Time Nick Message 02:10 benrob0329 Ok, now ive heard it twice 02:10 benrob0329 the default hand is getting droppe somehow 02:11 benrob0329 *dropped 06:46 * benrob0329 mumbles about pulling in changes fron Antarctica again 07:04 nerzhul merging #5603 thanks sofar 07:04 ShadowBot https://github.com/minetest/minetest/issues/5603 -- Plug two minor Leaks by sofar 07:05 nerzhul sofar, which static analysis tool do you use ? 07:08 sofar nerzhul: I've put it in scan.coverity.com 07:09 sofar I can invite people and grant them access to it - just pm me your email address 07:09 nerzhul i started to do it before for didn't get time to finish, it's nice, can we plug it in CI ? 07:09 sofar no 07:09 sofar I intend to just upload new master snapshots once in a while 07:10 nerzhul master snapshot ? coverity results you mean ? 07:10 sofar coverity doesn't work like that 07:10 sofar it takes like 15 minutes to do a scan 07:10 nerzhul yes i know :) 07:11 nerzhul i didn't tried cppcheck on minetest before, only on my own projects, trying it now to see what happens :) 07:14 nerzhul sfan5, sofar nore on #5475 if i add sqlite3 by default on newly created worlds is this okay for you ? 07:14 ShadowBot https://github.com/minetest/minetest/issues/5475 -- Player data to Database by nerzhul 07:29 sofar so, I'm ok with that but I would like sfan5 to agree to it as well 07:32 nerzhul sfan5, i hope you can review before feature freeze, ideal can be in 1 week maximum permitting to merge it as it now fits all requirements 07:37 nerzhul pushing a trivial fix pointed by sfan5 07:40 paramat can you link the diff or commit? 07:40 paramat and describe it? 07:41 nerzhul oh it's just a unused variable on android added yesterday 07:43 paramat ok can see it now 07:44 nerzhul celeron55, typo backend => player_backend 07:44 nerzhul thanks for notice it :) 07:45 celeron55 do you not test this stuff at all? 07:46 celeron55 that was the most obvious case i could come up with 07:46 nerzhul i tested all, but i forget to fix the typo in the help :p 07:46 celeron55 does it mean mean i need to modify world.mt to migrate players from files to sqlite3? 07:47 nerzhul yesterday i migrated my players from files => sqlite => pgsql => files => pgsql => sqlite => files 07:47 nerzhul yes, like when you want to migrate your map 07:47 celeron55 that seems weird, why doesn't it put player_backend = files in there automatically? 07:47 nerzhul because it will not be the case in existing worlds 07:47 celeron55 but i made a new world 07:48 nerzhul oh, the default is not set, right, when not found it assumes it's files 07:48 celeron55 i don't see why this shouldn't work without editing files 07:48 sofar 3 negatives in 1 sentence 07:48 nerzhul it works without editing, you only need to edit when you want to migrate 07:50 celeron55 well whatever, my point is that at some point it should put player_backend = files into world.mt so that i don't need to 07:50 nerzhul i'm doing it 07:51 nerzhul it's just not saved, adding it :) 07:51 sofar +1 for writing that out, yes 07:52 nerzhul no problem, i'm doing it, it should be done in case we choose sqlite3 by default because empty is files :) 07:53 nerzhul it's fixed now 07:54 * benrob0329 mumbles about Antarctica and the stk devs one more time before going to bed 07:58 celeron55 ok, my test case works fine now 07:58 nerzhul nice thanks for pointing that :) 07:59 celeron55 i don't particularly care whether the default is sqlite or files 08:00 celeron55 maybe there's no reason not to use sqlite3? 08:01 nerzhul sqlite3 is good, less files, less disk agressions, it's reliable, and for server owners use pgsql and plug you website on the db :D 08:01 celeron55 first-time server owners would be thankful i guess 08:05 celeron55 it would be really nice for the migration tool to ask "player_backend not defined. Migrate from files backend? y/n" in the case of old worlds 08:05 celeron55 :P 08:06 celeron55 maybe not; then we need a --noninteractive option and whatever 08:06 celeron55 it's fine now 08:15 celeron55 nerzhul: maybe it should write the database only when it sets player_backend? 08:16 celeron55 i mean, write world.mt 08:16 celeron55 otherwise 99.99% of the time it's writing it unnecessarily 08:19 nerzhul celeron55, for old world i wrote auto migration but it was removed, we will re-add it in a future release 08:19 celeron55 i'm referring to just the addition in the last commit :P 08:20 nerzhul yes, i can add you change if needed, i just copied model from ServerMap 08:21 celeron55 like... just move the updateConfigFile() into the !conf.exists("player_backend") block 08:25 nerzhul yes 08:25 nerzhul done 08:25 celeron55 i accidentally tested what happens when you open a player_backend = sqlite3 world in the current MT version: it just creates a new players/singleplayer file, which will be ignored by the PR version :P 08:26 nerzhul yes, we are not dangerous 08:26 nerzhul also player files are moved to players.bak when