Time Nick Message 00:39 kaeza Woah, there's game translations now 00:40 kaeza Spanish seems to be at 100% 02:19 hecks sfan5 what happened to the libs on your server? 04:03 Wuzzy kaeza: no, this is an unofficial fork. official MTG is untranslatable, of course. 04:05 kaeza Oh, I see... ☹️ 04:07 kaeza I should go now. Good night all. 09:15 sfan5 hecks: ? 09:16 hecks The URLs that wget uses in the build script, they're 404 09:17 sfan5 "GET /irrlicht--win64.zip HTTP/1.1" 09:18 sfan5 i'm not sure what's broken with your system, that should say irrlicht-1.8.4-win64.zip 09:19 hecks ohhh hold on 09:19 hecks my bad 09:19 hecks I missed the $irrlicht_version bit 11:00 nerzhul Wuzzy, rubenwardy: sorry for the break. Please note it was done in two commits. 11:00 nerzhul https://github.com/minetest/minetest/commit/c617526ecae0d147899036e95584489c63b62e56?diff=unified 11:00 nerzhul this is the pull from weblate 11:00 nerzhul https://github.com/minetest/minetest/commit/c2dabcff6ddd50520669f3a663b782ec28fc0094 this is the updatepo.sh 11:01 nerzhul our translation model is quite anoying i think, i know some are attached to weblate but it's not the best integrated thing for translations :( 11:01 hecks iconv.h:24:29: error: ISO C++ forbids declaration of '__declspec' with no type [-fpermissive] 11:01 nerzhul after the two commits, weblate pulled master on c2dabcff6 11:06 hecks retrying with master just to see if I'm not going crazy 11:07 hecks oh, now it errors on translations 11:07 hecks I suppose you guys are talking about this exact problem 11:09 hecks oh well I have my branch, that should compile 11:30 hecks nope, /src/util/string.cpp just won't compile 11:31 nerzhul i should take time to look at this 11:32 nerzhul will do it asap 11:32 hecks the translation related errors are something recent; the string.cpp thing must be something on my end 11:34 nerzhul currently compilation works 11:35 nerzhul maybe i should see on gitlab.com 11:35 nerzhul https://gitlab.com/minetest/minetest/pipelines/44931091 11:35 nerzhul only windows failed :o 11:35 hecks ...yup, windows 11:35 nerzhul i see why i didn't saw the problem 11:36 hecks yup, this is the error I was getting trying to build master 11:36 nerzhul it's a weblate failure 11:36 nerzhul at this line it's totally wrong 11:39 hecks now why does string/iconv crap out the build for me... I'm trying to compile a months old snapshot and it's still hosed 11:43 hecks if the bundled build sh doesn't work I'm officially giving up 11:45 sfan5 travis runs the buildbot for every commit and it works fine 11:47 sfan5 hecks: which mingw-gcc version are you using? 11:47 hecks c compiler ident is gnu 8.2.1 11:48 hecks but wait, now it's 8.2.0 11:48 sfan5 i have 7.2.1 here and travis uses 7.1.1, both of which work 11:49 hecks I don't get a choice, such is the life on arch 11:50 hecks well alright, final attempt 11:50 sfan5 try with -fno-permissive or whatever makes gcc stop doing that thing 11:53 hecks and where do I provide this magical argument in this cmake business 11:55 sfan5 -DCMAKE_C_FLAGS="" or -DCMAKE_CXX_FLAGS="" 11:58 hecks well it fixed itself without that 11:58 hecks every time i compile minetest https://a.uguu.se/tLZsxM7RTjKM_meseblock.png 12:07 nerzhul i pushed a fix 12:22 hecks figured out my build issue; toolchain path was incorrectly expanded and I think it was trying to make a linux build with mingw libs 13:32 nerzhul there are many files with this problem, i will push the all fix 13:36 nerzhul pushed 14:42 nerzhul translation issue fixed :) https://gitlab.com/minetest/minetest/pipelines/45093882 16:20 p_gimeno good job nerz :) 16:37 nerzhul not exactly because it broke but at least we are up to date and okay 18:30 paramat all 3 CDB-related blocker PRs just need 1 extra approval now 18:57 Wuzzy Hi. I just wrote a wiki article about the Minetest version. please roast it: 19:07 sofar haha 19:07 sofar actually reads just fine, I'd say 19:24 p_gimeno in "The full format" I'd note that the hash is not necessarily in increasing order, as it seems to be a natural expectation from a version number 19:26 Wuzzy good point 19:26 Wuzzy oh 19:26 Wuzzy btw 19:26 Wuzzy https://hosted.weblate.org/projects/minetest/minetest/#alerts 19:26 Wuzzy ↑ this might explain the weblate sadness lately... 19:38 p_gimeno where's that sed script located? 19:42 p_gimeno there's no sed anywhere in the repo 20:05 rubenwardy all 3 CDB PRs are also pretty trivial 21:31 sofar I'm looking to add an API method to make teleportation more reliable 21:31 sofar I need to be able to send some mapblock data to a remote player, essentially 21:32 sofar two problems, I'd like to be able to do it per player, so e.g. as a player method (send a player these specific mapblocks) 21:32 sofar from lua, since it's a special case 21:33 sofar I can add a player method, but marking mapblocks as dirty is in the client class? No idea how to get a backref to it 21:33 sofar on top of that, marking a block dirty won't do anything if the server doesn't send it because it's too far 21:34 sofar suggestions welcome, I've looked at this before and I think this would really help both ITB and CTF and similar games 21:47 Wuzzy sofar: wait, what? what u mean, "more reliable"? 21:47 Wuzzy are you saying that setpos() is broken? 21:48 paramat do you mean sending mapblock data of the destination to the client that is about to teleport there? do you mean the client would load in the destination before teleporting? 