not17:45
+ vagrantc (~vagrant@2600:3c01:e000:21:7:77:0:50)17:48
minutejosch: bizarro https://source.mnt.re/mntmn/reform-system-image/-/jobs/20448#L479917:50
minuteah maybe that happens when someone hits cancel17:50
minuteright, sorry for the noise17:50
jfredwoohoo, back to a fully functional pocket reform! :D17:51
minutejfred: noice, what did you do?17:51
jfredchrooted into my rootfs on disk, removed an old kernel image from /boot that wasn't on the rootfs, ran some package upgrades/update-initramfs/flash-kernel, installed firmware-iwlwifi which I didn't have installed yet17:54
jfred...not entirely in that order which took some troubleshooting, the leftover kernel image was blocking the reform-tools postinst script from running successfully17:55
jfredvery glad I was able to boot the stock image from SD to use as a rescue image17:57
joschjfred: do you have a theory why you had a leftover image? Maybe this is something we should be checking for in reform-check18:00
joschminute: yes, the problem with blender from unstable is, that it depends on libopenimageio2.5 and that shared library got recently rebuilt with a newer version of libopenexr than blender was compiled with. The issue would be fixed in unstable by triggering a rebuild of blender.18:02
minutejosch: right.18:07
jfredjosch: I think I hadn't run package upgrades on both machines recently enough/hadn't removed old kernel packages prior to the upgrade18:07
minutegnah i failed to select to build only the "any" image, so the build will take much longer and do a lot more uplaoding18:08
minuteuploading18:08
jfredIOW the two machines weren't sufficiently identical before I basically swapped /boot between them, heh18:08
jfredalso gotta say it is very nice having the speedy rk3588 in the pocket18:10
joschjfred: if you swapped /boot you probably used reform-emmc-bootstrap and then that tool failed to clean up?18:11
- g3ist (QUIT: Ping timeout: 248 seconds) (~g3ist@user/g3ist)18:13
jfredjosch: ah. no, I... had forgotten about reform-emmc-bootstrap, so I guess I did things the hard way XD18:13
jfredI still need to put the i.MX8MP into the Reform 2 so I'll try reform-emmc-bootstrap when I do that ^^;18:18
minutejosch: bmaptool is broken now on my amd64 machine > TypeError: 'GPGMEError' object is not subscriptable18:19
minute> File "/usr/lib/python3/dist-packages/bmaptool/CLI.py", line 167, in verify_bmap_signature_gpgme18:19
Svpgave a try to kmscon, it's somewhat handy, could be worth considering18:19
minuteSvp: yeah?18:19
minutewill try bmaptool copy --no-sig-verify18:20
minutehmm nope... > TypeError: 'NamedFile' object is not iterable18:20
minuteweird18:20
joschminute: can you give me the full invocation?18:22
Svptrue color, live font resize and custom fonts instead of fbcon are much welcomed18:22
Svpand has a mouse cursor instead of that dastardly cursor block!18:24
minutejosch: interestingly it works after downloading the img.gz and the .bmap manually18:27
minutejosch: invocation was > sudo bmaptool copy https://source.mnt.re/mntmn/reform-system-image/-/jobs/20449/artifacts/raw/reform-system-any.img.gz /dev/sda18:28
minuteSvp: oh noice18:28
- rick_ (QUIT: Remote host closed the connection) (~m-x42ph2@m.n0id.space)18:28
+ rick_ (~m-x42ph2@m.n0id.space)18:28
Svpgonna try to run it on my reform VTYs for a while but so far i like what i see18:29
- rick_ (QUIT: Remote host closed the connection) (~m-x42ph2@m.n0id.space)18:33
+ rick_ (~m-x42ph2@m.n0id.space)18:34
joschminute: thank you that should work and i'll put it on my todo list to try out bmaptool on a Debian unstable system later today18:37
- amospalla (QUIT: Quit: WeeChat 4.9.0) (~jordi@user/amospalla)18:38
rick_oh lol, til.. debian13 doesn't have the last command anymore and the replacement is wtmpdb18:38
minuteok after two little changes in the extlinux.conf on my fresh image, i can boot into our debian on 6490-next, over serial console18:42
minutehmm wait wat > Hardware name: MNT Pocket Reform with QCS6490 Quasar Module18:43
minuteah it ignored fdtfile=18:44
minuteoh, and there are no qcs6490-mnt-{pocket,next}.dtb on the resulting disk, that's werid18:44
minutewhy would that be so... i guess they need to be added to reform-tools/machines?18:45
