**** BEGIN LOGGING AT Mon Jan 24 02:59:56 2022 Jan 24 07:21:25 Morning! Jan 24 07:21:28 Tofe/JaMa: https://github.com/shr-distribution/meta-smartphone/pull/141 Jan 24 07:22:58 morning Jan 24 07:58:40 Tofe: Came across this from SFOS with regards to integration into OS: https://github.com/lal883/tama-sailfish-waydroid-instructios#7-modify-the-file-vendoretcvintfmanifestxml Jan 24 07:58:48 NOthing too much really, but might give some ideas Jan 24 08:19:32 Morning! Jan 24 08:19:44 JaMa: oh, you had newer images ? that's great :) Jan 24 08:22:15 JaMa: Nice :) Jan 24 08:29:09 I have a lot of stuff in 120G downloads dir :) Jan 24 08:29:31 good :) Jan 24 08:30:31 bumped some stuff in webos-ports-setup and re-sync honister and master branches in meta-pine64-luneos Jan 24 08:55:10 JaMa: Nice, now people should be able to build stuff again without jumping though hoops Jan 24 13:13:58 morning! happy monday! Jan 24 13:14:09 any luck with resurrecting the Touchpad? Jan 24 13:22:43 "happy monday", what an intersting concept Jan 24 13:23:08 codepoet: yes, we've solved the issues Jan 24 13:23:25 We don't have an updated build yet, though Jan 24 13:23:56 but I can pull the bits and roll my own build, ya? Jan 24 13:24:10 yep, you'll need: Jan 24 13:24:12 https://github.com/webOS-ports/meta-webos-ports/pull/481 Jan 24 13:24:22 thanks much, i'll give it a shot! Jan 24 13:24:36 https://github.com/Tofee/meta-smartphone-1/commit/c0628e638beecdbd0734f1ac3e4e78012a8f96c8 Jan 24 13:25:12 and probably https://github.com/Tofee/android_kernel_htc_msm8960-1/commit/6836f8db8c67183431aca7d0cf343904e0ef7d64 Jan 24 13:26:03 will running the bb build job do a fresh pull? or should I manually pull these changes? Jan 24 13:26:04 with these, I have a booting TP with wifi Jan 24 13:27:03 I've merged the first one, so a pull will get it Jan 24 13:27:22 the second PR, I think it should also come with a pull (but let me know if not) Jan 24 13:27:47 the kernel patch, no, it needs a SRCREV bump for the recipe -- and also it's not merged yet Jan 24 13:27:57 ^ Herrie: maybe we could merge it ? Jan 24 13:28:56 that would be nice :-) Jan 24 13:29:09 there's a lot of surface area to learn if I have to make changes manually! Jan 24 13:29:33 No problem, we'll handle that part :) Jan 24 13:32:39 thanks! let me know when its safe to fire up a build! Jan 24 13:33:54 ok ! Jan 24 14:21:47 there were quite a few changes in honister branch, so you can fire it now to update the rest and then rebuild only small part when the fixes are in Jan 24 14:22:30 that's a good tip -- but my build machine is old, so I think I'll just wait to do it once. in no hurry! Jan 24 14:23:53 you'll only need to build kernel in 2nd pass Jan 24 14:24:19 and how would I do that? Jan 24 14:24:46 update, build, update, build Jan 24 14:25:18 it will be probably already merged before your old buiders finishes updating qtwebengine Jan 24 14:25:35 Tofe: do you have PR for the kernel commit? Jan 24 14:26:00 I can merge it to kernel repo and do the SRCREV bump after my kirkstone build finishes Jan 24 14:26:44 aha it's in herrie's github Jan 24 14:29:17 JaMa: yes https://github.com/Herrie82/android_kernel_htc_msm8960-1/pull/2 Jan 24 15:40:40 all linux-* recipes migrated to use shr-distro repo and bump prepared, will merge after build test Jan 24 15:42:17 oh, thanks Jan 24 15:51:01 merged Jan 24 15:51:32 awesome! ready to build? Jan 24 15:58:10 codepoet: pull meta-{webos-ports,smartphone} and I think it should be ok yes Jan 24 16:00:46 ok, how do i do that? i only know how to do what's in the wiki (https://webos-ports.org/wiki/Build_for_Tenderloin) Jan 24 16:01:01 i dumped all its caches and downloads, so I'm assuming it'll pull everything fresh, ya? Jan 24 16:02:54 ah, that wasn't necessary Jan 24 16:04:00 in the wiki, what we propose corresponds to the step "update