Sunday, 2020-05-31

T42maffeen was added by: maffeen06:17
T42<maffeen> Hi06:17
rinigusMorning! Looks like only fingerprints added in settings can be removed, otherwise sailfish settings gets confused. I'll look into test app to support API fully06:18
riniguspiggz /\06:18
T42<adampigg> Rinigus : thanks06:35
T42<maffeen> Hi! Does anyone know if I can use Lineage OS 17.1 as base? I would like to port Sailfish OS to my device.07:00
r0kk3rznot yet07:05
T42<adampigg> R0kk3rz: you got any ports with a fp sensor ?07:10
r0kk3rzmaybe if they ever open up this nokia thing ive got07:12
T42<maffeen> thanks r0kk3rz, I think that I'll just wait07:12
riniguspiggz: I am moving `setState(FPSTATE_IDLE)` as enumerate changes the state as well. let me know if it's wrong and you moved it there for some reason that I am missing07:44
T42<adampigg> Rinigus: carry on, i can test later....working on project r5 this morning08:08
T42<adampigg> I agree gplv3 btw08:09
riniguspiggz: very good, will continue08:09
T42<adampigg> Rinigus: did you have to do any fixing at the hal level for your port?08:09
rinigus@adampigg: no, none08:10
rinigusas long as you don't use clear(), I was fine08:10
T42<adampigg> Oh gr8, but biometryd didnt work?08:10
rinigus@adampigg: sorry for delay. just submitted PR for review and one issue with a crash08:40
rinigus@adampigg, @erfanoabdi: looks like biometryd was failing on my device for two reasons. 1: clear() leading to remove(0,0) on android side seems to be not supported on android. this is a working theory, has to be confirmed if someone wants to investigate. on my device, removing one-by-one works08:43
T42<erfanoabdi> @rinigus [@adampigg, @erfanoabdi: looks like biometryd w …], removing fingerid 0 is clear all according to aosp api08:44
rinigus2nd issue: biometryd got somehow confused with future and promises. as a result, it was constantly crashing on my device.08:44
T42<erfanoabdi> If it's crashing the VTS also doesn't pass on your device08:45
rinigus@erfanoabdi: yes, it is. but looks like not all devices support it. sony devs told that at least in one case they have to loop over the fingerprints to remove them08:45
rinigusin practice, removal one-by-one is working and is used by android as well.08:46
T42<erfanoabdi> In biomertyd git history there's the clear all method with one by one removing08:46
T42<erfanoabdi> U can revert it08:46
rinigusas for VTS, no idea what is it even08:46
rinigus@erfanoabdi: probably we can revert it. should look into it at some point. however, its rather difficult to debug that biometryd, or I am missing something. in this respect, bare bones piggz daemon looks to be sufficient.08:49
T42<erfanoabdi> Yeah his daemon looks good for Sailfish08:49
rinigus@erfanoabdi: sony devs wanted to look into remove(0,0) issue. but nothing popped up when we checked just the beginning of the stack. its bit deeper where the error is issued08:50
rinigusI will check whether dropping clear from biometryd test will make it pass on my device.08:51
rinigus@erfanoabdi, @adampigg: when I disable clear() in the test, biometryd works as expected09:12
rinigusso, the issue was in the errors in clear() that somehow messed up the observer09:13
riniguspiggz: I will look now into test gui - as was planning earlier this morning09:20
T42<erfanoabdi> 👍👍Nice09:34
rinigus@erfanoabdi: thank you very much for your help!09:38
T42<erfanoabdi> rinigus: no problem, u guys did all the hard work ;)09:38
riniguspiggz did :) ^09:39
T42<adampigg> :) great code from erfan was the enabler though.... And thanks for the PRs this morning rinigus...while ive been rubber coating the undeside of my renault :D09:43
T42<adampigg> Im wondering about adding different implementations of the android side in the future, for older bases like LOS 14 for mido port, as that likely wont like the hidl interface?09:44
T42<RealDanct12> @adampigg is the src up?09:46
T42<adampigg> @RealDanct12 [@adampigg is the src up?], Sure,
T42<erfanoabdi> @adampigg [Im wondering about adding different implementa …], just build the hybris backend with legacy hal09:47
T42<adampigg> Ah, ok, ill try it latet09:47
T42<erfanoabdi> @adampigg [Ah, ok, ill try it latet], there's a similar code for gps hal on ubports that opens legacy hal and uses it09:49
T42<erfanoabdi> that can be useful to check as reference(thats what i did for fp)09:49
