**** BEGIN LOGGING AT Sat Jan 07 03:00:02 2017 Jan 07 04:22:38 In the long term, we should think about ways to address the issues raised in the article “How Technology Hijacks People’s Minds — from a Magician and Google’s Design Ethicist” https://getpocket.com/explore/item/how-technology-hijacks-people-s-minds-from-a-magician-and-google-s-design-ethicist-1300144185 Jan 07 08:04:30 morning Jan 07 08:34:01 Tofe: https://bpaste.net/show/84d7670f1549 Jan 07 08:34:09 This is on my 12.1 based Mako Jan 07 09:58:39 Herrie: seems we need to recompile android using sfdroid sources, we need at least modified android_frameworks_native and android_hardware_libhardware for graphics socket Jan 07 10:01:15 we might want to try apkenv as well, seems it's very simple to port (for Tofe :) ) Jan 07 10:01:28 i even have bb recipe for it somewhere Jan 07 10:02:54 nizovn: Ah OK Jan 07 10:03:03 That's what I thought, it already looked to easy like this Jan 07 10:03:14 They have a ton of Android repo's in sfdroid github Jan 07 10:04:33 yes :) Jan 07 10:34:24 Morning! Jan 07 10:34:42 Disclaimer: I am on a *very* jumpy connection :) Jan 07 10:37:06 Herrie|2: you have to take into account that sfdroid is originally made for Sailfish, meaning it supposes that the android image (and its init) and the classic glibc os are merged Jan 07 10:38:45 nizovn: to me, the main issue is still that our android init is run within a lxc container, pretty invisible to the main os Jan 07 10:43:59 Tofe: seems sfdroid and surfaceflinger communicate through socket, i'm not familiar with lxc, maybe we can just open the socket in android's filesystem (this was working using chroot method) Jan 07 10:45:10 nizovn: could be... and I totally missed the fact that some bits were hacked in the android stack itself, though it makes sense Jan 07 10:46:15 I musst say I'm not very much tempted in recompiling a big chunk of android Jan 07 10:46:34 yup, understand Jan 07 10:50:49 qh but wait, don't we already build that part for the hybris layer ? Jan 07 11:03:23 Tofe: I think it uses a seperate Android in the LXC compared to what the OS itself uses? I found this: https://bpaste.net/show/82489f780a0a Jan 07 11:08:12 Herrie|2: in the script we see that it tells you to extract the patched android directly in /, i.e. on top of /system Jan 07 11:08:34 I don't want to do that, if it can be avoided Jan 07 11:12:30 Also I think I made the lxc container too complex: it doesn't really need the hybris-hal layer in there, it is only needed by glibc binaries Jan 07 11:14:14 (I still need to think this through again) Jan 07 11:15:39 Tofe: Take your time :) Jan 07 11:44:05 Tofe: BTW my build didn't like the: rm /usr/share/applications/sfdroid*.desktop so I commented that out during my build Jan 07 11:45:07 "rm: cannot remove '/usr/share/applications/sfdroid*.desktop': No such file or directory" Jan 07 11:45:58 It would then of course give me a warning for the sfdroid*.desktop files that were there, but at least it would build. Jan 07 12:08:45 BTW: Testing a morty based CM 12.1 build for Mako now :) Jan 07 12:11:47 Well building is a better word, it'll take a while still I guess :P Jan 07 12:21:34 Tofe: On another subject: For QtWebEngine: Seeing that they're in 5.8 beta already. I guess we can skip 5.7 and go to 5.8 directly when we upgrade? Jan 07 12:22:10 Release of 5.8 is due in 2 weeks or so Jan 07 12:22:46 Would move to Chromium 53 vs 49 in 5.7 Jan 07 18:43:10 OK got Mako & Maguro building for morty, soon some tests :D Jan 07 19:10:51 OK Mako is still bootlooping... Jan 07 19:10:59 Tofe: How do I get & pull the log again for that? **** ENDING LOGGING AT Sun Jan 08 03:00:00 2017