| Oksana | Going mildly insane. After several zypper refresh & update, Sailfish OS 3.2.1.20 didn't boot on Fxtec Pro1 anymore - stuck on splash screen. Flashing with Factory Reset tool didn't help - still stuck, just splashscreen now reset to default. | 00:56 |
|---|---|---|
| Oksana | So, I fastbooted into twrp, and with File Manager and terminal reduced size of Data from 55GB to 11GB (removed all videos downloaded from Wikimedia Commons, deleted *.part and so on), and am now backing up to a microSD card. | 00:57 |
| Oksana | Hopefully, my audio recordings, epub books, browser bookmarks+session/tabs+logins/passwords , Notes and such are saved into the back up. | 00:58 |
| T42 | <ankaos> @elros34 [@ankaos be more verbose, what config, which fi …], https://paste.ubuntu.com/p/987PVySKJK/ my defconfig cyanogenmod_d802_defconfig | 01:01 |
| T42 | <elros34> yeah I don't believe that is your real .config (too old kernel to have almost everything enabled). I asked about config.gz. I can't help you. you are still showing me mer-kernel-check against cyanogenmod_d802_defconfig which is pointless as I told you many times | 01:09 |
| T42 | <ankaos> https://paste.ubuntu.com/p/6R4NrRHp8d/ you asked about config.gz | 01:13 |
| T42 | <ankaos> @elros34 [yeah I don't believe that is your real .config …], yea this link :) | 01:15 |
| Oksana | TWRP backup format is... strange. I expected a zip ball, not a few data.ext4.win00 files. | 01:15 |
| Oksana | Trying to install Sailfish OS from zip. Failed to mount/data , Device or resource busy. Failed to mount /vendor (invalid argument) | 01:16 |
| T42 | <elros34> @ankaos ok so you didn't enable audit like I told you many times but lets leave it right now. How is your dmesg and journal now? | 01:17 |
| T42 | <ankaos> I deleted all the files. because it doesn't work. | 01:17 |
| T42 | <elros34> @Oksana: if /data is already mounted then there will be warning that it can't mount /data | 01:18 |
| Oksana | Hmm, I didn't manually mount it, but maybe, backup did mount it. | 01:19 |
| Oksana | Anyway, Sailfish OS installed fine, brand new, but, now I have got to somehow extract files out of TWRP back up. | 01:25 |
| T42 | <leewp14> Is anyone able to clone sensorfw? I keep getting rpc failed result=52. It worked fine yesterday. | 02:01 |
| T42 | <birdzhang> @leewp14 of course, rpc error means your network is not good | 02:07 |
| T42 | <leewp14> Thanks. I will investigate on my own… | 02:14 |
| r0kk3rz | Oksana: what did i tell you about zypper update :P | 04:09 |
| DylanVanAssche | mal (IRC): @irc_piggz:dylanvanassche.be So updating mesa is fine for you? You can keep a lower version in the latte repos? | 05:11 |
| DylanVanAssche | mal (IRC): I updated some packages in native:common to the latest version and fixed the gst-plugins-good package. | 07:13 |
| DylanVanAssche | Removing `talloc` builddep from `mesa` seems to result into a build failure: https://build.sailfishos.org/package/live_build_log/home:dylanvanassc/mesa/sailfish_latest_armv7hl/armv8el | 07:13 |
| DylanVanAssche | I have to go now. | 07:13 |
| DylanVanAssche | mal (IRC): Good news! Updating `libdrm` and `mesa` to the latest release fixed the building issue 🎉 | 08:41 |
| mal | DylanVanAssche: looks quite ok now, gst-vaapi still needs updating | 11:00 |
| mal | DylanVanAssche: wondering about mesa updating, in a way using same mesa version as official one would be nice but if native ports need newer then having newer makes sense | 11:01 |
| T42 | <leewp14> @adampigg : do you still have the sfos package for n7000? | 11:42 |
| deathmist | how does `/home` encryption work exactly? I'd like to include and test that on my port; is adding `sailfish-device-encryption` to patterns enough? | 11:46 |
| DylanVanAssche | mal (IRC): We have our own mesa in dontbeevil adaptation. I updated it to the latest mesa release. | 11:47 |
| DylanVanAssche | For lima driver it's the best way since it gets updates very frequently. | 11:47 |
| DylanVanAssche | mal (IRC): I didn't do gst-vaapi since it was disabled, was not sure | 11:47 |
| mal | DylanVanAssche: disabled? it's only used for x86 builds | 11:52 |
| mal | afaik latte uses it? | 11:52 |
| DylanVanAssche | Ah yes, it's only disabled for arm. | 11:53 |
| DylanVanAssche | I haven't resolved mesa on native-common since it's a version from piggz to stay latte compatible I think? | 11:53 |
