Time Nick Message 00:07 ShadowNinja Comments on #1129? 00:07 ShadowBot https://github.com/minetest/minetest/issues/1129 -- Add output and success return values to chatcommands by ShadowNinja 00:08 ShadowNinja I'd like to restore the order to (success, message). 00:08 ShadowNinja That fits in with other Lua functions much better. 00:26 LemonLake Hello from the actual Linux Mint! 00:58 ShadowNinja This switches the order back and fixes some style issues: http://sprunge.us/ODaV?diff 00:59 ShadowNinja And finally #943. 00:59 ShadowBot https://github.com/minetest/minetest/issues/943 -- Increase the amount of particles that breaking a node generates by ShadowNinja 01:01 ShadowNinja See #1149 from before too. I'd appreciate it if some of you looked at those and commented on them. BBL. 01:01 ShadowBot https://github.com/minetest/minetest/issues/1149 -- Add AppData file by davidgumberg 01:04 VanessaE 943 needs some tuning, ShadowNinja 01:04 VanessaE I like it, but it adds some FPS sagging 01:06 LemonLake RealBadAngel, I can't get over how cool this is, thanks a lot! 01:08 VanessaE (or they used to) 01:08 VanessaE LemonLake: after you've used it for a while, you'll begin to wonder why you waited for so long ;) 01:09 LemonLake ikr 01:09 LemonLake I figured out that it auto-starts anyway, so I don't have to be there to choose an OS 01:09 LemonLake I can mostly use this as my main OS, just gotta get a few apps on it :D 01:10 VanessaE you may also find, later, that you can move your Windows stuff to a VM inside of your Linux host, and ditch the real windows partition entirely 01:10 LemonLake Maybe 01:10 LemonLake But I might just leave it at dual-boot 01:15 LemonLake I can't get eclipse to run 01:15 LemonLake "The file "eclipse" has no known programs associated with it. If you trust the source of this file, and have sufficient permissions, you can mark it executable and launch it. Or, you can use the "Open With" dialogue to pick a program to associate it with." 01:15 LemonLake It thinks it's an ordinary file 01:15 LemonLake When I tell it to run as an executable, nothing happens 01:18 LemonLake Oh, I probably have to get java, lol 01:41 LemonLake Okay, now it's time to try minetest on the real mint 01:42 LemonLake Oh my god it works :DD 01:42 LemonLake And it runs waaay better than on Windows 13:32 spyjoshx I am interested in testing the minetest official android port 13:32 LemonLake The official one? There's an official one? 13:32 spyjoshx Thats what I heard 13:33 spyjoshx Re: Android port efforts 13:33 spyjoshx by stu » Mon May 12, 2014 4:31 pm 13:33 spyjoshx Sorry, I am no longer working on this as there is now an official android port in the pipeline. 13:33 spyjoshx Ask on irc freenode #minetest or #minetest-dev if you are interested in testing it. 13:33 LemonLake There's this: http://dev.minetest.net/Android 13:33 spyjoshx trying it 13:33 LemonLake There's also an illegal one, that's the only one I've actually heard of 13:33 spyjoshx whats it called\ 13:33 LemonLake The illegal one? 13:34 spyjoshx yes 13:34 LemonLake Buildcraft 13:34 spyjoshx just curios 13:34 spyjoshx oh 13:34 spyjoshx Is it illegal? 13:34 LemonLake That's what I just said. 13:34 spyjoshx gotta tell my dad that 13:34 LemonLake roflmfao 13:34 spyjoshx ? 13:35 spyjoshx nevermind thx 14:01 LemonLake Yo 14:45 ShadowNinja There is no official Android port yet, but sapier's port probably comes closest to being official. 14:47 ShadowNinja VanessaE: It was plenty low on my computer, but particles aren't very fast, they use client-side entities rather than Irrlight's particle system. 14:48 sfan5 ShadowNinja: -g 14:48 ShadowNinja (They also cause a segfault when reloading a singleplayer world after a mod crashes.) 