| T42 | <mintphin> should i try reclonining? | 00:01 |
|---|---|---|
| mal | usually no need to do that | 01:13 |
| T42 | <edp_17> @elros34 [you have it already installed. I am only sayin …], I've installed the patterns-sailfish-device-adaptation-hammerhead with zypper, then the ver --dup started to download packages.🤞 | 07:54 |
| T42 | <edp_17> Btw, when I did the OTA on my treltexx, all packages were installed but after the first start I had problems with starting apps (firejail issue). I manually needed to flash the hybris-boot.img to solve it, so it looks like the kernel wasn't installed as part of the OTA update. | 07:56 |
| T42 | <zinstack625> Did an aarch64 hybris-16 build and a switch from armv7hl to aarch64 seemingly broke camera, any news on that? | 11:41 |
| T42 | <zinstack625> Logcat/dmesg/journalctl are silent on that and GST_DEBUG warns about droideglsink not existing | 11:43 |
| T42 | <edp_17> Hi all, is there a way to debug a binary without source? I've tried strace to get more info but nothing useful is there. | 11:45 |
| mal | @zinstack625 are mini*service process running? | 12:12 |
| T42 | <zinstack625> mal: minimedia and minisf, don't think there are others | 12:13 |
| T42 | <zinstack625> Video playback and audio is fine, just camera | 12:15 |
| T42 | <zinstack625> Android side seems fine too, as logcat is silent (except the regulars like fingerprint and pmi) | 12:18 |
| T42 | <Verevka86> @zinstack625 [mal: minimedia and minisf, don't think there a …], Missings libandroidicu.so, this not work 😔 - | 12:27 |
| T42 | <Verevka86> https://github.com/SailfishOS-sagit/droid-config-sagit/commit/5a9cc0515375e479f98088c85f2da0a31dd34f0e | 12:27 |
| T42 | <zinstack625> @Verevka86 there is no apex in hybris-16 | 12:31 |
| T42 | <Verevka86> @zinstack625 [@Verevka86 there is no apex in hybris-16], thought it was about 17.1😊, 16.0 all good for me | 12:32 |
| T42 | <zinstack625> Yeah, 16.0 worked until I did aarch64 | 12:33 |
| T42 | <Verevka86> @zinstack625 [Yeah, 16.0 worked until I did aarch64], haven't checked yet | 12:33 |
| T42 | <zinstack625> I think 17.1 misses an sepolicy at system, but am too lazy to check | 12:34 |
| T42 | <lackingcheesepuffs> video recording is broken on my port, how do I debug it? | 12:36 |
| T42 | <zinstack625> @lackingcheesepuffs GST_DEBUG=3 (higher is more verbose, lower is less), logcat and journalctl, see if audio dies at a switch to video, try strace, probably some other ways too | 12:42 |
| T42 | <lackingcheesepuffs> @zinstack625 [@lackingcheesepuffs GST_DEBUG=3 (higher is mor …], how do I launch logcat in terminal? | 12:52 |
| T42 | <zinstack625> @lackingcheesepuffs /system/bin/logcat | 12:52 |
| T42 | <lackingcheesepuffs> ah thank you, I will try | 12:53 |
| T42 | <mintphin> same issue still | 13:13 |
| T42 | <zinstack625> Android 10 tree based builds can occupy just a slight over 8 gb, mine utilizes zram to keep alive. Check if you have enough ram | 13:40 |
| T42 | <zinstack625> Oh, and mind you, I turn everything off, my ram usage at start of build is 150 mb, and it still manages to swap | 13:42 |
| T42 | <zinstack625> When it starts actually building, memory usage is like 3gb I think | 13:43 |
| T42 | <mintphin> ok yeah i used zram and it's working fine now | 14:32 |
| T42 | <edp_17> @elros34 : I've applied the mousetracker manually on the n7000 and lipstick is complaining: file:///usr/share/lipstick-jolla-home-qt5/compositor.qml:24:1: module "com.mousetracker" is not installed import com.mousetracker 1.0 | 15:34 |
| T42 | <edp_17> log: https://paste.ubuntu.com/p/44PwgCZXVM/ | 15:34 |
| T42 | <edp_17> compositor.qml: https://paste.ubuntu.com/p/WmCvvNTZg3/ | 15:34 |
| T42 | <edp_17> other info: https://paste.ubuntu.com/p/6Wy9XPMScy/ | 15:34 |
| T42 | <edp_17> On the treltexx, it works. | 15:35 |
| T42 | <elros34> Did you install rpm? It has c++ plugin which provides that com.mousetracker | 15:35 |
| T42 | <elros34> ah according to logs you did | 15:36 |
| T42 | <edp_17> No, I only copied the files and modified compositor.qml. | 15:37 |
| T42 | <elros34> are you sure you copied it to correct path? com/jolla/ looks wrong | 15:38 |
