Time Nick Message 10:04 p_gimeno @Krock @mtda I maintain an up-to-date Irrlicht mirror with all branches, https://gitlab.com/pgimeno/irrlicht-mirror 12:15 mtda hello 12:16 mtda remind me which commit do i have to revert to 12:19 Krock p_gimeno: nice. is mirroring a gitlab feature or does this happen using local scripts? 12:19 Krock http://irc.minetest.net/minetest-dev/2019-08-09#i_5575884 12:19 Krock ^ mtda 12:19 mtda thanks, i'll check 12:41 mtda ok had to leave for a while... now compiling from that commit 12:53 mtda still died in guiButton.cpp 12:53 mtda same error 12:58 mtda termbin.com/1q5u in case it's any different 13:12 Krock well so: reset your branch to d1a1c5cb~1 (one commit prior that) because that's the commit that introduced this enum 13:12 Krock if that works, you can still apply the OGLES patches on top 13:13 Krock what's the status now is the shader path variable now found in the git mirror clone? 13:14 mtda renderingengine and everything that has to do with gles compiles 13:31 Krock > 200 FPS in singleplayer with vsync on 13:31 Krock current master. I doubt that's a minetest issue? 13:32 Krock First I thought my profiler changes might have touched it but apparently vsync just broke at some point in a recent update. Any other occurrences? 13:39 mtda well, i'm compiling in a very unusual environment 13:40 mtda so it might indeed be something to do with ARM or GLES, not minetest 13:41 mtda the commit you've provided is having much more luck though, it already went way past where it previously died 14:02 mtda 67%.. make your bets when will it crash 14:05 mtda have to say, this older commit throws considerably less warnings. maybe my build environment is outdated 14:13 mtda i'm sorry, but termbin.com/c6w4 14:15 sfan5 missing -lEGL 14:15 mtda elaborate? 14:16 sfan5 irrlicht was not linked with libEGL despite requiring it 14:16 sfan5 but you can just fix this while linking minetest using -DCMAKE_EXE_LINKER_FLAGS="-lEGL" 14:17 mtda what consequences would that have? 14:17 mtda i'm new to most of this 14:18 mtda ok it's done 14:19 mtda i'll do a smoke test without minetest-game 14:20 mtda it started up! 14:21 mtda minimal dev test runs, produces decentish fps 15:05 Krock but that's OpenGL, I suppose? 15:05 mtda yes 15:09 mtda ok, now wat do to make it GLES instead of OpenGL 15:10 deltasquared I recall there being a setting for this, however most builds if you try to set it to gles will complain that it's not enabled/compiled in and falls back to regular opengl. 15:11 deltasquared no idea if the video_driver property in advanced settings will do what you want. 15:11 mtda i found the setting and set it to ogles2. might have been a mistake, as after a restart all i see is blue 15:12 deltasquared o.o 15:12 mtda manually changed to ogles1 15:12 deltasquared mtda: try turning off all shaders 15:12 mtda picture is back 15:12 deltasquared including the shader tickbox itself 15:12 mtda they are unavailable now 15:12 mtda apparently i don't have ogles2 support? 15:13 deltasquared hrm... turn them off anyway and try again? I've found ogles2 will allow you the shader settings, but they're all fail broken. 15:13 deltasquared all I get for instance in many cases is a black screen with just nametags and HUD elements. 15:13 deltasquared (this being on my android device using ogles2 there) 15:13 mtda i'm not sure raspi supports ogles2 15:14 mtda 3b+ at least 15:14 deltasquared mtda: try running es2info. 15:14 mtda not found 15:14 deltasquared hmm... wait, was it called something else 15:14 deltasquared oh wait, es2_info sorry 15:14 mtda doesnt exist 15:15 deltasquared it should come along with whatever package has glxgears... 