Thursday, 2019-10-17

T42<BusterBg_18> xreactx: are you sure that your device sources are good?00:16
xreactx[m]__Yes, it's officially supported by lineage. The vendor blobs are on TheMuppets00:25
xreactx[m]__I'm attempting to build it for the OnePlus 7 Pro.01:01
T42<DylanVanAssche> @piggz Browser is not just white, it crashes :0 Webcat with updated QtWebkit works fine though04:18
sledgesPSA: SFOS community IRC meeting at #mer-meeting in just over 5mins06:53
sledges1hr5mins also qualifies for "just over 5mins" in valve time06:55
T42<adampigg> well, im not having much luck with the meetings and my mobile coonnection!08:34
nitin03I have successfully made boot.img plus sailfish os zip file but it is not booting ...I can't debug it as mer debug not showing up..I can't telnet not my device12:31
nitin03*I can't telnet into my device12:33
T42<birdzhang> nitin03: 16.0 base ?12:52
T42<Akatsu %lastname%> Is there any script which will make full setup of dev workspace?13:07
T42<BusterBg_18> @birdzhang is it common for 16.0 based ports to not show telnet then?13:28
deathmistmal: I heard I needed changes in dcd hybris init files for hybris-16.0 at what should I do about vold etc?13:29
malvold can be disable because the patch needed for disabling that is on hybris-patches repo13:31
deathmistmy current dcd sources for 16.0 are a direct clone of the 15.1 ones13:31
T42<birdzhang> @BusterBg_18 i have the same issue, and no telnet13:31
deathmistmal: cool, then I should add the rest of these to my own disabled_services.rc right?13:33
malprobably yes13:35
T42<BusterBg_18> @birdzhang  On Xiaomi_sirius I have the same issue it just never brings telnet up and stays on bootlogo for around a minute13:35
T42<BusterBg_18> I have played around with selinux stuff but still fails to boot13:36
malare you sure your device doesn't dmverity or something like that enabled?13:37
malxperias had that and needed to revert a commit from kernel sources13:38
masha11mal: Hello. my sensors are unstable. My log failed sensors: , successful sensors: My hw-settings.ini: What can be done with these?13:48
mal@adampigg looks like some fixing is needed for cam app, not sure why I didn't notice that bug before, some other things also related to the menus need fixing13:50
malmasha11: I don't have time to check those right now13:51
masha11mal: ok13:51
malusually such things are some timing issues or something like that13:53
T42<adampigg> mal:/yeah, seen14:09
T42<adampigg> im sure a margin will fix it14:09
mal@adampigg in finnish translation some texts are too long for portrait menu and don't have the proper margin in right side and also no fading which is the correct way based on silica style guide14:13
mal@adampigg I think we should talk a bit later about the menus some more14:14
T42<adampigg> mal: if you can catch me tonight! i have a parents evening, a trip to look at a car, and a pool game, inbetween making supper for family!14:22
T42<adampigg> but, i have a 1hr bus trip in about 30 mins when im good to talk14:23
T42<adampigg> on mido, in portrait, the first menu item is slightly off the top of the screen, but is selectable still14:24
malsame on fp2, it's usable but not nice14:24
T42<adampigg> if it was just anchored corretly it would help14:26
malyes, also some complaints from me about margins and stuff in general14:27
malwe'll see if I have time to fix those tonight14:27
T42<adampigg> and do a quick 0.6.1?14:28
T42<adampigg> mal, one of my jobs tonight now cancellled,/so will have a bit more time15:26
mal@adampigg found some other issues in menus also16:18
xreactx[m]__mal: any ideas what could be stopping my vendor.img from building. The error hangs then fails at 99%16:30
malcomment it out from hybris/hybris-boot/Android.mk16:33
malassuming you really don't need it16:33
xreactx[m]__Does sfos make any modifications to the vendor.img?16:38