riniguspiggz: as I have no idea when Verify is called, I smuggled debug message into it as well. pushed it into PR waiting your review11:57
ggabrielo/ sorry for the n00b question... but I'm getting ninja: error: 'fstab.common' - I know you have to copy fstab.common to out, but I don't know exactly where. Anybody can help?13:33
ggabrielnever mind... I needed fstab.common in . as opposed to out14:03
ggabriel(that must be a mistake in the ninja file)14:03
riniguspiggz: submitted PR for "test" app. so, there are 2 PRs for you to review: one for daemon and one for client. I will wait for reviews and do something else...14:07
T42<ot_kurva> Problem: droid-config-jasmine_sprout-1-1.armv7hl obsoletes ofono-configs-mer provided by ofono-configs-mer-1.23+git12-1.armv7hl14:14
T42<ot_kurva> How to fix?14:14
piggzrinigus: all merged \o/14:49
rinigusthanks! :)14:49
riniguspiggz: just saw that we still have "default cover" for the app... will fix that too.14:51
rinigusI'll file few issues for GUI as well14:51
piggzok ... later on i will start working through the issues, starting with ading a license and readme14:52
piggzi need to see if it will work on mido and pro114:52
piggzi just need to finish up with my car, after taking  a break to fix my parents car!14:52
piggzi have no idea how a 150A fuse can blow!!!!!14:52
piggzi think it must have shorted while removing the battery14:53
T42<Verevka86> @ot_kurva [How to fix?],
riniguspiggz: I wonder if you can reproduce
rinigusfor me its quite reproducible, unfortunately. sometimes it takes one FP fine, but the second leads to the reboot.15:40
rinigusso, in this respect, test app is working better. never seen that with it15:40
riniguspiggz ^15:40
T42<adampigg> Rinigus, ive not had an issue adding lots of fp, even when it fills up15:48
T42<adampigg> Rinigus, maybe we could add a quirk for it?15:49
zinstackHello. Building hybris and stuck at being unable to boot, obviously. My device is moto g7 (river), sdm632 chip, hybris-16.0 and lineage-16.0 as a base. Fun stuff happens in hybris-boot.img. I know selinux should be disabled, either by kernel command (looks like a preferred method) or by removing it from kernel altogether. In my case, neither work. With selinux, I can boot into init-debug with real rootfs and all,15:54
zinstackbut as soon as I do anything about selinux, I get thrown into bootloader immediately after booting. In this case, no usb devices show up in dmesg. Here's my messy kernel config: Kernel config check results in a few warnings, but nothing that catches the eye. Please ignore SELINUX_DISABLE, it has no effect on the whole picture. I've been using calebccff's port to enchilada as my15:54
zinstackreference. Does anybody know what should I do?15:54
T42<elros34> this is wrong: CONFIG_SECURITY_SELINUX_DISABLE=y. YOu must disable selinux via bootparam15:55
T42<elros34> How fast does it reboot?15:59
riniguspiggz: let's see what your other devices do. I can also recommend 'test' app16:03
T42<adampigg> Rinigus wdym recommend?16:05
zinstackelros34: This doesn't disable selinux, oddly, only permits disabling it in post-boot (don't know how though, it wasn't there before anyway): CONFIG_SECURITY_SELINUX_DISABLE=y. It reboots instantly, as if boot image is corrupt (I think, not entirely sure).16:05
T42<elros34> For sure to get sfos working you must have CONFIG_SECURITY_SELINUX_DISABLE=n. Which exactly config cause instant reboot?16:07
T42<elros34> Did you remove parsing of skip_initramfs in kernel?16:08
zinstackThe same config with that exact line, but with bootparam selinux=0. About skip_initramfs, not sure, how'd I do it?16:09
T42<elros34> it will probably not help but instead selinux=0 you can use CONFIG_SECURITY_SELINUX_BOOTPARAM_VALUE=016:15
zinstackWill try that with bootparam first, enchilada port used bootparam and that patch, but I haven't paid enough attention to it. Thanks beforehand, will report after about half an hour16:17
T42<elros34> that skip_initramfs patch is required, otherwise hybris-boot intrams could not be used16:19
riniguspiggz: I mean that users could use it for fingerprint management, not Sailfish Settings16:34
rinigus@adampigg: as for quirk, we would have to figure out what is causing the reboot, I guess.16:35
T42<adampigg> Rinigus, the quirk would be to only allow adding one finger16:40
