| ohyash | anyone who's facing auto shutdown with salifish 3? | 06:21 |
|---|---|---|
| T4 | <elros34> @ohyash: yes, reboots/kernel panic in 3.0.1.11 due to new firewall rules. Check your dmseg (/dev/last_kmsg or pstore) | 08:32 |
| piggz_ | mal: for the UI, should we stick to the light-on-dark icons, or also have dark-on-light? | 09:08 |
| T4 | <adampigg> @eugenio_g7 hows the sensors? | 09:38 |
| T4 | <eugenio_g7> @adampigg they work if I flash my self-compiled droid-hal, but I need to restart sensorfwd anyways (so there is also a timing issue) | 09:52 |
| T4 | <eugenio_g7> With the dhd in OBS the iio devices don't get created at all | 09:52 |
| T4 | <eugenio_g7> Will clean up things and retry a kernel build tonight | 09:53 |
| T4 | <adampigg> @eugenio_g7 ah, i wonder if there is an old hal on obs, i will upload my current one | 09:54 |
| T4 | <adampigg> @eugenio_g7 huh, no, the version number matches | 10:04 |
| T4 | <eugenio_g7> Yea, the git rev seems correct | 10:10 |
| mal | piggz_: need to see how official icons are done, maybe those have some other way | 10:18 |
| mal | piggz_: I think it's possible to color the icons programmatically | 10:18 |
| kimmoli | iconprovider can do the trick | 10:20 |
| masha11 | mal: Hi. I check the contents of the files fstab in the folder out/. This is https://pastebin.com/sedDTtgs. What can you say about it? | 12:25 |
| mal | masha11: those are not ok, you might need to run something like "make fstab.griffin" in HABUILD_SDK | 12:36 |
| *** OhYash1 is now known as ohyash | 12:44 | |
| piggz_ | kimmoli: so, using this https://together.jolla.com/question/44325/iconbutton-how-to-use-own-icons-with-highlight/ ? | 12:59 |
| piggz_ | do i need to have mutiple icons in each color style? | 13:00 |
| kimmoli | would need to check how it is done on native apps | 13:00 |
| kimmoli | *jolla | 13:00 |
| kimmoli | but with that custom icons gets highlighted color, but how is the normal theme color passed (?) | 13:01 |
| kimmoli | icons are transparent+white with alpha, that just converts white to whatever is passed after ? argument in filename | 13:01 |
| piggz_ | kimmoli: good, my custom icons are transparent white with alpha... | 13:11 |
| mal | piggz_: kimmoli colorScheme in https://sailfishos.org/develop/docs/silica/qml-sailfishsilica-sailfish-silica-theme.html/ | 13:19 |
| mal | piggz_: kimmoli there are also dark/light colors which can then be used based on color scheme | 13:20 |
| kimmoli | aa there are methods | 13:20 |
| piggz_ | mal: i saw that yesterday, and wastn sure if it was a bool or an enum! | 13:21 |
| mal | piggz_: I can have a look later today | 13:21 |
| kimmoli | colorScheme is bool. the description does not make sense for a boolean? | 13:21 |
| piggz_ | kimmoli: my thoughts! | 13:21 |
| mal | piggz_: do we want the menu to change color based on light/dark ambience | 13:24 |
| mal | kimmoli: it's actually an enum with only two values which is same as boolean, comparing colorScheme is done against Theme.DarkOnLight and Theme.LightOnDark | 13:26 |
| mal | like this Theme.colorScheme == Theme.DarkOnLight | 13:26 |
| kimmoli | pl | 13:27 |
| kimmoli | ok | 13:27 |
| mal | in a sense having it descibed as boolean is a bit wrong | 13:27 |
| piggz_ | mal: ideally we might want to change the whole menu ... but, tbh, in this instance, i think just the dark sceme works well, whichever scheme is set for other apps | 13:27 |
| piggz_ | and, as the background is the viewvinder, the chances are that the dark buttons will be more visible, more of the time | 13:28 |
| masha11 | mal: Please tell me how to run correctly "make fstab.griffin". I get an erro: https://pastebin.com/ULP0RUzT | 13:32 |
| T4 | <elros34> @masha: use: "make modules | grep fstab" to figure out correct command | 13:40 |
| masha11 | T4: thank you. I got the following: "fstab.qcom libfstab libfstab_32" | 13:43 |
| masha11 | T4: Should I run it: "make fstab.qcom"? | 13:45 |
| T4 | <elros34> I guess | 13:47 |
| T4 | <elros34> I see only fstab.qcom here https://github.com/LineageOS/android_device_motorola_griffin/blob/lineage-15.1/rootdir/etc/fstab.qcom | 13:48 |
