**** BEGIN LOGGING AT Sun Jun 16 02:59:58 2019
Jun 16 04:56:20 <Herrie> Tofe: OK PR's are done for when you're awake and had your first coffee. Relevant bits also PR-ed upstream.
Jun 16 05:19:30 <Herrie> Tofe: OK a few more to follow after some clicking around ;)
Jun 16 05:24:21 <Herrie> OK also pushed fixes for Browser & Email apps
Jun 16 06:58:25 <Herrie> Tofe: Seems there's one more keymanager issue to fix as well. Not really sure how & where yet, would need to add some debugging in the .js I guess to figure out which one is offending: 2019-06-16T05:19:33.797898Z [32135] user.err keymanager [] legacy-log  TypeError [ERR_INVALID_ARG_TYPE]: The "value" argument must not be of type number. Received type number     at new from (buffer.js:215:11
Jun 16 06:58:25 <Herrie> )     at javascript/utils/KeyStore.js:60:38     at Array.forEach (<anonymous>)     at Object.copyKey (javascript/utils/KeyStore.js:57:19)     at Object.getKeyRawByName (javascript/utils/KeyStore.js:81:32)     at Object.getKeyDecryptedByName (javascript/utils/KeyStore.js:100:35)     at FetchKeyAssistant.run (javascript/assistants/FetchKeyAssistant.js:17:14)     at Future._docall
Jun 16 06:58:25 <Herrie> (/usr/palm/frameworks/foundations/version/1.0/node_module.js:1:10066)     at Future.now (/usr/palm/frameworks/foundations/version/1.0/node_module.js:1:8725)     at c.runCommand (/usr/palm/frameworks/mojoservice/version/1.0/node_module.js:1:9779)
Jun 16 06:58:49 <Herrie> Seems similar to https://github.com/nodejs/node/issues/26741 and https://github.com/the-road-to-graphql/fullstack-apollo-express-postgresql-boilerplate/issues/46
Jun 16 07:08:14 <Herrie> So probably something in https://github.com/webOS-ports/keymanager/commit/64b90ead778fe5cc127dc1c1fd720df9e4300f36 needs some tweaking still
Jun 16 07:09:18 <Tofe> Morning!
Jun 16 07:09:34 <Tofe> wow, you've already been busy !
Jun 16 07:12:07 <Tofe> Herrie: something I really wonder, is why does db8 need outbound permission for some apps ? shouldn't db8 be app-agnostic ?
Jun 16 07:13:02 <Tofe> PRs merged, it's what I would have done too
Jun 16 07:22:38 <Herrie> Tofe: Well db8 has 2 permission models it seems :S
Jun 16 07:22:55 <Herrie> dynamic & static
Jun 16 07:23:02 <Herrie> Just I'm not really sure which one does what :S
Jun 16 07:24:04 <Herrie> Tofe: 5:30 this morning someone decided it was a good time to wake up :P
Jun 16 07:24:21 <Herrie> I added some debugging to my keymanager to see if I can track it down
Jun 16 07:25:55 <Tofe> ah, the joy of children
Jun 16 07:26:24 <Herrie> Hmmz I have emails now in my email app :)
Jun 16 07:26:27 <Herrie> So that's good ;)
Jun 16 07:26:37 <Tofe> ah great
Jun 16 07:27:00 <Herrie> Contacts still no luck
Jun 16 07:27:05 <Herrie> But we'll see after a new build
Jun 16 07:30:30 <Herrie> Ah some new issues in emails :S
Jun 16 07:30:35 <Herrie> CRITICAL: 03:29:44.992: Error: invalid parameters: caller='com.palm.app.email' error='property not allowed - '$activity''
Jun 16 07:33:04 <Herrie> Tofe: Also https://github.com/webOS-ports/configurator/pull/3 and https://github.com/webOS-ports/filecache/pull/3
Jun 16 07:33:09 <Herrie> Gave you permissions to merge
Jun 16 07:36:47 <Tofe> thanks
Jun 16 07:39:09 <Tofe> I'm a bit surprised LG let these outbound error in OSE
Jun 16 08:02:41 <Herrie> Tofe: Me too, but could also be due to our specifics
Jun 16 08:02:50 <Herrie> They use very little activities in OSE it seems
Jun 16 08:02:53 <Herrie> We use quite a bit more
Jun 16 08:03:00 <Herrie> Could be that some only appear for us
Jun 16 08:04:03 <Tofe> Maybe yes
Jun 16 08:08:38 <Herrie> Tofe: For keymanager: This is the offending line: https://github.com/webOS-ports/keymanager/blob/64b90ead778fe5cc127dc1c1fd720df9e4300f36/service/javascript/utils/KeyStore.js#L60
Jun 16 08:08:51 <Herrie> Herrie 4b: typeof from[name].length number
Jun 16 08:09:03 <Herrie> TypeError [ERR_INVALID_ARG_TYPE]: The "value" argument must not be of type number.
Jun 16 08:12:11 <Tofe> mmmh I wonder what that line meant before... pre-allocating space in the buffer ? interpret length as a string ?