14:49 ShadowNinja sfan5: s|-g|s/g/c/| 14:49 sfan5 oh right 14:49 sfan5 someone help me: http://dev.minetest.net/Android#Getting_stuff_to_work_without_libCrystax 15:33 Megaf_ Hi, I got a crash 15:33 Megaf_ http://paste.debian.net/102065/ 15:36 lemonlake_ was it just that once? 15:37 Megaf_ I'm not sure 15:38 Megaf_ I never stop my server, so when it stops itself I update it 17:09 sapier comments on #1200? I'd like to merge that one this week as it's one of the last pieces to android support 17:09 ShadowBot https://github.com/minetest/minetest/issues/1200 -- Add support for exiting formspecs by doubleclicking outside by sapier 17:11 proller again codestyle change mixed with new feature 17:12 sapier if you believe removing a useless temporary variable is codestyle yes 17:13 sapier but if you feel better I can split it to two commits ;-) 18:22 proller sapier, https://github.com/minetest/minetest/pull/1200/files#diff-65f34680878a6bd86f3a59ebc0c06c6dL2342 18:24 sapier wait a moment I accidentally remerged on fixing some comments ;-) 18:24 sapier what's wrong at that line anyway? 18:28 proller split sapierizing code and adding new features or changing code 18:28 sapier sorry I thought you had a valid issue 18:31 proller nobody can show you valid issues hidden in 1000 space changes 18:33 sapier proller almost everything you assume to be a "space change" is a real code change 18:33 LemonLake ugh that reformat doe 18:34 LemonLake horrible 18:34 proller ) { -> )\n{ - code change? 18:34 LemonLake ^ 18:34 proller sapier, bad news for you.. 18:34 sapier well if you don't understand the difference between "almost" and "all" you wont find bugs anyway 18:35 proller its very hard in your changes 18:35 ShadowNinja sapier: Seems good. 18:36 ShadowNinja sapier: Usual style issues though. 18:37 proller and usual hidden bugs and asserts 18:37 sapier there ar 19 lines due to coding style and 95 code changes proller you should really stop drinking if you see 1000 space changes 18:38 sapier ShadowNinja: where? 18:39 ShadowNinja sapier: Missing spaces after commas and )\n\t{ throughout. 18:40 sapier ok I'll recheck 18:40 sapier I still want a tool to use it's ridiculous how much time is wasted due to lack of automated fixing 18:42 ShadowNinja sapier: http://sprunge.us/ZLUe?cfg 18:42 sapier where to use? 18:45 ShadowNinja sapier: That's an astyle config file. 18:45 sapier if I use astyle uppon that file proller will jump hexagons due to code style changes ;-) 18:46 ShadowNinja sapier: Note that it isn't perfect, when it breaks lines it uses spaces to align the second line. 18:46 ShadowNinja sapier: Yes, I didn't say that you should do that, but it's available. 18:47 sapier I don't think that config is gonna work ... whats the tabwitdth good for anyway? 18:48 sapier ok I think I fixed all coding style issues for new code and some (but not all) for this file 19:27 eythen hi guys ... i have a question to redis migration ... can someone help me to find the problem which occurs by migrating from sqlite to redis db? 19:43 sapier guess sfan5 would be best to be asked 19:44 eythen well i made a issue for this 19:45 sapier reminding time by time helps (at least if it was me) ... but don't ask once an hour ;) 19:46 eythen well it is not that critical so ... i can wait til someone will takevoer ... its just playground but i was curious how redis would perform of an exisiting map 19:47 sapier if I do remember correct megaf did play around with redis too but I could mix something up there 19:47 sapier well theoretically redis should be slower for almost all real life scenarios 19:48 eythen thats the question ... redis is statet on wiki at minetest for better performance choice rather tah nsqlite3 19:49 sapier without benchmark results I'd not believe this to be true 19:50 eythen then minetest shouldnt state such argument on wiki if there is no real performance test made doh 19:51 eythen but i could imagine it would be perform a bit better than sqlite3 19:51 LemonLake ummm 19:51 sapier well there's lot minetest shouldn't do ;-) that's the way it is 19:51 eythen sqlite3 19:51 eythen sqlite3's bottle neck is the storage beneath the db file 19:52 eythen redis lives from ram 19:52 sapier I don't see a reason why a db adding additional network overhead is supposed to be faster 19:52 eythen ram performance most times better than any storage device except solid states 19:52 sapier redis has to save it's data to disk too 19:52 sapier if it's not saved it's worthless 19:52 eythen yes it does ... but not while using the informations ... redis makes period saves 19:53 sapier well you can tell sqlite do to so too but that's not how we use it