malno, but it might need things from vendor build in droid-hal packaging stage16:41
mallike fstab uevent*.rc files etc16:41
malotherwise udev rules and mount services are not correctly created16:42
malso instead of removing that I suggest you have a look at output of "make modules" and see if you have anything related to fstab or ueventd in there16:42
xreactx[m]__Thanks for the tip16:44
T42<BusterBg_18> mal: speaking of fstab I took a look at dmverity on my device tree and its already disabled17:00
T42<BusterBg_18> What's the kernel repo of Sony hybris? I want to look for that commit you speak of17:00
malthe sony kernel in mer-hybris github17:03
T42<adampigg> mal: hit me with the issues17:14
mal@adampigg fixing already, mostly done17:14
mal@adampigg just wait a moment and I'll make a PR17:15
T42<Akatsu %lastname%> Is there any way to use repo init without account? I have issues getting repo init on mer-hybris17:23
malthere is never need for account for repo init17:25
malsome people have had issues if the git username or something has some special characters17:25
T42<Akatsu %lastname%> gpg gives me no public key "can;t check signature"17:26
T42<Akatsu %lastname%> and then error: `cloning the git-repo repository failed`17:26
maltry the fix in the answer there17:29
malnot sure if that is the error you have17:29
mal@adampigg sorry for a small delay, fixing translations also17:32
T42<adampigg> mal: it will have a thorough review!17:38
rinigusHi! I have trouble booting using flashed hybris-boot on xperia xz2.17:38
malquite trivial fixes17:38
rinigusAfter flashing images, I am getting "Your device is corrupt. It cannot be trusted and will shutdown in 5 seconds" instead of SONY logo17:38
rinigusThis is also after disabling vbmeta (fastboot --disable-verity --disable-verification flash vbmeta vbmeta.img) and generation of dtbo image corresponding to kernel.17:38
rinigusLooks like its something before dtbo/vbmeta checks as I can get to sony logo with aosp boot image and all other sailfish images17:38
rinigusany ideas on what to test?17:38
malI didn't understand most of that17:39
rinigusmal: :( bit more specific?17:40
vknechtrinigus, just wondering, did you disable encryption (ie. reformat) userdata before flashing ?17:40
malno idea what vbmeta means and what you even did17:41
rinigusvknecht: yes, I did17:41
malthere shouldn't be need for such things if you revert the verity commit from kernel17:41
vknechtok ; must say I also occasionnaly get that17:41
rinigusmal: just a sec17:41
malrinigus: did you flash kernel or just use fastboot boot?17:42
malto me it seems you are doing something overly complicated things17:42
rinigusmal: I did flash the kernel. as far as I understood, tama platform is the first with dtbo and vbmeta partitions17:43
rinigusmal: if you look at you'll see dtbo images for xz2, xz2c and xz317:44
rinigusmal: what's a commit I am expected to revert to disable verity? I would expect its already done if I use
rinigusvknecht: is there any way to check for encryption on userdata? although I expect that check a bit later...17:47
rinigus... during a boot17:47
malrinigus: why aren't you just using that kernel and adapt the defconfig?17:48
malthat is latest 9.0 kernel17:49
rinigusmal: I thought I did... letme check17:49
malah, sorry17:49
malI misread your link17:49
malrinigus: you should rebase17:50
rinigusmal: I did already at
T42<Akatsu %lastname%> @mal —no-repo-verify worked as temporary fix17:51
malrinigus: mer-hybris also got latest changes two days ago17:51
vknechtbtw, what's the best git practice to update forked kernel: git rebase, or git merge/pull ?17:53
rinigusmal: that's right, I'll merge them. but I suspect its something else. something that flags bootloader and it just refuses to boot17:53
rinigus... but I will surely merge17:53
malrinigus: maybe it needs signed boot image?17:53
