Time Nick Message 00:08 LipkeGu hi 11:17 Calinou http://minetest.net/old_releases 11:17 Calinou Ubuntu One is shut down, there is a dead link 11:17 Calinou (the installer ones) 13:00 Zeno` will apply https://gist.github.com/Zeno-/e7e28439f396ee7a70df in a few minutes (trivial fix) 15:22 SudoAptGetPlay Settings(path):remove(key) always returns true even if it's not deleting the key (I didn't forgot Settings(path):write() ) 15:23 Zeno` maybe it's a bug 15:24 T4im hm given that I found another method that had the same problem, maybe there's a general lua api problem with returning success? ;) 15:25 T4im #2015 that is 15:25 ShadowBot https://github.com/minetest/minetest/issues/2015 -- minetest.dig_node(test_position) does not fail when expected -- always returning true 15:26 SudoAptGetPlay In my case I try to remove "default:sword_diamond" from a items.conf and remove hapilly returns true but the key 's still there :/ 15:27 T4im should I try to reproduce? 15:28 T4im hm.. wait.. do you call write after remove? 15:29 T4im ah 15:29 T4im you wrote that already :D 15:29 T4im sorry 15:29 SudoAptGetPlay http://pastebin.com/Fbs7WDcQ 15:39 SudoAptGetPlay Should I create a github issue ? 15:39 T4im sure 15:41 T4im yea, I can reproduce this 15:42 SudoAptGetPlay #2264 15:42 ShadowBot https://github.com/minetest/minetest/issues/2264 -- Settings(path):remove(key) always returns true 15:43 T4im well actually, the problem seems to be one earlier 15:43 T4im the fact that it doesn't persist the removal 15:47 T4im https://github.com/t4im/mtt/blob/master/api_test/data_handling.lua#L37 <-- fails here 15:48 T4im the removal happens correctly in ram 15:48 T4im apparently 15:51 T4im I'd say :write() is the one that should not return true 15:51 SudoAptGetPlay But I don't test it 15:51 T4im I did ;) 15:53 SudoAptGetPlay oh I see now 15:53 T4im as you see it does not fail the assertion after the remove, both to_table and get correctly show the removal 15:54 SudoAptGetPlay indeed, write is guilty 15:55 T4im ha actually 15:55 T4im the last segment fails too 15:55 T4im so that's the third function returning true that shouldn't 15:56 SudoAptGetPlay So Issue shall become 'Settings(path):write() still saves removed keys' or something ? 15:56 T4im no you can actually keep it 15:56 T4im settings:remove("cobble") should have failed, but din' 15:56 T4im didn't 15:57 SudoAptGetPlay Only certain strings are a problem then ? 15:57 T4im no… :D 15:57 T4im cobble was a key that didn't exist 15:57 T4im and it STILL returned true 15:57 T4im when else but then should it fail? 15:58 SudoAptGetPlay Yeah pretty meesed up 15:58 SudoAptGetPlay *messed 16:01 T4im its more likely that it fails to actually write… so that might be 2 different problems there… the one of failing to communicate fail and then to fail to write to the file 16:33 Zeno` #2265 16:34 ShadowBot https://github.com/minetest/minetest/issues/2265 -- Increase MapBlock::actuallyUpdateDayNightDiff() performance by 2-8x by Zeno- 16:34 Zeno` night 16:34 Zeno` haha! that's almost a pun considering the function I was looking at 16:34 kilbith not 50000% :( i'm disapointed 16:34 Zeno` no, not 50000% :( 16:36 Zeno` oops! 16:36 * Zeno` amends 16:37 nrzkt Zeno can you commit #2214 and #2209 please ? 16:37 ShadowBot https://github.com/minetest/minetest/issues/2214 -- Performance improvement -> Craftdef.cpp: Improve loop and mathematics for CraftDefinitionShaped::check by nerzhul 16:37 ShadowBot https://github.com/minetest/minetest/issues/2209 -- Little performance improvement: use getPlayer(peer_id) instead of getPlayer(playername) by nerzhul 16:38 Zeno` I agree with both of those PRs 16:38 Zeno` I can't merge them (including my own) until the feature freeze is over, though :/ 16:39 nrzkt erf 16:39 Zeno` hmmmm, are you here? 16:40 Zeno` The problem is that I think Wayward_One and others are still having problems with simplesingleplayer mode 16:40 Zeno` despite yesterday's advances 16:40 hmmmm hello 16:40 Zeno` So I'm not really sure what's happening 16:40 Zeno` hi, hmmmm 16:40 hmmmm i'm here but not quite here 16:41 hmmmm cool performance increase 16:41 hmmmm I think the day/night system is whack and that's the first thing to change when refactoring lighting 16:41 hmmmm also expireDayNightDiff is useless iirc 16:50 Zeno` yeah I was looking at that function as well 16:50 kilbith guys, Gettext did not worked on the latest ubuntu 14.04 build 16:51 kilbith ^ Tesseract 16:51 Zeno` nrzkt, I want this performance thing with simple singleplayer fixed as much as you do so we can move forward 16:53 Zeno` but I'm out of ideas. The only thing I can think of (suspect) is the mesh update thread 16:53 Zeno` kahrl discussed the possibility of reducing that thread's priority the other day but I'm unsure if it makes much difference 16:54 Zeno` it makes a difference for me, but... 16:54 * Zeno` cries 17:01 nrzkt okay Zeno, 17:01 nrzkt i hope this release will be available soon 17:02 Zeno` can I have a tissue? 