Time Nick Message 07:49 xavierg pgimeno_: yeah, I was wondering whether openbox had kept some damned information about the Minetest window (presumably based on title) but I found nothing about this kind of behaviour and I have only one application-specific rule in my OpenBox configuration. 07:49 xavierg (and it does not impact minetest, of course) 07:55 xavierg pgimeno_: There is indeed a significant difference. 07:56 xavierg While in-game, on the problematic X user session, Alt+left-click moves the window, Alt+right-click resizes the window. 07:56 xavierg On the X user session where everything works fine, Alt+left-click moves the window too... 07:56 xavierg but Alt+right-click, instead of resizing the window... moves it. 07:57 xavierg This is obviously due to the fact that the mouse pointer remains centered on the session which works fine. 07:58 xavierg Therefore, I have to explain why the mouse pointer is offset in my session. 07:58 xavierg First things first, let's measure the offset 08:02 xavierg The offset is non-constant; it diminishes as I make the Minetest cross closer to (0, 0) by moving the window and increases in the opposite direction 08:05 xavierg (at this stage, this is officially an OpenBox issue, but if it can help explain #4629... 08:26 pgimeno_ xavierg: do you have a means to change the WM? in Debian there's a menu for that in some WMs 08:28 xavierg pgimeno_: yes, I can try that; any decent replacement in mind? 08:29 xavierg btw, I have just tried running that other user's openbox in my X session; it does not fix the issue. 08:30 pgimeno_ well, for an experiment any will do 08:30 pgimeno_ fluxbox also allows changing that back 08:32 pgimeno_ hmm, something in your .Xsession maybe? 08:33 xavierg I have replaced OpenBox with FluxBox and still encounters the issue. 08:33 xavierg (this is getting frustrating) 08:34 xavierg I wish it happened in all sessions, I would have blamed SDL or something :') 08:41 xavierg I do not have a custom .Xsession; both users rely on the system-wide one. 08:46 xavierg I think I am out of ideas; there are many potential culprits (minetest, SDL, the window manager, possibly environment variables) but I cannot pinpoint any that could explain that only my session is impacted. 08:47 xavierg s/SDL/Irrlicht/ 09:00 pgimeno_ could listing the dotfiles help shed any light? 09:01 pgimeno_ does renaming out .minetest change anything? 09:03 xavierg .minetest was deemed innocent yesterday. I suspect some difference in the way the cursor is configured. 09:03 xavierg (that would indeed end up in some dotfiles) 09:05 xavierg Hmm, both sessions use the same cursor theme and afaict they have the same rendering. 09:07 xavierg To shed some light on the situation: I have two users on that machine. "xavier" is my historic user, with an atrocious history of dotfiles piling up years after years. 09:07 xavierg Some time ago, I started writing SaltStack formulas to recreate "xavier"'s desktop in a clean, well-managed and reproducible way. 09:08 xavierg I used a second user named "xaviert" (t for test) to test this clean setup. 09:09 xavierg so, as a solution, I could probably migrate from "xavier" to "xaviert" right now and forget that issue. 09:09 xavierg but, eh, I am sooooo curious to know what triggers the issue in the first place :') 09:15 pgimeno_ do you have some other software that captures the mouse, e.g. VirtualBox? 09:15 pgimeno_ running, I mean 09:15 xavierg nope 09:16 xavierg Yesterday evening, I tried to run with as little side-software as possible to check that, but that gave no results. 09:17 xavierg Therefore I am falling back on the following hypothesis : somewhere in my dotfiles is a damn directive that messes things up 09:18 pgimeno_ after discarding the WM, I can only think of two potential culprits, the DE and X11 09:20 xavierg As you can guess from my SaltStack-related explanations, both users have the same DE, the same WM and run on the same machine (i.e. they share the same X.org/X11 stack) 09:21 xavierg hence the "mystery dotfile" hypothesis. 