**** BEGIN LOGGING AT Wed Mar 02 02:59:56 2022 Mar 02 08:23:21 Morning! Mar 02 10:19:11 Morning! Mar 02 10:27:29 Herrie: looks like I won't even need to rebuild webruntime in the new branch Mar 02 10:37:40 Tofe_: That's what I thought already but wasn't 100% sure Mar 02 10:38:13 I wonder what is this mksnapshot tool Mar 02 10:39:17 seems related to nodejs or so Mar 02 10:39:57 Seems to build the V8 JS Mar 02 10:40:46 ah ok, this might take some time then Mar 02 10:46:02 oh, do_rootfs already, good Mar 02 14:38:22 Tofe: Seem I'm running into some CORS for Enyo 1.0 apps: Mar 02 15:31:51 qemux86-64 WebAppMgr[785]: [] [pmlog] wam.jsconsole ERROR {} E[785:785:CONSOLE(0)] "com.palm.app.accounts- \"Access is blocked to resource: file:///usr/palm/public/accounts/com.palm.othermail/images/othermail32.png\", source: file:///usr/palm/applications/com.palm.app.accounts/index.html (0)\n" Mar 02 14:39:16 Herrie: ah, that would be why we don't have images near the account labels Mar 02 14:40:05 I don't really see the blocking logic here, both source and resource are local Mar 02 14:40:48 Tofe_: Well it could help if I add https://github.com/webOS-ports/com.webos.app.imageviewer/blob/webOS-ports/webOS-OSE/webos-meta/appinfo.json#L16 for example Mar 02 14:40:50 Testing that now Mar 02 14:41:44 Seems not ;) Mar 02 14:41:50 too bad Mar 02 14:43:35 Not sure what causes it Mar 02 14:45:00 Also, I couldn't access the remote debug homepage in chromium, how do we activate it on the new WAM ? Mar 02 14:49:36 Tofe_: Also not sure on how to enable it Mar 02 14:56:53 Tofe: https://www.webosose.org/docs/tools/sdk/cli/cli-user-guide/#ares-inspect Mar 02 14:57:08 I guess we need to check the code of the cli tools to see what it does Mar 02 15:33:58 Tofe: https://github.com/webosose/ares-cli/blob/main/lib/inspect.js#L23 Mar 02 15:36:54 And https://github.com/webosose/ares-cli/blob/main/lib/inspect.js#L405-L485 Mar 02 15:37:06 So it might be as easy as adding an additional port forward for 9998 Mar 02 15:37:29 yes, very possible Mar 02 15:39:09 Seems more is needed somehow Mar 02 15:39:57 I could understand if it's disabled by default Mar 02 15:40:42 found this: https://github.com/webosose/wam/commit/426376594d15d9af7d8f42fcbd763e708a5bfe8d Mar 02 15:43:39 Tofe: Ah OK Mar 02 15:43:41 That makes sense Mar 02 15:46:48 Now getting this in logs: "com.palm.webappmanager" does not have sufficient outbound permissions to communicate with "com.webos.lunasend-990" Mar 02 15:46:53 But I know how to fix that at least Mar 02 15:54:00 OK seems it's on HTTP and not HTTPS but got something at least now Mar 02 15:57:21 Let me PR some changes here at least, so we have them Mar 02 15:57:39 I'm not 100% sure the /var/luna/preferences/debug_system_apps is required, let me remove it first Mar 02 15:59:28 Seems it's needed Mar 02 15:59:37 Should we put it in our recipe by default for now? Mar 02 15:59:47 Since we're not on production devices really Mar 02 16:02:45 yes, in the -dev recipe Mar 02 16:02:57 we already have another file like that Mar 02 16:03:07 Yeah Mar 02 16:06:12 https://github.com/webOS-ports/wam/commit/04fd13f2a909ffc65f3a3b6d4ba56b551f3b0e9d part 1 Mar 02 16:09:09 Should we add port 8080 right away? Mar 02 16:09:42 For NodeJS services Mar 02 16:10:10 you mean in the appliance ? Mar 02 16:10:41 in this case, yes Mar 02 16:11:13 Well seems for services there's a NodeJS inspector that runs on 8080 port Mar 02 16:11:36 We might map them onto a differnt port locally though? Since on local machine peopel might have 8080 used for web server? Mar 02 16:13:23 I have 8080 available, but it's often taken, so let's choose another one Mar 02 16:13:34 the one you prefer is fine :) Mar 02 16:14:00 maybe something 55xx like the 5522 port ? i.e. 5580 ? Mar 02 16:14:52 OK I was thinking the same Mar 02 16:16:40 I guess this should do it: https://github.com/webOS-ports/meta-webos-ports/commit/4701ea1792bda8a02807c345d65d6401c50efa49 Mar 02 16:17:31 Need to build test it still, but at least it seems to work when I manually add these bits here Mar 02 17:46:54 Herrie: luna-appmanager isn't bumped yet, right? Mar 02 