metadata": Jan 24 16:04:07 # go one level up to the webos-ports-env dir Jan 24 16:04:07 $ cd .. Jan 24 16:04:08 $ make update Jan 24 16:04:17 ok perfect Jan 24 16:04:39 or, what I proposed what just to go into the subfolders and do "git pull" Jan 24 16:04:48 both should end up with similar results Jan 24 16:05:01 s/what/was/ Jan 24 16:12:19 looks like there was a messy merge Jan 24 16:12:43 luneos.inc has some "<<<<<" git merge notes in it, so it doesn't parse. I'll clean-up locally, but probably needs to be fixed in git Jan 24 16:22:00 Tofe: you would need to pull oe-core and bitbake as well to get latest and at that point using make update is safer as it uses the revisions specified in webos-ports-setup (and at least build tested by me) Jan 24 16:23:15 it looks good in git, possibly some your local changes Jan 24 16:34:30 ya could be. build is rolling, back in 4-5 hours! Jan 24 17:12:25 ok, it built (with a bunch of warnings) and I deployed it. same result, so I think my build didn't go well. Jan 24 17:12:49 the warnings were all Package version for package XXX went backwards which would break package feeds Jan 24 17:12:53 it also built really fast Jan 24 17:16:30 oh wow, it actually booted eventually! but it says "no wifi support" during OOBE, so I must have done something wrong Jan 24 17:34:08 can't turn it on in control panel either Jan 24 17:39:15 maybe your build directory wasn't updated correctly, check the output of make update that no layers are updated now Jan 24 17:42:09 codepoet: maybe my fix isn't enough; you can try to do the following: in qmlterm, do "rmmod ath6kl ; modprobe ath6kl" Jan 24 17:44:12 also, something I didn't PR yet is to disable systemd-udev-settle.service Jan 24 17:46:14 codepoet: you can also try this: "systemctl mask systemd-udev-settle.service" Jan 24 17:46:32 I'll try to retest my PR Jan 24 17:58:16 alright, i'm on wifis! Jan 24 17:58:37 ok great Jan 24 17:58:52 I guess my wifi fix wasn't complete :/ Jan 24 17:59:18 codepoet: did you just reload the ath6kl module ? or also disable udev-settle ? Jan 24 18:00:30 i did both Jan 24 18:00:42 and the wifi still works after a reboot ? Jan 24 18:00:48 (that's my case) Jan 24 18:02:11 oh i haven't dared to reboot yet! Jan 24 18:02:18 want to try my musuem app Jan 24 18:02:27 :) please do Jan 24 18:02:35 (the latter) Jan 24 18:04:00 its installing! Jan 24 18:08:08 everything works but using it to install an app. I use a PalmService request to palm://com.palm.applicationManager to tell it to launch PreWare and do an install. Jan 24 18:08:51 maybe the appid is different for Preware2, I'll see if I can solve. Jan 24 18:09:31 you mean preware2 refuses to install your ipk? Jan 24 18:09:50 not necessarily. i'm not sure if i'm launching it right Jan 24 18:10:02 should I be able to use the palm sdk command line tools to install/debug apps? Jan 24 18:10:13 like palm-run and palm-log? cause they're not working Jan 24 18:10:26 JaMa: https://git.yoctoproject.org/poky/plain/meta/recipes-support/rng-tools/rng-tools/rngd.service would need some patching I guess, to not rely on systemd-udev-settle.service anymore Jan 24 18:11:25 codepoet: I don't understand, why not use the UI ? Jan 24 18:13:49 Tofe: you mean for LuneOS or in general? IIRC we use rngd.service only for qemu to initialize /dev/random? Jan 24 18:13:59 otherwise qemu boot time is terrible Jan 24 18:14:21 @Tofe, the UI isn't useful for writing apps, I need the command line tools so I can read log messages Jan 24 18:14:26 JaMa: well, on tenderloin boot time is terrible because of systemd-udev-settle.service being triggered Jan 24 18:14:43 do i do adb instead? Jan 24 18:15:15 codepoet: ah ok, I understand; then using a luna-send call should be the way to go Jan 24 18:15:25 rng-tools is only in QEMU_RDEPENDS, can you check what pulls rng-tools into the