**** BEGIN LOGGING AT Wed Mar 16 02:59:58 2016
Mar 16 05:08:54 <Andolamin> Oh! Oh! Oh!
Mar 16 05:09:02 <Andolamin> Tofe, Herrie: ping
Mar 16 05:09:09 <Andolamin> I know it's early but maybe?
Mar 16 05:17:29 <Herrie> Andolamin: No up already
Mar 16 05:17:48 <Andolamin> I think it just clicked why the builds might be failing
Mar 16 05:18:47 <Andolamin> QT_PACKAGE_ARCH is MACHINE_ARCH, which means it builds raspberrypi2 and raspberrypi3 respectively. BUT several components (including webos-keyboard) are built with qmake and aren't MACHINE_ARCH.
Mar 16 05:20:02 <Andolamin> Both the pi2 and the pi3 build cortexa7hf-vfp-vfpv4-neon so when you build e.g. the pi2 it builds correctly (qmake in raspberrypi2, webos-keyboard in cortexa7hf...)
Mar 16 05:21:07 <Andolamin> BUT if you then follow up and build the pi3 it builds qmake into raspberrypi3 but then tries to build webos-keyboard from sstate-cache and it fails because it's configured for qmake in raspberrypi2
Mar 16 05:21:15 <Andolamin> Same holds in reverse, of course
Mar 16 05:22:05 <Andolamin> SO, I think I just need to set QT_PACKAGE_ARCH to raspberrypi2 for the raspberrypi3 since the builds are supposed to be identical for these components.
Mar 16 05:22:39 <Andolamin> We'll have to change it back to MACHINE_ARCH when we get a proper raspberrypi3 tune, but then webos-keyboard et al wouldn't be building from cortexta7hf anyway
Mar 16 05:22:45 <Andolamin> Sound plausible?
Mar 16 05:23:24 <Andolamin> Also, morning! ;)
Mar 16 05:25:06 <Andolamin> Oh, and I think there's something off with the cleanup script for the raspberrypi2, as the wrong files showed up at http://build.webos-ports.org/luneos-testing/images/raspberrypi2/
Mar 16 05:27:50 <Andolamin> Well, I'm off to bed. I have a fresh build running to test my theory so hopefully I'll know for sure in the morning.
Mar 16 05:28:09 <Herrie> Andolamin: Yeah I noticed the images are not populating
Mar 16 05:28:12 <Herrie> Checking the logs now
Mar 16 05:28:26 <Andolamin> The build finished successfully though, so yay!
Mar 16 05:28:27 <Herrie> Suspect its: https://github.com/webOS-ports/webos-ports-setup/blob/master/scripts/staging_sync.sh#L22
Mar 16 05:28:35 <Herrie> Checking now
Mar 16 05:30:30 <Andolamin> *Looks* like that just creates the md5sums. Does look like it needs to be updated to look for *.rpi-sdimg, though.
Mar 16 05:31:38 <Andolamin> Anyway, bed now. I'll pop back on in the morning. Hopefully both the Pi3 and Pi2 build will have gone through.
Mar 16 05:35:53 <Herrie> Good night!
Mar 16 06:09:34 <Herrie|Laptop> JaMa: ping
Mar 16 07:33:14 <Tofe> Andolamin: too early for me, I'm no father :)
Mar 16 07:52:41 <Herrie|Veer> JaMa: ping
Mar 16 07:53:34 <Herrie|Veer> Tofe: morning
Mar 16 07:57:51 <Tofe> Morning!
Mar 16 07:58:19 <Tofe> Herrie|Veer: build is running for qemu :)
Mar 16 07:58:26 <Herrie|Veer> Tofe: Rpi2 image isn't rsyncing yet so changed some script and retrying now.
Mar 16 07:58:38 <Herrie|Veer> Build went OK though :)
Mar 16 07:58:49 <Herrie|Veer> Tofe: nice
Mar 16 09:44:19 <JaMa> Herrie|Veer: pong
Mar 16 10:25:07 <Herrie|Veer> And he disappeared LOL :P
Mar 16 10:26:45 <Tofe> Herrie|Veer: I just remarked... I forgot to switch to Qt 5.6 ! What an idiot :p
Mar 16 10:28:47 <Herrie|Veer> Tofe: I blame lack of coffee :P
Mar 16 10:30:35 <Tofe> Indeed I started the build at 7:00am... *before* coffee. Terrible mistake.
