**** BEGIN LOGGING AT Mon Oct 16 03:00:02 2017
Oct 16 06:41:53 <Tofe> Herrie: are you sure you updated webos-ports-setup/testing ?
Oct 16 06:55:04 <Herrie|Pre3> Morning
Oct 16 06:57:05 <Herrie|Pre3> JaMa didn't, cancelled his build, I should have done it
Oct 16 06:57:59 <Herrie|Pre3> had some issues with node-gyp packages. reverted that commit
Oct 16 06:59:04 <Herrie|Pre3> Restarted build, tenderloin ok. We'll need another build for N5 and wifi module fix for N4
Oct 16 07:25:01 <Tofe> I'm starting a hammerhead at home
Oct 16 07:25:11 <Tofe> +build
Oct 16 07:25:33 <Tofe> Herrie|Pre3: what wifi module fix ?
Oct 16 07:26:02 <Herrie|Pre3> Tofe: The kernel module
Oct 16 07:26:12 <Herrie|Pre3> I need to see what I did locally
Oct 16 07:26:32 <Herrie|Pre3> Could be I forgot a small bit in my PR
Oct 16 07:27:27 <Herrie|Pre3> kernel-module-wlan
Oct 16 07:27:30 <Tofe> the kernel module is missing ?
Oct 16 07:27:33 <Herrie|Pre3> Probably is a defconfig flag
Oct 16 07:31:21 <Herrie|Pre3> Couldn't find anything to satisfy 'kernel-module-wlan'
Oct 16 07:31:37 <Tofe> Herrie|Pre3: it's probably not built at all
Oct 16 07:33:20 <Tofe> Herrie|Pre3: you removed the defconfig altogether for mako ? that's a bit strange, isn't it?
Oct 16 07:33:44 <Tofe> ah ok, you reuse as-is the one from ubports
Oct 16 07:33:58 <Herrie|Pre3> Tofe: Yeah
Oct 16 07:34:14 <Herrie|Pre3> I cannot connect to my home network for some reason. Maybe IP switched :P
Oct 16 07:34:22 <Herrie|Pre3> Need to setup some dyndns or something
Oct 16 07:34:40 <Herrie|Pre3> Provider used to give me static but seems to change quite frequently now :s
Oct 16 07:34:48 <Tofe> https://github.com/ubports/android_kernel_google_msm/blob/cm-14.1/arch/arm/configs/mako_defconfig#L468 ah yes they chose not to have the wifi code as a module
Oct 16 07:35:01 <Tofe> They didn't even activate modules in that kernel
Oct 16 07:35:05 <Herrie|Pre3> Tofe: Yeah I used their defconfig to test locally and I had the same issue
Oct 16 07:35:13 <Herrie|Pre3> But fixed it somehow
Oct 16 07:35:16 <Herrie|Pre3> Just need to check how
Oct 16 07:35:47 <Herrie|Pre3> It's a bit weird we require it as a module but that was due to fixed mac address right?
Oct 16 07:35:56 <Tofe> right
Oct 16 07:36:10 <Herrie|Pre3> I could add a simple patch for this value in our recipe
Oct 16 07:36:11 <Tofe> so I wonder if that still works
Oct 16 07:36:27 <Herrie|Pre3> We should be fine with rest of their defconfig
Oct 16 07:36:42 <Herrie|Pre3> Or we should see how the other guys get persisting mac
Oct 16 07:39:30 <Tofe> you didn't get the same error for N5 ? the situation seems to be identical
Oct 16 07:40:18 <Herrie|Pre3> Tofe: nope
Oct 16 07:40:26 <Tofe> strange
Oct 16 07:40:38 <Herrie|Pre3> Well Hammerhead didn't build yet
Oct 16 07:41:00 <Herrie|Pre3> I'll let this set of builds finish and restart another hammerhead just
Oct 16 07:42:18 <Herrie|Pre3> It failed with initially with the node-gyp-native which I now reverted
Oct 16 07:42:47 <Herrie|Pre3> There's something weird going on with node-gyp-native and I wasn't able to find much leads with some hours of digging
Oct 16 07:42:54 <Herrie|Pre3> Might poke JaMa for that
Oct 16 07:43:13 <Herrie|Pre3> We need those for node-sqlite3 and some other bits like pmloglib it seems
Oct 16 07:54:37 <Herrie|Pre3> Update isn't that urgent but would be good to have at some point
Oct 16 07:55:39 <Herrie|Pre3> So we have node-sqlite3 back which is used by keymanager
Oct 16 07:57:21 <Herrie|Pre3> I had both Hammerhead & Mako building locally so need to see what I did
Oct 16 07:57:30 <Herrie|Pre3> Had a bit of a frankenbuild
Oct 16 07:57:38 <Herrie|Pre3> So could be I forgot a small bit
Oct 16 07:57:48 <Tofe> Herrie|Pre3: virtual/kernel build will succeed, but not luneos-dev-package
Oct 16 07:59:05 <Herrie|Pre3> Well I had whole images building at home :P
Oct 16 07:59:15 <Herrie|Pre3> So could be I forgot something small really
Oct 16 07:59:19 <Herrie|Pre3> Need to check
Oct 16 07:59:23 <Tofe> ok :)
Oct 16 07:59:35 <Herrie|Pre3> Or I used a local defconfig still :s
Oct 16 07:59:39 <Herrie|Pre3> Anyway will fix it
Oct 16 07:59:44 <Herrie|Pre3> Should be easy patch
Oct 16 08:02:30 <Tofe> well, using a local defconfig isn't that bad :p
Oct 16 08:05:30 <Herrie|Pre3> Tofe: Yeah but would be easier if we could use upstream but well
Oct 16 08:05:46 <Herrie|Pre3> Not maintaining kernel code already improvement
Oct 16 09:31:21 <Tofe> https://www.krackattacks.com/ not good...