image? Jan 24 18:15:41 so none of the sdk tools work? Jan 24 18:15:49 JaMa: will check yes Jan 24 18:16:13 these were the command line tools for webOS: http://sdk.webosarchive.com/docs/docs.html#dev-guide/tools/command-line-tools.html Jan 24 18:16:42 codepoet: I didn't say that, but I never actually tested it I think; Herrie, did you? Jan 24 18:17:13 basically the LS2 APIs are still there, maybe with some exceptions Jan 24 18:17:30 how do you guys install and debug apps? Jan 24 18:18:40 Tofe: looks like openssh depends on it now Jan 24 18:18:57 That's the point, I think, we don't do that enough; Herrie tested the installation of plenty of legacy apps some years ago, and the result was quite ok (not sure which method he used though) Jan 24 18:19:44 About debugging apps, I'm not the right guy here, as I never developped an Enyo app :p (I did a bit of debug here and there but that's all) Jan 24 18:19:55 just had a forced reboot event. wifi is still working! Jan 24 18:20:28 Tofe: added as PACKAGECONFIG to openssh in oe-core fe99349, I guess we should remove that PACKAGECONFIG Jan 24 18:21:09 JaMa: but doesn't it make sense to have rngd for openssh ? Jan 24 18:22:36 removing udev-settle will probably break rngd a bit, not sure which option is worse (possibly broken rngd or no rngd at all) Jan 24 18:23:03 oe-core e9714d42 added settle also to Wants Jan 24 18:23:25 mmh Jan 24 18:26:13 thing is, udev-settle will probably be removed in a future systemd release anyway Jan 24 18:26:53 I understand the rationale behind the commit though, and it's quite fine Jan 24 18:31:02 but rngd should wait for a udev path (or something like that), not be based on a snapshot of the system's udev state Jan 24 18:31:20 ( ^ that's why udev-settle is deprecated ) Jan 24 18:34:49 I've read 2 RH bug reports related to rngd start with udev and I'm not very clever from it, dunno what's the right solution here Jan 24 18:35:32 well I'm not sure either Jan 24 18:35:44 I'm reading a bit this thread here https://lists.linuxfoundation.org/pipermail/virtualization/2014-August/027060.html Jan 24 18:38:55 I'd be fine with a tenderloin-specific solution, but I don't want to make rngd-tools MACHINE specific of course Jan 24 18:40:43 as only openssh rrecommends it, we can make it MACHINE_ARCH and mark it ABISAFE Jan 24 18:40:59 ah, interesting: https://github.com/nhorman/rng-tools/blob/master/rngd.service <-- no udev settling in there Jan 24 18:50:20 One of my apps works! Jan 24 18:50:21 https://wosa.link/image.php?am9ud2lzZXxtYmUwZmFjZQ Jan 24 18:50:44 app is clean enyo with a webservice call Jan 24 18:51:26 you can download it here...but i don't know how you'd install it: http://appcatalog.webosarchive.com/showMuseumDetails.php?app=1005797 Jan 24 18:53:30 codepoet: is your website also providing an app feed? Jan 24 18:54:30 no, the Museum used a different payload structure. i did tinker with a feed, but there's over 3000 feeds, so it didn't work well. Jan 24 18:54:37 i'd have to make multiple feeds. Jan 24 18:55:05 the Museum app actually mostly works. once I figure out how to debug PalmSystem calls, I think I can get install working on LuneOS Jan 24 18:56:11 *over 3000 apps (not feeds) Jan 24 18:57:48 pushed with adb to media/internal/downloads then installed with Preware2 Jan 24 18:58:18 great, now I can follow the telescope going to L2 :) Jan 24 18:58:32 ok, so you use adb tools. i'm rusty but i've used before Jan 24 18:58:46 and yes, the image changes periodically too! Jan 24 19:01:24 ok, adb install doesn't work because it wants the app to be something.apk and webOS packages are something.ipk Jan 24 19:16:30 Tofe: is this something you were seeing with gst? Jan 24 19:16:33 Jan 24 19:15:40 rpi4jama gst-plugin-scan[787]: type name '-a-png-encoder-pred' contains invalid characters Jan 24 19:16:36 Jan 24 19:15:40 