because you risk loosing data 19:53 eythen true 19:53 eythen so i am wondering how you use redis then? ;-) 19:54 sapier most ppl don't use redis 19:54 eythen bcause of reliability 19:54 sapier it's supported like leveldb but actually official backend is sqlite ... everything else is on your own risk 19:55 eythen if the server crash, the redis db informations are gone ... and if you are lucky, then only the last 3 minutes 19:55 sapier that's responsibility of server owner 19:55 eythen depends how you setup the redis 19:55 LemonLake http://pastebin.com/7y1ijjRq My intention here is for the current element to be removed, and all the other to move upwards. 19:56 LemonLake However it doesn't quite work that way 19:56 sapier "all the others"? 19:56 sapier you know you shouldn't remove elements of other mods? 19:56 sfan5 eythen: I'm here now, do you need help? 19:57 eythen hi sfan5 i am stuck within the migration thing of redis 19:57 LemonLake Sorry, not sure what happened there 19:57 eythen it cores dump probably because of a too large sqlite db 19:57 sfan5 in that case compile a debug build and reproduce the problem while in gdb 19:58 eythen ah ok 19:58 sapier ethyn you should check for dhud[name].deaths[id] prior accessing it if it's possible to be nil 19:58 sapier sorry meant LemonLake 19:58 sfan5 eythen: you already did? that is good 19:58 sfan5 but well 19:58 sfan5 it should be using a debug build 19:58 LemonLake What if I use v instead? 19:59 LemonLake Well if it was nil, it'd complain. 19:59 LemonLake So it's not nil. 19:59 sfan5 and "bt full" output would be much more helpful 20:02 eythen bt full does not gives without debug compile any more infos 20:02 eythen the debug compile arg is like this? cmake . -DRUN_IN_PLACE=1 -DBUILD_CLIENT=0 -DUSE_REDIS=1 -DDISABLE_LUAJIT=1 -DENABLE_CURL=1 -DCMAKE_BUILD_TYPE=debug 20:02 sfan5 yes 20:04 eythen and then run in gdb again 20:07 eythen ok i will give you the bt full in 45 minutes ... *hmpf* 20:07 eythen takes that long to reach 54% 20:09 eythen is there any compile switch for speed up mograting? like -DMIGRATE_TYPE=Speed? :D 20:11 sfan5 eythen: currently not 20:11 eythen sfan5, JJ - have to wait 20:12 dsimon eythen, both redis and sqlite have various safety and performance settings related to how often they write to disk 20:12 eythen sfan5: i never got mysql backend with lua to work ... you remember the sort of code you made for us? 20:13 sfan5 yes 20:13 sfan5 it would be pretty easy to make a mysql backend in C++ 20:13 eythen didnt figured out how to pass the blocks to mysql 20:14 dsimon what's the motivation for using mysql here? 20:14 eythen well as long you can read and write c++ sure ... if not your lost ;-) 20:14 dsimon it seems like a bad choice 20:14 eythen dsimon: motiviation is to have all in one db and not stored in several files or database 20:15 dsimon eythen, all worlds? 20:15 eythen dont know how much perfomance mysql will bring with but an earlier build from someone with mysql was performaning much more better 20:16 eythen dont like to hang too much out of the window ... 20:16 dsimon if mysql was performing much better than sqlite that indicates a serious problem with the sqlite usage code :-| 20:16 sfan5 if you would use MySQL for minetest I would backup the minetest world seperately 20:16 eythen why sfan5 ? 20:16 eythen whats the bad choice with mysql? 20:17 eythen same with leveldb? 20:17 sfan5 if you need to restore a backup for one db (that is not the world) you'd need to wait for all world data to get reapplied 20:17 eythen ummm ... lets say you have 5 worlds 20:17 dsimon eythen, i strongly expect that if you benchmarked it, it would be slower 20:17 eythen would you startup each with different dbs? 20:17 dsimon mysql provides a lot of features that minetest doesn't need, and those features have performance costs 20:17 eythen i mean within mysql 20:18 eythen mysql is an example 20:18 * VanessaE peeks in 20:18 eythen by the way we use mysql for internal mods we use on our server 20:18 * LemonLake also peeks in 20:20 eythen one question dsimon: how do you erase a special area on sqlite with an cli command? 20:20 dsimon eythen, you're implying that you can do so in mysql with a query? 