Wednesday, 2019-02-27

masha11Hello. Help me, please. How do I do this step? HADK 5.3: In initrd we then have to specify hardcoded /dev/mmcblkXpY nodes for /boot and /data partitions. My device: Motorola Moto Z (griffin). My fstab: My fixup-mountpoints:
T4<elros34> @masha11: your fixup-mountpoints looks ok. make hybris-hal or hybris-boot and changes should be included in initramfs08:49
piggzmal: dhv changes made ... now, if you change your mind again...... !!"$$£%£""10:20
masha11T4: After I make hybris-boot, should I rebuild
malyou need to run at least -d11:52
malunless you manually flash the kernel for testing, meaning flashing hybris-boot.img from out/ using fastboot11:53
malbut to include new kernel to your next build -d is needed11:54
masha11mal: ok11:54
masha11mal: According HADK 5.3: In initrd we then have to specify hardcoded /dev/mmcblkXpY nodes for /boot and /data partitions. Does this mean checking a file fixup-mountpoint or something else?12:00
malmasha11: those are hardcoded automatically by build scripts12:08
malusing fixup12:08
masha11mal: ok12:09
masha11mal: Further: "After initrd, systemd needs to mount all other required partitions (such as /system, /firmware, / persist, /config, . . . ) for the HAL layer to work." Should I mount partitions or does systemd do it automatically?12:13
malmasha11: again those are created automatically from fstab using fixup12:15
masha11mal: ok12:15
malwhat are you trying to fix?12:17
masha11mal: I checked the mount points, they look fine. I'm trying to get an initrd download. I do not have a message "Running Mer Boat Loader" in dmesg. Maybe you need to add any of configuration to defconfig?12:36
malmasha11: which android base?12:45
masha11mal: android base 8, lineageos 15.112:47
malshow your defconfig12:48
T4<elros34> masha11: try to enable CONFIG_USB_CONFIGFS_RNDIS and disable CONFIG_RNDIS_IPA also change these NETFILTER*=m to =y13:21
masha11T4: ok13:22
masha11When I install in TWRP, then I have the following output:, is this normal?13:23
T4<elros34> not sure, I never looked at zip installation log. For now flashing hybris-boot is enough. Latest release is
T4<elros34> you can chek in twrp, sailfish is installed in /data/.stowaways/sailfishos/13:29
vknechtfunny... in h-a-c, ratio is strange for some resolutions eg. 4096x2160 is 256:135 instead of 9:5 like in camera-resolutions.json generated from droid-camres16:00
malvknecht: I noticed that also, needs some code the make those aspect ratios better, ping piggz16:01
vknechtnah, it's his day off :-)16:10
vknechtmight have a look some day, or file an issue to keep it in mind16:11
piggzmal: vknecht: because 256:135 is technically more correct than 9:5 ;)17:02
piggzi got the code of stackexchange for calculating the nearest ratio17:03
vknechtah ! all is well then :-)17:03
kimmoliisnt there a 'fuzzycompare' in camres?17:04
*** OhYash1 is now known as ohyash20:01
vknechtany tip/example about passing all of a ListView model content as an array ? line 34 was a naive try I guess...
heroic_1piggz: where do you get droid-hal-mido-img-boot from?
