| Mister_Magister | oki i've slept through almost entirety of my work today time to do some portin' | 17:31 |
|---|---|---|
| Mister_Magister | lemme dig up logs so that mal will stop being angry at me | 17:31 |
| Mister_Magister | mal: https://github.com/mer-hybris/hybris-patches/pull/31 | 17:38 |
| Mister_Magister | oh shiet wait i messed up the patch | 17:38 |
| Mister_Magister | its fixed | 17:38 |
| Mister_Magister | so its done, i made the PR i notified ginko about the PR, i reverted hybris debug patch and now i have to fix fajita and enchilada zipsa | 17:41 |
| Mister_Magister | i got beer in one hand and keyboard in the other time to do some work :P | 17:42 |
| mal | Mister_Magister: merged, thanks | 17:53 |
| mal | it seems all other bases have that already so everything should be good now | 17:55 |
| Mister_Magister | mal: don't tell me you haven't done it yourself because then you couldn't review it yourself and then you would have to bother somebody so that they can review it so that you can merge it? | 18:05 |
| Mister_Magister | he probably like "not at all!" xd | 18:11 |
| Mister_Magister | mal: how… busy are you? I have a problem where in the zip aspect ratio is not being set, but if you update from older zip it works fine and i have no clue how to go about debugging it | 18:15 |
| mal | well I have pushed changes to lineage base branches in hybris-patches without review before | 18:17 |
| mal | reason why I didn't do that was that I didn't seem to have local checkout of hybris-16.0 source tree and to be sure that the patch applied without issues I would have wanted to have it | 18:18 |
| Mister_Magister | mal: but i literally just cloned the repo out of source to make the patch xd | 18:18 |
| mal | yeah, that probably would have been fine also | 18:19 |
| Mister_Magister | well anyway its done | 18:19 |
| Mister_Magister | now i need to figure out why this https://build.sailfishos.org/build/nemo:testing:hw:oneplus:fajita/sailfishos_4.6.0.13/aarch64/droid-config-fajita/_log doesn't set pixel_ratio | 18:19 |
| Mister_Magister | could it be because i added notch cutouts? | 18:19 |
| Mister_Magister | https://github.com/sailfish-oneplus6/droid-config-enchilada/blob/hybris-16.0/sparse/etc/dconf/db/vendor.d/screen-config.txt | 18:19 |
| mal | is the /etc/dconf/db/vendor.d/silica-configs.txt file correct in the zip? | 18:20 |
| mal | that config is wrong | 18:20 |
| Mister_Magister | wait | 18:21 |
| Mister_Magister | yeah that config is wrong | 18:21 |
| Mister_Magister | i lack the [] part | 18:21 |
| mal | afaik each file needs the proper sections for configs | 18:21 |
| mal | [desktop/sailfish/silica] | 18:21 |
| Mister_Magister | yeap | 18:21 |
| Mister_Magister | [desktop/sailfish/silica] | 18:21 |
| Mister_Magister | we literally arrived at the same thing at the same time xd | 18:21 |
| mal | what is the display resolution of that device? | 18:22 |
| Mister_Magister | something by something | 18:23 |
| Mister_Magister | um 1080 x 2340 pixels, 19.5:9 ratio (~402 ppi density) | 18:23 |
| mal | I think the notch config is not entirely accurate then, it does probably work though | 18:24 |
| mal | the x offset is a bit small | 18:24 |
| Mister_Magister | its scaled | 18:24 |
| Mister_Magister | @notkit said so, no clue why though | 18:24 |
| mal | what? | 18:24 |
| Mister_Magister | ye | 18:24 |
| mal | on fp5 I have cutouts = [[568, 0, 88, 103]] | 18:25 |
| Mister_Magister | huh | 18:25 |
| T42 | <Mister_Magister> @NotKit a word | 18:26 |
| Mister_Magister | its not like theres documentation for it | 18:26 |
| Mister_Magister | mal: while i have you here… how do i debug videos playing at 20fps | 18:37 |
| mal | do logs show anything suspicious? does gst report any dropped frames for example? | 18:43 |
| Mister_Magister | i'll try the GST_DEBUG as i haven't tried that one yet but nothing in journal | 18:46 |
| mal | also nothing in logcat? | 18:47 |
| Mister_Magister | sorry i meant logcat | 18:48 |
| Mister_Magister | DPI is back :) | 18:56 |
| Mister_Magister | mal: any info on "Device supports multiple back cameras, please define dconf /apps/jolla-camera/backCameraLabels"? | 18:59 |
| mal | does the device have more than 1 back camera? | 19:11 |
| mal | if so then add some label for them like this https://github.com/mer-hybris/droid-config-sony-murray/blob/master/sparse/etc/dconf/db/vendor.d/jolla-camera-hw.txt | 19:12 |
