**** BEGIN LOGGING AT Tue Jan 31 03:00:02 2017 Jan 31 08:53:27 EricBlade is kind of cursed, when it comes to red N5s... Jan 31 08:53:33 Morning! Jan 31 09:12:50 morning Jan 31 09:41:28 is it possible to try LuneOS with MultiROM, or I surely need to install old Android as the main ROM? Jan 31 09:41:40 on Nexus 4 Jan 31 09:42:36 TheKit: I never tried yet; if I understand correctly, there are a couple of options to activate in our kernel first, like kexec Jan 31 09:43:01 kexec patches are needed only for the main kernel, I think Jan 31 09:43:04 *kexec-hard Jan 31 09:43:08 ah ok, yes makes sense Jan 31 09:44:07 I didn't look at it too much; where does it store/install the various OSes ? Because it seems to be able to boot two concurrent Androids, and I can't see how Jan 31 09:45:21 TheKit: also, on mako we are now using CM12.1 Jan 31 09:46:10 there are different storage options, for example in folder inside /data/media/0 Jan 31 09:47:18 so latest stable is Android 4.2 based and nightly is CM12.1-based? Jan 31 09:47:45 TheKit: for LuneOS, we are using folders /data/luneos and /data/luneos-data, and we chroot at boot time Jan 31 09:49:32 TheKit: nightly is too unstable, prefer testing, which is also based on CM12.1 and works ok - though I did an important fix two days ago for audio call Jan 31 09:49:55 stable is still based on 4.2.1 yes Jan 31 09:52:54 TheKit: very probably the way SailfishOS was integrated in MultiROM would work for us with a few tweaks Jan 31 09:57:24 MultiROM, I think, first flashes corresponding CM base, then modifies boot.img to mount root from MultiROM folder Jan 31 09:57:35 Android base is copied into /system inside SFOS root Jan 31 10:01:45 So basically it creates a fake /system Jan 31 10:02:06 ah no way, not only that Jan 31 10:02:42 Why didn't they document that a bit ?! Their open-source project is like a black box... Jan 31 10:04:01 oh, there's a wiki. Hehe. Jan 31 10:04:27 https://github.com/Tasssadar/Team-Win-Recovery-Project/blob/android-6.0/multirom/multirom.cpp#L2406 Jan 31 10:04:32 https://github.com/Tasssadar/Team-Win-Recovery-Project/blob/android-6.0/multirom/multirom.cpp#L2453 Jan 31 10:06:13 thanks Jan 31 10:13:49 TheKit: I see two obstacles for using LuneOS with MultiROM as-is: our initrd script (looks like https://github.com/webOS-ports/meta-webos-ports/blob/master/meta-luneos/recipes-core/initrdscripts/initramfs-boot-android/init.sh ) and our fstab entry ( https://github.com/shr-distribution/meta-smartphone/blob/krogoth/meta-lg/recipes-core/base-files/base-files/mako/fstab ) Jan 31 10:13:59 I can't find the part where it actually modifies script inside boot.img, but the modified function is https://github.com/Tasssadar/sailfish_multirom_packer/blob/master/zip_root/post_install/sailfishos-init#L121 Jan 31 10:26:45 TheKit: still, I'll give a closer look when I've finished fixing two bugs we have currently (sensors with CM12.1 hammerhead not working, and some systemd service order mistake) Jan 31 10:39:35 oooh, google earth entreprise will go Apache2 in March :) Jan 31 11:34:37 I could boot it with MultiROM by manually extracting files, symlinking /data/webos and commenting out /system in fstab Jan 31 11:35:15 what should the button in the bottom of screen do? my mako touchscreen is a bit broken Jan 31 11:36:02 TheKit: The white small bar? It's a virtual gesture area. I.e. back or forward swipe ;) Jan 31 12:23:29 Herrie|Pre3: you'll get two mails from me, it's just that I forgot to do "reply all" the first time :) Jan 31 12:56:26 thanks Herrie, Tofe, lots to read there! Jan 31 13:09:56 thanks Tofe, those steps were very straightforward & are complete now Jan 31 13:10:56 Tofe: OK :) Jan 31 13:11:07 Good it will be sorted shortly :) Jan 31 13:41:51 Glad it worked that well :) Jan 31 13:41:59 so am I! Jan 31 13:58:03 It's running so will probably take a while. Jan 31 14:53:19 novaldex: hey, did the jenkins slave rebuild started last night? today I see both kwaj and abaco as offline Jan 31 14:57:56 hey JaMa, not aware of anything so i'll go take a look Jan 31 15:17:26 novaldex: ah, nearly! --> Your version is: javac 1.8.0_66. The required version is: "1.7" Jan 31 15:18:42 novaldex: can we install openJDK 7 on that machine? Jan 31 15:19:29 Tofe: that's a funny one, who on earth would've upgraded Java on that machine? far as I know it's not had any work Jan 31 15:20:36 novaldex: Might be the standard version in this Ubuntu? Jan 31 15:21:10 Bonaire is still 14.04.3 LTS.. Jan 31 15:21:33 1.7 is there, /usr/lib/jvm/java-1.7.0-openjdk-amd64 Jan 31 15:22:08 true, "HOST_OS_EXTRA=Linux-3.13.0-74-generic-x86_64-with-Ubuntu-14.04-trusty" Jan 31 15:22:22 can you manually/explicitly define the java exec? Jan 31 15:22:29 I don't know Jan 31 15:24:26 okay, I can force it to use 1.7, will that break anything else? Jan 31 15:24:55 JaMa: still looking into Kwaj/Abaco Jan 31 15:26:21 novaldex: we could prepend the path to the correct JDK in