Oct 16 09:38:30 <Herrie|Pre3> Tofe: No indeed :'(
Oct 16 09:38:42 <Herrie|Pre3> Long wait for updates it seems will start :'(
Oct 16 09:39:39 <Tofe> yup, seems like wpa_supplicant should be updated urgently
Oct 16 09:41:47 <Herrie|Pre3> Tofe: Yocto is usually quite quick with such CVE updates
Oct 16 09:45:00 <Herrie|Pre3> Seems they're on it on WPA already ;)
Oct 16 09:45:15 <Tofe> hehe
Oct 16 09:45:20 <Herrie|Pre3> http://w1.fi/cgit/log/
Oct 16 09:45:37 <Tofe> well, the number of LuneOS client devices in the wild is pretty small ;)
Oct 16 09:45:43 <Herrie|Pre3> I guess we could include some of the patches as a band aid :P
Oct 16 09:47:21 <Herrie|Pre3> But I suspect they'll push out 2.6.1 or 2.7 shortly
Oct 16 09:47:29 <Herrie|Pre3> To properly address this.
Oct 16 09:48:52 <Tofe> yes, I guess so too
Oct 16 10:30:50 <JaMa> Herrie|Pre3: it will take a while until my bitbake world will check the qt5 change, last build failed because of other change I was testing
Oct 16 10:31:43 <Herrie|Pre3> JaMa: OK
Oct 16 10:32:20 <Herrie|Pre3> We switched to my meta-qt5 for testing until you have changes merged
Oct 16 10:33:09 <JaMa> yes, I've noticed
Oct 16 10:38:47 <Herrie|Pre3> Since we know it'll probably take a while. I tested it with our layers and seems fine for hammerhead, tenderloin, qemux86 and raspberrypi3-64
Oct 16 10:38:56 <Herrie|Pre3> So you should be fairly OK
Oct 16 11:02:50 <Tofe> Herrie|Pre3: my hammerhead build just finished
Oct 16 11:12:04 <Herrie|Pre3> Tofe: OK :)
Oct 16 12:38:02 <Herrie|Pre3> Tofe: On Jenkins it also finished :)
Oct 16 12:41:56 <Tofe> Ah great
Oct 16 12:42:00 <Tofe> testing time tonight then
Oct 16 12:44:14 <Herrie|Pre3> Tofe: Yup
Oct 16 13:00:37 <Herrie|Pre3> MartinHov: Therre should be a new emulator image later today with some more significant changes ;)
Oct 16 13:00:53 <Herrie|Pre3> JaMa: Not sure you saw my comments on node-gyp-native upgrade.
Oct 16 13:01:09 <Herrie|Pre3> I tried to debug quite some hours on Saturday but with little luck
Oct 16 13:06:52 <MartinHov> Herrie: the new emulator bits (testing) have a new time stamp on the directory but the files are still 12. Oct and version 0-483 !  Did the build fail ?
