Saturday, 2019-01-12

*** ChanServ sets mode: +v T402:43
*** ChanServ sets mode: +v T408:02
Benclark006Hi, when i run build packages, it's not using the wrong model name, how do I change that?10:40
malwhat do you mean?10:40
Benclark006when i run it returns Fatal: 'rpm/droid-hal-hammerhead.spec' doesn't exist (and could not be made from a yaml)10:41
Benclark006even though echo $DEVICE returns mako10:42
T4<It_sMike> mal I just reinstall my platform sdk and now face with this problem while building with mic : , guess the problem cames from mer tools10:43
Benclark006turns out i just screwed up my environment variables, it's working now10:48
malBenclark006: but you are not building for hammerhead10:48
malBenclark006: so what did you do to break it?10:49
Benclark006i have no clue lol10:49
Benclark006but i just redid chapter 4.1 and it works neato now10:49
mal@It_sMike what is trying to install?10:51
T4<It_sMike> Installing nothing! just want to build image with mic and it returns me this error...10:51
Benclark006ok so when i run it get Fixing mount-points for device mako10:52
Benclark006sed: can't read tmp/units/*: No such file or directory10:52
Benclark006error: Bad exit status from /var/tmp/rpm-tmp.niss6I (%build)10:52
Benclark006here's my fixup mountpoints
malBenclark006: did you properly run
malI mean make hybris-hal also10:59
malbefore that10:59
Benclark006i think so10:59
T4<It_sMike> Fixed... if everyone face with this problem while building image ( with mic ) make sure that you are using last version of platform sdk.11:12
lbtPSA Mer git is going down for an upgrade11:23
T4<adampigg> Lbt, will include 3.0.1?12:09
r0kk3rzprobably not12:10
* lbt goes to check if that's on public release yet12:15
r0kk3rzstill ea afaik12:16
*** ChanServ sets mode: +v T414:16
malpiggz: I think I figured out a clean way to fix the SDL crash issue16:12
piggzmal: go on16:22
piggzshow me the code :)16:23
piggzits nice to understand these things16:23
malpiggz: just a moment16:29
malpiggz: so the issue was that the display variable was provided to wayland as the user data but that was not the same pointer that is actually reallocated by SDL_AddVideoDisplay so any subsequent calls to any of the callbacks would cause a crash16:35
malpiggz: so basically the user data pointwr provided to wayland was freed which is not good16:36
piggzmal: great, so this will bring upgraded wayland to sfos?16:50
malpiggz: hopefully yes16:54
malpiggz: I need to ask from SDL people if they think that is correct16:54
malpiggz: argh, who uses "index = _this->num_displays++;"17:00
malto operator orders are always difficult17:01
malpiggz: does that patch seem logical to you?17:15
malpiggz: for reference this is how adding display is done
piggzmal: is sdl_zero needed if display is on the stack not the heap?17:27
malpiggz: hmm, not sure17:29
Benclark006Hi, so when I build, I get this here error: . Here's my
lbtPSA: Mer git is back18:53
piggzBenclark006: that isnt a full output is it?18:59
Benclark006that should be everything19:00
malBenclark006: are you building 14.1 now?19:00
Benclark006becuase there's a bunch of stuff pre-made for it19:01
Benclark006so i thought it was my best bet19:01
maloh, you probably shouldn't use that, it would make life more complicated19:01
Benclark006oh, ok19:02
malfor example camera and stuff would probably be pain in 10.119:02
malif your device has 14.1 images available then that would be my suggestion19:02
malBenclark006: last time anyone built 10.1 was probably years ago19:03
Benclark006i guess19:04
malpiggz: now I added also display orientation handling to SDL when using wayland backend19:11
piggzmal: feature complete then! you have a game engine using htat dont you?19:20
piggzive got a weird algorithm to implement for sleep detection19:21
malpiggz: I haven't decided if I should use SDL backend for scummvm or the Qt backend I made for it19:23
malpiggz: the new patch
malI need to test that on some device with landscape display to make sure the conditions are correct20:09
Benclark006hi again, doesn't have a lineage-14.1 branch20:09
Benclark006is there a mirror for it?20:09
malBenclark006: there is the old cyanogenmod repo20:15
Benclark006I just thought of that, i'll have a look20:16
malBenclark006: wait20:16
malBenclark006: this is the lineage repo for mako kernel different name than for 15.120:17
Benclark006oh, great20:17
Benclark006thanks :)20:17
*** ChanServ sets mode: +v T420:57
UmeaboyI need help with the hello world script because it gives me this error: sb2: Error: Invalid target specified, aborting.21:28
UmeaboyI did exactly as it said in the HADK.21:29
r0kk3rzthe error isnt obvious enough?21:30
r0kk3rzinvalid target, yo target name is wrong21:31
UmeaboyWhere do I change it?21:31
UmeaboyIn hadk.env?21:31
r0kk3rzin the sb2 command21:32
r0kk3rzthe -t flag is the target name21:32
Umeaboyr0kk3rz: As you can see here it didn't make any difference changing the line:
maldid you name the the target correctly when adding the target, the default commands in wiki use different naming21:36