piggzheroic_1: ist built by d-h-d20:12
heroic_1as far as I can see it only builds droid-hal-img-boot, not the device-specific one20:12
piggzheroic_1: no, the device specific one 'Provides' the generic pacakge20:14
heroic_1ok nevertheless, where is the device-specifc one built? I only see "%package img-boot"20:15
malheroic_1: what is the problem you are trying to solve?20:20
heroic_1I have switched all of my spec files over to use the "community-adaptation"-way you recommended, including getting rid of the "official" kickstart files20:21
heroic_1Now I am wondering where you get "droid-hal-kagura-img-boot" from20:22
malheroic_1: was there some reason for that? didn't the old way work?20:22
heroic_1mal: the "offical" way is a bit cumbersome as it requires users to wipe userdata entirely20:22
heroic_1I have this package which provides droid-hal-kagura-img-boot
heroic_1But that one uses jolla-recovery from hybris-initrd, which expects an lvm setup20:23
heroic_1jolla-init from pulls in
malheroic_1: if you remove the have_custom_img_boot from spec then img-boot is automatically created from hybris-boot.img built in hybris-hal build20:23
heroic_1ok and I should remove the requirement on the device-specific droid-hal-$DEVICE-img-boot then in patterns?20:24
heroic_1piggz: but I'm still wondering where you build droid-hal-mido-img-boot, I can't see any repo like "droid-hal-img-boot-mido" under your github account20:25
malheroic_1: read what I just said20:25
heroic_1mal: now I'm confused, you now showed me another device which seems to use a device-specific img-boot20:26
heroic_1is there some rpm magic going on that created "droid-hal-kagura-img-boot" from "%package img-boot" ?20:26
malheroic_1: base name of droid-hal packages is droid-hal-$DEVICE and when you add img-boot to it you get the device specific package20:28
malit shouldn't be that difficult to understand20:29
_svenheroic_1: what device are you working on?20:29
heroic_1mal: thank you. it is not particularly difficult to understand, just not documented clearly that the basename is automatically prepended. a section in the wiki about how patterns/spec/inc files work would go a long way20:31
heroic_1_sven: sony xperia xz, codename kagura20:31
malonce again you are overthinking20:31
piggzheroic_1: i would guess that the %package macro expands to droid-hal-$device + img-boot20:32
piggzbut, as mal says, you're over thinking .... i just happens as part of d-h-d, i didnt do anything else20:33
heroic_1piggz: yes, I understand it now, thank you20:33
heroic_1there's just a lot of magic going on, and it's not good for grep-a-bility20:33
_svenheroic_1 i tried that one 1,5 years ago,20:34
heroic_1_sven: I have it working, ril, audio, camera, bt, pretty much everything except usb20:34
heroic_1_sven: but the port is ugly, it uses the "sailfish x" way of packaging which makes little sense for community adaptations20:35
heroic_1(pretty much because I kanged nearly everything from the jolla nile port)20:35
_sveni still got the phone, lended it out to someone, maybe its time to get it back20:36
malvknecht: ^ check the PR above20:38
malthat PR became a bit large again20:39
malheroic_1: it's basic spec feature that main package name is prepended to subpackage names, main package name is defined here
heroic_1ye Imma go write up some docs on that, if only for myself20:44
heroic_1again mal, for someone not immersed in this it's completely alien20:45
heroic_1and I don't think it's useful to expect people to know all about rpm packaging when they only need a specific subsection20:46
vknechtmal, now with selfies ? :-)20:49
malvknecht: yes, although I need to figure out how to mirror the viewfinder20:55
malheroic_1: you really shouldn't have to think about packaging much when porting20:56
malvknecht: I also have preliminary code for dynamic orientation handling but it has some viewfinder orientation issues20:57
vknechtthat's cool, I'll try tomorrow ; for gallery, I guess dbus'showPhotos', undefined) would do for now, unless we really want to keep the slideshow ?20:59
vknechtthe 'showImages' call sounds like it could open only a limited list of files, but not able to use it atm21:01
heroic_1mal: took me 3 minutes
heroic_1any way to get that info added to the wiki?21:12
heroic_1mal: remember the slowness to refresh repos I told you about? turns out the sailfish chroot does not like systemd-resolved21:43
heroic_1switched NetworkManager's dns resolver on the host to dnsmasq and inexplicably it is way faster now21:44
heroic_1the chroot is somehow misconfigured, because ALL network calls are extremely slow when using systemd-resolved21:44
heroic_1even pinging takes ~10 seconds and more21:45
heroic_1this is going to be a problem soon-ish because of systemd's relentless creeping prevalence in distros21:47
heroic_1because arch is so bleeding edge it's kind of a harbinger for what's to come on ubuntu etc :(21:47
malpiggz: heh, I forgot we have grid layout, I was wondering how to do that22:03
piggzmal: ha, it was pretty simple ;)22:03
malpiggz: so it seems :)22:04
malpiggz: will also help with my dynamic orientation code22:04
piggzmal: i put a comment on the one lime i expect to get commented on ;)22:04
malpiggz: maybe do anchors like this for readability at least when there are more than 3 of those22:07
piggzmal: ok22:07
malpiggz: qt qml guide suggests that way22:07
malgrouped properties in there22:09
piggzmal: updated22:11
malpiggz: I looked at anchors and wonder do we need that one with different value, anchors.margins would set all at the same time22:13
piggzmal: oh yes, forgot that...22:13
malpiggz: approved22:15
malpiggz: good fix22:16
piggzmal: coolio ... thats me for the evening, gnight..22:16
malpiggz: I'll try to get the dynamic orientation finished tomorrow22:17
piggzand i'll do the scene icons22:18

Generated by 2.17.1 by Marius Gedminas - find it at!