| rinigus | piggz_: you could use colorScheme as bool as well. true corresponds to light background, false dark. background itself should be available via Theme.overlayBackgroundColor , other colors as usual. in pure maps, I just have light (white) and dark (black) icons and swap between them depending on Theme.colorScheme | 13:52 |
| piggz_ | rinigus: is it possible to 'force' a scheme/icon set for the app? | 13:53 |
| masha11 | Thank you. Now my fstab looks like this: https://pastebin.com/eYtbMfjr | 13:54 |
| piggz_ | im not sure light icons wil dark text will work well against a viewvinder, esp ot doors | 13:55 |
| mal | piggz_: yep, we probably need to keep some parts of the UI dark always | 13:58 |
| *** OhYash1 is now known as ohyash | 13:59 | |
| piggz_ | mal: yup, thats my thoughts ... so, i either need to override the icon theme in the case where I get icons from the system, or, i could just copy the icons i need | 13:59 |
| piggz_ | not sure if thats allowed | 14:00 |
| rinigus | piggz_: not that I am aware of. in some respect maps are probably similar to camera app - but it does work ok on light ambience. if you want to force a scheme, just use the theme colors if they are according to the dark ambience and otherwise set your own | 14:00 |
| rinigus | ... ahh, you use system icons. then not sure if it will work | 14:00 |
| piggz_ | rinigus: its not the colours that are the issue, its when I get the icon using image://theme/,,,,, | 14:00 |
| piggz_ | yup :) | 14:01 |
| rinigus | piggz_: make an open call somewhere and ask for icons. | 14:01 |
| piggz_ | or just hard-code the full path to the dark icon ;) | 14:02 |
| mal | piggz_: it's trivial to override theme | 14:03 |
| mal | we can just use darkPrimaryColor etc | 14:03 |
| piggz_ | mal: its the icon selection we need to override | 14:03 |
| mal | piggz_: I'll do some testing later today | 14:04 |
| piggz_ | k | 14:05 |
| masha11 | To build rpm/dhd/helpers/build_packages.sh , the version of Target and Tooling must match the version of Sailfish? For example 3.0.0.8. | 14:07 |
| masha11 | mal: Now my fstab looks like this: https://pastebin.com/eYtbMfjr . This is normal? | 14:08 |
| mal | yes | 14:08 |
| masha11 | mal: ok. What about the versions Target/Tooling and Sailfish? | 14:10 |
| mal | just read what the warning box in hadk pdf says | 14:12 |
| masha11 | mal: ok | 14:15 |
| mal | in chapter 6 | 14:15 |
| piggz_ | how did i not know about https://maemo-leste.github.io/ :D | 14:16 |
| mal | piggz_: I heard about that some time ago | 14:19 |
| piggz_ | just realised mido has bluez5 listed in the ofono noplugin config ... what would that effect? | 14:37 |
| *** ChanServ sets mode: +v T4 | 14:47 | |
| mal | piggz_: do you have something like this https://github.com/mlehtima/droid-config-fp2-sibon/blob/master/rpm/droid-config-fp2-sibon.spec#L18 | 14:53 |
| piggz_ | mal: sort of, not not as detailed, I only enable hfp_ag_bluez5 | 14:56 |
| piggz_ | and dont disable any | 14:56 |
| piggz_ | i wonder if that contributes to my PBAP not working | 14:57 |
| piggz_ | i will test as soon as I get my car back from the garage! | 14:59 |
| mal | piggz_: xa2 has the same as fp2 so maybe try that | 15:04 |
| piggz_ | mal: will do, missin pbap is been a pain ... however, still people complain about it not working on official devices, there is a port today on tjc | 15:05 |
| piggz_ | for me, on jollaC, it seemed to stop working at 2.1.1 | 15:06 |
| piggz_ | but, maybe it will work | 15:06 |
| mal | piggz_: also maybe try updating droid-config submodule, ref https://github.com/mer-hybris/droid-hal-configs/commit/9f92d40ee2a81253cf43734adc8e43d443c4c590 | 15:12 |
| piggz_ | mal: on it | 15:17 |
| mal | piggz_: there is also another commit related to that config | 15:18 |
| piggz_ | mal: does pbap work on fp2? | 15:19 |
| mal | piggz_: no idea what pbap is :) | 15:20 |
| piggz_ | mal: where you connect to a car, and ti displays your contacts | 15:21 |
| piggz_ | phone-book-access-profile | 15:21 |
| mal | piggz_: I don't have a car | 15:21 |
| heroic_1 | hey, any idea why I'm missing "droid-system"? I have these installed: https://del.dog/oxewuwarof.sql | 15:43 |
| heroic_1 | and they should provide it: https://github.com/mer-hybris/droid-system-device/blob/860e7526d705e98f29974890c8fc73725970d7c0/droid-system.inc#L19 | 15:43 |