| mal | DylanVanAssche: so just removing the talloc dependency wasn't enough? | 11:54 |
| mal | that is just an obsolete dependency that does nothing | 11:55 |
| DylanVanAssche | mal (IRC): Not really, I removed it and I had a build failure about some missing .so files. I took a wild guess and just updated mesa + libdrm, builds fine now :D | 11:56 |
| DylanVanAssche | I checked the release notes of libdrm and it seems that these .so files were available in a newer version than we have. | 11:56 |
| mal | DylanVanAssche: maybe you used wrong version, it was built from 19.1.4+git1 tag not from master | 11:56 |
| DylanVanAssche | mal (IRC): Hmmm maybe, but updating it wouldn't hurt I guess? | 11:57 |
| mal | well if latte works then updating is fine | 11:58 |
| mal | piggz: ^ | 11:58 |
| DylanVanAssche | I updated it in the dontbeevil repo, not in the native common. one If piggz can test the latest mesa, we can update it in native common as well | 11:59 |
| mal | sure, let's wait for that, you can fix gst-vaapi, should be simple submodule update | 11:59 |
| DylanVanAssche | mal (IRC): If I want to fix it, I need to change the build URL from piggz gitlab to mine. Maybe let piggz do that? | 12:01 |
| DylanVanAssche | mal (IRC): There is something weird now going on, after some time everything freezes on the screen since the mesa update. I would wait | 12:02 |
| mal | DylanVanAssche: ok, piggz can do it | 12:04 |
| deathmist | mal: any idea what's needed for `/home` encryption enablement on a SFOS port? installing `sailfish-device-encryption`, adding https://github.com/mer-hybris/droid-config-sony-ganges-pie/commit/52e273b7c63487ddb3ed16308e2be7801225e6e4 and enabling encryption from system settings leads to https://i.imgur.com/AbrXzs1.jpg on my device | 12:12 |
| DylanVanAssche | deathmist (IRC): AFAIK, you need a separate home and root partition. Not sure if that's the case for your port | 12:13 |
| deathmist | so I'd need Btrfs subvolumes because this is just a droid and there's no way I'll mess partition tables on this thing | 12:16 |
| rinigus | Deathmist: or use LVM, as xperias do | 12:23 |
| deathmist | .-. I have no clue how either of those work though | 12:23 |
| rinigus | Deathmist: as encryption is just a pin, it was very simple to decrypt. wasn't there someone who showed that its just few seconds/minutes if data is extracted to PC? | 12:25 |
| rinigus | ^ in the sense that is using pin to encrypt data | 12:25 |
| rinigus | ... that's as far as I understood from chats - not checked the code nor have deep understanding of how it works | 12:26 |
| deathmist | meh, I'll look into encryption later then | 12:53 |
| deathmist | if others here have already moved to 3.3.0.16, can you sign into Jolla accounts on initial setup screen if connecting to internet via WLAN instead of cellular data? for me the system date & time are just incorrect (and not auto-updating by default) and I get "Problem with network" | 12:54 |
| T42 | <elros34> for me auto update time is always checked at first boot but I didn't try to sign into to Jolla account. If I disable it once it doesn't use correct timezone anymore (AFAIK known bug) | 13:13 |
| piggz | mal: my firewall fix also worked for birdzhang | 13:21 |
| T42 | <birdzhang> yeah, thank you ;) | 14:01 |
| T42 | <birdzhang> @deathmist same issue on my two device, network problem | 14:02 |
| mal | hmm, very strange | 14:05 |
| deathmist | this wasn't an issue on 3.2.1, I've always connected to WLAN on the initial setup screen | 14:12 |
| mal | deathmist: afaik home encryption assumes the official way for the port meaning separate home which is not the case for community ports | 14:13 |
| mal | deathmist: btrfs is not used in new devices like sonys, those use the normal ext4 | 14:14 |
| mal | and lvm | 14:15 |
| mal | deathmist: I haven't yet built a 3.3.0 image so I need to test that | 14:16 |
| DylanVanAssche | mal (IRC): When I run `zypper ref` of a repo, the repo is not updated. Zypper says it is up to date but if you look on OBS, the package has a different version there. Any idea? | 14:58 |
| DylanVanAssche | nvm it seems that it needed some propagation. However, that's the first time that it takes so long | 15:02 |