09:21 pgimeno_ yeah, I was thinking DE configuration 09:21 xavierg I mean: the same DE *AND* DE configuration 09:22 pgimeno_ replacing the DE isn't so easy as replacing the WM, but maybe it's worth a shot 09:22 xavierg well, yes and no 09:22 xavierg since xaviert's minetest works perfectly using the same DE, same WM, same DE configuration and same WM configuration. 09:22 xavierg I suspect a dotfile that would mess with X11 though 09:23 xavierg I have just checked .Xresources and .Xdefaults 09:23 xavierg but that does not solve anything 09:23 xavierg I am going to create a third user 09:28 xavierg and done; I now have three users, all of them with LxQt+OpenBox; one has the default configuration provided by Debian and minetest works fine; one has my custom, SaltStack-controlled configuration and minetest works fine. 09:28 xavierg One has my custom, SaltStack-controlled configuration + years of piled up dotfiles and does not work. 09:28 xavierg It HAS to be something in my home :) 09:29 xavierg That being said, I cannot afford to spend all day on this issue, I am supposed to achieve some work. I'll keep you updated :) 09:30 pgimeno_ I need to go as well. Good luck :) 09:36 xavierg [10:28:39] < xavierg> It HAS to be something in my home :) 09:36 xavierg Correction: is HAS to be a file specific to my user (e.g. it could also be some file under /var/lib) 12:00 xavierg Extraordinary. 12:01 xavierg My next experimentation consisted in replacing my home with a fresh, empty one then copying all dotfiles from the old home into it. 12:02 xavierg and guess what: with all the dotfiles from the home where minetest does not work... minetest works. 12:03 xavierg so either there is some non-dotfile interfering (that would be a first, right?) or something weird happened within, say, .cache. 12:03 xavierg of course, if I revert to the former home directory, minetest does not work. 12:04 xavierg so at least we have something consistent here :D 12:14 xavierg Ok, I think I have the culprit(s) 12:14 xavierg I just need a little more time to refine the diagnosis 12:18 xavierg Aaaand fully diagnosed. 12:20 xavierg in my .config/autostart, I have a configure-logitech-mouse.desktop that executes ~/bin/configure-logitech-mouse -- this explains that the issue required both dotfiles and non-dotfiles. 12:20 xavierg ~/bin/configure-logitech-mouse is: 12:20 xavierg xinput --set-prop 'Logitech USB-PS/2 Optical Mouse' 'Coordinate Transformation Matrix' 1.600000, 0.000000, 0.000000, 0.000000, 1.600000, 0.000000, 0.000000, 0.000000, 1.000000 12:22 xavierg modifying the coordinate transformation matrix of an input device is a convenient way to configure its behaviour but it (presumably) drives Irrlicht crazy :) 12:22 sfan5 make sure to post that in the issue 12:22 xavierg I am going to add all of this to the Github issue 12:23 xavierg I guess openbox was innocent all along. 12:23 xavierg We can now make a Hollywood movie out of how we harassed it. 12:40 xavierg sfan5: Fixer: pgimeno_: rubenwardy: done: https://github.com/minetest/minetest/issues/4629#issuecomment-362253640 13:33 MinetestBot 02[git] 04nekoswag -> 03minetest/minetestmapper: put color values in quotes 13cca7072 https://git.io/vNFv2 (152018-02-01T13:30:36Z) 15:52 MinetestBot ThomasMonroe: Feb-01 10:00 UTC hi 16:52 pgimeno_ xavierg: great! I do use xinput to adjust my mouse too, but I only adjust "Device Accel Constant Deceleration" so I never experienced that issue 17:20 rubenwardy Zyrou: I like it 17:20 rubenwardy where are you planning on putting each team? 17:24 rubenwardy Zyrou: I like it 17:24 rubenwardy where are you planning on putting each team? 17:24 Zyrou You see those red and blue glass buildings? 17:24 Zyrou That was my thought but tbh you can put them wherever you wish 17:24 rubenwardy ah, I see 17:25 Zyrou Somewhere inside, whichever floors 17:29 rubenwardy looks good 17:29 rubenwardy they're missing stairs though 17:30 rubenwardy I guess that's in the todo? 17:32 ThomasMonroe lol 17:41 xavierg pgimeno_: yep, that makes sense; my laptop's touchpad is configured using synclient and properties so the issue was specific to my historic Unix user on my workstation. 