Jun 16 08:12:43 <Tofe> I'll look a bit in that code
Jun 16 08:13:04 <Herrie> It should be enough to cast it into a string I guess
Jun 16 08:13:40 <Tofe> ok, so "dest[name] = new Buffer.from(String(from[name].length));" then
Jun 16 08:14:02 <Tofe> I can PR that
Jun 16 08:14:31 <Herrie> Let me test it first
Jun 16 08:18:46 <Herrie> Seems we need something else :P Error: error:0606506D:digital envelope routines:EVP_DecryptFinal_ex:wrong final block length     at Decipher._flush (internal/crypto/cipher.js:134:28)     at Decipher.prefinish (_stream_transform.js:141:10)     at Decipher.emit (events.js:189:13)     at prefinish (_stream_writable.js:630:14)     at finishMaybe (_stream_writable.js:638:5)     at endWritable
Jun 16 08:18:46 <Herrie> (_stream_writable.js:649:3)     at Decipher.Writable.end (_stream_writable.js:589:5)     at _crypt (javascript/utils/KeyStore.js:40:16)     at decrypt (javascript/utils/KeyStore.js:46:16)     at Object.rawCB (javascript/utils/KeyStore.js:115:25)
Jun 16 08:19:40 <Tofe> https://nodejs.org/api/buffer.html#buffer_new_buffer_size
Jun 16 08:19:51 <Tofe> let me propose a real fix :)
Jun 16 08:21:16 <Tofe> ah no wait I was looking at the wrong constructor
Jun 16 08:22:21 <Tofe> mmmh weird code... maybe we modified this before during a nodejs upgrade?
Jun 16 08:23:37 <Tofe> yep https://github.com/webOS-ports/keymanager/commit/64b90ead778fe5cc127dc1c1fd720df9e4300f36#diff-6be2fd50b818c5c05c2ec75fdf71875eL60
Jun 16 08:23:40 <Herrie> Yes I did with this one: https://github.com/webOS-ports/keymanager/commit/64b90ead778fe5cc127dc1c1fd720df9e4300f36
Jun 16 08:23:53 <Herrie> But it was merely what was done before ;)
Jun 16 08:24:00 <Herrie> Just with new API
Jun 16 08:24:10 <Tofe> can you try "dest[name] = new Buffer.alloc(from[name].length);"
Jun 16 08:27:06 <Tofe> and as I'm quite confident on that one: https://github.com/webOS-ports/keymanager/pull/6
Jun 16 08:30:55 <Herrie> Testing now
Jun 16 08:33:08 <Herrie> Yup works
Jun 16 08:33:23 <Tofe> :)
Jun 16 08:52:01 <Herrie> Well have a bunch of other permission errors now with the IMAP not being able to write stuff to filecache for example
Jun 16 08:52:08 <Herrie> But it's progress at least
Jun 16 08:56:08 <Herrie> Bit by bit we'll get there
Jun 16 08:56:30 <Herrie> Will head out for a bit with Mrs and little ones, back in some hrs I guess
Jun 16 08:58:55 <Tofe> ok ! :)
Jun 16 11:36:37 <saidinesh5> Tofe: Herrie do you know what drm-egl-server does in QtWayland?
Jun 16 11:38:35 <Tofe> saidinesh5: mmh no... maybe a wayland extension ?
Jun 16 11:38:57 <saidinesh5> oh okay
Jun 16 11:39:35 <saidinesh5> PACKAGECONFIG[wayland-brcm] = "-feature-wayland-brcm,-no-feature-wayland-brcm,virtual/egl"
Jun 16 11:39:35 <saidinesh5> #PACKAGECONFIG[drm-egl-server] = "-feature-drm-egl-server,-no-feature-drm-egl-server,libdrm virtual/egl"
Jun 16 11:39:35 <saidinesh5> #PACKAGECONFIG[libhybris-egl-server] = "-feature-libhybris-egl-server,-no-feature-libhybris-egl-server,libhybris"
Jun 16 11:39:55 <saidinesh5> this stuff is commented out in the qtwayland_git.bbappend
Jun 16 11:39:58 <saidinesh5> so was curious
Jun 16 11:40:47 <saidinesh5> (earlier it was whining about ERROR: qtwayland-5.12.3+gitAUTOINC+bcdc00bcdc-r0 do_configure: QA Issue: qtwayland: invalid PACKAGECONFIG: drm-egl-server [invalid-packageconfig] )
Jun 16 11:49:31 <saidinesh5> Also any idea what's the difference between EGL and DRM EGL drivers for qtwayland?
Jun 16 11:50:25 <saidinesh5> https://bpaste.net/show/9305082019b8
Jun 16 12:38:07 <Herrie> Well DRM ones will allow DRM protected content I guess such as Netflix, Amazon Prime etc
Jun 16 12:38:23 <Herrie> At least that's my guess. Tofe or JaMa might have some more insight
Jun 16 12:48:47 <Tofe> Herrie: in this context, I think it's Direct Rendering Management, i.e. mainline kernel rendering
Jun 16 13:03:07 <saidinesh5> but EGL in mesa is implemented with DRM itself right?
Jun 16 13:04:15 <saidinesh5> Herrie: the protected content is mostly done with libwidevine or something .. and i don't think that cares much about wayland or x11 or anything iirc
Jun 16 13:10:52 <saidinesh5> bah ran out of disk space
Jun 16 13:10:59 <saidinesh5> had like 160GB!!!
Jun 16 13:12:55 <Tofe> saidinesh5: llvm + llvm-native take 41GB only by themselve
Jun 16 13:13:07 <saidinesh5> Tofe: what needs llvm?
Jun 16 13:13:13 <Tofe> mesa + gallium
Jun 16 13:13:17 <saidinesh5> that took so long to compile too
Jun 16 13:13:18 <saidinesh5> Ohh
Jun 16 13:14:31 <Tofe> saidinesh5: but you pointed out something interesting here; I'll try activating "wayland-drm-egl-server-buffer" for pinephone, which also uses Mesa/Gallium
Jun 16 13:14:48 <Tofe> (iirc, it's the new name of wayland-drm-egl-server)
Jun 16 13:15:00 <saidinesh5> Ahh go ahead! i will have to figure out where i can make some space up from
Jun 16 13:15:08 <saidinesh5> ohh
Jun 16 13:15:17 <saidinesh5> i thought it was some developer thing
Jun 16 13:15:48 <Tofe> I'm not sure yet
Jun 16 13:16:12 <saidinesh5> one of those server things was developer thing not meant to be in production devices
Jun 16 13:16:13 <Tofe> but the alternative is using this "Shm emulation server buffer integration" which looks like a software fallback
Jun 16 13:17:15 <saidinesh5> https://github.com/qt/qtwayland/tree/5.12/examples/wayland/server-buffer
Jun 16 13:17:58 <saidinesh5> "Note: the shm-emulation-server integration does not actually share graphics
Jun 16 13:17:59 <saidinesh5> buffers, so it will not give any graphics memory savings. It is intended solely
Jun 16 13:17:59 <saidinesh5> for testing during development and should never be used in production."