Oct 16 13:09:18 <Herrie|Pre3> MartinHov: It still needs to sync
Oct 16 13:09:30 <Herrie|Pre3> Might take some hrs still since some builds are still running
Oct 16 13:15:03 <MartinHov> Herrie: ok - I downloaded the qumex86 bits some 3h ago after the directory was time-stamped today - but still the bits are from 12. Oct only
Oct 16 13:17:00 <Herrie|Pre3> MartinHov: Due to qt 5.9.2 release upgrade builds take longer because they need to rebuild qtbase + qtwebengine for all target architectures ;)
Oct 16 13:17:09 <Herrie|Pre3> Next build should go a lot quicker ;)
Oct 16 13:18:53 <Herrie|Pre3> With a lot quicker I mean < 20 mins per target instead of 3,5 hrs :
Oct 16 13:18:54 <Herrie|Pre3> :P
Oct 16 13:23:37 <JaMa> Herrie|Pre3: nope, looking now
Oct 16 13:44:20 <Herrie|Pre3> JaMa: Thnx
Oct 16 13:44:34 <Herrie|Pre3> I'm not really sure what the issue is to be honest
Oct 16 13:44:43 <Herrie|Pre3> It seems somehow related to the offline registry
Oct 16 13:50:30 <JaMa> Herrie|Pre3: halfhalo actually might know more about this one, he added --no-registry for some reason somewhere else
Oct 16 13:50:43 <JaMa> Herrie|Pre3: but with pyro I'm not seeing any build issue with this
Oct 16 13:50:50 <Herrie|Pre3> I assume this might have been addressed internally as well @ SVL
Oct 16 13:51:07 <Herrie|Pre3> JaMa: That's because we use the old gyp-native
Oct 16 13:51:12 <Herrie|Pre3> Which is still 0.12 based
Oct 16 13:51:33 <Herrie|Pre3> If you update to 4.0.0 you'll need to update the registry as well due to version requirements it seems
Oct 16 13:51:40 <JaMa> you mean the registry?
Oct 16 13:51:42 <Herrie|Pre3> When I attempt to do that I get these url issues
Oct 16 13:52:01 <Herrie|Pre3> JaMa: Old gyp-native uses old registry
Oct 16 13:52:24 <JaMa> node-gyp-packages is still at 0.12.2 internally :)
Oct 16 13:52:30 <Herrie|Pre3> The packages in the registry aren't compatible with gyp-native 4.0.0 and need updating.
Oct 16 13:52:32 <Herrie|Pre3> JaMa: Ah....
Oct 16 13:52:48 <Herrie|Pre3> I.e. build fails when I don't update them
Oct 16 13:52:49 <JaMa> but also with old node-gyp-native
Oct 16 13:53:04 <Herrie|Pre3> I update them from npm and just paste the json from there
Oct 16 13:53:06 <JaMa> hmm not sure why it works in unstable builds
Oct 16 13:53:13 <Herrie|Pre3> Could be something is wrong with that
Oct 16 13:53:34 <Herrie|Pre3> JaMa: Maybe you disabled the components using it :p ?
Oct 16 14:34:42 <Herrie|Pre3> Tofe: Seems wpa_supplicant has a set off patches ready. I can test those tonight to see if they build and push them.
Oct 16 14:58:01 <Tofe> Herrie|Pre3: well, it'll be in our next release anyway, so it's not so much of a rush for us
Oct 16 15:01:59 <Herrie|Pre3> Tofe: Nope but good to have it closed :P
Oct 16 15:02:04 <Herrie|Pre3> So will have a quick look
Oct 16 16:43:10 <Tofe> Herrie|Pre3: signature verification failed when sending the zip via sideload, so I think the zip is incorrect
Oct 16 16:55:08 <Herrie|Pre3> Tofe: That happens sometimes with me too
Oct 16 16:55:15 <Herrie|Pre3> Let me try when I'm home
Oct 16 16:55:25 <Herrie|Pre3> I have that happening sometimes
Oct 16 16:55:32 <Herrie|Pre3> Rebuilding it fixes it usually
Oct 16 17:18:34 <Tofe> First time I see it
Oct 16 17:33:08 <JaMa> Herrie|Pre3: quite a few failures from 5.9.2 change: http://errors.yoctoproject.org/Errors/Build/47537/
Oct 16 17:55:09 <Herrie> JaMa: I only tested with our image, so it could be that oe-world will give such errors
Oct 16 17:55:13 <Herrie> Let me see what I can address
Oct 16 18:03:47 <Tofe> Herrie: looks like I wasn't using the latest meta-smartphone in my build; I fixed it, relaunched a build (should be almost only the kernel and libhybris stuff)
Oct 16 18:28:37 <Herrie> Tofe: OK
Oct 16 19:08:23 <MartinHov> Hmm, the qemux86 testing image is still not renewed .. # is the build still running ??
Oct 16 19:11:19 <Herrie> MartinHov: Yeah
Oct 16 19:11:24 <Herrie> Last one now
Oct 16 19:11:29 <Herrie> RaspberryPi3-64 bits
Oct 16 19:11:33 <Herrie> Takes a while
Oct 16 19:11:38 <Herrie> 3,5 hrs each due large rebuild
Oct 16 19:34:38 <MartinHov> AH - ok - so builds run one at a time - not in parallel ..
Oct 16 19:47:26 <Tofe> damn, downloading a kernel at 50KB/s isn't fun
Oct 16 19:57:19 <Herrie> MartinHov: Yes, because they reuse bits form each other
Oct 16 19:57:31 <Herrie> I.e. ARMv7 targets like N4, N5 and TP reuse bits where possible
Oct 16 19:57:41 <Herrie> So we only build qtbase & qtwebengine once
Oct 16 19:57:52 <MartinHov> ok - I see
Oct 16 19:58:09 <Herrie> But we have qemux86, qemux86-64 and raspberrypi3-64 as well
Oct 16 19:58:18 <Herrie> So that's 4* 3,5 hrs
Oct 16 19:58:43 <MartinHov> qemux86 and qemux64 are build independently .. - right ?