migrated 08:27 nerzhul i go buy some food, see you, thanks for your test, don't hesitate to comment the pr :) 08:47 paramat will merge #5519 in 10+mins 08:47 ShadowBot https://github.com/minetest/minetest/issues/5519 -- Sneak: Add option for old move code by paramat 09:11 paramat merging in a moment 09:19 paramat merged 10:10 celeron55 i'm now attempting to strip out FarMap from the FarMap + map transfer + render improvement PR 10:11 celeron55 not looking like a very easy rebase 10:20 celeron55 after that i need to work on the bugs... it works very nicely, until it doesn't 8) 10:25 nerzhul take your time celeron55 i think it's better to do a such change after release :) our changeset is pretty big we need features for next release 10:30 celeron55 i'm not aiming for release 10:31 nerzhul nice :) 10:31 nerzhul any other comments on player to db ? 10:35 celeron55 i won't do a code review but the concept is fine and the important parts are clearly working :P 10:38 celeron55 and the sqlite3 database schema seems fine 10:38 nerzhul yeah it's not a perfect db scheme but it's sufficient for MT needs 11:46 celeron55 merging #5595 after 5 minutes 11:46 ShadowBot https://github.com/minetest/minetest/issues/5595 -- Cache MapBlock data in the mesh generator thread for less stutter by celeron55 11:52 Krock afterwards, there's this compiling fix for MSVC: https://pastebin.com/raw/kg5tCAB2 11:53 Krock I'm not sure if it wouldn't better to add an overload for it 11:53 Krock inline int floor(int &v) { return v; } 11:53 Krock but that's quite nonsense 11:55 celeron55 wait what, lol 11:55 Krock https://github.com/minetest/minetest/commit/34d32ce55ae4f3f29d7b645075dc8efacb2c96d2#commitcomment-21783097 11:56 Krock but there are further errors, got to fix them 11:59 Krock funny enough that all errors were caused by one commit 12:01 celeron55 it was an old changeset taken from a fork i think 12:02 celeron55 https://github.com/minetest/minetest/pull/4664 12:10 Krock Pushing https://pastebin.com/raw/gUaUTJGu in 10 minutes 12:10 Krock (unless there are other opinions, for sure) 12:30 nerzhul strange that MSVC doesn't follow standard functions... or not 12:37 paramat i'll make the .conf extra PR, discussion: http://irc.minetest.net/minetest-dev/2017-04-16#i_4877699 . are people ok with the name 'minetest.conf.example_extra'? 12:39 nerzhul - is better than _ for a file extension, i don't know how windows will work if we have a such extension 12:40 nerzhul Krock, can you test #5475 on windows using sqlite ? 12:40 paramat ah that must be why c55 suggested '.' 12:40 ShadowBot https://github.com/minetest/minetest/issues/5475 -- Player data to Database by nerzhul 12:41 paramat i'll go with '.' 12:41 nerzhul paramat, i think yes 12:41 paramat ok thanks 12:41 nerzhul why not extra.example isntead of example.extra ? extension is example not extra on the main file :) 12:47 celeron55 i think .example.extra looks nicest in a file listing 12:47 celeron55 next to .example 12:48 celeron55 it's clear that it's extra to the example 12:48 nerzhul [ex]tra [ex]ample :) 12:48 nerzhul as you prefer 12:48 celeron55 not an example of an extra config or something 12:49 paramat yes i agree with c55 12:55 Krock nerzhul, sure. *grabs diff* 12:58 Krock I agree with c55 that it should be .example.extra. The filename gets quite long but it's consistent with the file naming 13:00 Krock nerzhul, database-sqlite3.cpp(569): error C2589: '(': Invalid Token on the right side of '::' 13:00 Krock same in L570 13:00 Krock apparently it doesn't know std::min, just go with MYMIN ^^ 13:25 paramat #5606 celeron55 13:25 ShadowBot https://github.com/minetest/minetest/issues/5606 -- Conf.example: Move some lines to minetest.conf.example.extra by paramat 13:26 celeron55 you should also run the generation script 13:26 celeron55 it modifies the header in minetest.conf.example a bit from what it is now 13:31 Krock nerzhul, successfully migrated 1 player.. seems to work :) 13:37 paramat c55 ok 13:50 Krock wow paramat. Well done. Your pull for adding the old sneak back in doesn't result in any conflict with my sneak changes :3 13:53 celeron55 what the f, who removed getTextureSource() from IGameDef? 13:53 nerzhul celeron55, maybe me because it's a pure client feature not a server feature 13:54 celeron55 IGameDef isn't a server-only thing 13:55 Krock removed in https://github.com/minetest/minetest/commit/8e7449e#diff-01baf55dd9338d866a8b4186cda8fbc0 13:55 nerzhul yes i know, it's why texture source is available on client not on server 13:55 celeron55 well then there needs to be a client version of IGameDef for the client 13:56 celeron55 or do we consider Client to be that 13:56 celeron55 maybe it is then 13:57 nerzhul yes it's owned by client, no need for an interface, it's the only owner 14:17 celeron55 what should the behavior be when the user increases the view range to such a value that client_mapblock_limit is too small for the amount of blocks needed for the range? 