Jun 16 13:18:06 <saidinesh5> so apparently it is for saving graphics memory
Jun 16 13:18:16 <saidinesh5> i bet i'd need that soon enough
Jun 16 13:19:37 <Tofe> I don't know that part of QtWayland well
Jun 16 13:19:55 <saidinesh5> yeah me neither
Jun 16 13:20:19 <saidinesh5> what size hard drive do you use btw?
Jun 16 13:20:25 <Tofe> Herrie: if we bump all the recipes we've merged recently, I'll start a new pinephone build locally
Jun 16 13:20:56 <Tofe> saidinesh5: 300GB, but I also use it for Halium
Jun 16 13:21:02 <saidinesh5> Ahh
Jun 16 13:21:05 <Tofe> Usually two images fit in there
Jun 16 13:21:22 <saidinesh5> i had like 160GB yesterday but can't finish building 1 image either
Jun 16 13:21:26 <Tofe> but pinephone doesn't share much
Jun 16 13:21:48 <saidinesh5> deleting some old halium stuff i abandoned to free up some space.. let's see
Jun 16 13:21:58 <Herrie> Tofe: putting kids to bed, then on my todo
Jun 16 13:22:05 <Tofe> ok, take you time :)
Jun 16 13:22:13 <Tofe> I can begin it too
Jun 16 13:22:31 <Herrie> Tofe: Also fine
Jun 16 13:22:53 <Herrie> Make sure to drop db8 patch since I merged it now in repo
Jun 16 13:23:00 <Tofe> yep
Jun 16 13:24:08 <saidinesh5> Aand my old anbox stuff
Jun 16 13:25:57 <saidinesh5> okay so 30GB free now.. let the bitbake begin!
Jun 16 13:26:04 <PureTryOut[m]> Ah Anbox, did you guys get that working on WebOS? We're trying to on postmarketOS, but are having issues so far
Jun 16 13:26:31 <saidinesh5> PureTryOut[m]: oh that was my anbox clone when i worked on it for sailfishos
Jun 16 13:26:45 <PureTryOut[m]> Clone?
Jun 16 13:27:17 <saidinesh5> yeah for sailfish os we forked anbox to make it run with native graphics drivers etc.. and create wayland windows directly instead of doing it the anbox way
Jun 16 13:27:27 <saidinesh5> gave us much better performance
Jun 16 13:28:45 <saidinesh5> https://github.com/sfdroid/anbox
Jun 16 13:28:53 <saidinesh5> (mal's fork is more updated )
Jun 16 13:30:13 <saidinesh5> https://github.com/mlehtima/anbox/tree/master-sf-new
Jun 16 13:31:10 <Herrie> PureTryOut[m]: We have it building and more or less in the images, but didn't really get it to run AFAIK
Jun 16 13:31:20 <Herrie> Issue was mainly our "old" kernels I'd say
Jun 16 13:31:24 <Tofe> PureTryOut[m]: we have **theoretically** anbox installed on our qemux86-64 image, but I doubt it works
Jun 16 13:31:28 <PureTryOut[m]> Sounds like the same state as we are in
Jun 16 13:31:33 <Herrie> But we haven't tried it too much
Jun 16 13:31:38 <PureTryOut[m]> As in, we have it compiling too, but we fail to run it atm
Jun 16 13:31:59 <Herrie> We got sidetracked by the whole webOS OSE stuff released by LG which has our priority over Anbox ;)
Jun 16 13:32:06 <Herrie> Though we'll pick it up at some point again
Jun 16 13:32:22 <Tofe> I could run it (it started the image, and so on), but never succeeded to launch an app
Jun 16 13:32:24 <PureTryOut[m]> Yeah makes sense
Jun 16 13:32:46 <saidinesh5> Herrie: yeah kernels < 3.18/3.10 were a total pain with anbox .. spent so long trying to add patches one after another for namespaces etc.. but each patch pulls in more patches.. and became a nightmare eventually
Jun 16 13:33:32 <Tofe> PureTryOut[m]: are you using snap to install it?
Jun 16 13:33:51 <PureTryOut[m]> Hell no
Jun 16 13:33:54 <PureTryOut[m]> Built from source
Jun 16 13:33:56 <saidinesh5> lol
Jun 16 13:33:57 <Tofe> :) same status here
Jun 16 13:34:20 <PureTryOut[m]> I am talking about postmarketOS here, we don't support Snap lol
Jun 16 13:34:39 <saidinesh5> | mv: target '/src/LuneOS/webos-ports-env/webos-ports/tmp-glibc/work/intel64_tablet-webos-linux/luneos-dev-image/1.0-r0/deploy-luneos-dev-image-image-complete/luneos-dev-image-intel64-tablet-0-0.rootfs.wic' is not a directory
Jun 16 13:34:39 <saidinesh5> | WARNING: /src/LuneOS/webos-ports-env/webos-ports/tmp-glibc/work/intel64_tablet-webos-linux/luneos-dev-image/1.0-r0/temp/run.do_image_wic.23916:1 exit 1 from 'mv "$out/$(basename "${wks%.wks}")"*.direct "$out.rootfs.wic"'
Jun 16 13:34:39 <saidinesh5>  hmm...
