Tuesday, 2020-12-08

deathmist_also how come on a SFOS 3.3 -> 3.4 upgraded install my nemo user got dropped from the system droid group and I can't add myself to it anymore? I'd really not like to bother with figuring out exactly what is messing up the permissions for my /proc/touchpanel files (needed for https://github.com/sailfishos-oneplus5/gesture-daemon to work)01:07
T42<elros34> deathmist_ use additional_ha_groups: https://github.com/mer-hybris/droid-hal-device/commit/1dac6f1ff3b58215003af61ba52bd91c5b17721902:47
T42<adampigg> deathmist_: thats not being very security conscious ;)09:24
T42<ankaos> PlatformSDK [enes@enes hadk]$ ubu-chroot -r $PLATFORM_SDK_ROOT/sdks/ubuntu10:58
T42<ankaos> bash: ubu-chroot: command not found10:58
T42<ankaos> latformSDK [enes@enes hadk]$ rpm/dhd/helpers/build_packages.sh —droid-hal12:20
T42<ankaos> * Building rpm/droid-hal-oxygen.spec12:20
T42<ankaos> sb2: Error: Tools directory '/srv/mer/toolings/xiaomi-oxygen-armv7hl' does not exist.12:20
T42<ankaos> No journal files were found.12:20
T42<ankaos> sb2: Error: Tools directory '/srv/mer/toolings/xiaomi-oxygen-armv7hl' does not exist.12:20
T42<ankaos> No journal files were found.12:20
T42<ankaos> Failed to seek to cursor: Invalid argument12:20
T42<ankaos> * Check /home/enes/hadk/droid-hal-oxygen.log for full log.12:20
T42<ankaos> mal: how fix this problem?12:22
deathmist_just read the HADK and make notes12:22
deathmist_@elros34 oh so that's what it's for... I thought it was broken because it didn't work on a fresh image so I dropped those changes12:23
deathmist_@adampigg I know, but it seems changing the permissions in vendor.lineage.touch@1.0-service.oneplus_msm8998.rc for /proc/touchpanel stuff to input indeed works, not sure what I tried before... now I just need a working bind-mount for the file (resides in /vendor) and I can drop the system group I hope12:27
deathmist_oh yeah I think the "@" in the filename actually made the bind-mount impossible via the unit, that's why I never fixed this properly; let me test again12:34
T42<ankaos> maybe fix bluetooth12:35
deathmist_oh yeah this is the issue: "<whatever>.mount: Where= setting doesn't match unit name. Refusing." the filename is vendor.lineage.touch@1.0-service.oneplus_msm8998.rc which makes creating the bind-mount unit name difficult, I tried "vendor-etc-init-vendor.lineage.touch\x401.0\x2dservice.oneplus_msm8998.rc.mount" as per "systemd-escape --path12:49
deathmist_/vendor/etc/init/vendor.lineage.touch@1.0-service.oneplus_msm8998.rc" with no luck12:49
T42<adampigg> deathmist_ : i just do my bind mounts in a script12:53
deathmist_hmm, I should probably do the same; I already feel like I have too many unit for those12:54
T42<adampigg> it works if turn off default dependancies ... thats the key to getting the service to run as part of local-fs12:55
T42<ankaos> deathmist_:https://paste.ubuntu.com/p/h9zyXNSXTk/ how fix?13:00
deathmist_update your submodules ¯\_(ツ)_/¯13:05
deathmist_@adampigg mind sending the script you use? I'd like a bit of reference :)13:30
deathmist_mal: could nfcd-dbuslog-plugin be added to :hw:common for 3.4.0 on OBS? my CI build failed because of it: https://gitlab.com/sailfishos-porters-ci/cheeseburger-ci/-/jobs/89571124415:34
deathmist_piggz: https://github.com/sailfishos-oneplus5/droid-config-cheeseburger/commit/c863033 and that's how easy it was after all :)15:53
piggzdeathmist1: ah, so you had a script already in place for mounts16:09
piggzdeathmist1: are you working on mainline 5t support?16:10
deathmist_well, many mount units anyway; I migrated them to the new script "framework" in the parent of the commit I linked :p16:10
deathmist_and yes I am, though not super actively currently16:11
piggzhow far are you? can it boot sfos?16:11
piggzif i got bored, and wanted somethine else to do, its the same soc as the pro1 isnt it?16:12
deathmist_I've not tried; currently the GPU isn't working as I've tried so I'm stuck with framebuffer graphics without acceleration, but there is some awesome work being done on https://github.com/SoMainline/linux/commits/angelo/somainline-msm8998-bringup which I'm definitely keeping my eye on16:13
piggznice, will watch16:13