14:18 celeron55 i'm thinking probably just to consider client_mapblock_limit to be a minimum and not use it as a limit when more is needed 14:19 Krock I'd say the smallest value counts. 14:20 celeron55 the current default of 5000 is enough for maybe a range of 180 14:21 celeron55 for more than that, i think it is silly that you couldn't just use + to increase the view range but instead have to go to advanced settings to raise this weird value called client_mapblock_limit, and you don't even know where to raise it 14:22 celeron55 currently the system just acts completely silly: the server sends more than client_mapblock_limit blocks and the client keeps deleting them after a while, loading itself, the network and the server completely unnecessarily 14:23 celeron55 the system i'm taking from farmap behaves well, but it reveals the sillyness of client_mapblock_limit 14:23 paramat hm good points 14:24 paramat but we'll still need a limit, but set to the volume of the currently visible sphere? 14:25 paramat .. with 5000 as the minimum limit 14:25 celeron55 that's what i propose (well, the farmap code uses a cube, but whatever) 14:25 paramat seems good 14:29 paramat i mentioned a sphere because now blocks are sent according to a spherical radius instead of a cubic radius 14:32 celeron55 i don't even remember which way this block sending works 14:32 celeron55 there's almost nothing left of whatever was changed during this time 14:36 celeron55 i think i'll make sure they're limited to a sphere if possible 14:41 celeron55 yeah it's already done 14:41 celeron55 good 14:53 celeron55 the client-side unload timeout doesn't really make any sense with this either 14:54 celeron55 actually it hasn't made any sense since client_mapblock_limit really 14:54 celeron55 again the only thing it does is cause unnecessary network load as the server wants the client to have the blocks in the area around the player, and the client hits the unload timeout when it just isn't looking at a mapblock for a while 14:55 celeron55 absolutely silly 14:57 celeron55 i propose getting rid of the client_unload_unused_data_timeout as it doesn't make sense 14:57 celeron55 actually i'll just do it as part of this, why am i even asking 14:59 celeron55 or hmm 15:00 paramat #5606 updated, i regenerated conf.example, not much changed hehe just spaces in 1 line 15:00 ShadowBot https://github.com/minetest/minetest/issues/5606 -- Conf.example: Move some lines to minetest.conf.example.extra by paramat 16:48 paramat celeron55 nerzhul Krock please can anyone approve #5606 ? 16:48 ShadowBot https://github.com/minetest/minetest/issues/5606 -- Conf.example: Move some lines to minetest.conf.example.extra by paramat 16:50 nerzhul Krock, i will fix the min problem, i hope in C++11 we can rmove this crap 16:51 Krock nerzhul, when we're going to C++11 I'll most likely not care about MSVC anymore 16:51 nerzhul oh ? why ? 16:51 Krock that's quite my milestone for switching from windows xp to linux 16:51 nerzhul your MSVC build is only for Windows XP ow ! 16:52 celeron55 i'm pretty sure microsoft doesn't support C++11 on windows xp 8) 16:52 Krock no, it works on all windows versions, nerzhul (WXP and upwards) 16:52 celeron55 it does? that seems crazy 16:52 Krock celeron55, Visual C++ 2010 already contains pretty much of it 16:52 nerzhul yes, but i would mean, you did it only for XP :p 16:53 Krock nerzhul, no. You misunderstood me. It doesn't matter what MSVC version is used. these errors will also occur on a W7 computer with 2012 version 16:53 Krock since VC++ in general doesn't seem to contain certain functions 16:59 Krock I think if I hadn't fixed that stuff, addi would come with a new PR after a while 17:06 sfan5 paramat: what is example.extra for? 17:07 nerzhul strange 17:08 Krock sfan5, to document stuff that doesn't fit into the settingtypes file 17:08 sfan5 huh 17:08 Krock apparently some noise explanations 17:43 paramat sfan5 there are links to the IRC discussions that explain 20:38 Fixer newest engine https://i.imgur.com/sZpBZS6.png 20:41 Fixer nothing new, but what is interesting is drawtime spike not associated with mesh updates? o_0 20:43 Fixer i rarely see such interesting spikes 21:10 sofar kewl 21:10 sofar that's called progress 21:10 ShadowNinja Meeting notes from the last meeting are now posted on the Doxygen site. 22:01 Hijiri when is the code freeze again? 22:01 Hijiri also what's a better name for tool caps override, set_tool_capabilities_override or set_tool_capabilties? 22:01 Hijiri the first is more accurate but it's really long 22:01 Hijiri or maybe set_toolcaps_override 22:08 sofar override_tool_capabilites ? 22:11 Hijiri that sounds good 22:11 Hijiri only four characters shorter but it feels much shorter than that because it's less words 22:22 paramat freeze 21st May 22:25 Hijiri okay, thanks 23:38 Fixer wth is this jitter? https://i.imgur.com/u8ea9R2.png 23:39 Fixer fuuuck, nevermind, two clients running at the same time