17:47:30 Tofe: Ehm let me check Mar 02 17:47:34 I think I did that too Mar 02 17:48:20 Ah bumped, but not to latest yet Mar 02 17:48:25 Let me do thta now Mar 02 17:48:41 ok ! Mar 02 17:48:50 Done Mar 02 17:49:02 it will help me for integrating JustType correctly Mar 02 17:49:38 Yeah Mar 02 17:49:40 ok, pulled, now I'll refresh the image and start improving our cardshell Mar 02 17:50:52 and at some point, a big refactoring of the cardshell should be considered, to match the MVC proposed by LSM Mar 02 17:51:41 also there is now a "WindowGroup" type, and we should use it at some point Mar 02 17:51:49 for the window stacking I mean Mar 02 17:57:01 Tofe: Yeah I guess some rework there would be good Mar 02 17:57:53 I think it would help getting some improvements for free, like better input management or multiple outputs Mar 02 17:58:32 Tofe: Yeah probably Mar 02 17:58:43 I'll try and see if I can get a bit further with maliit later today Mar 02 17:58:48 Just a lot going on at the moment Mar 02 17:58:58 Seems I need to add some env variables mainly Mar 02 17:59:10 Last time I tried it was missing XDG_RUNTIME_DIR it seems Mar 02 17:59:16 That's LG's maliit Mar 02 17:59:36 note that I'm not filtering out maliit's window in the cardshell yet, so the keyboard might very well appear as a card Mar 02 17:59:58 Tofe: Sure Mar 02 18:00:12 It's more that it was still not looking good in logs at my side Mar 02 18:00:33 ok Mar 02 18:07:01 Once the logs look OK I guess I might have something more to say about how it will look on UI side Mar 02 18:12:50 Tofe: New image the remote inspector seems to work out of the box Mar 02 18:12:52 So that's good Mar 02 18:12:59 great Mar 02 18:13:19 I've push a little commit for luna-appmanager to get rid of the window decoration Mar 02 18:14:33 Tofe: Ah yes that was a bit ugly Mar 02 18:14:38 Also it seems QML apps are stretched Mar 02 18:16:03 that's because I can't get hold of the API to request a client-side resize of the buffer Mar 02 18:22:44 Herrie: do you know what kind of LS2 call I should do to open an URI in a card ? Mar 02 18:23:02 wait, let me google that Mar 02 18:27:25 Looks like there is palm://com.palm.webappmanager/launchUrl Mar 02 18:28:22 Tofe: But we have no browser yet Mar 02 18:28:36 At least not that's working so not sure that will work? Mar 02 18:28:51 in a card, I mean Mar 02 18:29:04 it's for firstuse app, to open the links Mar 02 18:30:05 I thought webengine was used for something else there but could be I'm mixing it up with something else Mar 02 18:30:48 we use WebEngine in FirstUse for showing some license details, and for the links in the Finish page Mar 02 18:31:03 but it's a really basic UI we propose Mar 02 18:31:12 so showing a card will be equivalent here Mar 02 18:31:54 Tofe: Well links should work by default since they have a mime-type/handler if I'm correct Mar 02 18:32:13 but, what will Qt do? Mar 02 18:32:13 The displaying of license is probably more problematic Mar 02 18:32:46 we don't have any "xdg-open" stuff on luneos Mar 02 18:33:24 Tofe: I am not sure to be honest really, I guess trial and see? Mar 02 18:34:40 I mean we should make the link clickable and have it make the ls2 call, rest should be taken care of by luna-sysmgr/sam Mar 02 18:34:59 Based on the registered handlers for the URL Mar 02 18:35:07 At least that's my understanding Mar 02 18:35:30 yes, it's the ls2 call I'm looking for Mar 02 18:35:52 I think a simple open should do Mar 02 18:38:18 Seems "open" indeed: https://kylemaas.github.io/luna-sysmgr/com_palm_application_manager.htm#com_palm_application_manager_open Mar 02 18:38:33 Specify the url in "target" Mar 02 18:44:59 thanks, that's the one I was looking for Mar 02 18:45:49 ok, well, currently it doesn't do anything but return "success !" Mar 02 18:46:05 Tofe: Well that's logical probably Mar 02 18:46:09 Anything in logs Mar 02 18:46:11 ? Mar 02 18:46:26 It's probably still associating those urls with org.webosports.app.browser Mar 02 18:47:30 https://github.com/webOS-ports/luna-init/blob/webOS-ports/master/files/conf/command-resource-handlers.json#L37 Mar 02 18:47:34 At least