Monday, 2021-06-14

TheHexagon@elros34 I do not see libfdt here in 17 base https://github.com/LineageOS/android_hardware_samsung/tree/lineage-17.1/dtbhtool . My dumbness, sometimes scares me.06:11
TheHexagonHi guys. When I compile hybris-hal and droidmedia, I get this error. I do not know where this would come from: https://paste.ubuntu.com/p/VtJn5WffF6/07:55
T42<elros34> so what was the culprit? Did you use wrong repo?  skia is in default hybris-17 manifest but I am still not sure which base do you use.08:09
TheHexagoni had the wrong revision of hardware/samsung08:09
TheHexagonbtw i fixed the above error08:09
TheHexagonso leave it alone08:09
TheHexagonbruh i just realised i didnt have ccache installed08:11
TheHexagonbecause we chroot into the habuildsdk08:11
T42<b100dian> Hi, I have managed to make changes to ngfd so that it produces some vibrations with my weird drv.18:37
T42<b100dian> However I've only tested by copying an .so over18:37
malcan you show how those changes look like18:37
T42<b100dian> when building the image, I get ' conflicts with file from package ngfd-settings-sailfish`18:37
T42<b100dian> Immediately18:37
T42<b100dian> actually the first commit is pushed https://github.com/b100dian/ngfd/commit/8c723005a3a5d68f76a597380cc6082cccbe785c18:40
T42<b100dian> The config and other changes 'wip' are not clean yet, basically FF_RUMBLE is still detected on my device so I commented it out18:40
T42<b100dian> I will make a PR when I have something compatbile with existing version, my question now if there are any perils in replacing ngfd-settings-sailfish with the one packaged by the default ngfd spec18:41
T42<b100dian> or is there an alternative way to build the ngfd rpm without the settings18:41
T42uOtRSMvVnP98120 was added by: uOtRSMvVnP9812019:03
TheHexagonHi guys. I get this error which says that a certain version of glibc "GLIBC_2.23" is not found. Can anyone help. I am using Lineage 17.1 as a base. https://paste.ubuntu.com/p/mvrhy4wTmM/19:36
TheHexagonI have alreadey tried sudo apt update + sudo apt upgrade to no avail19:36
T42<elros34> @b100dian what you are trying to install so it conflicts with ngfd-settings-sailfish?19:51
T42<b100dian> Hi @elros34 , ngfd built from sources19:53
T42<b100dian> actually ngfd-settings-basic19:53
T42<b100dian> (is the one conflicting)19:54
T42<elros34> so don't, that package is not installed by default19:54
T42<elros34> frankly you should not change anything in patterns after building your custom ngfd19:54
T42<b100dian> ok, so this is probably a case where I shouldn't have used —mw but —build ..?(I'm using the former)19:55
T42<elros34> so when you have conflits? When you build image or when you build package, that is not same thing19:56
T42<b100dian> you're right, the error is on building with mw, sorry it sounded to me like an image creation error (Im looking at this for some time:)19:59
T42<b100dian> this is the full output, even with —build20:00
T42<b100dian> https://pastebin.ubuntu.com/p/hvmwgf2MyC/20:00
TheHexagonhey, for my error above, can i just compile the latest version20:00
T42<elros34> @b100dian then build_packages.sh --help should lead you to --do-not-install option20:01
T42<b100dian> thanks, this indeed doesn't give me the error, - but will the package be included?20:03
T42<elros34> whether package  will be included depends on package version, it doesn't matter whether is installed or not on target20:05
T42<b100dian> Thanks, I'll try this way. What I did was adding a version requirement in patterns (before you advised not to) and was seeing this error which made me believe that wont work:  "nothing provides ngfd>1.2.5 "20:11
T42<b100dian> Btw, thanks for pointing me to qt5-feedback-haptics-droid-vibrator  the other day @elros34 , this is next in line for constant and periodic+custom force-feedback support:)20:17
T42<elros34> I think you meant qt5-feedback-haptics-ffmemless? Anyway have you tried droid plugins? Maybe they will work good enough and you don't need any changes to memless plugins20:19
T42<b100dian> Sorry, ffmemless one20:21
T42<b100dian> droid plugins were loading /vendor/lib64/hw/vibrator.default.so and then failing to open `/sys/class/timed_output/vibrator/enable` and `/sys/class/leds/vibrator/activate` which I don't have20:35
T42<b100dian> It seems android is using `/vendor/bin/hw/vendor.qti.hardware.vibrator@1.2-service` which connects to /dev/input/event2 where I straced the actuall force-feedback calls from20:36
T42<elros34> so I guess new implementation is needed to get that vibrator@1.2 working20:46
T42<b100dian> I'm not sure, probably binder based? I'm not familiar with what qmi is20:47
T42<b100dian> I found it much easier to talk to the kernel directly - I find more documentation on linux than on android itself - or more readable?20:48
T42<b100dian> But don't hesitate to nudge me in the right direction :)20:49
T42<elros34> sure, nothing wrong using it directly. I am not that experienced but mal will probable know how to use tgat vendor.qti.hardware.vibrator@1.2-service20:50
T42<b100dian> Well, in this particular case, Xiaomi seems to have made an android specific kernel driver and only their vendor binary knows how to talk to it to make it vibrate. The usefulness of my method is not easily extensible to other devices, but then, between vendor blobs and kernel blobs, we have the source for the latter (in some cases)20:52
malit's always possible to add binder backend to ngfd plugin and qtfeedback21:00
malbut it's very rare to have that strange kernel drivers so not sure if it's easier to expand the support for memless drivers21:05

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