21:48 sofar yes 21:48 Wuzzy aaaaahhh 21:49 Wuzzy so your problem is... player teleports, but area is completely not loaded yet? 21:49 sofar as in, I send a mapblock, wait a small bit of time, then teleport 21:49 sofar yep, so player falls and gets stuck in the ground 21:49 Wuzzy ??? 21:49 Wuzzy why should the player fall? isn't ignore walkable? 21:49 sofar or just falls in void 21:50 sofar could even be that the area has significantly changed since the last time and you know the client cache of the mapblock is outdated 21:50 sofar e.g. map reloads in CTF 21:50 Wuzzy sounds tricky. you need to be able to tell the future. :D 21:50 Wuzzy because how do you know when the player *wants* to teleport? 21:51 Wuzzy esp. when its an insta-teleport (gameplay-wise) 21:51 sofar in both cases for itb and ctf you'd know ahead 21:51 Wuzzy let alone /teleport... 21:51 sofar this isn't for /teleport 21:51 Wuzzy fixed locations? 21:51 sofar never fixed 21:51 Wuzzy :( 21:51 sofar if it was fixed I didn't need this 21:51 Wuzzy how are you going to preload if you dont know the location? 21:52 sofar I know the location 21:52 sofar I also know the client doesn't 21:52 sofar those are the conditions 21:52 Wuzzy thats what i meant with "fixed location". the destination location is fixed to a known position 21:53 Wuzzy heh. funny. i had similar thoughts a long time ago when I messed with riventest... but didnt think it thru 21:53 sofar well sure 21:53 Wuzzy (riventest has teleporters in form of books) 21:53 Wuzzy so i agree the problem is real 21:53 sofar in ITB case, player steps on a teleport, gets teleported to newly created box, but the client just sees nothing and falls 21:54 paramat player won't fall in unloaded or ungenerated areas, but idea is still valid for other reasons 21:54 Wuzzy are you sure you need more API for this? what about anchors? 21:54 Wuzzy e.g. in on_construct of teleporter node (or whatever), force-load destination area? 21:54 sofar ITB can't use anchors, anchors are server side 21:54 Wuzzy then i cant follow you :( 21:55 Wuzzy i thought the server knows where to teleport the players... in ITB 21:55 sofar paramat: in the case of ITB, the player falls all the time, irregardless 21:55 sofar there's no such thing as a client-side anchor 21:55 Wuzzy sofar: thats probably because the destination is not ignore anymore :) 21:55 Wuzzy i never claimed theres a client-side anchor. 21:56 sofar the server keeps the mapblock loaded so nodetimers tick in it, but that's all server side 21:56 sofar teleport to an anchor location and the client still has to load it somehow 21:56 sofar the server doesn't send map updates for anchor blocks to clients 21:56 Wuzzy aaaaah 21:56 Wuzzy i see it now!! 21:57 Wuzzy the server knows the destination but the client doesn't (yet). thats the problem? 21:57 sofar even so, I would have anchor blocks all over since ITB can place boxes in any random location 21:57 Wuzzy yes :(((( 21:57 sofar lol, yes, that's the problem 21:58 paramat ah ok 21:58 Wuzzy ITB seems to push Minetest really REALLY hard to the edge tbh. i mean, you basically added a full-blown db in it 21:59 sofar we could have done mod_storage but it would be fairly large 22:00 sofar going straight to sqlite really helped 22:00 Wuzzy do you think it was a limitation of Minetest? 22:00 Wuzzy or is this SQLite unique enough to ITB that u dont wanna see it in other games? 22:00 sofar no, it was purely a choice for performance 22:01 Wuzzy i still have not managed to make ITB run on my machine :( aaanyway 22:01 sofar having sqlite available made a lot of things easier 22:01 Wuzzy the first question we need to answer maybe is how the client should be educated about "shiny new mapblocks" ... and when... 22:01 sofar in the case of ITB it's just when a player goes into a new box 22:01 Wuzzy would the client get these mapblocks automagically? via lua callback? or how? 22:02 sofar player steps on a series portal, chooses a box, the box gets built by vmanip, mapblocks get sent, then player pos updated to box start spot 22:02 Wuzzy btw. how many players can ITB (theoretically) support? 22:03 sofar max_players is 25, we've hit 15 but it didn't really bother the server. I don't even know what the limiting factor would be 22:03 sofar afk a sec, brb 22:18 sofar since ITB doesn't really promote multiplayer, mostly it's a come and go of players at random times 22:19 sofar so if you see 5 players online at the same time, it's a lot 22:19 sofar I have been toying with the idea of making more games that would need or benefit from some sort of forced mapblock sending 22:21 sofar Wuzzy: because of the way that ITB separates players in space when they play or build, there is no real scaling factor since one player causes only updates to that one player 22:21 sofar players are always beyond the map_send_distance or whatever the thing is 22:22 sofar so the load is lineair per player, and not exponential like normal maps where players can all be in the same mapchunk making changes 22:25 sofar Wuzzy: I recently made sure itb runs on 5.0, but I had little setup issues - if you want maybe at some point I can walk you through it 22:25 Wuzzy cool 22:25 sofar also, maybe at some point I'll publish more data from the database 22:25 Wuzzy oh, i HAVE TO use 5.0.0-dev for it to work? 22:26 sofar no, 0.4.17 is what it's on right now 22:26 sofar both should work. I think the only issue is the worldmods that are required 22:26 sofar perhaps I should update docs 22:26 sofar afk