| mal | DylanVanAssche: which OBS project? | 15:14 |
| DylanVanAssche | mal (IRC): Dont be evil: https://build.merproject.org/package/show/nemo:devel:hw:pine:dontbeevil/mesa | 15:55 |
| DylanVanAssche | I'm just not sure what the difference is between the list of packages on OBS (when you surf to the repo URL) and what zypper sees | 15:55 |
| deathmist | mal: any news on fingerprint MW for community devices btw? | 16:09 |
| mal | no news | 16:10 |
| DylanVanAssche | mal (IRC): Appearently, the flashlight plugin of Jolla Settings is closed source and hardcoded to droid stuff. | 16:42 |
| DylanVanAssche | Now for the PinePhone, this won't work. It tries to create a gstreamer source element which only works on droid devices. | 16:42 |
| DylanVanAssche | Do you have an idea on how we can fix this? Like writing such plugin ourselves? But no idea how to do that | 16:42 |
| mal | DylanVanAssche: yeah, it needs to be fixed somehow | 16:46 |
| mal | needs some discussion internally | 16:47 |
| DylanVanAssche | Or maybe OSS that plugin 😁 heheheh | 16:47 |
| DylanVanAssche | Okay, do I need to bring it up during a SFOS community meeting? | 16:47 |
| mal | DylanVanAssche: does camera work in pinephone yet? does flash work via gstreamer? | 16:48 |
| DylanVanAssche | mal (IRC): Cameras are not working yet (they work, but not via gstreamer yet). Flash is controlled via /sys/class/leds or using V4L2 APIs (more likely if you are using multimedia stuff) | 16:50 |
| mal | DylanVanAssche: first we need to see which way to use flash on that device | 16:55 |
| DylanVanAssche | mal (IRC): I would go for the V4L2 API, you need to implement the camera stuff anyway as a V4L2 API (which QtMultimedia supports). The GPIO interface is used for scripting stuff and such. | 16:56 |
| DylanVanAssche | That's the same way AFAIK as the hybris devices | 16:57 |
| abranson | looks to me like it would work through v4l2-request>libva>gst-va>qtmm | 16:58 |
| DylanVanAssche | Yep something like that | 16:58 |
| deathmist | https://wiki.merproject.org/ has been down for 8 hours at least today, doesn't look like maintenance stuff either | 17:01 |
| DylanVanAssche | mal (IRC): Mesa is fixed for the PinePhone. We need to use the 'master' branch, not the point releases. The LIMA driver fixes are not backported to the releases, so a release can be broken for LIMA. Master works fine for the PinePhone 🎉 but not suitable I guess for native-common? | 17:45 |
| mal | DylanVanAssche: maybe pinephone should use its own mesa for now then, hopefully next release will work for both, need to get information from piggz about latte first | 18:20 |
| T42 | <adampigg> argh, too much to do! | 18:25 |
| DylanVanAssche | mal (IRC): Sure :) | 18:38 |
| deathmist | mal: https://wiki.merproject.org/ is still down :/ | 19:13 |
| mal | deathmist: yeah, I informed the relevant people already | 19:13 |
| piggz | mal: experimenting with new mesa on latte.... | 19:32 |
| piggz | last working was 19.1.4 | 19:32 |
| mal | piggz: so the issue was never solved? | 19:40 |
| piggz | nope | 19:47 |
| mal | piggz: update bluetooth-rfkill-event-hciattach in latte project | 19:54 |
| piggz | ok | 19:55 |
| piggz | mal: bt rfkill event has a conflict with a system package now... | 20:03 |
| piggz | File /usr/sbin/bluetooth_rfkill_event | 20:03 |
| piggz | from install of | 20:03 |
| piggz | bluetooth-rfkill-event-hciattach-1.0.8-1.3.1.jolla.i486 (adaptation-community) | 20:03 |
| piggz | conflicts with file from package | 20:03 |
| piggz | bluetooth-rfkill-event-1.0.6+bluez5.20190210090636.1.g35e927a-1.2.1.jolla.i486 (@System) | 20:03 |
| piggz | mal: oh, this looks like fantastic news | 20:05 |
| piggz | @eugenio_g7 mesa 20.1.0-rc1 works .... | 20:06 |
| piggz | now mal will make me check the latest point release :/ | 20:06 |
| mal | :D | 20:09 |
| mal | piggz: why do you have two of those rfkill event packages installed | 20:10 |
| mal | I assume that other one is some test version | 20:10 |
| T42 | <eugenio_g7> piggz: \o/ | 20:11 |
| piggz | rebuilding 20.0.6 | 20:11 |
| piggz | mal: no idea atm ... ill look in a sec | 20:11 |
| piggz | mal: yeah, that version is gone now, it didnt have a repo | 20:13 |