15:15 deltasquared dunno what that is on debian 15:15 mtda i do have gears tho 15:15 deltasquared hmm 15:17 mtda it's possible i have horribly misconfigured graphics though 15:17 mtda can't even run openarena 15:18 mtda so yeah, on default settings with ogles1 and no smooth lighting it's playable 15:18 mtda unless you go to spawn of a public server 15:23 mtda i hope this "stress testing" of mine helps 15:23 mtda running on unusual underpowered hardware and software 15:40 Krock !tell mtda see debug.txt. I think it cannot find the shaders 15:40 ShadowBot Krock: O.K. 16:13 sfan5 Krock: i think player movement would overwrite any "prediction" done by other clients 16:14 sfan5 the other problem is that it doesn't apply in free_move mode, which is not something the server can tell 16:25 Krock right 16:25 Krock too bad that there's so much latency 16:56 sfan5 Krock: your suggestion tricked me, checking after multiplication is intentional 16:57 sfan5 |d| == 1 in fact usually means that each vector component is somewhat less than 1 17:06 Krock yes it is, thus k needs to be checked against lower values 17:06 Krock the movement distance will always be 1 * k (but in 3d) 17:11 sfan5 yes, but I am intentionally not checking the distance 17:13 sfan5 I could check the distance instead, but it should be more like < 1.5 17:14 Krock mhm okay 17:19 Krock nerzhul, nore, rubenwardy, sfan5, sofar: Would you be available for a meeting in like 40 minutes? https://dev.minetest.net/Meetings#2019-07-X 17:19 sfan5 i guess 17:22 Krock other contributors are certainly also welcome to give inputs 17:32 paramat i'm around from now on 17:38 sfan5 merging #8685, #8776 in 5 minutes 17:38 ShadowBot https://github.com/minetest/minetest/issues/8685 -- Player knockback and add_player_velocity API by sfan5 17:38 ShadowBot https://github.com/minetest/minetest/issues/8776 -- Finish getNode cleanup by osjc 17:45 Krock paramat: hi. Will you be available for a meeting in ~15 minutes? 17:46 sfan5 did github just create a merge commit 17:46 sfan5 again? 17:46 sfan5 or is the description just wrong? 17:49 sfan5 looks like it's just the message 17:50 sfan5 I did set it correctly but github said "merge failed" after the merge and upon reloading the pull request was already merged 17:50 sfan5 maybe github should fix their stupid software... 17:54 paramat yes i can do meeting, but have very few thoughts on the PR topics 17:55 Krock force push 17:55 Krock nvm, too late (>5 mins) 17:56 paramat or, my thoughts are already present as comments in the PRs 18:03 Krock Okay, let's start. 18:04 Krock #6688 is a massive PR, but needs rebase and design re-considerations 18:04 ShadowBot https://github.com/minetest/minetest/issues/6688 -- Custom particle generators for particle spawners by raymoo 18:05 Krock should it be closed? I don't see any hope there right now, and maybe it could be done using CSMs somewhen 18:07 paramat my opinion is clear in the comments, this is a very bad idea 18:08 Krock so close? 18:08 paramat i think so yes 18:09 Krock rubenwardy apparently likes it. well so, reopen on rebase is still an option 18:13 Krock what to do with #8382? close or merge? 18:13 ShadowBot https://github.com/minetest/minetest/issues/8382 -- Fix debug.txt growing bigger and bigger by HybridDog 18:14 paramat i find this one difficult to decide on 18:15 paramat i feel the implementation may need some reconsideration 18:15 DS-minetest huge debug.txt files are awful imo 18:16 rubenwardy it shouldn't be our concern 18:16 rubenwardy we should pass it off to systemd 18:16 rubenwardy the most we should do is trim the logs to be at most a certain length when opening MT 18:18 sfan5 well you can't really tell windows users to set up systemd just so their folder doesn't keep growing in size 18:18 paramat i would prefer a large debug.txt is trimmed (oldest entries trimmed) than delete if over size 18:19 Krock wrote a suggestion. any comments on zipping the first, then dropping it after the next limit overflow? 