malvknecht: it depends, if you plan to upstream then git rebase, if not then it doesn't matter much17:54
deathmistholy crap my hybris-16.0 rebase zip "boots" and I get telnet on port 2323 :D17:54
rinigusmal: maybe. don't know. how do you do that? I though that's what vbmeta does17:54
vknechtok thanks17:55
deathmistmal: droid-hal-init fails as I expected since I get no UI, could I be missing something in my dcd repo for hybris-16.0 ports?17:56
riniguswhere can I read what's on hybris-boot.img? maybe something easier than mount loop?17:56
vknechtsplit_bootimg hybris-boot.img17:58
vknechtgunzip hybris-boot.img-ramdisk.gz17:58
vknechtmkdir ramdisk && cd ramdisk17:58
vknechtcpio -id --no-absolute-filenames < ../hybris-boot.img-ramdisk17:58
vknecht(something like that)17:59
rinigusvknecht: thanks!18:01
mal@adampigg it seems french translation has line number but others don't maybe I'll remove those18:05
T42<adampigg> mal: ah, that explains why the frech one keeps getting updated18:05
T42<adampigg> qtcreator manages that though doesnt it?18:05
vknechtwas wondering also ; maybe because made with Qt linguist ?18:05
malyes, no need to have those afaik, at least other languages work18:05
deathmistmal: here's also logs, journal dmesg
deathmistoh I see: droid-hal-init: Could not read properties from '/plat_property_contexts': No such file or directory and good old "Failed to initialize property area"18:13
maldeathmist: show what you have in out/..../root in ANDROID_ROOT18:17
mal@adampigg finding new issues all the time18:18
malwell this last one is not very critical18:18
malin portait mode the menu height is not half of the screen height but instead it is width of the screen18:19
maldeathmist: on device check where you have that plat_property_contexts is it in /odm /vendor or something else "find / -name plat_property_contexts"18:26
deathmistmal: it found /system/etc/selinux/plat_property_contexts, guessing symlink that to / ?18:28
malhmm, wondering why it didn't find it from there, symlink is the first hacky way18:29
maldeathmist: you need at least there in straggler files: bt_firmware, firmware, dsp, persist18:29
malobviously as /dsp for example18:30
T42<adampigg> mal: looks good.... no anchoring issues?18:54
malinterestingly no18:55
mal@adampigg I tried to fix the height also but that caused anchoring issues similar to what we had now before that fix18:56
T42<adampigg> let em test it quickly ....18:56
vknechtmal, "Grille :" is not incorrect, in french two-part punctuation (like ? ; !) requires space before and after18:56
* vknecht winks at Mister_Magister 18:58
Mister_Magisteru need something?18:58
vknechtno, just a sort of private joke explained :-)18:59
T42<adampigg> mal: yep, works19:00
malvknecht: really? odd, I'll change it back19:00
Mister_Magistervknecht: whatcha talkin bout tho19:00
malvknecht: reverted now19:01
Mister_Magisterah translation19:02
Mister_Magistervknecht: well lemme answer with another meme "gtfo i'm playing minecraft!"19:02
mal@adampigg are there any other critical issues in cam app that should be fixed now, it seems german translation wasn't finished19:04
T42<adampigg> well, the resolution is slightly off centre in portrait mode ... but thats not critical :)19:05
malI almost tried to translate that but decided to leave ti for others, I have only studied german language for 5 years in school19:05
T42<adampigg> yeah, im best leaving translation to others19:06
malhopefully we'll get updates to those soon19:06
mal@adampigg the resolution is not very nice looking anyway19:06
malit would be better near the top in portrait mode19:07
T42<adampigg> yeah, im thinking that19:07
mal@adampigg I can do some testing19:07
T42<adampigg> ill push out 0.6.1 anyway as im out soon, so no great rush, i know how busy you are19:08
vknechtmaybe at the same place/along same edge, but drawn/written vertically ?19:09