| Mister_Magister | yes it does | 19:14 |
| Mister_Magister | wish it hadn't | 19:14 |
| Mister_Magister | cause the other one is no different xd I think its used for depth | 19:15 |
| mal | ok, then adding the config might not be useful | 19:15 |
| Mister_Magister | yeah | 19:16 |
| Mister_Magister | mal: how bad of an idea would be to copy home from one device to another | 19:56 |
| mal | hmm | 20:11 |
| Mister_Magister | at least like .config and .local | 20:12 |
| Mister_Magister | maybe without overwrite | 20:12 |
| Mister_Magister | or i could just copy everything starting with harbour and my apps | 20:12 |
| Mister_Magister | that would be probably the safest | 20:13 |
| T42 | <b100dian> I am finally debugging video recording on A11 / codec2 and camera2 branch, with the newly-found HYBRIS_ENABLE_LINKER_DEBUG_MAP | 20:22 |
| T42 | <b100dian> And I have a crash that is a plain function call in bionic https://hastebin.skyra.pw/oxiluvoheb.swift | 20:23 |
| T42 | <b100dian> basically this https://github.com/LineageOS/android_frameworks_av/blob/lineage-18.1/media/codec2/sfplugin/C2OMXNode.cpp#L173 | 20:23 |
| T42 | <b100dian> given that the call comes from a thread spawned inside stagefright, could it have any relationship with hybris, maybe a hook is needed for that function too? | 20:26 |
| mal | could be, checking what that does | 20:54 |
| T42 | <b100dian> Presumably in this case it just lowers the API30 level from error to warning only (or it should) | 20:57 |
| T42 | <b100dian> "In API level 29, fdsan warns when it finds a bug. In API level 30, fdsan aborts when it finds a bug". https://android.googlesource.com/platform/bionic/+/master/docs/fdsan.md . I'm on 30 | 20:58 |
| T42 | <b100dian> I would like to understand is if hybris is in action when stagefright plugin calls what seems to be a bionic-only function | 20:59 |
| mal | @b100dian add a dummy hooks like the ones for other fdsan things like this https://github.com/libhybris/libhybris/blob/master/hybris/common/hooks.c#L2876 | 21:06 |
| T42 | <b100dian> on it | 21:06 |
| mal | and then the other part like this https://github.com/libhybris/libhybris/blob/master/hybris/common/hooks.c#L3338 | 21:07 |
| mal | this is how the function is defined https://github.com/LineageOS/android_bionic/blob/c9fb1e8640d34e81966f3ff874bc310cabc151bf/libc/bionic/fdsan.cpp#L339 check the parameters etc from that | 21:08 |
| T42 | <b100dian> Thanks. | 21:16 |
| T42 | <b100dian> Unrelatedly, trying to fork libhybris, and finding out I already had, and there's a commit that makes test-hwcomposer run on this adaptation https://github.com/libhybris/libhybris/commit/4699aaf42efe70334d1bc7b51cdec385c9c804dd, presumably by disabling/skipping HWC1 | 21:16 |
| mal | @b100dian any luck with libhybris? | 22:15 |
| T42 | <b100dian> had some trouble using directly `enum android_fdsan_error_level`, went for char because little-endian and that enum doesn't seem big. Now I have local packages, maybe I should just scp libhybris rpm? (As there are so many). Or just OBS build and zypper up | 22:20 |
| T42 | <b100dian> the android types are not immediately available, maybe I should have included some "Q" header, but I have not seen that in the current hooks | 22:21 |
| T42 | <b100dian> I'll grab libhybris-common.so.1.0.0 | 22:23 |
| T42 | <b100dian> Looks like the crash is not there, now there's a lockup of the viewport (and the whole app if you interact with it). Lemme build on OBS, to make sure there's no interdependencies between .so files, because gdb also doesn't show anymore symbols ;) | 22:39 |
| mal | check logcat for anything waiting for something | 22:42 |
| T42 | <b100dian> Ok, there is something. Recording starts around [FTS_TS] message (touched the screen) and ends in some DroidMediaCamera: setup_capture_session failed. But that's progress ! | 22:51 |
| T42 | <b100dian> https://hastebin.skyra.pw/raw/dofakurega | 22:51 |
| mal | btw, feel free to make a PR to libhybris for that hook | 22:56 |
| T42 | <b100dian> libhybris/libhybris, not sailfish-mirror I suppose | 22:59 |
| mal | yes, libhybris/libhybris | 22:59 |
| T42 | <b100dian> https://github.com/libhybris/libhybris/pull/569 | 23:06 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!