| T42 | <edp_17> On this forum, I don't think that need to mention no blobs. If I was happy with what current android offered, I wouldn't be here. :) | 06:11 |
|---|---|---|
| T42 | <edp_17> mal, @elros34 : Using focal chroot, the hybris-hal building has completed. I got some warnings and this error "cp: bad 'out/target/product/treltexx/boot.img.lok': No such file or directory" but "build completed successfully". | 07:22 |
| T42 | <edp_17> @elros34 : When building droid-hal, I got this https://paste.ubuntu.com/p/KZw6C8Z6gB/ | 07:38 |
| T42 | <edp_17> I've seen you recommended to comment out /default.prop from droid-hal-device.inc. | 07:38 |
| T42 | <edp_17> Is that really the solution? If so, why this wasn't a problem on other android bases? | 07:39 |
| Thaodan | I see you run a 3.10 kernel that can be an issue later down the line since older kernels usually have bugs when it comes to mount bind | 07:49 |
| T42 | <edp_17> Thaodan: Yeah, I know that is an old kernel. Fingers-crossed it will be okay. | 07:51 |
| T42 | <edp_17> When I commented that /default.prop in droid-hal-device.inc, I got other errors: https://paste.ubuntu.com/p/7Yn5RzW3Tv/ | 07:52 |
| Thaodan | check if straggler files is up to date | 07:52 |
| T42 | <edp_17> Thanks. | 07:54 |
| T42 | <edp_17> It looks like there are some straggler_files defined. Perhaps I need to add more from this error. | 07:55 |
| T42 | <edp_17> Hmm, except the default.prop, all those are already added to straggler_files. | 08:17 |
| T42 | <edp_17> Okay, I've found the solution. Comment /default.prop in rpm/dhd/droid-hal-device.inc AND remove all straggler files from rpm/droid-hal-$DEVICE.spec. This way, droid-hal has built. 👍😃 | 08:28 |
| jusa | piggz: commented to your commit | 09:36 |
| T42 | <adampigg> Ooo, thx | 09:37 |
| piggz | jusa: still goes to speaker, and toggling doesnt put to earpiece | 09:56 |
| jusa | hmh can you paste pulseaudio logs? | 10:02 |
| piggz | jusa: ofc... | 10:03 |
| piggz | jusa: here is everything https://gist.github.com/piggz/514062ed4afc9d9bf48163ba3553ef1b .... starts with a journalctl -b, then a journalctl -f as i make an incoming call .... there are some gaps where I have annotated such as << about to press speaker button >> | 10:13 |
| piggz | mal: Thaodan: any idea why lipstick would fall back to swrast with new mesa? some devs trying to tell me its a compostor issue. https://paste.mozilla.org/atocH50y#L61 | 11:01 |
| Thaodan | Why is it a compositor issue? | 11:25 |
| Thaodan | piggz: ^ | 11:25 |
| piggz | Thaodan: *shrug* | 11:25 |
| piggz | they say the compositor chooses the driver | 11:26 |
| piggz | personally i dont agree, but ive not much experience with all these egl internals | 11:27 |
| piggz | Thaodan: so, lipstick itself is rendering with LIMA i think ... but, then when apps open, they use SWRAST, would that be right? | 11:30 |
| piggz | because, if i remove swrast-dri.so, lipstick still opens, but apps cant load | 11:30 |
| piggz | and they complain about missing swrast | 11:31 |
| T42 | <elros34> maybe missing some env variable | 11:31 |
| jusa | piggz: I guess journal rate limiter is on? :/ | 11:32 |
| Thaodan | yeah | 11:33 |
| Thaodan | piggz: try opening fingerterm and then run env | 11:33 |
| Thaodan | then you can see if something is missing | 11:33 |
| jusa | piggz: hey, could you use this to get the logs? https://gist.github.com/jusa/91a86d589b0c2c58e95c9d63247686e8 | 11:33 |
| piggz | jusa: sure.. | 11:34 |
| jusa | run once, reboot, try the calling thing a bit, then re-run the scirpt | 11:34 |
| piggz | jusa: Thaodan: battery died, both jobs stalled whil i take car to garage! | 11:39 |
| piggz | Thaodan: fingerterm env https://paste.mozilla.org/jVxDp7de | 13:05 |
| piggz | jusa: https://www.piggz.co.uk/index.php/s/X3zK7jA9wFqnSmT | 13:10 |
| mal | piggz: have you checked strace for the mesa issue? | 13:15 |
| Thaodan | piggz: Did you start the app from fingerterm? | 13:15 |
| piggz | Thaodan: i ran env from fingerterm yes... and it was painful! | 13:16 |
| piggz | mal: do irun strace on an app, or on lipstick? | 13:16 |
| mal | not sure :) | 13:17 |
| Thaodan | piggz: I mean the example app from there | 13:17 |
| Thaodan | strace on the example app | 13:18 |
| Thaodan | lipstick would be much harder to decifer | 13:18 |
| piggz | Thaodan: which example app? | 13:18 |
| piggz | if a run strace on fingerterm, i can see swrast_dri.so get loaded | 13:20 |
| Thaodan | piggz: the app you got the mesa output from | 13:21 |
| Thaodan | Is there no other dri present? | 13:22 |
| piggz | ah, eglinfo | 13:22 |
| Thaodan | yes that one | 13:22 |
| Thaodan | I assumed a.out = example app | 13:22 |
| piggz | yeah, i just ran gcc on the source file... | 13:23 |