Jun 16 13:35:00 <Tofe> .wic is a file, yes
Jun 16 13:35:32 <Tofe> I think your "${wks%.wks}")"*.direct expands to multiple files
Jun 16 13:35:45 <saidinesh5> here's the full log: https://bpaste.net/show/f1e8bbe8eefb
Jun 16 13:35:56 <PureTryOut[m]> If any of you are interested, our progress for Anbox is documented here https://gitlab.com/postmarketOS/pmaports/merge_requests/432
Jun 16 13:35:57 <saidinesh5> and i don't think i even touched the .wks bits
Jun 16 13:37:02 <saidinesh5> PureTryOut[m]: see why the anbox rootfs fails to mount?
Jun 16 13:37:27 <Tofe> saidinesh5: what's the content of /src/LuneOS/webos-ports-env/webos-ports/tmp-glibc/work/intel64_tablet-webos-linux/luneos-dev-image/1.0-r0/deploy-luneos-dev-image-image-complete/ ?
Jun 16 13:37:53 <saidinesh5>  $ ls /src/LuneOS/webos-ports-env/webos-ports/tmp-glibc/work/intel64_tablet-webos-linux/luneos-dev-image/1.0-r0/deploy-luneos-dev-image-image-complete
Jun 16 13:37:53 <saidinesh5> luneos-dev-image-intel64-tablet-0-0         luneos-dev-image-intel64-tablet-0-0.iso            luneos-dev-image-intel64-tablet-0-0.rootfs.ext4      luneos-dev-image-intel64-tablet-0-0.testdata.json  luneos-dev-image-intel64-tablet.hddimg  luneos-dev-image-intel64-tablet.manifest       luneos-dev-image-intel64-tablet.testdata.json
Jun 16 13:37:53 <saidinesh5> luneos-dev-image-intel64-tablet-0-0.hddimg  luneos-dev-image-intel64-tablet-0-0.qemuboot.conf  luneos-dev-image-intel64-tablet-0-0.rootfs.manifest  luneos-dev-image-intel64-tablet.ext4               luneos-dev-image-intel64-tablet.iso     luneos-dev-image-intel64-tablet.qemuboot.conf
Jun 16 13:38:10 <saidinesh5> oh there's an  iso i can test out!
Jun 16 13:38:13 <PureTryOut[m]> saidinesh5: you're asking if I know why the rootfs fails to mount? No I do't
Jun 16 13:38:15 <PureTryOut[m]> *don't
Jun 16 13:38:54 <saidinesh5> PureTryOut[m]: yeah that could be a good start to  dig up. iirc anbox creates an lxc conf to mount the anbox rootfs to that path and then simply start the android's init in lxc
Jun 16 13:40:17 <saidinesh5> i can't remember what anbox's default rootfs image is part of.. but for sailfish fork of anbox, it was squashfs + overlayfs
Jun 16 13:40:30 <saidinesh5> and so kernel config needed to have squashfs enabled
Jun 16 13:41:05 <PureTryOut[m]> I have no clue how to debug it tbh
Jun 16 13:41:54 <Tofe> PureTryOut[m]: you may need squashfs
Jun 16 13:42:27 <PureTryOut[m]> Ah thanks, I'll check if we have that
Jun 16 13:42:29 <Tofe> ah, that's what he said :)
Jun 16 13:42:37 <PureTryOut[m]> Lol
Jun 16 13:43:49 <Herrie> Tofe: merged
Jun 16 13:44:12 <Tofe> great
Jun 16 13:44:19 <Tofe> I've already started pinephone build here
Jun 16 13:44:46 <saidinesh5> PureTryOut[m]: so anbox = basically modified android emulator image running inside an lxc container..
Jun 16 13:45:16 <saidinesh5> modified image = basically modifications to send the render commands and audio data to the SDL anbox app running outside the container using some IPC mechanism
Jun 16 13:45:19 <PureTryOut[m]> Yeah I know
Jun 16 13:46:46 <PureTryOut[m]> Ah I don't have overlayfs enabled, let's see if that helps
Jun 16 13:48:22 <saidinesh5> overlayfs was needed only for sailfish anbox iirc(to share the graphics drivers of the host android system with the anbox )
Jun 16 13:48:54 <saidinesh5> helps with debugging too though
Jun 16 13:56:19 <Tofe> Herrie: I forgot keymanager :)
Jun 16 13:57:13 <Herrie> Tofe: oops ;)
Jun 16 13:59:08 <Herrie> Tofe: Also I noticed App Menu doesn't work in Enyo 1 apps (C+DAV Sync App), Calendar, Email
Jun 16 14:00:11 <Herrie> The logs look OK though, so probably just rendering that's an issue
Jun 16 14:00:42 <Tofe> ok
Jun 16 14:02:29 <Herrie> Tofe: I mean this is the log: https://bpaste.net/show/d4c23da1af4d
Jun 16 14:02:32 <Herrie> That looks OK-ish to me
Jun 16 14:04:07 <Tofe> yes, nothing seem wrong there
Jun 16 14:07:14 <Herrie> So could be just rendering or focus or something like this
Jun 16 14:10:44 <Tofe> hard to tell like this, but maybe
Jun 16 14:22:59 <Herrie> Tofe: Well we had this before at some point. Probably something again in Qt that broke stuff :P
Jun 16 14:26:30 <PureTryOut[m]> Are there plans of upstreaming the majority of the Qt patches required for LuneUI or are you guys planning on patching Qt indefinitely?
Jun 16 14:28:59 <Herrie> Tofe: Any objections with the following: https://github.com/webOS-ports/mojoservice-frameworks/commit/9d725d8cf390b1771644b73f3ededbedb48cb22b ?