Oct 16 19:58:43 <Herrie> But as long as we don't touch qtwebengine/qtbase next builds usually finish in 20 minutes per device/target
Oct 16 19:58:49 <Herrie> MartinHov: Yup
Oct 16 19:58:52 <Herrie> 32 and 64 bits
Oct 16 19:58:58 <Tofe> ok, build finished, flashing seems to be going fine
Oct 16 19:59:24 <Herrie> JaMa: https://github.com/Herrie82/meta-qt5/commit/6936319942f9b411defc829ed34ecaed6cccd2b6
Oct 16 19:59:27 <Herrie> That should work better
Oct 16 20:00:03 <Herrie> I moved back to 1.3.10 for Tufao due to 1.4.x needing boost.http
Oct 16 20:00:10 <Herrie> Which I couldn't really find
Oct 16 20:00:14 <Herrie> Just regular boost
Oct 16 20:04:57 <Tofe> Herrie: N5 is all fine for me ! no regression here.
Oct 16 20:05:21 <Tofe> Though I quite don't understand how the wifi module got built :p
Oct 16 20:07:32 <Tofe> ah now I understand, I was looking at the wrong defconfig. Ok, all clear then :)
Oct 16 20:08:58 <Herrie> Tofe: Yeah
Oct 16 20:09:01 <Herrie> They have 2
Oct 16 20:09:08 <Herrie> I ahd the wrong one initially too :P
Oct 16 20:09:48 <Herrie> Tofe: Only issue is BT I guess currently in terms of regression
Oct 16 20:09:52 <Herrie> But that's kinda expected
Oct 16 20:10:02 <Herrie> I guess we should tackle the BlueZ5 upgrade sometime soon
Oct 16 20:11:07 <Tofe> Herrie: BT worked before on N5 ?
Oct 16 20:11:14 <Herrie> Tofe: Well a bit I guess
Oct 16 20:11:18 <Herrie> I didn't test it in detail
Oct 16 20:11:26 <Herrie> Anyway the BlueZ5 would be good
Oct 16 20:11:34 <Herrie> Just Settings would be an issue
Oct 16 20:11:43 <Herrie> BlueZ5 required 3.4 kernel at least
Oct 16 20:11:48 <Tofe> Settings ?
Oct 16 20:11:56 <Herrie> Settings app with the C++ plugin
Oct 16 20:12:07 <Herrie> THough not sure how much work it would be to update the plugin
Oct 16 20:12:08 <Tofe> ah, nevermind, we'll have a QML one :p
Oct 16 20:12:20 <Herrie> Tofe: Well that needs polishing, but yes
Oct 16 20:12:43 <Tofe> Herrie: polishing, you're being gentle, there's still 95% of the app to write :D
Oct 16 20:13:58 <Tofe> But updating the C++ plugin might not be that much of a problem, I think. I'd need to check more carefully though.
Oct 16 20:19:11 <JaMa> Herrie: I've already dropped it from meta-qt5 to get more successful build, but will try to include after next round is finished building
Oct 16 20:20:13 <JaMa> Herrie: hmm in some .patch files you're changing author, you cannot do that
Oct 16 20:20:32 <JaMa> add your +Signed-off-by: but keep original author
Oct 16 20:21:43 <Herrie> JaMa: Ehm sorry not my intention let me fix that
Oct 16 20:24:39 <Herrie> JaMa: Fixed
Oct 16 20:24:55 <Herrie> JaMa: This all builds with Hammerhead for me
Oct 16 20:25:08 <Herrie> Tested the components 1 by 1 to make sure
Oct 16 20:25:22 <Herrie> I guess I didn't pick some up because we don't use them in our layers
Oct 16 20:25:38 <Herrie> Tofe: Well, don't want to demotivate ;) But yes Settings needs quite some work
Oct 16 20:27:29 <Herrie> Tofe: For migration of BlueZ this should be helpful: http://www.bluez.org/bluez-5-api-introduction-and-porting-guide/
Oct 16 20:29:48 <Herrie> JaMa: I guess for the gyp-native I'll just try to add a single version of each package to the registry instead of replacing the whole file
Oct 16 20:29:54 <Herrie> That might lead to less issues
Oct 16 21:17:19 <JaMa> Herrie: ok, thanks
Oct 16 22:39:37 <MartinHov> testing qemux86.zip is touched now - but not the right file size. I will look next day ..
**** ENDING LOGGING AT Tue Oct 17 03:00:02 2017