vknechtperhaps with each character rotating ? XD19:09
vknechtjust an idea...19:10
mal@adampigg we could wait until tomorrow for the release19:10
T42<adampigg> too late its on openrepos :)19:10
mal@adampigg I think the rotation code now is a bit overly complicated, I need to thing a bit how ot make it more simple19:21
deathmistmal: my /system is not mounted (neither is /vendor properly anyways, there's just bt_firmware dsp & firmware_mnt in /vendor), maybe that's why it can't read /plat_property_contexts :/  maybe that's why stuff isn't working :v19:30
maldeathmist: show your fstab19:31
maland droid-hal spec19:32
malwhat device is that?19:33
deathmistmal: cheeseburger (OnePlus 5) dhd spec and I figured out the actual fstab seems to be instead of , what is the recovery fstab even used for? doesn't seem to be very useful for Sailfish anyways19:39
maldeathmist: we need to think how to do the mounting, maybe add custom mount services for now, /vendor needs to to be mounted before those other ones are mounted on top of it19:41
deathmistI've been changing both (recovery.fstab & fstab.qcom) for hybris-15.1 as well and it threw me off that the older fstab.qcom wasn't in  init/vendor/etc/fstab.qcom on 16.0 like on 15.1
maldeathmist: do this for both system and vendor (correct device of course in there and names otherwise)19:43
maldeathmist: in vendor.mount also add "Before: vendor_dsp.mount" or something like that (check correct name) and similar lines with the other vendor*.mount services19:44
deathmistI think I could fix the actual fstab.qcom too since I was editing the useless one, is this a better choice instead?19:44
maldeathmist: for system.mount sure, not for vendor.mount19:45
malbecause of the required mount order19:45
deathmistah, okay I'll create proper mount services then :)19:45
maleasier to test with manual services first19:45
malnote that it's a symlink in, not a file, in that commit I linked19:46
deathmistmal: thank you, I did it completely wrong on my 15.1 port and I have no clue how I even got graphics and stuff working properly :D  now /vendor and /system have proper units created, they're active and have been mounted, DHI still fails unfortunately, journal
maldeathmist: check your /usr/libexec/droid-hybris/system/etc/init/servicemanager.rc on the device20:28
deathmistmal: what about it?
xreactx[m]__mal:  "ueventd.rc" and "fstab.qcom" are listed20:32
maldeathmist: missing the setenv LD_PRELOAD /usr/libexec/droid-hybris/system/lib64/libselinux_stubs.so20:38
maldeathmist: it seems I forgot the patch from hybris-16.020:39
malxreactx[m]__: so run "make fstab.qcom ueventd.rc"20:40
malxreactx[m]__: and check vendor folder again20:40
maldeathmist: check the other servicemanager rc files in that folder for example, I will need to fix things it seems20:41
maldeathmist: hopefully that helps, if not then we need to revert one commit20:42
piggzmal: are there any predfined 'good' settings for /desktop/sailfish/silica' sizes for a low res (720p) large screen (10") ?20:44
xreactx[m]__mal: the out/***/guacamole/vendor directory has /bin /etc /firmware /lib /lib64 /radio and /usr directors. build.prop, default.prop, and ueventd.rc are listed20:44
malxreactx[m]__: so do you now get the same error you got before?20:46
xreactx[m]__run "make -j$(nproc --all) hybris-hal" again? sorry for the noob questions20:47
xreactx[m]__ok, running now20:47
malpiggz: hmm, that is similar to my galaxy tab 420:52
piggzmal: got a settings .txt file for that?20:53
xreactx[m]__mal: new error
piggzive been playing with various values, and its a lot better, but soem quirks20:53
deathmistmal: ooh ok, I added it below the first "service servicemanager" line and restarted but sadly DHI still fails :/  updated journal
deathmistyou sure that nothing is wrong in my /usr/libexec/droid-hybris/system/etc/init/*.rc files?
