**** BEGIN LOGGING AT Thu Nov 17 02:59:57 2022 Nov 17 06:31:24 Morning! Nov 17 06:31:31 Tofe: Seems that patch killed WAM Nov 17 06:32:45 Ah after manual restart WAM works again Nov 17 06:37:22 Tofe: Seems the patch works :) Nov 17 06:37:31 I get a version number back from Appinfo.json now Nov 17 07:21:11 ah, good! Nov 17 07:23:23 morning Nov 17 08:02:00 Tofe: So I guess we can merge it :) Nov 17 08:02:25 With regards to the "black" windows that appear (at least in qemux86-64), it seems this might be OSE's Volume App Nov 17 08:02:46 I saw something that resembled a slider this morning for a few short moments before I got a black overlay window Nov 17 08:03:12 ah, ok; mmh I wonder how this kind of app is integrated in the usual OSE compositor Nov 17 08:03:32 as there isn't much flexibility wrt display of cards Nov 17 08:03:58 Herrie: merged Nov 17 08:08:30 Tofe: I'm not sure, it seems they might have appeared in their current state after your recent changes Nov 17 08:09:38 ah, might be the change in SYSTEM_UI alerts Nov 17 08:09:49 I don't recall seeing them before in this shape at least Nov 17 08:09:52 Tofe: Yeah might be Nov 17 08:09:55 might not be the right filtering Nov 17 08:10:24 It's system_ui type: https://github.com/webosose/com.webos.app.volume/blob/master/webos-meta/appinfo.json#L20 Nov 17 08:10:36 ok, then this comment https://github.com/webOS-ports/luna-next-cardshell/blob/herrie/qt6/qml/Notifications/AlertWindowsArea.qml#L43 was on spot :) Nov 17 08:10:46 And https://github.com/webosose/com.webos.app.notification/blob/master/webos-meta/appinfo.json#L12 Nov 17 08:11:14 so... do we have other alerts than com.palm.systemui, on our side ? Nov 17 08:11:28 What do you mean with "alerts" ? Nov 17 08:11:35 All apps could technically send alerts I guess Nov 17 08:11:45 power menu, mainly, also there is the disk full alert Nov 17 08:12:26 we don't have the WindowType we used to have, unfortunately Nov 17 08:13:04 and we don't know yet how to handle non-fullscreen windows Nov 17 08:13:28 that's also why I'm curious how volume app should show up Nov 17 08:15:37 I guess we could check OSE Emulator somehow? Nov 17 08:15:46 Or what it shows on LG's TV's Nov 17 08:19:19 https://github.com/webosose/luna-surfacemanager/blob/master/base/qml/WebOSCompositorBase/views/base/SurfaceView.qml#L84 seems it's either fullscreen (and scaled if needed), or centered on screen Nov 17 08:19:35 but to me it seemed like no app could have custom size Nov 17 08:19:53 it would be great if I'm wrong Nov 17 08:21:49 there is a "transparent:true" in appinfo.json, must be part of the trick Nov 17 08:43:47 I don't see any index.html on github, is it generated at build time? Nov 17 08:50:42 Tofe: I see and index.js Nov 17 08:50:51 So I guess so Nov 17 08:51:08 ok I see it on the qemu image Nov 17 08:51:18 not much magic, at first sight Nov 17 09:20:43 Tofe: I can close them for now and they're gone from the WebInspector and don't interfere anymore, but I guess the handling should be updated somehow Nov 17 09:21:33 Herrie: I see them as an opportunity to have a similar way to handle dashboards and such Nov 17 09:30:28 Tofe: Yeah Nov 17 10:45:16 Tofe: BTW now looking into the cookies issue with the apps, it's really weird. It's probably some Chrome quirk. I enabled --enable-file-cookies, but in console I see now cookie being written Nov 17 10:45:30 If I change the Enyo code to use localstorage instead of Cookies, all seems written fine Nov 17 10:45:40 No errors in console though either it seems Nov 17 10:46:37 I initially thought it would have to do with SameSite or domain fields in the cookies but adding those don't seem to do much either Nov 17 12:37:52 I guess we could update Enyo to use localStorage where it's available instead of cookies, that should solve a lot of the issues I guess Nov 17 13:28:52 Tofe: Suggestion would be to use localStorage when available otherwise cookies. Should