| T42 | <elros34> try rpm -U --nodeps mousetracker*.rpm | 15:39 |
| T42 | <edp_17> I extracted the files from the .cpio in the .rpm. Then copied those into the path where they were in the .cpio. I'll double check it again. | 15:41 |
| T42 | <edp_17> @elros34 : Yep, the correct location of the mousetracker folder was /usr/lib/qt5/qml/com. It works there now. Thanks. | 15:50 |
| T42 | <mintphin> it's failing now | 15:53 |
| T42 | <mintphin> `FAILED: ninja: 'out/target/product/laurel_sprout/kernel', needed by 'out/target/product/laurel_sprout/boot.img', missin | 15:53 |
| T42 | <mintphin> g and no known rule to make it` | 15:53 |
| T42 | <mintphin> do i need to provide offsets? | 15:55 |
| T42 | <mintphin> wait | 15:57 |
| T42 | <mintphin> found the issue in the faq | 15:57 |
| T42 | <lackingcheesepuffs> tried building aarch64 with hybris-16.0, when running —mic it returns this | 16:09 |
| T42 | <lackingcheesepuffs> Unable to find package: patterns-sailfish-device-configuration-jasmine_sprout | 16:09 |
| T42 | <lackingcheesepuffs> but armv7hl was building fine | 16:09 |
| T42 | Ivan %lastname% was added by: Ivan %lastname% | 16:19 |
| T42 | <zinstack625> @lackingcheesepuffs try manually installing that with "sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -m sdk-install -R ./droid-local-repo/$DEVICE/droid-configs/patterns-sailfish-device-configuration-$DEVICE* | 16:29 |
| T42 | <zinstack625> I still haven't resolved the issue with camera on 16.0 aarch64, but killing minimediaservice seemed to have triggered that: https://pastebin.com/xCSuJKSG | 16:43 |
| T42 | <zinstack625> I hope that could have some info | 16:43 |
| T42 | <zinstack625> Hmm, why is minimediaservice 32-bit when all the system is 64? | 16:46 |
| T42 | <zinstack625> minimediaservice is 32-bit and libdroidmedia is 64, should that be an issue? | 16:48 |
| rinigus | @zinstack625: it could be OK. as I was told, some androids use 32bit bits for it | 17:51 |
| T42 | <zinstack625> rinigus: the silence everywhere is sort of worrying | 17:52 |
| rinigus | news from elsewhere: Purism L5 community member just made first VoLTE calls, from the look of it. those who are on matrix can check it out in #community-librem-5:talk.puri.sm room | 17:53 |
| rinigus | all was done by just enabling volte in the modem using AT commands | 17:53 |
| rinigus | tested in germany, one of the local providers | 17:54 |
| T42 | <elros34> @lackingcheesepuffs do not install that pattern in target. Update your droid-hal submodule instead | 17:56 |
| T42 | <zinstack625> Just looked at camera provider file and it is indeed 32 bit | 17:57 |
| T42 | <zinstack625> So should I build droidmedia 32 bit? If yes then how to do not hacky? | 17:58 |
| mal | @zinstack625 I remember some issues with camera on aarch64, you need to show logs and output from camera | 18:53 |
| T42 | <zinstack625> mal: logcat is silent when camera starts, there is nothing in dmesg and journal, only GST_DEBUG | 18:54 |
| T42 | <zinstack625> If GST_DEBUG is reasonable, how verbose should I set it? | 18:55 |
| mal | so you anything about missing libs | 18:55 |
| mal | *do you see | 18:55 |
| T42 | <zinstack625> No, and both 32 bit and 64 bit droidmedia libs are present | 18:56 |
| mal | have you tried to remove the gst cache? | 18:56 |
| T42 | <zinstack625> How do I do that? | 18:57 |
| mal | in ~/.cache/gstreamer-1.0 or something like that | 18:57 |
| mal | it's important you do that and then get GST_DEBUG to see what happens | 18:58 |
| T42 | <zinstack625> Oh, wait | 18:58 |
| T42 | <zinstack625> Segfault at libgstdroid.so | 18:58 |
| mal | if it sees some plugin loading fails it keeps the information in cache | 18:58 |
| mal | check also logcat now | 18:59 |
| T42 | <zinstack625> Logcat is silent, gst_debug revealed it's blacklisting libgstdroid.so | 19:04 |
| mal | anything helpful in gst_debug? | 19:16 |
| mal | or have you tried using gdb to check the crash? | 19:16 |
| T42 | <zinstack625> gdb doesn't stop at a segfault, idk why | 19:18 |
| T42 | <zinstack625> currently searching through 75k lines of gst_debug | 19:18 |