rpi4jama gst-plugin-scan[787]: g_type_set_qdata: assertion 'node != NULL' failed Jan 24 19:16:39 Jan 24 19:15:40 rpi4jama gst-plugin-scan[787]: g_type_set_qdata: assertion 'node != NULL' failed Jan 24 19:16:51 I just noticed this in luna-next on rpi4-64 with kirkstone Jan 24 19:32:48 here's another important app from the Museum that mostly works! http://wosa.link/image.php?am9ud2lzZXx0YmUyZmQ3OA Jan 24 19:39:23 JaMa: no, unfortunately it doesn't seem related Jan 24 19:49:02 Seems I missed a bit, let me catch up Jan 24 19:49:41 lots of Luna crashes, especially after unlocking Jan 24 19:49:47 Tofe: Merged the kernel commit Jan 24 19:50:52 any way to do a screen shot? Jan 24 19:51:17 codepoet: swipe right on the gesture bar Jan 24 19:53:42 got it. here's the Museum with the LuneOS filter on. Three apps tested to work so far! http://wosa.link/image.php?am9ud2lzZXx2YWZmMzdjNw Jan 24 19:56:21 codepoet: I replied you on webOS Nation Jan 24 19:56:47 I think you can SSH/SCP into a Android device too, never tried it myself, but works on PinePhone, PinePhonePro and VirtualBox Jan 24 19:57:09 P.s. For quick development cycles and debugging the VirtualBox image is pretty nice Jan 24 19:57:12 I use it a lot myself Jan 24 19:57:20 saw it. adb shell works, but i don't know how to watch logs Jan 24 19:57:22 It doesn't do touch, but it does pretty much everything else Jan 24 19:57:34 Herrie: we don't activate usb network for android devices; so it's possible to ssh over wifi, but not over usb Jan 24 19:57:47 Tofe: Yeah I mean over wifi Jan 24 19:57:49 On Android Jan 24 19:57:53 I do that on PP and PPP too Jan 24 19:58:04 Connect to WiFi, then connect to the device's IP via port 22 Jan 24 19:58:08 ssh over wifi, adb shell over usb is good enough Jan 24 19:58:20 codepoet: You can SCP as well Jan 24 19:58:25 Via WiFi, beats ADB Jan 24 19:58:33 adb is not always very stable in my experience Jan 24 19:58:51 I recall JayCanuck got WOSQI working at some point, not sure if that still works Jan 24 19:59:00 4.6.0 If I remember correctly Jan 24 20:00:18 for logs, I usually use "journalctl -f -l", but that might be too verbose for your app debugging Jan 24 20:00:39 There is also still some loggin in /var/log/messages and /var/log/legacy-log Jan 24 20:04:07 ya, need to find where enyo logs to Jan 24 20:04:24 no way to debug without that. in a browser it just writes to the console.log Jan 24 20:04:45 maybe i can pull apart my enyo SDK and see where it writes if it detects PalmSystem Jan 24 20:04:55 Well you can technically open the browser on the TP, load 127.0.0,1:1122 and you get the concole there Jan 24 20:05:10 trying Jan 24 20:05:21 If it does a call to non existing bits, should be in journalctl really Jan 24 20:05:25 PalmSystem is there Jan 24 20:05:29 We use it in various places Jan 24 20:05:42 lol, Luna crashed when I entered that URL Jan 24 20:05:44 Stuff like Google C+DAV with Cross App stuff is pretty advanced Enyo and works Jan 24 20:05:51 Hmmz that's weird Jan 24 20:06:15 We never really got it to port forward, ideally you should be able to debug it from desktop browser Jan 24 20:06:23 now it won't launch the browser, i'll reboot Jan 24 20:06:40 "systemctl restart luna-next" should do it Jan 24 20:06:44 Normally Jan 24 20:06:53 Stability might need some tweaks here and there Jan 24 20:07:08 I didn't see much with my other devices as such, so could be something TP specific Jan 24 20:07:19 It's a bit of a "weird duck" with all it's things Jan 24 20:08:00 Seeing it was proper webOS Linux before and then got "Android-ified" and we use that part Jan 24 20:08:11 So it might not be as stable as some of the other devices Jan 24 20:08:42 For Preware on LuneOS the appID is: org.webosports.app.preware Jan 24 20:08:51 That should launch Jan 24 20:08:58 Most other legacy bits just work as is Jan 24 20:09:31 