18:25 blaise can we use tubes with technic reactor core yet? 18:39 blaise no comment? 18:40 ThomasMonroe not that i am aware of no 18:42 Krock blaise, comments may take hours on IRC 18:42 blaise hrmm 18:42 blaise werent they talking about integrating tubes for the nuclear reactor like 6 or 7 years ago? 18:49 progysm talking is not doing 19:54 blaise progysm: oh yeah? 20:09 Out`Of`Control hi 20:19 blaise helo 20:19 blaise hello 20:19 blaise Viper: how are you? 20:20 Viper good. how are you? 20:23 blaise I'm well thanks.. 20:25 Hijiri blaise: Problem that I imagine is that the reactor needs to be full to be on, and then it's hard to queue up more fuel rods for injection 20:26 blaise yeah, only one water block can be missing for the hv cable 20:26 Hijiri two 20:26 Hijiri right? 20:27 blaise no, just one 20:27 Hijiri oh 20:27 Hijiri but the other layers can have two missing 20:27 blaise if it was two, then a tube would be possible 20:27 Hijiri other layer* 20:27 blaise yeah, one for the cable and another as a man hole 20:27 Hijiri one water block missing for the hv cable 20:27 blaise so you can climb in there and refuel it manually 20:27 Hijiri or one water block missing for node breaker with bucket loaded 20:27 Hijiri :P 20:28 blaise say what? 20:28 Hijiri :I 20:29 blaise sorry, I'm trying to design a centerfuge cascade 20:29 blaise lmao 20:29 Hijiri something fancier than 34 centrifuges in a line? 20:29 Hijiri or was it 33 20:29 blaise 34 20:29 Hijiri ok 20:29 blaise and yeah, 20:30 blaise I dont wanna do the line dance thing 20:30 Hijiri one time I had a one centrifuge cascade 20:30 Hijiri but I never completed a 6 rod set 20:30 blaise HEH 20:30 Hijiri I used luacontrollers and diglines to control which stacks got put in 20:31 blaise I was thinking of using the teleport tubes.... idk though 20:31 Hijiri teleport tubes are kind of expensive 20:31 blaise yeah 20:32 Hijiri putting them in a line is already efficient unless you are trying to go for 34 centrifuges 20:32 Hijiri I guess if you can remove the extra tubes you put for input on some centrifuges and just have the sorting tube directly output into the machine 20:33 Hijiri or if you reduce the # of sorting tubes by combining the sorting of adjacent centrifuges 20:33 Hijiri tubes are cheap though 20:36 blaise I'm using same tube for input and output 20:42 Hijiri It never occurred to me to try that, I guess then you basically have the minimum number of needed things 20:42 Hijiri except maybe if you want to try reducing the number of sorting tubes 20:43 Hijiri you could put four centrifuges around a sorting tube and have the highest possible output go up, and the lowest go down 20:43 Hijiri and the side direction just have the dusts corresponding to that centrifuge 20:43 Hijiri oh wait, but do they need to be powered from the bottom 20:43 Hijiri what would be neat is if you could just power them from the sides and have just stacks of centrifuges 20:44 Hijiri I think doing it this way would reduce the number of sorting tubes by a factor of 4 20:45 Hijiri now I want to try it myself 20:50 Hijiri just tested, they need to be bottom-powered 20:51 Hijiri You could still have a wire efficient setup where you connect two centrifuges to each sorting tube 22:36 blaise Hijiri: I'm running into a problem where if the centerfuge is facing a different way it wont output 22:37 blaise I tried making a square array instead of in a straight line 22:38 blaise kinda screws the whole thing up 22:38 blaise was working fine till I went around a corner 22:38 blaise they'll go in but not out 22:38 blaise wtf 22:43 Hijiri blaise: the side it outputs from changes depending on which way it faces 22:43 Hijiri on one axis it is the left side otherwise it is the right size 22:43 Hijiri weird old technic bug 22:43 blaise omfg 22:44 * blaise attacks some old people with a rubber chicken 22:45 * blaise flips out 22:45 Hijiri it's pretty dumb 22:52 blaise aaaaand, it wont accept/send from the top 22:52 blaise great