Mar 16 10:46:19 <Tofe> Ok, restarted with Qt 5.6. Hopefully the oe layers are still the same.
Mar 16 10:48:00 <Tofe> NB: there is a dependency on ninja-native and remove-libtool, which are in meta-oe and oe-core in jethro
Mar 16 11:04:15 <Herrie|Veer> JaMa: you're back :)
Mar 16 11:04:22 <Herrie|Veer> Need some help with the Pi
Mar 16 11:04:34 <Herrie|Veer> Build seems OK just rsync has some issues
Mar 16 11:04:55 <Herrie|Veer> Not 100% sure what causes it since I cannot see what's available file wise on build server
Mar 16 11:08:10 <JaMa> in luneos-testing there are currently only images for these: grouper/ maguro/  qemux86/
Mar 16 11:11:14 <Herrie|Veer> JaMa: The naming for Pi images is different
Mar 16 11:16:31 <Herrie|Veer> JaMa: We should be getting a luneos-dev-image-raspberrypi2.rpi-sdimg when we run MACHINE=raspberrypi2 bb luneos-dev-image
Mar 16 11:17:09 <JaMa> I've started the build to see
Mar 16 11:17:21 <JaMa> do we need to rsync kernel images and dtbs?
Mar 16 11:17:28 <JaMa> there are like 30 of them in deploy dir
Mar 16 11:17:29 <Herrie|Veer> But it could be jenkins.job.sh somehow removes it, it shouldn't as per lines 410-417
Mar 16 11:17:46 <Herrie|Veer> We only need to rsync the .rpi-sdimg
Mar 16 11:18:29 <Herrie|Veer> jenkins-job.sh should remove the others as per above lines
Mar 16 11:19:46 <JaMa> Herrie|Veer: some webos-ports-setup changes are missing in fido branch
Mar 16 11:19:50 <JaMa> e.g. staging_sync: Add .md5 files also for rpi-sdimg files
Mar 16 11:20:04 <JaMa> all changes should go to "fido" first, then testing branch should be rebased on top of that
Mar 16 11:20:08 <JaMa> and similarly stable branch
Mar 16 11:20:10 <Herrie|Veer> JaMa: OK
Mar 16 11:20:21 <Herrie|Veer> I just did that as a quick test this morning
Mar 16 11:20:50 <Herrie|Veer> To see if it made a difference but seems it doesn't. When we run testing from Jenkins it takes the testing branch right?
Mar 16 11:21:43 <JaMa> do we need to rsync bcm2835-bootfiles ?
Mar 16 11:23:44 <Herrie|Veer> JaMa: No only the .rpi-sdimg afaik
Mar 16 11:23:58 <Herrie|Veer> That will be put on the SD card and that's it.
Mar 16 11:25:13 <JaMa> ok, will update jenkins-job.sh to remove it and "Image"
Mar 16 11:26:35 <Herrie|Veer> JaMa I thought maybe luneos-package.inc#l22 might be the issue somehow? That we need to cater for .rpi-sdimg there as well?
Mar 16 11:26:56 <Herrie|Veer> This is in meta-luneos/recipes-core
Mar 16 11:28:47 <Herrie|Veer> Also jenkins-jobs.sh#l220 for the sstate shouldn't we expand that with the other targets as well?
Mar 16 11:29:25 <Herrie|Veer> it only has "qemux86 maguro grouper" now
Mar 16 11:30:00 <Herrie|Veer> Were just a few of my observations this morning after 1 coffee :P
Mar 16 11:31:09 <JaMa> Herrie|Veer:  we should keep only 3 in different order "maguro grouper qemux86" and then we can add --analyze parameter which is supported in fido
Mar 16 11:31:53 <Herrie|Veer> JaMa: OK, just want to understand the logics as to why :)
Mar 16 11:32:21 <JaMa> or we can call it twice, 2nd time with "raspberrypi2 raspberrypi3 mako"
Mar 16 11:32:33 <JaMa>   --analyze
Mar 16 11:32:34 <JaMa>         Show the differences between MACHINEs. It assumes:
Mar 16 11:32:34 <JaMa>         * First 2 MACHINEs in --machines parameter have the same TUNE_PKGARCH
Mar 16 11:32:36 <JaMa>         * Third optional MACHINE has different TUNE_PKGARCH - only native and allarch recipes are compared).