Only thing that changed with newer LS2 security permissions is that you need to specify in appinfo.json the permissions it requires Jan 24 20:09:43 I.e. to do various system calls, db access etc Jan 24 20:09:49 LG locked that down quite a bit Jan 24 20:09:58 That didn't exist in legacy webOS at all Jan 24 20:10:35 So if something in an app doesn't work just do a "journalctl | grep LS_HUB" Jan 24 20:10:42 And you'll probably get some permission error Jan 24 20:11:14 Some example of permissions: https://github.com/Herrie82/core-apps/commit/01773d3dfeb5b991799d13530428f01e291c712b Jan 24 20:11:18 ok, it won't respond remotely from http://touchpadip:1122, but it did load something at localhost:1122 Jan 24 20:11:30 Ah OK Jan 24 20:11:30 a list of "inspectable pages" Jan 24 20:11:36 Yeah that should be it Jan 24 20:11:41 It should show you the various apps Jan 24 20:12:10 only apps installed in /usr/palm/applications Jan 24 20:12:32 i'll try manually going to cryptofs apps Jan 24 20:14:20 https://imgur.com/a/RLuL15t this is what I get on VirtualBox Jan 24 20:14:40 I think we only install in /usr/palm/applications for now Jan 24 20:15:39 Once you open an app it should show Jan 24 20:16:58 In the list that is Jan 24 20:18:09 ya, i see the list. none of my apps that i've installed and run show up in it Jan 24 20:19:15 Hmmz that's weird though Jan 24 20:20:02 If I refresh the 127.0.0.1:1122 at my end I get your WebbTracker Jan 24 20:20:51 maybe i have to re-launch after reboot? Jan 24 20:21:02 that got it Jan 24 20:21:16 Yeah you need to launch the app for it to appear in the list Jan 24 20:21:20 and there's my logs! Jan 24 20:21:26 It's a list of currently running apps Jan 24 20:21:33 ok, this is something... Jan 24 20:21:46 they must be getting written somewhere I can tap into on the filesystem... Jan 24 20:22:01 Your problem is probably it tries to do some luna call for which it doesn't have permissions Jan 24 20:22:25 Which should show in journalctl or /var/log/messages or /var/log/legacy-log Jan 24 20:22:28 yes, i think you're right. would either of these let me ask another app to launch? "applications.internal", "applications" Jan 24 20:22:52 Do a "journalctl | grep LS_HUBD" Jan 24 20:23:01 It should give you some permission errors I guess Jan 24 20:23:19 Which permissions to give is not always very easy to establish... Jan 24 20:24:14 It's defined in the component you're trying to call: Ie: https://github.com/webOS-ports/luna-appmanager/blob/master/files/sysbus/luna-appmanager.api.json Jan 24 20:24:42 So for launch both "applications" and "applications.internal" should work or even applications.launch Jan 24 20:24:43 nothing from that grep Jan 24 20:25:01 Maybe it was LS-HUBD Jan 24 20:25:02 k, i need the applications permission then Jan 24 20:25:18 Yeah this was a pain to migrate as you can imagine Jan 24 20:25:33 We needed to do this for all apps & services we had and it aren't even that many Jan 24 20:25:43 Took me a couple of months to get untangled really Jan 24 20:25:54 i'll bet. will need to come up with a faster way to iterate. adb push, launch preware, install, run, open browser... Jan 24 20:25:55 Because all webOS components needed permissions added Jan 24 20:26:13 so is that something LG added? Jan 24 20:26:53 Yeah we updated a lot of the Open webOS components (basically webOS 3.0.5 "open source") to the ones LG put out with webOS OSE Jan 24 20:26:53 (its a good idea, though. very android-like) Jan 24 20:27:16 They also dropped the private and public bus things in LS2 Jan 24 20:27:40 Just the problem is that LG did a code drop with no history and there was 5 years of development in between Jan 24 20:27:53 ok, how about this. i'll whip up a test app that intentionally throughs an enyo.error with a special string, then grep the whole FS looking which file contains