deathmist_touchscreen (up to 10 points) and a bunch of other things are working nicely already though, so it has potential to become usable on mainline as a phone when the modem can properly be hooked up with userspace :)16:16
deathmist_any news on a better solution to https://github.com/sailfishos-oneplus5/nemo-qml-plugin-systemsettings/commit/527ea89 still? I can't get my custom 0.5.52 package version installed on the device with the OBS repo: http://repo.merproject.org/obs/nemo:/testing:/hw:/oneplus:/cheeseburger/sailfishos_3.4.0.24/armv7hl/?C=M;O=D16:54
deathmist_"No update candidate for 'nemo-qml-plugin-systemsettings-'. The highest available version is already installed."16:56
deathmist_ah seems this worked on tama, I didn't put the "3" release number in the tag: https://github.com/sailfishos-sony-tama/nemo-qml-plugin-systemsettings/releases/tag/
rinigusDeathmist: no, nothing better so far...18:05
rinigusmal: would you mind to raise the issue with nemo-qml-plugin-systemsettings on devices with different internal storage notation in the office?18:06
rinigus... as was done few times already in 2020?18:06
deathmist_I'm still puzzled as to how this was not hit on any of the officially supported Sony device ports either; afaik every somewhat recent QCOM device has ~70 partition spanning sda-sdf most likely on UFS internal storage18:11
piggzpro1 has same fork :D18:13
piggzwe need a PR that describes the paritions anming in a file18:14
rinigusdeathmist_: they are on older soc18:23
deathmist_obviously different OEMs do different things (not all of them have 70 partitions like OnePlus xD), but the regex always matches all UFS internal storage partitions anyway (sd*); instead of a simple regex could we potentially check /sys/block/* with udev and make sure something is a USB stick/SD card etc that way?18:26
deathmist_if I run "udevadm info --query=property --path=/sys/block/sda | grep ID_BUS" for example that prints "ID_BUS=scsi" as it's a part of my UFS internal storage18:27
deathmist_plugging in a USB stick via OTG and checking /sys/block/sdg in my case returns "ID_BUS=usb"18:30
T42<ankaos> @ot_kurva [Sorry,  https://github.com/mer-hybris/hadk-faq …], not work bluetooth19:09
T42<ankaos> (Photo, 720x1280) https://irc.thaodan.de/.imgstore/sKAqckl4xF.png Bluetooth screen19:19
T42<ankaos> sh-3.2# /usr/bin/anbox-shell.sh20:46
T42<ankaos> ERROR: Cannot attach to container as it is not running20:46
maldeathmist_: I added that package to common repos now21:05
T42<ankaos> link please21:11
T42<ankaos> mal: and how install anbox?21:11
T42<ankaos> sh-3.2# anbox.sh21:13
T42<ankaos> [ 2020-12-08 21:12:47] [daemon.cpp:61@Run] Failed to connect to DBus21:13
deathmist_mal: thanks (nfcd-mce-plugin was needed too), but I mentioned nfcd-dbuslog-plugin in my message which also depends on libdbuslog21:18
maldeathmist_: things get a bit messy when people use too new nfcd21:36
deathmist_mal: https://gitlab.com/sailfishos-porters-ci/cheeseburger-ci/-/jobs/898354638 hmm I'm unsure what packages I should pull in now from my metapackage, these were all added in https://github.com/mer-hybris/droid-hal-configs/commit/09a089a and it works locally because everything is from git master22:05
maldeathmist_: common had the nfcd version which is in 3.4.0 release, if you use that version then you should not add those new nfc packages to meta-packages22:32
deathmist_mal: even nfcd-mce-plugin? it sounds really useful for battery life savings :/22:39
maldeathmist_: https://git.sailfishos.org/mer-core/nfcd/commit/4f1051c37cf1ec3a95ba6fd7144799d5d0b72c0d22:41
maldeathmist_: those packages used to be in the main nfcd but were separated22:41
maldeathmist_: and https://git.sailfishos.org/mer-core/nfcd/commit/9112838f96188459b6375dfd9d4b95034207173e22:42
deathmist_ah cool! I'll drop those changes then; I'll probably write up some build_mw() extension to allow building packages on a specific version and most likely make a PR, it's useful in general because otherwise you now need to reset --hard and build_packages --offline ... to avoid using latest sources again22:43
malyeah, build_packages.sh is a bit problematic when devel has moved forward too much22:44
T42<Leonardo %lastname%> Fwd from ™^°: https://t.me/binarycommunitycojFX22:58

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!