Mar 16 11:32:39 <JaMa>         * Next MACHINEs are ignored
Mar 16 11:33:22 <Herrie|Veer> JaMa: OK. I thought by adding the others we'd have better reuse of sstate or it doesn't take care of that?
Mar 16 11:33:35 <JaMa> commit e73e1261879d9154d89cec35669ba22b499d8331
Mar 16 11:33:35 <JaMa> Author: Martin Jansa <martin.jansa@gmail.com>
Mar 16 11:33:35 <JaMa> Date:   Wed Jun 25 19:06:36 2014 +0200
Mar 16 11:33:35 <JaMa>     sstate-diff-machines.sh: Return non-zero return code when there was some failure detected
Mar 16 11:33:38 <JaMa>     
Mar 16 11:33:40 <JaMa>     * add --analyze option, which is useful when using from jenkins job
Mar 16 11:33:43 <JaMa>       where you don't want to read output just to dectect how bad it was
Mar 16 11:33:45 <JaMa>     * I was always using something like this inside jenkins job, but better
Mar 16 11:33:48 <JaMa>       to share it in original script
Mar 16 11:33:50 <JaMa> no this is only about detecting the issues for sstate reuse
Mar 16 11:34:00 <JaMa> someone needs to read the output and fix the issues manually
Mar 16 11:35:02 <Herrie|Veer> Ah ok
Mar 16 11:39:53 <JaMa> Herrie|Veer: only tar.gz images are created, because local.conf has:
Mar 16 11:39:54 <JaMa> IMAGE_FSTYPES_forcevariable = "tar.gz"
Mar 16 11:40:19 <JaMa> so you'll have to add override for rpi MACHINEs
Mar 16 11:40:22 <JaMa> should I do it?
Mar 16 11:50:27 <JaMa> Herrie|Veer: 4 fixes pushed to jenkins-job, lets see how it works in next build
Mar 16 12:00:13 <Herrie|Veer> JaMa: thnx!
Mar 16 12:00:29 <Herrie|Veer> Sorry was on the road heading home
Mar 16 12:11:21 <Herrie|Veer> Feel free to kick off a testing build
Mar 16 12:11:49 <Herrie|Veer> I'm on the way for a couple of hrs so it will be finished then :)
Mar 16 12:11:55 <Herrie|Veer> When I'm back
Mar 16 12:12:53 <JaMa> it's already finishing the build
Mar 16 15:15:08 <DougReederPhone> Hey, is there a way to get the wi-fi MAC addr from the gui?
Mar 16 15:48:49 <Herrie|Veer> DougReederPhone: Not really but you could use FingerTerm app to query connman I guess :P
Mar 16 15:54:46 <Tofe> Don't we have ifconfig ?
Mar 16 15:55:11 <Herrie|Veer> I think we do :)
Mar 16 16:13:14 <DougReederPhone> Conman does not appear to be available, and ifconfig responds with 0 lines of output
Mar 16 16:13:47 <DougReederPhone> Ifconfig -a respons no sich device
Mar 16 16:13:55 <Herrie|Veer> DougReederPhone: try: connmanctl services
Mar 16 16:13:58 <Herrie|Veer> for example
Mar 16 16:14:19 <Herrie|Veer> DougReederPhone: What you want to achieve in general?
Mar 16 16:14:30 <JaMa> Herrie|Veer: I've updated the compare-signatures task as discussed, now it shows bunch of issues
Mar 16 16:14:33 <JaMa> http://jenkins.nas-admin.org/view/luneos-testing/job/luneos-testing_workspace-compare-signatures/6/console
Mar 16 16:14:47 <JaMa> I guess I'm the only one who understands what they mean and how to fix them..