17:14 JakubVanek hello, I think client-server communication protocol should be reviewed in next release 17:15 JakubVanek for example there is s16 in sao damage but u8 in player damage 17:16 Krock there can be mobs with more than 255 hp 17:33 VanessaE mod profiler apparently needs to use wider datatypes for some fields :-) --> http://pastebin.ubuntu.com/10094757/ (lines 78 and 80) 17:37 VanessaE (and yeah, we're working on solving the cause of the ridiculous times that exposed this bug) 17:43 Krock VanessaE: total | | -4294829465 18:15 nrzkt JakuVanek 18:16 nrzkt JakubVanek, i'm working on protocol rework, on next release we need a protocol breakup for: fix some wrong packets 18:16 nrzkt if you tell the damage packet is wrong i think it will be fixed 19:25 est31 nrzkt: will the issue that when the client recieves packets in wrong order crashes be fixed? 19:26 Krock wrong order? 19:26 nrzkt no because i don't care about this network layer. It will be rewriten for future release (it's in progress), and use a proper TCP connection with proper sessions instead of this ugly layer 19:27 est31 good. 19:27 est31 so nothing speaks against the switch to tcp, great 19:28 Krock why TCP? 19:28 nrzkt in fact est31 i look at packets which need to be sure, and it's 95% of MT packets, then TCP is a good option 19:28 Krock I thought UDP never was a problem for the engine 19:28 nrzkt please look at the code 19:28 nrzkt and show me 19:29 est31 95 % are sent reliable? thought less 19:29 nrzkt ofc 19:29 T4im most are probably chat messages/notifications, inventory operations 19:30 est31 SAO moves 19:30 nrzkt in fact only movement doesn't need reliability 19:30 est31 every player knows every time the position of every other player 19:30 nrzkt i think we can use both TCP and UDP 19:30 T4im yea, but you can correct positiopn, est31 19:30 est31 lol 19:30 nrzkt no but at this time movement is setn as unreliable 19:30 T4im if you loose one movement/positional change, you update it the next tick 19:30 T4im lose* 19:31 est31 yea 19:32 T4im nrzkt: how about map updates? do they need to be reliable? 19:33 nrzkt i don't study this event in fact, i'm working on pure network layer at this time 19:33 T4im ah 19:33 nrzkt i rewrite packet read/write and it's finished, it tell me which packet is reliable or not in a table (from source) 19:36 nrzkt and tcp permit to secure protocol, like authentication and in the future add encryption 19:37 Calinou we could try making less stuff reliable 19:37 est31 for udp, there is dtls 19:37 Calinou kilbith, I think the translation issue is fixed. it is in latest Git 19:37 nrzkt network isn't a problem in MT, it's the engine itself. 19:38 est31 ? 19:38 nrzkt using UDP doesn't improve performance. nobody uses RNIS or satellite connections. 19:38 T4im you can get udp reliable too, its just a lot more work ;) and we don't have the latency requirements of a fps, so why not tcp? ;) 19:38 nrzkt and we don't need the latency gain offered by UDP 19:38 T4im ^^ ;) 19:38 nrzkt UDP reliable = TCP over UDP. 19:38 est31 yea 19:38 nrzkt why rewrite TCP over udp, kernel does it properly. 19:39 est31 with udp you always have to watch out, dont rewrite tcp 19:39 jin_xi nrzkt: are you nerzhul on github? 19:39 nrzkt ofc 19:45 nrzkt why ? :) 19:46 jin_xi anyways, rebased pull #1700, wip particles if anyone wants to help 19:46 ShadowBot https://github.com/minetest/minetest/issues/1700 -- wip irrlicht particles by obneq 19:46 nrzkt jin_xi i tested it 19:46 nrzkt performance are ugly on intel GPU 19:46 nrzkt and there are bugs in collision calculs, some particles fly with high speed 19:47 jin_xi yes, there are many problems and i am an inadequate programmer ;) 19:48 nrzkt i rebased RBA water shader and it works perfect and fix water issues 19:49 VanessaE did you fix the flowing glitches? 19:49 nrzkt flowing glitches ? 19:49 VanessaE it doesn't stretch the texture properly on vertical flow, and the movement of the water does not follow how the water moves without the shader. 19:49 VanessaE or it didn't when I tried it last. 19:53 VanessaE (in fact as I recall, "flowing" water didn't appear to actually flow at all :) ) 21:57 Calinou I don't understand what nrzkt wants to do; first, he rewrote the Minetest networking system, and now he wants to switch to TCP? 21:57 Calinou nrzkt> why rewrite TCP over udp, kernel does it properly. 21:57 Calinou because it lags more… 21:57 Calinou try writing a FPS that uses TCP (hint: this was Cube in its early days) 21:57 Calinou and was the reason eihrul wrote enet 22:36 Wuzzy Hi! Is Minetest using Lua 5.1, Lua 5.2 or whatever Lua the system ships with? 22:37 Calinou 5.1, Wuzzy 22:37 Calinou update to 5.2 is not planned AFAIK (there is an issue on GitHub) 22:37 Wuzzy ok thanks 22:37 Calinou most systems supply 5.1 anyway, so do Windows builds 22:37 Calinou 5.3 was recently released, too