our script for PATH, eventually Jan 31 15:26:40 or maybe set it in the jenkins job properties, if that's possible ? Jan 31 15:26:49 (no idea how that works :) ) Jan 31 15:27:01 full path: /usr/lib/jvm/java-7-openjdk-amd64/jre/bin/java Jan 31 15:27:46 i'm happy to try forcing the default version for now if you like, just bear in mind it may cause other issues Jan 31 15:28:14 JaMa, both servers are up & live, so not sure why they're reporting as being down Jan 31 15:29:03 novaldex: we can try forcing it to see if there will be other problems, yes Jan 31 15:30:09 okay, done, the default is now 1.7 Jan 31 15:30:36 JaMa: Kwaj was reporting an error so i'm going to give it a bounce & see how it goes Jan 31 15:37:55 novaldex: Should I kick it off again? Jan 31 15:38:28 Herrie|Pre3: sure, why not? Jan 31 15:40:18 OK running :) Jan 31 15:42:51 ... and failed :p still using javac 1.8 Jan 31 15:43:47 boo.. Jan 31 15:45:30 novaldex: you did an update-alternative for java ? or something else? Jan 31 15:45:49 Tofe: that's exactly what I did Jan 31 15:46:27 but there was a choice of 2 different Java7s Jan 31 15:46:35 one 'auto' and the other 'manual' Jan 31 15:46:40 so i've just flipped to the other Jan 31 15:47:08 there was one java8 and two java7 ? Jan 31 15:47:36 There are 4 choices for the alternative java (providing /usr/bin/java). Jan 31 15:47:37 Selection Path Priority Status Jan 31 15:47:37 ------------------------------------------------------------ Jan 31 15:47:37 0 /usr/lib/jvm/java-7-openjdk-amd64/jre/bin/java 1071 auto mode Jan 31 15:47:37 1 /usr/lib/jvm/java-6-openjdk-amd64/jre/bin/java 1061 manual mode Jan 31 15:47:38 * 2 /usr/lib/jvm/java-7-openjdk-amd64/jre/bin/java 1071 manual mode Jan 31 15:47:39 3 /usr/lib/jvm/java-8-oracle/jre/bin/java 1063 manual mode Jan 31 15:47:41 4 /usr/lib/jvm/jdk1.6.0_38/bin/java 1 manual mode Jan 31 15:47:55 it was on #3, so I swapped it first to #0 Jan 31 15:48:00 now i've just swapped it to #2 Jan 31 15:48:57 weird. Jan 31 15:49:03 very! Jan 31 15:49:17 Herrie|Pre3: another build ?... :) Jan 31 15:52:07 novaldex: thanks Jan 31 15:53:16 JaMa: welcome! to be on the safe side i'm running fsck on them both too Jan 31 15:53:25 novaldex: I cannot reach them over VPN, so maybe the jenkins master has the same routing/networking issue Jan 31 15:53:31 once they're back up should jenkins just detect them or do we need to give something a nudge first? Jan 31 15:54:13 Failed again :P Jan 31 15:54:20 I've logged to abaco after long delay, but kwaj still isn't responding to ssh over VPN Jan 31 15:54:53 kwaj is still running fsck Jan 31 15:55:20 now kwaj is back as well Jan 31 15:55:26 Herrie|Pre3: was it java again? Jan 31 15:55:27 both re-connected to jenkins Jan 31 15:55:37 [01/31/17 15:55:02] [SSH] Checking java version of /home/jenkins/jdk/bin/java Jan 31 15:55:40 [01/31/17 15:55:04] [SSH] /home/jenkins/jdk/bin/java -version returned 1.7.0_80. Jan 31 15:55:56 kwaj is using 1.7.0_80 already, but that's ubuntu 16.04 not 14.04 Jan 31 15:56:08 JaMa: glad they've reconnected, funny it happened in the first place Jan 31 15:57:23 novaldex: javac again... What does it say for "which javac" ? Jan 31 15:59:17 good question, 2secs Jan 31 15:59:28 just /usr/bin/javac Jan 31 15:59:53 which points to /etc/alternatives/javac Jan 31 16:00:04 so let me try update-alternatives with javac instead Jan 31 16:00:16 javac -version ? (or something like that maybe) Jan 31 16:00:35 aha, that only gives us 3 options, java-8-oracle & jdk1.6.0 Jan 31 16:01:03 well, at least the problem is located Jan 31 16:01:44 Sorry, I'll have to leave for a couple of hours Jan 31 16:02:13 thanks for your help Tofe, we're getting somewhere Jan 31 16:02:46 Herrie|Pre3, i've got another hour before i'll be offline for a while too Jan 31 16:03:12 just checking Java7 to see if it's got javac Jan 31 16:07:59 Tofe: if we drop it to 1.6, would that also cause problems (taking in mind i'm quite expecting the answer there to be 'yes') Jan 31 16:24:56 What uses Java? Jan 31 18:11:14 DougReeder: It's for our small CM builds for our Android based targets Jan 31 18:15:53 DougReeder: It's the small images on http://build.webos-ports.org/cm-wop-12.1/ that are then use by our build process Jan 31 18:16:05 Even when you build locally it would normally pull them from there Jan 31 18:16:53 Because of the ancient Android build bits we cannot really do this in 1 go in BitBake like our other components Jan 31 18:17:17 I.e. They use old gcc 4.7/4.8 while we're at 5.3/6.0 already Jan 31 18:20:43 Hrm. I have no idea which version of Java to use, then. Jan 31 18:23:18 Java 7 :P Jan 31 19:38:47 novaldex|away: no it has to be java 7 Jan 31 20:24:46 novaldex|away: it's strange, there *is* a javac binary in openjdk7 on trusty Jan 31 20:27:40 novaldex|away: looking at the package's postinst script, it seems there should be an alternative javac for openjdk7 too **** ENDING LOGGING AT Wed Feb 01 03:00:01 2017