rinigus@adampigg: that's assuming it goes bad somewhere below the daemon. from the logs, I don't see even enroll being called16:51
rinigus... but I would say its not critical as there is a workaround.16:52
rinigusin terms of using other software16:52
rinigusbtw, on sony, I can use fingerprint to enter gestures. such as sliding the finger is translated into the keyboard's arrow keys (up, down, left, right). all is hooked nicely and can be used in ... terminal!16:54
T42<adampigg> Rinigus: its all good fun this bringing fp to the porters!16:56
rinigusit is, indeed!17:00
rinigus@adampigg: I'll start then with android rpm bits . let's see if I will regret it17:02
zinstackOkay, having progress. Now it's not happy about /system not being mounted. Me too, but I think it's about fixup-mountpoint, am I right? My device is a/b, so there is no specific "by-name/system" partition, they have letters (Wow). I guess there should be one?17:48
zinstackOh, and I had to touch *_property_contexts files in rootfs to make droid-hal-init go. Bummer that I don't know how to make them appear automatically. Does anybody know?17:51
T42<elros34> Check whether system.mount system_root.mount in /lib/systemd/system are created and contains /dev/mmcblk or sdx path. About context files, Are you sure you use target and you added them to straggler files?17:53
zinstack@elros34 Well, files exist but they only reference fstab.qcom, that's because I didn't mention "by-name/system" specifically, only system_a and system_b. I didn't touch straggler, but I know what it is now and how to fix it17:59
T42<elros34> So probably you are using old target otherwise droid-hal will complain about missing files18:01
zinstackNo, my target is 3.2.1-19 I think, it's pretty recent I believe18:02
T42<elros34> latest is and you should use it. Here are my notes for most common issues:
zinstackWeird naming scheme or an abandoned branch? This thing is full of surprises18:10
zinstackOh, and how do I know which straggler files would I need? Is it trial and error? Goodness I hope not18:11
zinstackAlso, it appears that my eyesight is declining and I couldn't spot that 3 is bigger than 2. Or is it my sanity?18:15
rinigus@adampigg: Android bits packaged; PR submitted. Will probably stop here for now18:55
T42<hacker12455> ```* dconf permission issues:18:57
T42<hacker12455> Regenerate your ks file if you use custome one```18:57
T42<hacker12455> so I removed all kickstart files and rebuilt droid-config18:57
T42<hacker12455> I still get dconf errors with mic18:57
T42<hacker12455> except I have given address is empty errors18:58
rinigusany gurus: how do we make fingerprint android store as in
rinigusif we need to mess with init scripts, were do those changes go? or should we make them somehow differenty as suggested by Thaodan in
T42<Verevka86> @rinigus [any gurus: how do we make fingerprint android …], This?19:06
rinigusVerevka86: as we need it rather late, maybe systemd would actually work. with its runtime dir support. will see the mans19:09
piggzrinigus: let me get up to speed with the days work, just getting to sit down!19:17
riniguspiggz: enjoy!19:24
piggzrinigus: are you around for comments?19:25
riniguspiggz: yes, for 5-10 mins19:34
piggzjust those on the pr19:38
riniguspiggz: replied and going away for some time19:40
piggzrinigus: np, thx20:09
riniguspiggz: readme looks nice!20:56
piggzrinigus: thx, its a start!20:56
T42<adampigg> @erfanoabdi building the android part on my mido port fails, with error "pre oreo not supported" .... does pre oreo need a different implementation?20:57
riniguspiggz: its late over here, I am off for today20:57
piggzrinigus: ill try and fix a few more packaging bits20:57
riniguspiggz: thanks!20:58
Kabouik@elros34, would you happen to know how one can force landscape orientation using the compositor for a Wayland app?21:45
KabouikI'm asking you because I know you used some qxwayland tricks for sailfish_chroot, but maybe that would be different for Wayland apps?21:45
KabouikI'm asking candidly without know much about the details, but I'm trying ot help a dev porting a desktop app to SFOS, and it's stuck to portrait at the moment21:46