it Jan 24 20:27:59 ouch Jan 24 20:29:26 So we had to compare each components code, see if we could merge without issues or come up with "migration plan" Jan 24 20:30:27 That's where this came from: https://webos-ports.org/wiki/LuneOS-webOS-OSE_Migration_Status Jan 24 20:32:26 wow Jan 24 20:34:05 alright, well i'm not smart enough to help with the OS-level stuff, but I think I might be able to hack together a tool chain for enyo app dev Jan 24 20:36:09 Well I'm also not, but still managing here and there Jan 24 20:36:18 I'm just HTML & JS guy like you mainly Jan 24 20:36:28 Can do a bit of QML, but anything C(++) breaks my head Jan 24 20:38:23 ... and that's where I begin my part :) Jan 24 20:39:26 Herrie: wasn't there something to activate to get the inspector visible from outside of localhost ? Jan 24 20:40:35 Tofe: Ideally yes... Jan 24 20:40:42 Well codepoet says it seems to work for him? Jan 24 20:40:48 On localhost:1122 ? Jan 24 20:40:52 for localhost yes Jan 24 20:41:01 but not over wifi Jan 24 20:41:09 I never got that to work no... Jan 24 20:41:24 There should be some port forward, we have it for VBOX but it doesn't work Jan 24 20:41:46 QTWEBENGINE_REMOTE_DEBUGGING Jan 24 20:42:02 from https://doc.qt.io/qt-5/qtwebengine-debugging.html Jan 24 20:42:49 could be put in our qtwebengine.conf somehow Jan 24 20:43:31 Tofe: Yeah I think thye might have added that not too long ago Jan 24 20:44:22 QTWEBENGINE_REMOTE_DEBUGGING=0 Jan 24 20:44:22 0 Jan 24 20:44:25 oops Jan 24 20:44:35 QTWEBENGINE_REMOTE_DEBUGGING=0.0.0.0:1122 might work for instance Jan 24 20:44:39 if i can find the log file (assuming it exists) i can do something like this to watch it from console https://stackoverflow.com/questions/2957041/how-i-can-simulate-tail-command-for-file-on-the-android-file-system Jan 24 20:44:42 Well we're supposed to be doing that already? https://github.com/webOS-ports/luna-webappmanager/blob/b95d031f27eebee0deb7c9660a8080e487738c9b/src/main.cpp#L140 Jan 24 20:45:35 Herrie: if it's just a port number, the debug can only be local Jan 24 20:46:04 codepoet: you're looking for LS2 errors, right Jan 24 20:46:31 i don't know what that is. just the enyo log Jan 24 20:47:14 one moment Jan 24 20:47:21 Tofe: Well Qt's documentation as usual not very clear Jan 24 20:47:32 Herrie: yes, quite so Jan 24 20:47:37 codepoet: Did you check /var/log/messages and /var/log/legacy-log? Jan 24 20:48:15 yes. lots of messages in there, but wayy more than i can follow Jan 24 20:48:25 "journalctl -f -l -u ls-hubd" Jan 24 20:49:25 @Tofe that has "MojoService" messages, but nothing from my app Jan 24 20:49:31 i thought Mojo wasn't in? Jan 24 20:49:49 Mojo framework not, but MojoService is to make it more confusing Jan 24 20:49:58 ooh you mean when your app does console.log etc Jan 24 20:50:19 yes, i think so. enyo.log seems to direct to console.log Jan 24 20:50:28 then "journalctl -f -l -u luna-webappmanager" Jan 24 20:50:31 i'm grepping the whole file system Jan 24 20:50:35 ok, i'll try that Jan 24 20:51:00 I'll try the remote web debug in the meantime Jan 24 20:51:24 yes, that has a few extra messages, but its not too bad Jan 24 20:53:23 this is good. i can use this Jan 24 20:54:01 codepoet: you can also add "QTWEBENGINE_REMOTE_DEBUGGING=0.0.0.0:1122" at the end of /etc/luna-next/qtwebengine.conf to enable remote debugging via wifi Jan 24 20:54:08 Tofe: Yeah seems you might be right, when I search for this env variable on GitHub code, I come across some samples with just port and also some with 0.0.0.0:port like: https://github.com/nickgermaine/Eden-Browser/blob/e7d560c1626b409c7effee4bad45a3dd0b4ec98c/main.cpp#L14 Jan 24 20:54:17 for me it worked well Jan 24 20:55:00 Herrie: I suppose we should patch our luna-webappmanager to do that in dev mode, otherwise it's really not friendly for devs Jan 24 20:55:28 