Time Nick Message 16:56 rubenwardy please may someone who can reproduce the error pick up #11466 16:56 ShadowBot https://github.com/minetest/minetest/issues/11466 -- Use scoped app storage on Android by rubenwardy 16:57 Krock I can re-test on Android 6.0 if that's what you're asking 16:57 Krock nvm. seems like you're asking for adopting 17:01 rubenwardy I think that the problem may be due to the apache commons io lib not working on older versions, given the stack trace in the PR 17:18 sfan5 just pick a different lib? 17:19 sfan5 or, stupid but workable idea: call /system/bin/cp -r 17:21 sfan5 another one: add a wrappe around fs::CopyDir and call that from Java 17:21 sfan5 wrapper* 17:27 MTDiscord is https://github.com/minetest/minetest/issues/11669 that could be added to the 5.5.0 milestone(since it is a regression)? 17:29 Krock This seems to be a nice side-effect without major problems though 17:29 Krock *niche 17:30 Krock still, there should be a 5% gap around the screen 17:30 MTDiscord its a major problem for any formspec that wants to be full screen, say perhaps if the main menu redesign wants a full screen element 17:30 MTDiscord which in fact, a few proposed ideas did 17:30 Krock thing is that there's no real fullscreen option, and the same thing happened for too tall formspecs in the past 17:31 MTDiscord 5.3 and below you can make a full screen formspec 17:31 Krock should be fixed nonetheless I guess 17:32 MTDiscord in fact, the src says a full screen in real cords is about 15 units, which is what i got from testing in 5.3 and below 17:33 MTDiscord https://github.com/minetest/minetest/blob/master/src/gui/guiFormSpecMenu.cpp#L3286 17:34 MTDiscord in theory if you rip out https://github.com/minetest/minetest/blob/master/src/gui/guiFormSpecMenu.cpp#L3261-L3266 and replace references to it to mydata.screensize.x/y the problem should be resolved 17:35 MTDiscord funny thing is krock, you suggested this regression here https://github.com/minetest/minetest/pull/10144#pullrequestreview-462771993 17:36 Krock .. which was then addressed in the next commit 17:42 MTDiscord anyways, if "should be fixed nonetheless I guess" does this mean it can be added to 5.5.0 milestone? 17:42 MTDiscord since there is no way for mods to work around this? 17:57 Krock added 17:59 Krock hmm 17:59 Krock could it be that gui_scaling must be taken into account there as well? (L3269) 18:00 Krock i.e. padded_screensize.X * gui_scaling / mydata.invsize.X 18:01 Krock or rather. divide by gui_scaling because that increases the final invsize