| heroic_1 | yet mic complains: "repo problem: nothing provides droid-system needed by droid-hal-kagura-0.0.6-201902181359.armv7hl" | 15:44 |
| mal | normally ports use the system partition and not the packages system | 15:44 |
| mal | *packaged | 15:44 |
| heroic_1 | I'm using the rootfs on the userdata partition via lvm | 15:46 |
| heroic_1 | This is more me being puzzled at how rpm works | 15:46 |
| heroic_1 | Shouldn't I see a sort of meta-package for "droid-system" when I search via zypper? | 15:46 |
| mal | droid-system is optional | 15:51 |
| heroic_1 | https://github.com/marina-bay/droid-hal-sony-tone/blob/sfos-3.0.1.14-android-8.1/droid-hal-kagura.spec#L59 | 15:51 |
| heroic_1 | I need it to pull in my own droid-system-kagura | 15:51 |
| mal | heroic_1: https://github.com/mer-hybris/droid-config-sony-nile/blob/master/rpm/droid-config-h4113.spec#L32 | 15:51 |
| heroic_1 | I know, but I want to try to use the lvm method | 15:52 |
| heroic_1 | Just to get the hang of it. I have already succeeded in building, this is more me wondering why when a package "Provides: " something, rpm still complains it is missing | 15:52 |
| mal | that is not related to lvm | 15:52 |
| heroic_1 | how can that be? | 15:53 |
| mal | you are misunderstanding it, even if sfos rootfs is using lvm then system partition can be either packaged to a rpm or mounted the usual way | 15:53 |
| mal | heroic_1: lvm is about sfos rootfs/home stuff | 15:54 |
| heroic_1 | yes, of course | 15:54 |
| mal | not about system partition | 15:54 |
| heroic_1 | but I want to try to package it | 15:54 |
| mal | I have never done that | 15:54 |
| mal | maybe take one step at a time, first make lvm build work and only after that try packaging droid-system | 15:55 |
| heroic_1 | just forget about the specifics of it for now. just in general, how can it be that the package "Provides: " something, is installed(on the target) and rpm still complains that it is missing | 15:55 |
| mal | also droid-system is a bit problematic in community ports, it might contain vendor libs and distributing those might cause issues | 15:56 |
| heroic_1 | I have already made the lvm build work by just keeping the obs_build guard and manually copying | 15:56 |
| heroic_1 | for sony open devices all vendor blobs are on the /odm partition | 15:56 |
| heroic_1 | everything in /system and /system/vendor is OSS | 15:56 |
| heroic_1 | we don't pull shit like lineage where they extract files from stock | 15:57 |
| mal | ok | 15:57 |
| heroic_1 | It takes 5 minutes for mic to refresh the repos. Is this the case for you as well? Is jolla throttling connections for community builds? | 15:58 |
| heroic_1 | It should only be checking the index file hashes, I mean how long can that take? | 15:59 |
| heroic_1 | In arch linux, it's instant if nothing changed | 15:59 |
| mal | no, refresh is very quick | 15:59 |
| mal | maybe you have something wrong in your .ks, pastebin the file | 15:59 |
| heroic_1 | it's a bit dirty https://del.dog/favojemeti.sh | 16:00 |
| mal | why @jolla-configuration-f8331 ? | 16:00 |
| mal | that's not supposed to be like that | 16:01 |
| heroic_1 | it wouldn't find the @Jolla Configuration f8331 one | 16:01 |
| mal | did you regenerate patterns properly? | 16:01 |
| heroic_1 | yes | 16:02 |
| heroic_1 | "build_packages.sh --configs" and then "process_patterns.sh" | 16:02 |
| mal | then something is wrong, @Jolla Configuration f8331 should always work | 16:02 |
| heroic_1 | let me try again with @Jolla then | 16:02 |
| heroic_1 | but it's gonna be 5 minutes before I can tell you whether it worked | 16:03 |
| mal | are you sure you don't have some issue with the naming, you seem to have a mixture of kagura and f8331 | 16:03 |
| mal | for example is droid-local-repo/kagura the correct one, and not f8331? | 16:04 |
| heroic_1 | yes, the local repo path is correct | 16:04 |
| heroic_1 | do you have some sort of credentials for the repos set up? the delay when refreshing is making development incrediably frustrating | 16:05 |
| heroic_1 | when I was building the "mw" packages, it took 5 minutes in between *each* package. who has the patience to tolerate that? | 16:06 |