Jun 16 14:29:31 <Herrie> The new activitymanager in OSE requires a boolean for the requirements. I figured internetConfidence in 2019 is a lot less of an issue v.s. 2009 ;)
Jun 16 14:29:38 <Herrie> So decided to change it like this
Jun 16 14:29:47 <Herrie> There's a few other places where this would need to be done as well
Jun 16 14:36:04 <saidinesh5> Tofe Herrie how do i simply disable the wic image?
Jun 16 14:36:34 <saidinesh5> also the iso is interesting.. when i try to boot to it i get the Cannot find rootfs.img file in /run/media/* dropping to a shell
Jun 16 14:36:59 <saidinesh5> but the silverlining is it actually gave me the efi shell to select between boot/install... so that's progress!
Jun 16 14:37:31 <Herrie> PureTryOut[m]: Well we upstream what we can, however we also rely on meta-qt5 layer which has a lot of specifics for the Yocto build system which are not suitable for upstreaming
Jun 16 14:37:36 <Herrie> Where possible we do upstream though
Jun 16 14:43:51 <Herrie> Tofe: This internetConfidence is mainly used for activities where it syns stuff on a regular basis (calendar, contacts, check for updates etc)
Jun 16 14:51:51 <PureTryOut[m]> Herrie: specific to the Yocto build system is different from patches required for the actual UI to run once installed
Jun 16 14:51:51 <PureTryOut[m]> The latter is what I personally carea bout
Jun 16 14:51:54 <PureTryOut[m]> *care about
Jun 16 14:52:13 <Herrie> PureTryOut[m]: Well those are minimal really currently
Jun 16 14:52:38 <Herrie> Our main ones are really specific for LuneOS (the whole Palm Bridge extension) is the largest ones. For the rest we don't have a whole lot really
Jun 16 14:52:49 <PureTryOut[m]> Do you have a list of those maybe?
Jun 16 14:53:12 <PureTryOut[m]> As I remember when we hade LuneUI packaged on postmarketOS in the past, we needed like 15 patches to qt-qtbase 5.12
Jun 16 14:53:19 <PureTryOut[m]> s/5.12//g
Jun 16 14:53:25 <Herrie> Yup: https://github.com/webOS-ports/meta-webos-ports/tree/master/meta-luneui/recipes-qt/qt5/qtbase, https://github.com/webOS-ports/meta-webos-ports/tree/master/meta-luneui/recipes-qt/qt5/qtquickcontrols, https://github.com/webOS-ports/meta-webos-ports/tree/master/meta-luneui/recipes-qt/qt5/qtwayland and https://github.com/webOS-ports/meta-webos-ports/tree/master/meta-luneui/recipes-qt/q
Jun 16 14:53:25 <Herrie> t5/qtwebengine
Jun 16 14:55:25 <PureTryOut[m]> Ouch qtwebengine, that is the one
Jun 16 14:56:00 <Herrie> There's not a whole lot we could upstream there unfortunately :(
Jun 16 14:56:24 <Herrie> Most is very LuneOS specific, others were dropped by Qt previously and we needed to add it back
Jun 16 14:56:44 <saidinesh5> seems to be coming from: https://github.com/openembedded/openembedded-core/blob/master/meta/recipes-core/initrdscripts/initramfs-framework/setup-live
Jun 16 14:57:04 <PureTryOut[m]> Can the LuneOS specific stuff not be patched out in LuneUI instead? Seems like the better approach to me
Jun 16 14:57:23 <Herrie> Though at some point we MAY decide to use LG's WAM with Chromium directly instead of QtWebEngine, but that's undecided and for future
Jun 16 14:58:02 <Herrie> PureTryOut[m]: Not really.... Because we use webapps that need specific functionality we need to patch webengine/chromium ourselves in order to support this
Jun 16 14:58:55 <Herrie> This is only interesting for LuneOS, not for Qt themselves. So I believe our chances when upstreaming this are close to 0
Jun 16 14:58:59 <Herrie> To get it merged
Jun 16 14:59:15 <PureTryOut[m]> I guess that is the issue of basing an UI on webtechnologies... Browser engines aren't made to support that
Jun 16 14:59:51 <Herrie> LG in webOS OSE took a different route. They use Chromium 68 (currently, was 53 or 55 before) in their webappmanager, where we use QtWebEngine
Jun 16 15:00:00 <Herrie> So they forked Chromium and added stuff there
Jun 16 15:00:08 <Herrie> And upstreamed bits
Jun 16 15:00:22 <PureTryOut[m]> If it can be installed alongside regular Chromium, that is fine to me
Jun 16 15:00:43 <Herrie> Well our UI is QML, but apps are HTML/JS that need hooks on the QtWebEngine side for stuff to work
Jun 16 15:00:49 <Herrie> That's the tricky part
Jun 16 15:01:19 <Herrie> We might consider LG's approach, but we haven't looked into it yet. As you can imagine it's pretty deep on a system level and will require a lot of tailoring in various bits
Jun 16 15:01:25 <Herrie> Not a quick swap & replace
Jun 16 15:01:48 <PureTryOut[m]> If the patched QtWebengine could work alongside a non-patched version of QtWebengine, I would have less problems with supporting LuneUI in postmarketOS. Although then you still have the issue of wanting to upgrade Qt but first having to rebase the patches to QtWebengine first
Jun 16 15:02:19 <PureTryOut[m]> I like QML UI's personally though. Plasma Mobile is QML and works well
Jun 16 15:03:31 <saidinesh5> Herrie: hooks like?