| mal | remember to remove the cache again before doing that | 19:18 |
| T42 | <zinstack625> Looks like it tries to load h263parse just before the error message | 19:24 |
| T42 | <zinstack625> Fun thing is that the "segfault" is a message in stdout | 19:25 |
| T42 | <zinstack625> or maybe a type name 'fraction' | 19:28 |
| mal | just to be sure, which gst.droid version are you using? | 19:34 |
| T42 | <zinstack625> mal: zypper info gstreamer1.0-droid says that version is 0.20201104.0-1 | 19:37 |
| T42 | <zinstack625> Not full gst_debug log is 7.5MB and 91k lines, can't really see anything suspicious there | 19:38 |
| T42 | <zinstack625> I could share it if anyone wants | 19:38 |
| mal | have you first tried less versbose | 19:42 |
| mal | *verbose | 19:42 |
| mal | anything in logcat when you do that test | 19:42 |
| T42 | <zinstack625> With GST_DEBUG=4, the only thing in logcat is some Adreno messages about libegl loading | 19:45 |
| T42 | <zinstack625> GST_DEBUG=3, some warnings about audio, but audio should be fine https://pastebin.com/JJxh2ab9 | 19:47 |
| T42 | <elros34> just an idea: did you clean mw sources (gst-droid) when you switch from 32 to 64 bit? | 19:49 |
| T42 | <zinstack625> I recloned hadk before building aarch64 | 19:50 |
| T42 | <zinstack625> But there are libraries present in both lib and lib64 | 19:51 |
| T42 | <zinstack625> Each different arch respectively, ofc | 19:51 |
| T42 | <elros34> so maybe gdb will show where it segfault | 19:52 |
| mal | it's quite difficult to help without seeing any output from anything | 19:53 |
| T42 | <zinstack625> gdb just doesn't stop: https://pastebin.com/S7HqJ6uT | 20:05 |
| T42 | <zinstack625> Logcat doesn't have anything: https://pastebin.com/SHuNP4fg | 20:06 |
| T42 | <zinstack625> Not a mention of camera in journalctl or dmesg | 20:06 |
| mal | @zinstack625 why you not reading the output of gdb | 20:07 |
| mal | "run with --gst-disable-segtrap --gst-disable-registry-fork and debug." | 20:08 |
| T42 | <zinstack625> "gdb —args jolla-camera --gst-disable-segtrap --gst-disable-registry-fork" is the same | 20:09 |
| mal | that's not how it works | 20:10 |
| T42 | <zinstack625> ? | 20:10 |
| mal | those are parameters to gdb | 20:10 |
| mal | not to jolla-camera | 20:11 |
| T42 | <zinstack625> And do I just pass them? | 20:12 |
| T42 | <zinstack625> gdb doesn't seem to know what any of those are | 20:12 |
| mal | ah, wait | 20:12 |
| mal | you can't use jolla-camera to debug the issue | 20:13 |
| mal | sorry | 20:13 |
| mal | you need to install gstreamer1.0-tools and then use gdb --args gst-inspect-1.0 --gst-disable-segtrap --gst-disable-registry-fork droidcamsrc | 20:16 |
| mal | @zinstack625 ^ | 20:19 |
| T42 | <zinstack625> stagefright, as it seems | 20:19 |
| T42 | <mintphin> Hey, I'm now configuring the kernel and there's this option I'm unsure about | 22:22 |
| T42 | <mintphin> I ran the kernel checker and it's throwing CONFIG_FW_LOADER_USER_HELPER as invalid, while it's not set | 22:23 |
| T42 | <mintphin> However CONFIG_FW_LOADER_USER_HELPER_FALLBACK is set | 22:23 |
| T42 | <mintphin> (to y) | 22:23 |
| T42 | <mintphin> Should I disable it? | 22:23 |
| T42 | <mintphin> Ah wait, I'm reading some previous messages and it seems to not affect functionality. | 22:24 |
| T42 | <elros34> keep both enabled | 22:28 |
| T42 | <mintphin> I'm getting some errors compiling the kernel | 23:02 |
| T42 | <mintphin> https://paste.ubuntu.com/p/5SmhMSZwVY/ | 23:02 |
| T42 | <elros34> did you enable some crypt options? | 23:05 |
| T42 | <mintphin> I just enabled the options mer-verify-kernel-config mentioned | 23:06 |
| T42 | <mintphin> https://paste.ubuntu.com/p/tRb3jFfgdc/ | 23:07 |
| T42 | <elros34> you mean optional extra filesystem? Do you need it for anything? If not do not enable any of them | 23:07 |
| T42 | <mintphin> ah ok | 23:08 |
| T42 | <elros34> fix errors, enable namespaces and don't worry about rest now. You can always enable more options later when you will have working gui | 23:12 |
| T42 | <mintphin> ok cool. it's compiling now | 23:38 |
| T42 | <mintphin> same issue | 23:50 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!