T42<elros34> I am not aware of any way to make lipstick rotate window for you. You need to do it in app itself and update content orientation property (IIRC this is how it's called in qt) so edge swipes works correctly. This is why I make qxcompositor so I could easy rotate xwayland (or any other app) to landscape without modifing sources.22:16
T42<elros34> Take a look at kodi: it rotates its content to landscape but doesn't update content orientation so swipes are incorrect. Another nother not sillica/qml app is hextool22:19
KabouikInteresting @elros34, thanks a lot. Did you see ishimaru's question just before in #sailfishos?22:28
KabouikIt's the more accurate question worded by a dev who knows what he is talking about, as opposed to my question22:29
KabouikFor the record, we were trying to force landscape orientation only, not autorotation; perhaps that would make things easier?22:31
T42<elros34> I rarely use irc. So to sum up AFAIK client needs to rotate rendering22:31
Kabouik<ichimaru> With lipstick, in order to get landscape orientation, does the client app have to rotate all it's rendering? Or is it possible for the compositor to do the transform while the client renders as normal?22:32
KabouikBut I if I understand correctly, that's answer B22:32
KabouikErm I meant A22:33
KabouikI will forward to him your answer elros34, thanks again.22:34
T42<elros34> yes A or if you need only landscape then lazy way via nested compositor22:35
KabouikWe would only need landscape yes22:44
T42<ankaos> Elros fixed bootloop23:02
T42<ankaos> Thanks23:02
T42<edp_17> @ankaos : What was the reason and how did you solve it?23:06
T42<ankaos> Ofono service do mask23:06
T42<ankaos> Ofono service not use23:06
T42<edp_17> Oh, I thought it was something else. Thanks.23:09
T42<edp_17> @elros34: Can you give me an example for how I can run a command when an evdev_trace action (certain button press) happens, please?23:10
T42<edp_17> I know how to do something with a udev rule but I don't know whether that can be used for this or not. (I couldn't make it work with an udev rule bu I might did not do it correctly.)23:11
T42<ankaos> @edp_17 [Oh, I thought it was something else. Thanks.], otherwise I couldn't look more.23:13
T42<edp_17> Plus I have a question regarding the bluetooth. If I turn it on and do not do anything, after three minutes, it turns off automatically. Is this normal?23:13
T42<elros34> udev rules are different thing. Not sure if there is nice way to do this with bash. It's quite easy with qt23:14
T42<ankaos> @edp_17 [Plus I have a question regarding the bluetooth …], Yes.23:14
T42<ankaos> This way on Android23:14
T42<edp_17> @ankaos: I never experienced this on Android. My devices always kept the BT on until I manually turned it off.23:16
T42<ankaos> Config_bt_sleep23:16
T42<edp_17> @elros34: Oh boy! Qt? What I would like to achieve was to wake up the screen when I press the home button.23:17
T42<ankaos> is there anything like this?23:17
T42<ankaos> My port Bluetooth not work23:17
T42<edp_17> The evdev-trace shows the action but I don't know how can I trigger the 'mcetool -U' command with it.23:18
T42<edp_17> @ankaos: I do not have BT_SLEEP in this kernel.23:20
T42<elros34> Maybe lipstick already support such a thing. I remember there were support for 'back' touch key but not sure whether it was only for android apps or something else.23:21
T42<edp_17> @elros34: Think I found similar things to what you said.23:22
T42<edp_17> I have checked /usr/share/lipstick-jolla-home-qt5/compositor.qml and /usr/share/lipstick-jolla-home-qt5/compositor/HardwareKeyHandler.qml but couldn't get closer.23:22
T42<ankaos> mer-kernel-check/mer_verify_kernel_config at master · mer-hybris/mer-kernel-check · GitHub23:23
T42<ankaos> Miklos: please look23:23
T42<edp_17> Plus I have managed to add the tests of the buttons (volume up/down, back and home) into csd and they all work in there.23:24
T42<edp_17> @ankaos: Thanks, I am aware of this but as I said I don't have this in this kernel. I might need to add though. However it says "Causes problems with bluez thus disabling is recommended.", so I am not sure this is what I need.23:26
T42<ankaos> I am currently trying to deal with the wlan problem.23:27
T42<edp_17> What is the problem?23:27
T42<elros34> @edp_17 maybe this could be useful:
T42<ankaos> @edp_17 [What is the problem?], wlan module not found. ı dont have wlan.ko23:45

Generated by 2.17.1 by Marius Gedminas - find it at!