Jun 16 15:04:26 <Herrie> saidinesh5: Well we need to make it run webapps and integrate into various bits of the OS, that's not very straight forward
Jun 16 15:04:39 <Herrie> Though I believe our patches are pretty neat to accomplish such task ;)
Jun 16 15:04:44 <Herrie> Tofe did a great job there
Jun 16 15:04:54 <Herrie> WAsn't easy with all the Qt & Chromium updates
Jun 16 15:05:00 <saidinesh5> also Tofe any idea why it wouldn't mount my usb stick (from which i am booting) into some /run/media/* if it is expecting the rootfs to be there..?
Jun 16 15:05:19 <Herrie> Though Qt is now pretty close to upstream Chromium, so rebasing up patches "shouldn't" be that hard anymore
Jun 16 15:06:35 <saidinesh5> Herrie: out of curiosity, integration like?
Jun 16 15:08:13 <PureTryOut[m]> saidinesh5Tofe: thanks for suggesting enabling overlayfs. Sadly it did not help
Jun 16 15:09:56 <saidinesh5> PureTryOut[m]: oh that's a shame.. hold on. will look at the exact command anbox uses to mount it's rootfs .. you can try executing it manually and see why it fails
Jun 16 15:11:50 <Herrie> Additional stuff like "window features (size of popup, windows etc)", sync calls on qtwebchannel, forcing font to Prelude, provide an initial url to load, password echo, devicepixel ratio etc
Jun 16 15:14:32 <saidinesh5> Herrie: ah interesting.. thanks
Jun 16 15:17:12 <saidinesh5> PureTryOut[m]: you have squashfs enabled right?  ( auto m = common::MountEntry::create(loop_device, android_rootfs_dir, "squashfs", MS_MGC_VAL | MS_RDONLY | MS_PRIVATE); )
Jun 16 15:20:10 <PureTryOut[m]> I do yes
Jun 16 15:25:19 <saidinesh5> hmm
Jun 16 15:43:44 <saidinesh5> Herrie: Tofe are the virtual terminals disabled on luneos by default? (i just got systemd up and running!)
Jun 16 15:44:14 <Herrie> saidinesh5: I think so yes....
Jun 16 15:44:23 <Herrie> On qemux86 we can ssh in though
Jun 16 15:45:51 <saidinesh5> Herrie: oh okay.. any quick hacks to enable that?
Jun 16 15:48:25 <Herrie> saidinesh5: On qemux86 I can use ssh via: localhost:5522 user:root password: <empty>
Jun 16 15:48:35 <Herrie> And SCP as well for that matter
Jun 16 15:48:44 <Herrie> On Android targets we use ADB
Jun 16 15:49:15 <saidinesh5> Herrie: good idea!! will have to make this tablet connect to my wifi network using some systemd script i see
Jun 16 15:49:33 <saidinesh5> oor will have to give this one some kind of usb networking
Jun 16 15:50:43 * saidinesh5 tries to think of some kind of usb ethernet option
Jun 16 15:52:17 <saidinesh5> Herrie: is there any quick command line thing i can sneak into systemd scripts to connect to my wifi network?
Jun 16 15:52:36 <Herrie> saidinesh5: Not really I guess
Jun 16 15:52:43 <Herrie> You can do some connman magic
Jun 16 15:52:45 <saidinesh5> you guys use conman?
Jun 16 15:52:46 <saidinesh5> yeah
Jun 16 15:52:46 <Herrie> If it starts already
Jun 16 15:52:49 <saidinesh5> looking it up
Jun 16 15:58:52 <Tofe> damn, was a bit away, let me read the backlog :p
Jun 16 16:00:22 <saidinesh5> oh not much in the backlog. just that some udev or mdev or whatever they use for init was failing to mount my usb stick in which the rootfs.img exists.. so i created another partition and wrote the rootfs.img to that partition .. and passed root=/dev/sda2 in commandline
Jun 16 16:00:36 <saidinesh5> and that starts systemd and every other service without any error/fail
Jun 16 16:00:38 <saidinesh5> but i see no UI
Jun 16 16:00:45 <saidinesh5> and my keyboard/mouse stop responding
Jun 16 16:01:10 <saidinesh5> which makes me think some UI is running... so need some kind of way to access some kind of shell in this tablet now
Jun 16 16:05:51 <PureTryOut[m]> saidinesh5: oh actually, I think I've been using the wrong kernel, might not have had overlayfs enabled after all
Jun 16 16:07:01 <saidinesh5> PureTryOut[m]: although it looks like you've been getting that error when trying to mount the squashfs filesystem.. so you can try mounting that squashfs image somewhere locally and see if it works
Jun 16 16:07:23 <Tofe> saidinesh5: ah you use your own initrd?
Jun 16 16:08:04 <Tofe> it's true his irc connection is bumpy :)
Jun 16 16:08:41 <PureTryOut[m]> I'm not sure where that squashfs image is? Is it this one? https://gitlab.com/postmarketOS/pmaports/blob/feature/anbox/main/anbox-image/APKBUILD#L15
Jun 16 16:09:10 <saidinesh5> hmm
Jun 16 16:09:55 <Tofe> PureTryOut[m]: yes, that's the one
Jun 16 16:10:10 <Tofe> saidinesh5: do you use your own initrd?
Jun 16 16:10:43 <saidinesh5> Tofe: nah . not yet at least. using the thing that comes from openembedded  / meta-intel
Jun 16 16:10:51 <PureTryOut[m]> What's the command Anbox uses to mount it anyway?
Jun 16 16:12:26 <Tofe> saidinesh5: you could be a try with https://github.com/webOS-ports/meta-pine64-luneos/blob/master/recipes-core/initrdscripts/initramfs-scripts-simple_1.0.bb but not sure it fits your case
Jun 16 16:13:26 <Tofe> it makes the assumption you have a "luneos-rootfs" partition
Jun 16 16:13:39 <Tofe> and was only tested on pinephone
Jun 16 16:14:40 <Tofe> saidinesh5: to get rid of .wic, you'll have to change the IMAGE_FSTYPES in your machine conf file
Jun 16 16:15:32 <saidinesh5> Tofe: Ahh okay.. looking at initramfs
Jun 16 16:15:44 <saidinesh5> btw. i just commented out luna-next's execstart
Jun 16 16:15:50 <saidinesh5> and replaced it with /usr/bin/ls
Jun 16 16:15:55 <saidinesh5> lol that failed a bunch of services
Jun 16 16:15:59 <saidinesh5> but i get keyboard working!!