that's what it should do Mar 02 18:47:43 ah, wait, WAM is crashing Mar 02 18:48:09 a reboot fixes it; it was probably the transition from firstuse that was erratic Mar 02 18:48:48 com.webos.app.enactbrowser- \"ReferenceError: chrome is not defined\" Mar 02 18:49:11 and a dark grey card is opened Mar 02 18:49:13 Tofe: Yeah that's what I'm getting when I launch the browser Mar 02 18:49:24 LG guys don't see very responsive Mar 02 18:49:42 meaning my LS2 call is working, just the browser doesn't work yet Mar 02 18:50:31 Tofe: Yes Mar 02 18:50:35 So that's good Mar 02 18:53:57 https://github.com/webOS-ports/org.webosports.app.firstuse/commit/4ffa3e3a5e5ea3b3e3057a98e936308a11c8b8f7 I'll put that in a proper branch name Mar 02 18:55:20 Tofe: Ah OK, that will at least work around those things Mar 02 18:55:32 Still a bit puzzled why the browser doesn't want to work Mar 02 19:15:53 I'll try to tackle that at some point, but the todo list is a bit long Mar 02 19:17:56 https://github.com/webOS-ports/org.webosports.app.firstuse/tree/wam here is it Mar 02 19:28:18 Tofe: Thn Mar 02 19:28:21 +x Mar 02 19:55:28 codepoet80: Seems you were right, we ARE running into some minor CORS issues it seems Mar 02 19:55:49 But I'm pretty sure we can tackle them. They're pretty much Enyo 1.0 only by the looks of it Mar 02 19:56:21 hi Mar 02 19:56:47 ya, i thought you might. there's a couple things I've learned that might be relevant. Mar 02 19:57:28 Enyo2 was deployed to webOS using Cordova, which had a few tricks to bypass CORS. Most of those tricks don't work in Android any more, but might still be working on whatever you're pulling from. Mar 02 19:58:28 when you compile (or launch?) the webview, you may need to include a flag that allows it to access file:// URIs, since that's now turned off by default Mar 02 19:59:41 codepoet80: OK any pointers that you have are welcome, we can try Mar 02 19:59:52 Seems we pass a whole bunch of flags anyway already Mar 02 20:00:35 is the qt-browser-whatever webkit based? or chromium based? Mar 02 20:01:42 Chromium but heavily patched Mar 02 20:01:55 https://github.com/webosose/wam/blob/master/files/launch/systemd/webapp-mgr.sh.in is the script used to run WAM and it has a ton of params ;) Mar 02 20:02:25 is "–allow-file-access-from-files" one of them? Mar 02 20:02:40 i don't see it Mar 02 20:03:43 for public internet CORS, you can also try "--disable-web-security" although that sounds scary. Mar 02 20:03:55 i'm not sure the full implications, but it definitely turns off CORS Mar 02 20:04:17 https://github.com/webosose/chromium91/blob/7deca0fa8aa01ec414cf689448b76b4f828a100e/src/content/public/common/content_switches.cc#L274 Mar 02 20:04:28 Seems these switches exist in Chromium 91 Mar 02 20:05:00 and the comment at 274 makes it seem less scary Mar 02 20:05:57 Yes Mar 02 20:06:09 We had this in our old webappmanager: https://github.com/webOS-ports/luna-webappmanager/blame/master/files/systemd/luna-webappmanager.service#L8 Mar 02 20:06:15 So I guess we can enable it in WAM too Mar 02 20:06:37 heart emoji Mar 02 20:11:26 Let me try that in a little bit Mar 02 20:11:36 Trying to get some tickets sorted now, so should focus LOL Mar 02 20:12:39 no rush. still learning enyo2, on here in the background most days, so just ping me when its time to starting testing again! Mar 02 20:15:11 codepoet80: Well a lot is still broken, but Enyo 2 and Enact mostly work for now Mar 02 20:15:17 Rest we need to tidy up Mar 02 20:15:24 Oh and yeah we don't have a working browser yet Mar 02 20:16:14 ha ha, that one's important Mar 02 20:16:26 But in general seeing how much we changed it's not that surprising we broke a few things Mar 02 20:16:31 Oh and no VKB either yet Mar 02 22:41:55 https://github.com/webOS-ports/luna-next-cardshell/commit/763082fd8aa463c8e177fa3d429ac7115349ea85 here we go, now off to bed :) Mar 02 22:52:38 Tofe: Ah nice will try that Mar 02 22:52:51 My experiments with WAM and Enyo 1.0 so far not that much success Mar 02 22:53:02 I added the flags but don't seem to do much somehow Mar 02 22:53:09 Anyway gn8 **** ENDING LOGGING AT Thu Mar 03 02:59:56 2022