r0kk3rzbut what is the name of the target sb2 knows about21:37
malsdk-assistant list can used to list the existing targets21:37
r0kk3rz-t should tab complete21:37
Umeaboysb2 -t and pressing tab button gives me this:21:38
Umeaboy /home/kristoffer/.scratchbox2 missing, create some targets with sb2-init!21:38
r0kk3rzsooooo you have no targets21:39
UmeaboyAs that page told me don't have to install any targets yet I left it be.21:39
r0kk3rzyeah read it again, you need targets to do anything with sb221:39
UmeaboyOK. Creating the targets now then.21:40
Umeaboysb2 -t SailfishOS-latest-armv7hl gcc main.c -o test went OK.21:46
UmeaboyDo I use sb2 -t for i486 as well?21:46
UmeaboyI see "Hello World" for armv7hl so the test worked :)21:48
UmeaboyIt's kind of annoying to see a row being set in the defconfig, but the mer kernel checker does not see it as set.22:09
UmeaboyAnything that can be done to solve that?22:10
malwhich flag?22:10
UmeaboyFor instance CONFIG_RD_GZIP22:12
piggzUmeaboy: sometimes a kernel does not support a flag, so it does not make it to the final config22:12
piggzfound in Linux kernels: 4.13–4.20, 5.0-rc+HEAD22:13
UmeaboyI added the ones that are missing.22:14
UmeaboyCONFIG_IKCONFIG_PROC is one of them as well.22:14
piggzso, make sure those are set22:15
malyep, some rare kernels don't have CONFIG_IKCONFIG enabled, or properly added as dependency so it's enabled when needed22:17
Umeaboypiggz: I changed every error and warning and reran the mer checker, but it still says the same. OK to continue?22:38
UmeaboyHere's the full output:
piggzUmeaboy: umm, no, it says the .config doesnt exist, have you checked that folder?22:39
Umeaboy /out/target contains a folder called maple, not g8142. Maybe I should rename the device inside of the hadk.env22:40
UmeaboyWould that help?22:40
piggzlikely :D22:40
UmeaboyAnd then what? source .hadk.env?22:41
piggzwell, you need the new environment....22:47
piggzyou can also just export DEVICE=.....22:47
malUmeaboy: how many times do I have to say, your device codename is maple, not g814222:47
UmeaboyI changed it now.22:48
maleverything is using maple except lunch command which uses g814222:48
Umeaboymal: I'm so used to using Lineage when building....22:48
malthat is the lunch part22:49
UmeaboyYeah and that's kind of confusing as I'm used to do lunch maple22:49
UmeaboyAs it's not officially supported by Lineage yet.22:49
UmeaboyI'm unsure about the SELinux part for CONFIG_AUDIT.22:54
maluse audit=y with bootparam22:56
malit clearly says that in bootparam description22:56
UmeaboyI guess it's OK to continue since it's only a warning and both are set to y.22:56
Benclark006Hi, when i try to build the kernel, i get this error ninja: error: 'kernel/lge/mako/arch/arm/configs/cyanogen_mako_defconfig', needed by '/home/ben/hadk/out/target/product/mako/obj/KERNEL_OBJ/.config', missing and no known rule to make it22:58
malis the kernel path in manifest correct i.e. does the kernel exist in kernel/lge/mako folder in $ANDROID_ROOT22:59
UmeaboyBenclark006: I always disable both NINJA and JACK when building, but that's me.22:59
Benclark006mal: yes, it exists23:01
malhow about that file?23:01
Benclark006it exists, but under a different name23:04
Benclark006instead of cyanogen_mako_defconfig23:04
Benclark006should i just rename it?23:08
Benclark006i feel like it could cause problems later down the line 🤔23:09
malnot sure if that will cause issues23:12
malwe'll see23:12
Umeaboymal: Apparently the Ubuntu chroot Java version is to old to run make hybris-boot.23:12
malBenclark006: which device repo do you use?23:13
malBenclark006: are you sure you didn't accidentially take the cyanogenmod device repo?23:14
malBenclark006: based on the error you might have taken it23:14
malUmeaboy: afaik there was a way to change it, something like sudo update-java-alternatives -s java-1.8.0-openjdk-amd6423:16
malBenclark006: check the local manifest, also remember the selinux=0 change to device repo23:16
Benclark006will do23:16
malBenclark006: lineage repo uses the correct defconfig
Umeaboymal: HABUILD_SDK [maple] kristoffer@localhost:~/hadk$ sudo update-java-alternatives -s java-1.8.0-openjdk-amd6423:19
Umeaboyupdate-alternatives: error: no alternatives for mozilla-javaplugin.so23:19
Umeaboyupdate-java-alternatives: plugin alternative does not exist: /usr/lib/jvm/java-8-openjdk-amd64/jre/lib/amd64/IcedTeaPlugin.so23:19
malso install it?23:20
malyou have apt in there23:20
Umeaboymal: sudo apt-get install openjdk-8-jdk says that it's already installed.23:22
Umeaboy* 3            /usr/lib/jvm/java-8-openjdk-amd64/jre/bin/java   1069      manual mode23:24
UmeaboySo it is selected as well.23:24
UmeaboyChecked with sudo update-alternatives --config java23:24
malso which version does the build want?23:31
Umeaboymal: According to the output 1.8.x23:41
UmeaboySee here:
UmeaboyI think it would be great to update to a more up to date version of Ubuntu when using chroot.23:55
UmeaboyBionic Beaver have an LTS.23:59

Generated by 2.17.1 by Marius Gedminas - find it at!