Tofe: Good idea Jan 24 20:55:40 And then the port forward in VBox will probably also work Jan 24 20:55:59 yes, most probably Jan 24 20:57:55 anyone know how to get out of journalctl without hitting Ctrl+C? Jan 24 20:58:03 Herrie: https://github.com/webOS-ports/luna-webappmanager/pull/91 Jan 24 20:58:04 ctrl+c kills my adb shell sessions Jan 24 20:58:34 Tofe: You tested this by building? Jan 24 20:58:38 oh, it shouldn't... I don't really know any other way, ctrl+c works well for me Jan 24 20:58:48 Herrie: mmmh.... wait a second ! Jan 24 20:58:56 I guess it will work anyway since the previously 1122 worked Jan 24 20:59:15 Might be good to double check we at least have /var/luna/dev-mode-enabled ;) Jan 24 20:59:30 ok i updated my qtwebengine.conf Jan 24 20:59:32 reboot required? Jan 24 20:59:51 codepoet: now, either reboot or start "systemctl restart luna-webappmanager" Jan 24 21:00:01 Herrie: build ok Jan 24 21:00:06 Tofe: Ok Jan 24 21:00:34 Herrie: I have that file on my TP Jan 24 21:04:39 On qemux86-64 as well Jan 24 21:04:42 OK merged Jan 24 21:05:27 yes, that worked Jan 24 21:06:18 codepoet: and you'll have easier access to logs too Jan 24 21:07:35 I don't get the app screencast though, not sure if that's normal here Jan 24 21:08:13 i don't either. just a green box Jan 24 21:08:15 but that's ok Jan 24 21:08:29 now if i can work an easier way to deploy, i'll be right at home :-D Jan 24 21:08:59 ok, good :) Jan 24 21:09:17 ah, to deploy, I misread Jan 24 21:10:57 codepoet: you can image something like this: Jan 24 21:11:12 1) adb push myipk.ipk Jan 24 21:11:19 sorry Jan 24 21:11:26 1) adb push myipk.ipk /tmp Jan 24 21:11:35 ya, i did the push, launch preware, browse, install Jan 24 21:11:36 2) adb shell "opkg install --force-reinstall --force-downgrade /tmp/*.ipk && rm /tmp/*.ipk" Jan 24 21:11:49 ok that's much better! Jan 24 21:12:06 that's installing the ipk as a system app though, but should do the trick Jan 24 21:12:21 oh Jan 24 21:12:46 Well you could I guess put a luna-send call instead of opkg install Jan 24 21:13:01 To do it via AppInstallD or whatever it's called again Jan 24 21:13:29 Herrie: ah, yes, wait a moment we had example ealier in the logs Jan 24 21:13:29 ya, maybe drop a script that takes the app as an argument and makes the right luna send call Jan 24 21:18:33 something like /usr/bin/luna-send -n 6 luna://com.palm.appinstaller/installNoVerify '{"subscribe":true, "target": "/tmp/myapp.ipk", "uncompressedSize": 0}' Jan 24 21:20:26 probably only "target" is required here Jan 24 21:20:59 ok, i can make that work. i'll whip up some scripts that work like the Palm SDK, to make it easier for developers to migrate (not that there are any developers left) Jan 24 21:22:37 The permission bits are easy enough to add... I guess we'd need to check on legacy webOS if the added tag doesn't create havoc there Jan 24 21:23:01 In case it doesn't we could just update the apps we know run on LuneOS to have a new appinfo and we can run the exact same IPK on legacy & LuneOS Jan 24 21:23:13 i will, but generally you can put anything you want in the appinfo json and it just ignores anything it doesn't understand Jan 24 21:23:33 Ah OK, that's what I kinda recalled, but wasn't sure, memory is fuzzy at times Jan 24 21:26:36 I recall there was a start of some SDK in the past let me check Jan 24 21:28:34 Ah there was this for installing it seems: https://github.com/webOS-ports/webos-ports-sdk Jan 24 21:28:54 Installing LuneOS I mean Jan 24 21:29:53 forked. i'll take it from there Jan 24 21:31:01 Any questions, guidance or help, we're usually around... Lots of things in our collective brain, not too much written down unfortunately Jan 24 21:31:11 Though most of legacy bits still apply and work Jan 24 21:31:19 And you know plenty of that **** ENDING LOGGING AT Tue Jan 25 02:59:56 2022