malxreactx[m]__: that is the same error20:58
malpiggz: I never got it to be perfect, one reason I never published the image20:58
piggzmal: classic you20:59
malpiggz: :)21:01
T42<BusterBg_18> mal: you were right, I still had verity enabled somewhere, it was on kernel21:03
maldeathmist: something to test: run "repo sync hybris-patches", then "cd frameworks/native", "git am ../../hybris-patches/frameworks/native/0006-hybris-use-libselinux_stubs-in-servicemanager.patch", "cd $ANDROID_ROOT/system/core", "git reset --hard HEAD~1" and "cd $ANDROID_ROOT" and finally make hybris-hal again21:18
malthen " -d" and build new image21:18
xreactx[m]__mal: at this point, should I just comment it out from hybris/hybris-boot/
malxreactx[m]__: maybe21:20
T42<adampigg> Sizes arnt too bad21:22
T42<adampigg> (Photo, 1920x2560)
T42<adampigg> (Photo, 1920x2560)
xreactx[m]__mal: which line(s) specifically?21:22
malxreactx[m]__: somewhere near the end you see "vendorimage" in some module lines21:22
mal@adampigg those icons are quite huge in top menu21:23
mal@adampigg top menu shouldn't be that wide, something is wrong21:23
xreactx[m]__mal: the entire line or just "vendorimage"21:23
maljust "vendorimage"21:23
xreactx[m]__should "HYBRIS_COMMON_ANDROID8_TARGETS" be updated Android 9 Targets?21:25
malxreactx[m]__: no, same works, that has "-ge" which mean those parts are used for android >= 821:26
piggzmal: this is what i using
malpiggz: ok, I'll need to boot my device some time and see what I have21:27
malpiggz: not today, maybe tomorrow or during the weekend21:28
xreactx[m]__mal: ok, build successfully completed after that... Is there anything else i should do prior to moving on to the next step?21:28
malxreactx[m]__: maybe check the "root" folder somewhere under out/ for things that could be needed in straggler files21:29
piggzmal: cool21:29
piggzi just made a change to the physical size env vairables, and....21:29
piggzthe top menu looks more normal now21:29
piggzbutu the launcer is crazy!21:29
malpiggz: though there must be some size missing21:31
malpiggz: so more tinkering then :)21:31
xreactx[m]__mal: do you have a link to ganges repo?21:31
malxreactx[m]__: I think it's better to have a look here, this has more suitable ones
malxreactx[m]__: are you using hybris-16.0 also?21:34
xreactx[m]__i am using hybris-16.021:34
xreactx[m]__haha, i think that repo was based on my failed attempt 2 years ago21:34
T42<adampigg> (Photo, 1920x2560)
T42<adampigg> (Photo, 1920x2560)
piggzmal: ^^ :D21:35
malxreactx[m]__: then you also need to do what I told deathmist a moment ago, the stuff starting with "something to test"21:35
malpiggz: I remember that, I had the same at some point :)21:35
xreactx[m]__mal: will do21:37
malalways a bad idea to make patches without actually testing things :) I never tested hybris-16.0 on a device, just made a test build21:37
T42<adampigg> Mal,.found a bad bug22:06
T42<adampigg> Switch to front camera, switch to back...resolution claims to be okx actual resolution is 320x24022:08
mal@adampigg camera swtiching is a mess, I should try to fix my fixes for that22:14
T42<adampigg> Switching is hard22:31
mal@adampigg it's annoying that the only serious issue I had was the initial values22:47
xreactx[m]__mal: strange thing just happened... I used the sailfishos-oneplus5 hybris-boot/ and his defconfig changes (listed at the bottom of his defconfig) and the vendor.img built successfully23:00
deathmistxreactx[m]__: interesting, glad my sources helped you :D  mal: tried again with the changes you listed and now systemd hangs on "starting" state for quite a while but at the end droid-hal-init still fails, another journal update
deathmistcould it be that you haven't synced your sources in 2 days? the patch to set an unset ANDROID_VERSION_MINOR to 0 for hybris-boot was just merged in a few days ago while I already merged the change from PR #169 in to my fork few weeks back when I was initially playing around with hybris-16.023:20
xreactx[m]__deathmist: I will be following you and letting you know what works for me.23:23
deathmistmal: so *servicemanager services all exit with code 1 currently and seem to get ended by a SIGKILL. I'll work on this tomorrow tho, gn all!23:35

Generated by 2.17.1 by Marius Gedminas - find it at!