| deathmist | mal: I tried logging in using cellular data instead of WLAN in initial setup, time & date got set correctly this time (and afterwards even had automatic update in system settings enabled) but I STILL get "Problem with network" and can't sign in, this is a new issue with 3.3.0, 3.2.1 did not have this issue | 20:20 |
| piggz | apparently, mesa 20 brought lots of intel fixes ... last version i tried was 19.3.3 | 20:20 |
| deathmist | and that older version could also update my time & date to be correct even if I choce WLAN during initial setup for networking instead of cellular data | 20:21 |
| piggz | mal: @eugenio_g7 mesa 20.0.6 is also safe for latte | 20:31 |
| piggz | mal: can bluetooth-rfkill-event-hciattach goto common? | 20:35 |
| mal | piggz: so if you have a good version of mesa updae native common | 20:42 |
| piggz | ok | 20:42 |
| mal | piggz: https://build.merproject.org/package/show/nemo:devel:hw:common/bluetooth-rfkill-event-hciattach | 20:44 |
| piggz | mal: gr8, can drop that.... | 20:44 |
| piggz | wonder what i should do about the broadcom-bluetooth-bluez5 .... | 20:54 |
| piggz | i could either PR it or drop it .... maybe it coul dbe dropped because it is just a build requres for bluez-libs | 20:54 |
| piggz | no, maybe its required | 20:55 |
| piggz | i should PR ir | 20:55 |
| mal | piggz: I replaced gst-vaapi in native-common | 20:59 |
| piggz | just seen | 20:59 |
| piggz | mal: at some point, mer-core should update its mesa and we should use that instead of mine ... mine is just a dumping ground of tags in random testing order! | 21:00 |
| mal | piggz: we would a separate spec for the driver version and that needs some discussion | 21:01 |
| mal | piggz: I'll replace libva also unless there is some good reason why it's old | 21:03 |
| piggz | mal: sure | 21:03 |
| piggz | or, | 21:03 |
| piggz | wait | 21:03 |
| piggz | i know the intel-vaapi-plugin is also v old | 21:03 |
| piggz | i think there was a reason, like build deps | 21:03 |
| mal | hmm | 21:04 |
| mal | I can revert it then | 21:04 |
| piggz | try it, if its no good we will find out | 21:06 |
| piggz | ive deleted mesa from latte repo now ... its a very slim repo! | 21:06 |
| mal | piggz: would be nice know what the issue with intel-vaapi-plugin was | 21:07 |
| piggz | mal: looking at the source ... ive made the version 1.5 for some reason | 21:08 |
| piggz | mal: its the libva and libdrm versions.... | 21:17 |
| piggz | those are a lot better than they used to be | 21:19 |
| piggz | so maybe less of an issue | 21:19 |
| mal | piggz: we seem to have new enough libdrm | 21:20 |
| piggz | mal: ok, if you update them, i will re-package intel-vaapi-driver on a rebased verison | 21:21 |
| mal | piggz: libdrm_dep = dependency('libdrm', version : '>= 2.4.52'), we have 2.4.97 | 21:21 |
| piggz | yeah, when i was packaging all this for latte originally, out libdrm was much older | 21:22 |
| mal | I'll update libva then | 21:22 |
| mal | to 2.6.1 | 21:22 |
| piggz | intel-vaapi-driver 2.4.0 requires libdrm 2.4.52 libva 1.4.0 | 21:23 |
| mal | ok | 21:24 |
| mal | hopefullythat new one works | 21:24 |
| mal | piggz: there has been quite much of mer-core updated so it's not that far behind anymore as it used to be | 21:24 |
| piggz | yup | 21:24 |
| piggz | mal: whats this doing https://build.merproject.org/package/live_build_log/nemo:devel:hw:xiaomi:latte/intel-vaapi-driver/sailfish_latest_i486/i586 | 21:37 |
| mal | piggz: libva-intel-driver-2.4.0+2.4.0.sailfish.20200503212748.ef1c1cd.tar.gz vs intel-vaapi-driver-2.4.0+2.4.0.sailfish.20200503212748.ef1c1cd | 21:39 |
| mal | note the name difference | 21:39 |
| piggz | ah | 21:40 |
| mal | piggz: btw, would be nicer to do that repo the usual submodule way | 21:41 |
| piggz | was thinking that | 21:41 |
| mal | why is the source not updating in that package | 21:51 |
| piggz | i dont know! | 21:52 |
| piggz | the package name matches the rpm name... it shouldnt be gotting intel-vaapi-driver from anywaher! | 21:53 |
| mal | piggz: look at the spec on OBS, that is the old one | 21:55 |
| piggz | mal: ill continue in the morning, its at least building now, and fialing later | 22:01 |
| piggz | mal: ok, i got it built | 22:06 |
| mal | good | 22:13 |
| T42 | <adampigg> Mal, oddly, had to specify revision | 22:28 |
| T42 | <elros34> IIRC obs doesn't like - in tags, that could be related | 22:40 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