Mar 16 16:15:25 <Herrie|Veer> JaMa: Hehe that might be... If you give pointers I'm happy to have a stab at it. I know you always have limited time :)
Mar 16 16:16:30 <Herrie|Veer> DougReederPhone: you can do "connmanctl services wifi_xxxxxxx_yyyyyy_security"
Mar 16 16:16:43 <DougReederPhone> Herrie, I need the wifi mac addr to register with a wifi network
Mar 16 16:16:57 <Herrie|Veer> This should give you a lot of info about the network.
Mar 16 16:17:45 <Herrie|Veer> DougReederPhone: Ah so you need the mac of the wifi adapter of the phone :)
Mar 16 16:19:34 <DougReederPhone> Yeah. Connmanctl services gives me that
Mar 16 16:19:54 <DougReederPhone> Damn, its hard to type in fingerterm
Mar 16 16:20:28 <Herrie|Veer> DougReederPhone: Yeah we should add this in Settings under a device info tab or something :P
Mar 16 16:22:38 <DougReederPhone> Thanks!
Mar 16 16:27:03 <DougReederPhone> I'm wondering if there's a good way to make it available in a drawer in wifi in preferences, but I'm thinking there's no good way to do that
Mar 16 16:28:37 <Herrie|Veer> DougReederPhone: Legacy had a device info app that held this info.
Mar 16 16:29:05 <Herrie|Veer> I think we expose it somewhere as well in webappmanager just don't use it anywhere currently
Mar 16 16:29:13 <DougReederPhone> Settings / About would probably be best
Mar 16 16:42:40 <Herrie|Veer> DougReederPhone: seems that it should be fairly straightforward. in About.js line 80 onwards. Seems there's a key called ro.mac.wifi that could be added there.
Mar 16 16:43:09 <Herrie|Veer> You might want to give that a try when you got some time. Similar one exists for BT called ro.mac.bluetooth :)
Mar 16 16:45:43 <Herrie|Veer> This appearently comes from Androids build.prop file so will only work on Android based targets I guess.
Mar 16 16:48:38 <DougReederPhone> Long term, it would be lovely to be able to specify that no unsecured connections could be made over a particular SSID
Mar 16 16:49:23 <Herrie|Veer> DougReederPhone: It depends what we get back from the network when it tries to connect I guess.
Mar 16 16:50:55 <DougReederPhone> that would probably require changes higher in the network stack 😞
Mar 16 16:54:52 <Herrie|Veer> DougReederPhone: It might actually return something useful in the error message already. elvispre has been toying quite a bit with wifi recently. Quite sure he knows :)
Mar 16 17:50:10 <Herrie> Andolamin: http://build.webos-ports.org/luneos-testing/images/raspberrypi2/ :D
Mar 16 17:50:17 <Herrie> JaMa fixed some bits today :D
Mar 16 17:53:26 <Andolamin> Awesome!
Mar 16 17:54:14 <Andolamin> At the pediatrician right now but will flash it when I get back
Mar 16 17:58:03 <Tofe> Herrie|Veer: fido+qt5.6 works well on qemu ==> it's specific to the hybris stack
Mar 16 17:58:51 <Tofe> Maybe we could try an older commit of hybris
Mar 16 18:02:46 <Herrie|2> Tofe: Ah nice, at least good to know 5.6 itself it
Mar 16 18:03:24 <Herrie|2> is pretty OK then. Can you check if they added the html5 notifications in 5.6? There's a test for it in Testr app ;-)
Mar 16 18:03:39 <Herrie|2> So should be easy enough to spot :-)
Mar 16 18:15:38 <Tofe> Herrie: denied
Mar 16 18:29:31 <Herrie|Veer> Tofe: hmmz but that could be permissions, can you share image with me?