18:19 sfan5 i was about to propose that 18:19 Krock nice 18:19 rubenwardy windows users shouldn't run servers :) 18:19 rubenwardy and don't need log rotation 18:19 Krock but they can 18:19 Krock and we should not prevent them from doing so if they want to 18:19 paramat that suggestion is fine with me 18:20 Krock great, so I guess that's settled for now 18:20 sfan5 don't even need to zip it, if debug.txt is over 50MB move it to debug.txt.1 (overwriting if already exists) and start with a new debug.txt 18:20 rubenwardy not having log rotation support doesn't prevent them from that 18:20 sfan5 ^ would've been my suggestion 18:20 Krock or that. even less io-consuming 18:21 Krock heading over to #6804. Is someone keen to adopt it? 18:22 ShadowBot https://github.com/minetest/minetest/issues/6804 -- Punchwear by sofar 18:22 Krock Actually I'm sceptical about the calculations in https://github.com/minetest/minetest/pull/6804/files#diff-b2e9183274c47c49c500e499eacd627eR223 18:23 paramat yes see my comments, something seems wrong 18:24 Krock nvm, it's correct. "f" is always < 1. that's the case if the punch happened only partially (not full interval) 18:25 Krock hence the wear proportionally increases to the punch/full_interval ratio 18:25 paramat well, i'm unsure whether there's a mistake or not 18:26 Krock question is whether to punish people who spam-click or just make it proportional to the time spent 18:26 Krock to be safe, this could be left in the current state and change it later 18:29 Krock so no volunteers. 18:30 Krock Why is #8611 on the meeting list? seems OK 18:30 ShadowBot https://github.com/minetest/minetest/issues/8611 -- Add support for set_formspec_prepend in main menu by rubenwardy 18:30 rubenwardy I didn't add that 18:31 paramat neither did i 18:31 rubenwardy Krock added it https://dev.minetest.net/index.php?title=Meetings&diff=4962&oldid=4959 18:31 Krock just seen that too. weird, nothing's to discuss there 18:33 Krock #8627 rebased it today. Are there any comments on the implementation? It works as expected 18:33 ShadowBot https://github.com/minetest/minetest/issues/8627 -- [Do not squash me!] Load CSM environment after the restrictions are known by SmallJoker 18:35 sfan5 seems fine to me 18:35 Krock rubenwardy: FYI. the insecure function has been removed and turned into a comment instead 18:40 Krock #8665 simple gameplay improvement for linux and windows. 18:40 ShadowBot https://github.com/minetest/minetest/issues/8665 -- Formspec: change the appeareance of the cursor on fields and co. by DS-Minetest 18:40 Krock Works fine, but was yet only tested on Linux. The code is very simple, thus I wonder why it did not get more attention yet 18:41 rubenwardy there's amerge conflict in master for some reason 18:41 Krock s/conflict/commit/ 18:42 Krock yes, according to sfan5 the "try again" on merging PRs creates merge commits even though they're disabled in this repo 18:42 rubenwardy yeah that 18:42 sfan5 tell github to fix their shitty software 18:42 rubenwardy oh I discovered that too 18:42 rubenwardy and then if you force push, it resets all the update timestamps of _all PRs_ 18:43 sfan5 if you merge two PRs quickly after eachother, the second will "fail" to merge but github will do it anyway with the default option of creating a merge commit even bypassing repo restrictions 18:43 Krock I had that "fail" too but it did not create a merge commit. must be a new bug 18:44 pauloue how is getmetatable insecure? 18:44 rubenwardy http://lua-users.org/wiki/SandBoxes 18:45 Krock you can get the table reference and alter that one, I guess 18:46 Krock comments on the cursor change PR? any approvals? 18:46 DS-minetest (note that the cursor pr has a little bug) 18:46 Krock that's not the kind of formspec changing PRs I had in mind but ok 18:46 Krock DS-minetest: which is that it does not work on mac 18:47 Krock that's not our problem. irrlicht must fix that 18:47 DS-minetest no, If you open a formspec and have your cursor on an element, it will stay normal until you move it away and back. 18:47 Krock so does the cursor if you keep it at the same place but scroll down in firefox 18:48 Krock weird. I've seen that for a while 18:48 Krock DS-minetest: I'd argue that's better than nothing 18:48 DS-minetest mhm 18:51 DS-minetest (btw, gtmetatable would probably be safe if you could just use it on tables) (maybe even the getmetatable("") example would not make problems for csm as it uses its own luastate and is not a sandbox that is maintained from lua) 18:56 deltasquared *reads backlog* 18:56 deltasquared wait 18:56 deltasquared player velocity!? 