| piggz | Thaodan: this is eglinfo ran from fingerterm. https://paste.mozilla.org/LZndLwFy ... you can see GBM, Surfaceless and Device platforms use lima, but wayland is using swrast | 13:26 |
| piggz | egl debugging looks like https://paste.mozilla.org/wRuo8Q9a | 13:29 |
| mal | piggz: do you have full build log of mesa build? | 13:30 |
| piggz | mal: yes, its on obs... | 13:31 |
| piggz | https://build.merproject.org/build/nemo:devel:hw:native-common/sailfish_latest_aarch64/aarch64/mesa/_log | 13:31 |
| mal | piggz: use build.sailfishos.org :) | 13:32 |
| mal | instead of that old url | 13:32 |
| piggz | mal: when i type in build. ... ff always siggests old url :) | 13:34 |
| mal | can you grab a strace of that same run you did just now? | 13:37 |
| Thaodan | https://doc.qt.io/qt-5/embedded-linux.html offers some debug helpers notably enabling these debug categories: qt.qpa.egldeviceintegration and t.qpa.eglfs.kms | 13:37 |
| piggz | mal: https://paste.mozilla.org/Z3nHXp2w | 14:06 |
| T42 | <elros34> in this strace Wayland platform is not loaded at all: wayland-0 -> ECONNREFUSED | 14:52 |
| piggz | yeah, sorry, i messd that one up... | 14:57 |
| piggz | @elros34: mal: https://paste.mozilla.org/o0KQzGZV | 14:58 |
| T42 | <elros34> so it makes connection to compositor and then decide to use swrast_dri.so | 15:04 |
| mal | where is the selection done? | 15:05 |
| T42 | <elros34> somewhere between 858-874, so you really needs to debug in at higher level than strace | 15:08 |
| T42 | <elros34> maybe it would be easier with https://github.com/sailfishos/qtwayland/tree/mer-5.6/examples/wayland/qml-compositor instead lipstick | 15:09 |
| T42 | <edp_17> Building hybris-hal has been running over 7 hours now. Is it normal? (los 17.1) | 15:50 |
| T42 | <Mister_Magister> depends on cpu | 16:03 |
| Mister_Magister | and ram | 16:03 |
| Mister_Magister | and memory speed | 16:03 |
| Mister_Magister | and disk speed | 16:03 |
| T42 | <edp_17> Thanks. Great. :( This process is killing my PC. (i3 - 4 cores, 8GB ram, sata hdd) | 16:19 |
| T42 | <edp_17> Building 14.1 takes few minutes but this... | 16:20 |
| mal | @edp_17 did you remember to apply patches? | 16:21 |
| mal | although I would assume it would break if patches were not applied | 16:22 |
| T42 | <edp_17> Nope. I mean not yet. | 16:23 |
| T42 | <elros34> so you are building android:) | 16:24 |
| T42 | <edp_17> Oh, you were talking about those in hadk-faq (hybris 16 and 17.1)? Yes, I've applied them. | 16:25 |
| T42 | <edp_17> I've already built this hybris-hal yesterday and it didn't take this long. | 16:25 |
| T42 | <edp_17> But I had to replace the trusty chroot with focal and updated something else in kernel, then this thing started to rebuild hybris-hal from scrach again. | 16:27 |
| T42 | <edp_17> Now it's completed. And the best part: "failed to build some targets (06:21:45 (hh:mm:ss))" | 16:28 |
| T42 | <elros34> you forgot about -jX? | 16:28 |
| T42 | <edp_17> Aahh, I can start it again. 😭 | 16:28 |
| T42 | <edp_17> No, I didn't forget that. I used: make -j$(nproc --all) hybris-hal droidmedia | 16:29 |
| mal | did you see any progress in the build or was it stuck? | 16:30 |
| T42 | <edp_17> I've seen the progress. (percentage counter counting and the files in compilation were shown) | 16:32 |
| T42 | <edp_17> Should I build audioflingerglue with command: "make -j$(nproc --all) $(external/audioflingerglue/detect_build_targets.sh $PORT_ARCH $(gettargetarch))"? | 16:34 |
| T42 | <edp_17> mal: @elros34 ^^ | 16:39 |
| T42 | <elros34> no | 16:43 |
| T42 | <edp_17> Great. Thanks. | 16:43 |
| T42 | <edp_17> Does it hurt if I did? 😊 | 16:45 |
| T42 | <edp_17> Well, it's just failed, so doesn't matter anymore. | 16:46 |
| *** piggz_ is now known as piggz | 17:16 | |
| deathmist | @slavamon https://github.com/mer-hybris/dummy_netd/pull/4 (& https://github.com/mer-hybris/bluebinder/pull/22) :) | 18:22 |
| T42 | <Hogir> Hello! Can you help me? | 18:48 |
| T42 | <edp_17> Hogir : It depends. :) With what? | 19:16 |
| piggz | deathmist: who would use a non systemd system??? :D | 19:19 |
| Thaodan | piggz: Esp on a mobile device.. | 19:19 |
| Thaodan | but I guess NIH people | 19:19 |
| deathmist | piggz: I do in fact and I prefer it that way for various reasons. I also would like to be able to do so on my phone in the very near future | 20:26 |
| T42 | <Hogir> I want to switch my phone system to Selfish (re @edp_17: Hogir : It depends. ...) | 20:38 |
| T42 | <edp_17> Hogir : What phone? | 20:52 |
| T42 | <edp_17> Oh, and that's Sailfish not selfish. | 20:54 |
| T42 | <Mister_Magister> i hate it when that happens | 22:36 |
| T42 | <Hogir> Redmi note 8 pro (begonia) (re @edp_17: Hogir : What phone?) | 23:31 |
| T42 | <Hogir> Sorry about the mistake | 23:33 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!