| mal | which of the repos is slow? | 16:08 |
| mal | refresh those manually using: sb2 -t $VENDOR-$DEVICE-$PORT_ARCH -m sdk-install -R zypper ref | 16:09 |
| mal | and see how it behaves | 16:09 |
| heroic_1 | each takes about 1 minute | 16:12 |
| heroic_1 | apart from the local repo, which is almost instant | 16:12 |
| mal | I just tried and all worked fine, no delay at all | 16:13 |
| mal | even forced refresh had no delay | 16:14 |
| heroic_1 | for me it's the same with the manual command, ~30s-1min each | 16:15 |
| heroic_1 | also, almost no network activity apart from a few spikes | 16:15 |
| heroic_1 | when I download the sdk tarballs, I have ~5mb/s speeds | 16:15 |
| heroic_1 | and the actual package downlaods are very fast as well | 16:15 |
| mal | no idea what could cause that | 16:16 |
| heroic_1 | I hope it's not something between the chroot and arch linux | 16:17 |
| heroic_1 | using NetworkManager on the host | 16:17 |
| heroic_1 | is there a flag I can give to zypper to tell me in more detail what it is doing right now? | 16:17 |
| mal | --verbose | 16:18 |
| mal | or -v | 16:18 |
| heroic_1 | let's see then whether it spends too much time waiting for connection or whether it is the post-processing that is slow | 16:20 |
| heroic_1 | I changed the "%package" back to "@Jolla Configuration f8331" and it seemed to work now btw | 16:22 |
| heroic_1 | the "Retrieving: repomd.xml" stage for the jolla repos takes forever | 16:43 |
| heroic_1 | the download from "repo.merproject.org" is almost instant however | 16:51 |
| heroic_1 | when I open the page in the browser, it loads instantly, e.g. "https://dvvwui2qyedk6.cloudfront.net/releases/3.0.1.11/hotfixes/armv7hl/repodata/repomd.xml" | 16:54 |
| heroic_1 | same for the linked .gz files | 16:55 |
| heroic_1 | spooky | 16:55 |
| vknecht | any thought about this udev/DEVLINKS error (14.1 base) ? offending line is the last one of https://pastebin.com/fyK4rGkR | 18:09 |
| vknecht | Jan 01 02:23:56 Sailfish systemd-udevd[507]: unknown key 'DEVLINKS' in /lib/udev/rules.d/999-android-system.rules:714 | 18:09 |
| vknecht | Jan 01 02:23:56 Sailfish systemd-udevd[507]: invalid rule '/lib/udev/rules.d/999-android-system.rules:714' | 18:10 |
| vknecht | journalctl : https://pastebin.com/mxrH6Nnc | 18:10 |
| vknecht | found similar error at http://merproject.org/logs/%23sailfishos-porters/%23sailfishos-porters.2018-12-02.log.html but not sure what to do | 18:11 |
| mal | vknecht: that file is automatically generated during build, have you modified it yourself? | 18:25 |
| vknecht | mal, I don't think so | 18:29 |
| vknecht | neither on device nor in sparse | 18:31 |
| mal | vknecht: does that cause some issues or just an error in logs? | 18:32 |
| vknecht | just log error, afaik | 18:32 |
| piggz_ | r0kk3rz: ping | 18:43 |
| heroic_1 | found the culprit for why "droid-system" was unavailable: https://github.com/mer-hybris/droid-hal-device/blob/master/helpers/util.sh#L239 | 18:46 |
| heroic_1 | I was building the spec files one after another, so when I built "droid-system-kagura.spec" it would wipe the rpm for "droid-system-kagura-f8331" | 18:47 |
| heroic_1 | Need to supply both spec files to the build_package.sh command at once | 18:47 |
| r0kk3rz | piggz_: moo | 19:29 |
| piggz_ | r0kk3rz: reply on telegram, not sfos related | 19:32 |
| piggz_ | mal: abranson: if scene-mode (exposure) is set to off, which I think is manual, should some sliders appear for setting apeture and exposure time | 20:12 |
| mal | piggz_: I have been thinking of investigating if such a thing is possible | 20:16 |
| mal | piggz_: I did some changes to UI again and I'm able to make it both follow the light/dark color scheme or just stay is default one, which one do we want, both look reasonable | 21:04 |
| piggz_ | mal: if both look reasonable, perhaps allow following the user scheme? | 21:04 |
| piggz_ | mal: where should we build it? | 21:17 |
| mal | piggz_: what do you mean? | 21:20 |
| piggz_ | mal: obs project | 21:38 |
| piggz_ | my personal, or common: | 21:39 |
| mal | piggz_: need to think | 21:58 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!