Jun 16 16:16:02 <Tofe> :)
Jun 16 16:16:03 <saidinesh5> in a tty
Jun 16 16:16:15 <saidinesh5> now trying to figure out the connman magic needed to connect to my homewifi
Jun 16 16:16:17 <saidinesh5> before i enable it
Jun 16 16:16:28 <Tofe> ah, for that I don't know
Jun 16 16:16:39 <Tofe> do you see it through USB ?
Jun 16 16:16:59 <Tofe> for pinephone I configured an ethernet over usb network with static IP
Jun 16 16:17:24 <saidinesh5> i dont have the usb cable for it
Jun 16 16:17:33 <Tofe> ok
Jun 16 16:17:58 <saidinesh5> like i am using usb jack for keyboard and mouse via otg
Jun 16 16:18:06 <saidinesh5> and the bootable usb stick
Jun 16 16:18:29 <saidinesh5> is there any opengl demo i can run?
Jun 16 16:19:33 <Tofe> well, luna-next can be used for this purpose
Jun 16 16:19:57 <Tofe> I don't think we install mesa demos by default
Jun 16 16:20:26 <saidinesh5> ohh
Jun 16 16:20:59 <Tofe> you just parse /etc/luna-next/environment.conf in your shell and then you can just start luna-next
Jun 16 16:21:11 <saidinesh5> as root?
Jun 16 16:21:14 <Tofe> yes
Jun 16 16:21:44 <saidinesh5> rebooting... because luna-next grabbed the input but no output
Jun 16 16:23:04 <Tofe> for reference, for pinephone (which also uses mesa gallium) I set EGL_PLATFORM=drm QT_QPA_PLATFORM=eglfs QT_QPA_EGLFS_INTEGRATION=eglfs_kms
Jun 16 16:24:52 <Tofe> for conman, have a look here: https://www.digikey.com/en/maker/blogs/2017/how-to-setup-wifi-on-the-beaglebone-black-wireless
Jun 16 16:26:49 <saidinesh5> either my connman is broken or something is off
Jun 16 16:27:25 <Tofe> no wifi?
Jun 16 16:27:34 <Tofe> "rfkill unblock wifi" maybe ?
Jun 16 16:27:49 <saidinesh5> nope..  some error when i say connmanctl enablewifu
Jun 16 16:28:06 <Tofe> weird
Jun 16 16:28:17 <saidinesh5> and yea tried the rfkill  unblock
Jun 16 16:28:22 <saidinesh5> noerror there
Jun 16 16:36:45 <Tofe> but the wlan device shows up in "ip a" ?
Jun 16 16:39:18 <saidinesh5> nope
Jun 16 16:39:52 <Tofe> then maybe it's a kernel module issue
Jun 16 16:39:55 <saidinesh5> there is a sit0@NONE
Jun 16 16:39:59 <saidinesh5> oh
Jun 16 16:40:20 <saidinesh5> but i also tried using a usb adapter too
Jun 16 16:40:46 <saidinesh5> usb wifi adapter
Jun 16 16:47:38 <Herrie> Tofe: OK new image ready to test locally
Jun 16 16:49:22 <Tofe> well, not mine :p
Jun 16 16:52:40 <Herrie> Tofe: Well seems I needed to update the other db8 role file :P
Jun 16 16:53:08 <Herrie> Not the dynamic but the static one it seems :S
Jun 16 17:01:56 <saidinesh5> bah
Jun 16 17:06:04 <PureTryOut[m]> saidinesh5: seems the Android image I'm using is faulty... Using one from like a year before mounts fine while all images inbetween fail
Jun 16 17:07:21 <saidinesh5> ohh... that could be it...
Jun 16 17:08:56 <PureTryOut[m]> Actually, that is a lie, that one is faulty too
Jun 16 17:09:09 <Herrie> Tofe: Bumped db8 a bit
Jun 16 17:09:13 <PureTryOut[m]> Should `mount android.img /var/lib/anbox/rootfs/ -t squashfs -o loop` do it?
Jun 16 17:10:42 <saidinesh5> that sounds right.. not sure if the order of the arguments matter or if there exists the target path
Jun 16 17:11:05 <saidinesh5> does the var/lib/anbox/rootfs exist?
Jun 16 17:14:24 <PureTryOut[m]> It does yes
Jun 16 17:14:27 <PureTryOut[m]> (that is where Anbox wants to mount it)
Jun 16 17:17:08 <saidinesh5> oh.. could be that something wrong with squash fs driver?
Jun 16 17:17:33 <saidinesh5> can you convert the image to  extfs image amd try mounting that?
Jun 16 17:36:22 <Herrie> Tofe: OK Calendar & Contact sync now "works", which means that all the "untested" mojo changes work. I don't see anything in Contacts & Calendar though, so will check logs now ;)
Jun 16 17:36:30 <Herrie> Probably some more permissions that need fixing
Jun 16 17:58:50 <saidinesh5> oh meh.. can't figure out why ifconfig/rfkill list isnt showing any wifi devices
Jun 16 17:58:53 <saidinesh5> but now the battery died
Jun 16 17:58:54 <saidinesh5> so meh
Jun 16 18:48:32 <Herrie> Tofe: seems contacts not showing up is due to some permission issues for contactslinker
Jun 16 18:48:42 <Herrie> Not really sure how to address that one
Jun 16 19:11:39 <Tofe> Herrie: do you have the error message ?
