**** BEGIN LOGGING AT Sun Apr 17 02:59:57 2022
Apr 17 08:19:09 <Tofe> Morning!
Apr 17 08:58:12 <HerrieTP> Morning!
Apr 17 09:16:14 <JaMa> morning
Apr 17 20:02:34 <Tofe> So, this getTimeZoneRules crash in WAM is still unsolved :/
Apr 17 20:02:58 <Tofe> But I tracked it down to this: 
Apr 17 20:03:02 <Tofe> (new PalmServiceBridge()).call("palm://com.palm.systemservice/timezone/getTimeZoneRules", '[{"tz":"Etc/GMT-0","years":[2022,2017,2018,2019,2020,2021,2022,2023,2024,2025,2026,2027]},{"tz":"Etc/GMT-0","years":[2022,2017,2018,2019,2020,2021,2022,2023,2024,2025,2026,2027]}]')
Apr 17 20:03:11 <Tofe> such call will crash WAM
Apr 17 20:03:43 <Tofe> it can be tried in any app from the inspector
Apr 17 20:05:10 <Tofe> shorter version: (new PalmServiceBridge()).call("palm://com.palm.systemservice/timezone/getTimeZoneRules", '[{"tz":"Etc/GMT-0","years":[2022]}]')
Apr 17 20:06:36 <Tofe> JaMa: about chromium91, I wasn't sure if python removal was already applied or not
Apr 17 20:06:47 <Tofe> python2 I mean
Apr 17 20:07:09 <Tofe> I know it was still WIP for some chromium versions
Apr 17 20:19:48 <JaMa> Tofe: it was applied, looks like opus is enabled only for arm builds
Apr 17 20:20:10 <JaMa> Tofe: I already had a backported patch laying around, not sure why I haven't added it in meta-webos-ports
Apr 17 20:20:34 <Tofe> ah, ok, I see
Apr 17 20:20:52 <JaMa> for timezone crash, check the .bbappend and .patch applied for tzdata in meta-webosose
Apr 17 20:21:23 <Tofe> JaMa: that one might be my own patch ?
Apr 17 20:22:27 <JaMa> python3 needs 30+ patches in webruntime-91 (already applied in meta-webos-ports), for webruntime-94 it will need just a few (but no ETA when 94 from LGE hits github, probably not with the next release)
Apr 17 20:23:08 <Tofe> for tzdata, I see the patch, it's indeed something we should also add on our side. However, I doubt it's the cause of the crash
Apr 17 20:23:26 <JaMa> https://github.com/shr-project/meta-webosose/blob/dunfell/meta-webos/recipes-extended/tzdata/tzdata/etcetera.patch
Apr 17 20:23:52 <Tofe> the LS2 call is working fine from the command line, for instance
Apr 17 20:24:02 <JaMa> yes, might not be directly related
Apr 17 20:24:50 <Tofe> it's unfortunate that debugging chromium's core is such a pain
Apr 17 20:26:01 <Tofe> I'll try adding from printf or so
Apr 17 20:36:15 <Tofe> ... and as often, just after I request help from LG, I think I understood where the crash is
Apr 17 20:37:26 <Tofe> https://github.com/webosose/chromium91/blob/master/src/neva/injection/renderer/webosservicebridge/webosservicebridge_injection.cc#L55= here we have a JSON array, but it doesn't check it, so the line 59 will crash
Apr 17 20:38:16 <Tofe> no, actually, it's dumber than that: the "&&" should be "||"
Apr 17 20:39:30 <Tofe> I'll test that right away
Apr 17 20:49:53 <Tofe> My fix worked :) ok I'll do a patch
Apr 17 20:58:33 <Tofe> Damn, I'm glad this issue is finally tackled
Apr 17 21:21:30 <HerrieTP_> Tofe: Silly mistake by LG devs by the looks of it?
Apr 17 21:21:37 <HerrieTP_> Weird only we see it though
Apr 17 21:21:43 <HerrieTP_> Could be edge case?
Apr 17 22:16:32 <Tofe> Herrie: most (if not all) of the APIs take a json object, not an array
Apr 17 22:17:05 <Tofe> but it could be a very simple way to crash a TV :)
Apr 17 22:28:32 <Tofe> now one remaining thing to do would be to better handle the "launchHidden" parameter. Shouldn't be hard, but I have to find the right place to put the logic
**** ENDING LOGGING AT Mon Apr 18 02:59:56 2022