work around all changes from Chromium 80+ with regards to cookies Nov 17 13:49:43 Only downside might be cookies which use "expires" tag that would need some additional handling, but that should be a small percentage really Nov 17 14:15:29 Sorry, I'm in a "meeting" day, so not very available until tonight :) Nov 17 14:43:56 Tofe: No worries :) Nov 17 14:44:15 BTW with your "fix" I'm seeing these in console, not the end of the work, but not very pretty either: "enyo-build.js:73 Could not parse /media/cryptofs/apps/usr/palm/applications/com.palm.app-museum2/framework_config.json SyntaxError: Unexpected token o in JSON at position 1 Nov 17 14:44:15 enyo-build.js:73 Could not parse /media/cryptofs/apps/usr/palm/applications/com.palm.app-museum2/appinfo.json SyntaxError: Unexpected token o in JSON at position 1" Nov 17 15:01:55 Tofe: For localStorage alternative for cookies: https://github.com/webOS-ports/enyo-1.0/commit/513bd38ee34e0b55f33846828b385fd2e7e157a3 Nov 17 15:02:36 Tested on runtime with Museum, seems to work :) Nov 17 15:07:25 Popup with version that should disappear on 2nd run, disappears Nov 17 18:25:40 Herrie: sorry, I didn't follow this cookie fix (If may say) Nov 17 18:25:55 but happy to merge the PR if you say it improve things Nov 17 18:52:38 Tofe: Short summary: document.cookie doesn't get populated (anymore) due to various changes in Chromium code. I've worked around this by using localstorage instead Nov 17 18:53:26 Seems cookies on file:// violates the spec and exception launch params were removed from all non-Android Chrome code Nov 17 18:53:57 So fixing cookie handling would not be pretty Nov 17 18:57:28 So this is a non-invasive workaround that should solve most of the issue and be future proof to any changes to Chromium code going forward Nov 17 19:10:07 ok I see Nov 17 19:10:18 yes using localStorage can be a nice workaround Nov 17 19:10:29 you already tested it then? Nov 17 19:11:44 Yes it works here Nov 17 19:13:18 Only on App Museum 2 but code isn't rocket science Nov 17 19:20:52 Tofe: These are a bit more worrying, but probably easy to fix: BTW with your "fix" I'm seeing these in console, not the end of the work, but not very pretty either: "enyo-build.js:73 Could not parse /media/cryptofs/apps/usr/palm/applications/com.palm.app-museum2/framework_config.json SyntaxError: Unexpected token o in JSON at position 1 Nov 17 19:20:53 [15:44] enyo-build.js:73 Could not parse /media/cryptofs/apps/usr/palm/applications/com.palm.app-museum2/appinfo.json SyntaxError: Unexpected token o in JSON at position 1" Nov 17 19:20:58 Came after your fix from yesterday Nov 17 19:33:56 Herrie: my guess is that this happens when the string is empty (not even "{}") and a const json was requested Nov 17 19:35:55 or, the json file isn't really a json Nov 17 21:18:15 Tofe: OK where to fix it is the question Nov 17 21:18:43 It's not really an error, code seems to be OK afterwards but looks sloppy in logs Nov 17 21:19:37 Herrie: could either be fixed in the little js code I committed, or in enyo Nov 17 21:20:04 Tofe: I guess in the js code you did is better so we do it in 1 place? Nov 17 21:20:12 And not in 2 different ones? Nov 17 21:20:27 but I realize I didn't read the second error Nov 17 21:20:49 because, well, /media/cryptofs/apps/usr/palm/applications/com.palm.app-museum2/appinfo.json does exist and isn't empty Nov 17 21:20:59 Yes Nov 17 21:21:00 so why would it begin with the "o" letter Nov 17 21:21:15 I can see tomorrow and see what's the full contents Nov 17 21:21:59 knowning what is returned by palmGetResource would be nice too Nov 17 21:22:48 Will add some debugging in the morning Nov 17 21:23:06 tomorrow I'll keep an eye on the ddr4 deals :) Nov 17 21:23:31 Tofe: me too have some notifications setup Nov 17 21:24:46 though I aim for something quite specific, so not sure at all that something will come up **** ENDING LOGGING AT Fri Nov 18 02:59:57 2022