Mar 16 18:34:07 <Herrie> nizovn: ping
Mar 16 18:34:48 <DougReederPhone> Html5 specifies that the user must allow a site to post notifications
Mar 16 18:35:10 <Herrie> DougReederPhone: Yeah I think they didn't implement the permissions in 5.5 yet
Mar 16 18:35:19 <Herrie> So could be we need to add them in 5.6 in the QML somewhere
Mar 16 18:35:21 <DougReederPhone> Run Tester in a desktop browser to see
Mar 16 18:53:43 <nizovn> Herrie: pong
Mar 16 18:59:52 <Tofe> Herrie|Veer: I can, but you'll have to be patient, it'll take about one hour to upload
Mar 16 19:02:07 <Herrie|Veer> Tofe: fine no rush
Mar 16 19:04:08 <Tofe> Well I'm not there this week-end, so better I share it now
Mar 16 19:07:22 <Tofe> Herrie|Veer: you've got mail
Mar 16 19:08:09 <Tofe> uh wait, I'm not sure the link works
Mar 16 19:33:00 <Herrie> nizovn: sorry for late reply was putting little one to bed
Mar 16 19:35:11 <nizovn> np
Mar 16 19:35:12 <Herrie> You see anyhting odd in https://github.com/webOS-ports/android-property-service/blob/master/src/property_service.c ? When I do a luna-send for com.android.properties/getAllProperties I get a proper reply, however when the same used in Enyo 2 app: Mar 16 18:52:02 mako android-property-service[9304]: *** Error in `/usr/sbin/android-property-service': free(): corrupted unsorted chunks:
Mar 16 19:35:12 <Herrie> 0x010b71a0 *** and Mar 16 18:52:02 mako kernel[698]: [ 4024.039614] type=1701 audit(1458150722.862:18): auid=4294967295 uid=0 gid=0 ses=4294967295 pid=9304 comm="android-propert" reason="memory violation" sig=6
Mar 16 19:35:39 <Herrie> If you have any ideas?
Mar 16 19:35:52 <Herrie> The reply is quite long in luna-send so could be a size issue?
Mar 16 19:36:18 <Herrie> https://bpaste.net/show/ff89b44e104b
Mar 16 19:38:47 <Herrie> Tofe: I'm getting 100 kb/s :P
Mar 16 19:39:13 <elvispre> Wi-Fi details show up in the system logs - including MAC address, I think - in a connman message from time to time.
Mar 16 19:39:30 <EricBlade> Herrie: that's not an exceptionally long ls2 response.  I've seen megabytes of data in single LS2 responses.  not on Lune, though.  Would be interesting to get a gdb trace on that.
Mar 16 19:39:58 <elvispre> I was thinking of trying to get hold of it for the Wi-Fi details panel which is conspicuously empty at the moment.
Mar 16 20:32:35 <Andolamin> Herrie: ping
Mar 16 20:34:47 <Andolamin> Herrie: https://github.com/webOS-ports/meta-webos-ports/blob/fido/meta-luneos/recipes-webos/luna-sysmgr-conf/luna-sysmgr-conf.bb is out of date, looks like it should point to bcbe46be9b841e5d8f9a59f7254d407cd74176ea
Mar 16 20:36:06 <DougReederPhone> Elvispre, the wifi details panel where?
Mar 16 20:37:33 <Andolamin> Herrie: Other than that issue (causes UI to be scaled wrong) the build looks good
Mar 16 20:46:00 <Herrie> Andolamin: We do those bumps manually, sometimes one gets missed out :S
Mar 16 20:46:03 <Herrie> Fixed now
Mar 16 20:46:12 <Herrie> How about the Pi3?
Mar 16 20:46:19 <Herrie> Can we start building it too?
Mar 16 21:13:24 <Andolamin> Herrie: Not as is
Mar 16 21:13:42 <Andolamin> Hoping to have the layers updated so that the builds will work within the next day or two
Mar 16 21:14:55 <Andolamin> Herrie: I think I have a solution to the problem, but would like JaMa's feedback before I try to integrate it
Mar 16 21:45:25 <nizovn> Herrie: this crash is caused by https://github.com/webOS-ports/android-property-service/blob/master/src/property_service.c#L97-L98 . will figure out why tomorrow.
Mar 16 21:57:11 <elvispre> DougReederPhone: You get a details panel if you tap the connected entry in the list. It has lots of headings but no values. Its only actual function is to let you forget that network again.
**** ENDING LOGGING AT Thu Mar 17 02:59:58 2016