Jun 16 19:19:29 <Herrie> Tofe: https://bpaste.net/show/296bd2d48d90
Jun 16 19:19:39 <Herrie> Seems that we don't have the prv/pub json files in our image
Jun 16 19:19:45 <Herrie> While it does exist on legacy webos
Jun 16 19:21:42 <Tofe> maybe generated from https://github.com/openwebos/app-services/blob/master/com.palm.service.contacts.linker/palm_bus_config.json ?
Jun 16 19:22:53 <Tofe> ah no sorry it's here https://github.com/webOS-ports/app-services/blob/webOS-ports/master/com.palm.service.contacts.linker/files/sysbus/com.palm.service.contacts.linker.json
Jun 16 19:23:58 <Herrie> Yeah something like that I have in legacy webOS Doctor files
Jun 16 19:24:27 <Tofe> mmh I see usr/share/ls2/roles/{prv,pub}/com.palm.service.accounts.role.json but no linker
Jun 16 19:26:17 <Herrie> Yes
Jun 16 19:26:21 <Herrie> Not sure why it's not there
Jun 16 19:28:28 <Herrie> Let me take a previous release image to see if it was there before
Jun 16 19:28:51 <Herrie> Just to make sure it's a regression somehow
Jun 16 19:29:03 <Herrie> Because we have a very small number of prv/pub roles in general I noticed now
Jun 16 19:29:10 <Herrie> V.s. legacy, but that could be normal
Jun 16 19:30:11 <Herrie> OK in Decaf release I have a lot more
Jun 16 19:30:37 <Herrie> Some quick food & children to bed ;)
Jun 16 19:42:07 <Tofe> :)
Jun 16 19:42:42 <Herrie> I need to check Doppio as well
Jun 16 19:42:58 <Herrie> Could be some OSE change in bb class that caused it
Jun 16 19:43:05 <Herrie> That's my first hunch
Jun 16 19:46:18 <Tofe> yes, it's quite possible, I agree
Jun 16 19:49:08 <Herrie> Or maybe I migrated some already to single file. Need to verify
Jun 16 19:57:46 <Herrie> OK in Doppio it's also there
Jun 16 20:01:09 <Herrie> Tofe: So the culprit is likely in: https://github.com/webOS-ports/meta-webos-ports/commit/48a534fd40c073403fb4711695f24ac610bd7173#diff-cd03be8797f023d071178be561e63dd4
Jun 16 20:02:58 <Herrie> To be honest I'm not sure what "    if [ ${#_LS_PUB[@]} -ne 0 ]; then" does... I just copied & pasted from OSE here
Jun 16 20:03:08 <Herrie> Same with         if [ -n "$tree" -a -d "$tree" ]; then
Jun 16 20:03:19 <Herrie> I suspect one of those is the problem
Jun 16 20:06:08 * Tofe clicks on the link, a 1500 pages bash book in the end
Jun 16 20:07:02 <Tofe> " if [ ${#_LS_PUB[@]} -ne 0 ]; " here means "if the 'find' command above found something, then..."
Jun 16 20:07:59 <Tofe> " if [ -n "$tree" -a -d "$tree" ]; " mean "if $tree isn't an empty string and if $tree is a directory then ...."
Jun 16 20:11:17 <Tofe> it does look alright, though I might have missed a mistake
Jun 16 20:17:55 <Herrie> OK I think I have found SOME problem ;)
Jun 16 20:18:08 <Herrie> https://github.com/webOS-ports/meta-webos-ports/blob/warrior/meta-luneos/recipes-webos/app-services/app-services.bb
Jun 16 20:18:25 <Tofe> mmh?
Jun 16 20:18:40 <Herrie> Points to webOS-ports/master branch
Jun 16 20:18:49 <Herrie> Which doesn't have the db8 migration commit from https://github.com/webOS-ports/app-services/commits/webosose
Jun 16 20:18:59 <Tofe> oh
Jun 16 20:19:11 <Herrie> I will create a new branch called webOS-ports/webOS-OSE in line with others
Jun 16 20:19:29 <Tofe> let me just PR webosose into master
Jun 16 20:19:38 <Tofe> ah, we could do that too
Jun 16 20:21:00 <Tofe> though in the end, we should choose a definitive branch, and either PR all to webOS-ports/master, or change our bbclass to point to webOS-ports/webOS-OSE
Jun 16 20:21:26 <Herrie> Tofe: Yeah, but that's for at end of migration ;)
Jun 16 20:21:33 <Tofe> but as long as the transition isn't finished, I'm fine with that yes
Jun 16 20:21:35 <Tofe> exactly
Jun 16 20:22:04 <Herrie> Tofe: OK the problem is that you migrated com.palm.service.accounts to ACG, but not the others in app-services
Jun 16 20:22:16 <Tofe> ah, oops
Jun 16 20:23:08 <Tofe> I don't think I'm fresh enough to do that tonight :)
Jun 16 20:23:31 <Herrie> Therefore the recipe installs the new ACG files but not the pub/priv ones anymore ;)
Jun 16 20:23:42 <Tofe> I understand
Jun 16 20:29:37 <Herrie> Tofe: Simple rename from the .json to .role.json might solve it as well ;)
Jun 16 20:31:50 <Herrie> Let me test that
Jun 16 20:33:46 <Herrie> Yup works
Jun 16 20:34:14 <Herrie> Files are in IPK at least
Jun 16 20:34:20 <Herrie> That's one part of the problem  at least
Jun 16 21:18:02 <Herrie> And another one bites the dust ;) https://github.com/webOS-ports/configurator/pull/4
**** ENDING LOGGING AT Mon Jun 17 02:59:57 2019