18:56 deltasquared FINALLY 18:56 deltasquared well that basically makes that hammer code I wrote obsolete 18:57 Krock it works but there's quite some delay from your punch to the other player's move 18:57 deltasquared I doubt that can be helped. 18:58 deltasquared seeing as it takes the *other* client to start applying the knockback presumably. 18:58 Krock punch packet -> to server -> to punched player -> player position packet -> to server -> to puncher 18:59 deltasquared not entirely sure but I don't think minecarp does any better from experience? 18:59 Krock they for sure do some prediction so it fells more natural 18:59 Krock *feels 19:00 Krock simple PR again #8656 review/approvals are welcome 19:00 ShadowBot https://github.com/minetest/minetest/issues/8656 -- Hide chat when console is open by ClobberXD 19:00 Krock no 19:01 deltasquared well, if the server could just *say* "this player is accelerated xyz, no arguments" the problem would be lessened (potentially only one round trip?), but... that's an established flaw at this point :P 19:01 Krock deltasquared: that was a suggestion, but problem here is that the server and client information can get out of sync 19:02 Krock hence the punched player moves backwards a bit after of what the puncher seees 19:02 Krock those movement packets by the punched player are also sent to the server (and fro there: distributed) 19:03 Krock so these two would conflict themselves by a server step and network lag in between 19:03 sfan5 maybe with server-side movement the server could predict what will happen and save two roundtrips 19:04 Krock yes, but Ben's PR is somewhat outdated by now and needs major rebasing 19:04 Krock if it's even worth picking that PR up again. starting fresh might be a faster option 19:05 paramat 8656 is merged already 19:05 Krock any comments on #8466? "won't add" or would someone like to see `add_list` to be an API function? 19:05 ShadowBot https://github.com/minetest/minetest/issues/8466 -- InvRef: Add convenience method add_list for bulk transfer of items by ClobberXD 19:05 Krock paramat: yes, I didn't check the PR beforehand 19:06 Krock for my part, the inventory functions are already feature-complete in this regard 19:07 Krock rubenwardy: how often do you need to add entire lists to another? 19:08 Krock for my mods this isn't used often, so it's probably the same for yours? 19:12 rubenwardy I don't mind too much 20:37 pauloue rubenwardy: why the global _ in dlg_create_world? 20:38 rubenwardy _ is a standard which means that the result is ignored 20:38 pauloue why is it global though? 20:39 pauloue I'm guessing it was intentional because it's ignored in .luacheckrc 21:01 sfan5 merging #8712 in 5 min 21:01 ShadowBot https://github.com/minetest/minetest/issues/8712 -- Fix some issues with minetest.clear_craft by pauloue 21:01 p_gimeno Krock: I'm using git-svn to mirror, using local scripts and a cron job 21:27 Krock will merge #8774 in ~5 minutes 21:27 ShadowBot https://github.com/minetest/minetest/issues/8774 -- Fix breakage of non-GLES2 setups by osjc 21:27 Krock sfan5: ping for the merge 21:27 sfan5 o yea 21:28 sfan5 done 21:28 Krock for comit number 8k 21:28 Krock next target: 9000 commits 21:29 sfan5 first we should finally make a release 21:30 Krock somewhere in between, yes. yet only 5.0.0 had >1000 commits 21:31 Krock *5.0.0 only had 21:32 Krock merging... 21:51 pauloue #8427 didn't get auto-closed 21:51 ShadowBot https://github.com/minetest/minetest/issues/8427 -- clear_craft() is reversing the recipes list 22:35 paramat a proposal and some questions for rubenwardy or whoever game#2438 22:36 ShadowBot https://github.com/minetest/minetest_game/issues/2438 -- Clean up use of 'gui_furnace_arrow_bg.png'