minuteyeah they're in /lib/modules/6.19.14-mnt-reform-arm64/dtb/qcom/ at least, that's good18:46
joschminute: right now, reform-tools only ships QCS6490 and QCS8550 for MNT Reform 218:53
joschPocket Reform, Reform Next and Reform Station probably need to be added?18:54
minuteo m g the display turned on on first try18:55
minutejust it's suffering because there's no qcom fw available lol18:55
minutejosch: station yeah but in general there are no station .dts yet18:55
minutejosch: but next and pocket for qcs6490 yes, and next for 8550, which i'll probably write later today (the dts)18:56
joschminute: wow huge success with having display working!18:56
minutejosch: yeah i didn't expect that to work after going through all the reform-debian-packages pipeline etc etc18:56
minute(patches)18:57
minuteinstalling firmware-qcom-soc...19:04
minutemhm, that has /lib/firmware/qcom/qcm6490 but not qcs6490, will symlink for now but maybe that has to be then explicitly set everywhere in the dts, hmhm19:06
minuteah no i'm wrong, it's there, all good19:07
+ wielaard (~mjw@gnu.wildebeest.org)19:16
- wielaard (QUIT: Remote host closed the connection) (~mjw@gnu.wildebeest.org)19:22
+ wielaard (~mjw@gnu.wildebeest.org)19:22
minutegnah how to exit usb gadget mode...19:33
- mjw (QUIT: Killed (tantalum.libera.chat (Nickname regained by services))) (~mjw@2001:1c06:2486:4600:5952:3a9:6e0e:555a)19:44
* wielaard -> mjw19:44
+ Guest6124 (~mjw@2001:1c06:2486:4600:5952:3a9:6e0e:555a)19:45
joschminute: is this what you saw including the "gzip: stdin: not in gzip format"? https://paste.debian.net/hidden/cb5fdff219:48
minutejosch: yes, but for me it was "pigz" saying something like that20:05
joschminute: i think this is a problem of https://source.mnt.re. Try it out by going here: https://source.mnt.re/mntmn/reform-system-image/-/jobs/20449/artifacts/file/reform-system-any.img.gz.bmap20:05
joschwhen you try downloading that you do not get a XML file20:05
minutejosch: i do get an xml file20:06
minutejosch: (downloaded in firefox)20:06
joschyes, but try this:20:06
joschcurl -L https://source.mnt.re/mntmn/reform-system-image/-/jobs/20449/artifacts/raw/reform-system-any.img.gz.bmap20:06
minutejosch: ahhh i guess i'm extremely silly and my repo is not public20:07
joschooooh20:07
minutejosch: sorry for wasting your time ;__;20:07
joschminute: not at all -- this failure mode of bmaptool is also very bad20:07
joschi'm fixing this and file a MR with bmaptool upstream20:08
joschif you were told that what bmaptool gets is not even an XML, then it would've been easier to diagnose20:08
minutejosch: that's true20:08
jfredoh, that's fun - I just tried connecting a pair of bluetooth headphones and the bluetooth controller seems to have died: https://paste.sr.ht/~jfred/cead7b2d0e50f850fd6992f4a12ef3bbdd614d0520:22
minutedamn > CONFIG_USB_QCOM_EUD20:25
minute> CONFIG_USB_QCOM_EUD is not set20:25
joschminute: your bmaptool problem made me found four other issues which are getting fixed now -- so great success :)20:37
minutejosch: oh nice20:38
vagrantchrm. i think my problems with ethernet are somewhere between the cable and the rj45 plug on the mnt/reform rather than software ... i seem to be able to nudge it in just the right way to get a connection20:49
vagrantchoping it is the cable, obviously... :/20:49
rick_vagrantc i had this already.. its probaply the cable.. a new adapter fixed it for me20:55
rick_and my second one currently acts up too.. these ix adapters doens't seem very 'industrial' ^^v20:56
vagrantcoh, this is a mnt/reform classic20:58
vagrantcso not the ix adapter20:58
rick_ohhh20:58
+ amospalla (~jordi@user/amospalla)20:59
vagrantcjust a typical rj45 port (as far as i know)21:00
- dodo (QUIT: Quit: dodo) (~dodo@user/dodo)21:16
+ dodo (~dodo@user/dodo)21:17
- pomel0 (QUIT: Ping timeout: 276 seconds) (~pomel0@user/pomel0)21:48
+ pomel0 (~pomel0@user/pomel0)21:48
- paperManu (QUIT: Ping timeout: 255 seconds) (~paperManu@modemcable141.205-200-24.mc.videotron.ca)22:05
- vyvv (QUIT: Quit: Leaving) (~vyvv@user/vyvv)23:04
+ paperManu (~paperManu@204.244.197.237)23:39

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!