20:20 eythen or lets say, how do you search on sqlite3 20:20 dsimon if so, you should be able to run the same or a similar query in sqlite 20:20 dsimon using the sqlite console 20:20 dsimon or any of the languages with binding support 20:21 dsimon which is pretty much all the languages :-) 20:21 eythen i am not saying that you could do it with the current exisiting structure how minetest stores informations to the db 20:21 eythen its just a mess to be honest 20:21 eythen only in point of search and seek 20:21 dsimon eythen, i'm not sure what you're talking about specifically 20:22 eythen data structure 20:22 dsimon eythen, okay, let me clarify: how do you erase a special area on MySQL? 20:22 eythen dsimon: again - with the current situation - equals to sqlite not 20:23 dsimon wait, the table schemas are different in mysql and sqlite? 20:23 eythen ^-^ 20:23 eythen nope 20:23 eythen sqlite3 is most likely similiar to mysql 20:24 eythen i was pointing to something different 20:24 dsimon i'm confused 20:25 eythen lets say ... i you could store coords on database and your blocks sepertely - wouldnt it take only few seconds instead scrampling the whole db for the entry point to find it? 20:25 dsimon well, that depends on how the indices are set up 20:26 eythen yep 20:26 dsimon but both mysql and sqlite support basic indices 20:26 sapier problem about storing blocks is the mapsize 20:27 eythen dsimon: i didnt say sqlite3 cant do it ... i just sayed ... the current method to store those informations are a bit .... globby 20:27 eythen you have 2 tables right now ... right? 20:28 eythen and only in onne table is tored linewise everything 20:28 eythen *stored 20:28 dsimon linewise? 20:28 sapier each additional byte of overhead per block causes about 65GB of additional mapsize (4096^3) 20:28 dsimon sapier, does that not depend on how many blocks have actually been travelled to and generated? 20:29 sapier of course but you still have to keep the worst case scenario in mind 20:29 dsimon hm, yeah 20:29 dsimon compression then maybe? 20:30 sapier on the fly compression on load unload? wasn't this about improving performance? ;-) 20:30 dsimon well, some compression libraries are very fast 20:30 dsimon like, they disappear in comparison to the disk/network overhead of the actual load 20:31 dsimon google's snappy for example 20:31 sapier still it's hard to predict the resulting speed gain/penalty 20:31 VanessaE sapier: jumping into the middle here, a worst case like that makes no sense when the maximum size of a map is bigger than the biggest single hard drive you can even buy to begin with... :) 20:32 dsimon VanessaE, yeah, but that max size is actually not even that large 20:32 VanessaE don't consider map sizes from the standpoint of storage, no one *really* cares about that. consider it from the standpoint of CPU usage to read/write it, and network bandwidth 20:32 dsimon well, hm, actually, i suppose it is considering how deep it goes 20:32 sapier those are completely different issues VanessaE ;-) 20:32 dsimon is it actually possible ATM to dig down for 65 km? 20:33 VanessaE dsimon: yes, people have actually done that 20:33 VanessaE well for the ~31 km it takes to get from the usual surface. 20:33 dsimon oh, surface at the midline makes sense 20:33 eythen dsimon: someone who patched minetest for using mysql did this that way: http://pastebin.com/7shhqaQe 20:33 RealBadAngel btw, about databases, mods need a generic way to store their data in a database 20:35 eythen RealBadAngel: we had to choose mysql because of all others we are generating on mysql so the database was set with the cms we build around 20:35 eythen it yould also be mariaDB or whatever php is able to read 20:35 RealBadAngel php? what for? 20:35 dsimon php can read pretty much anything at this point, with the proper libraries 20:36 dsimon it's almost like it's a real language nowadays :-) 20:36 LemonLake I thought it was? 20:36 RealBadAngel im talkin now about mods data 20:37 dsimon LemonLake, it has had some very serious language design issues for a long time, only recently some of those are being resolved 20:37 RealBadAngel by now mods have to save own data to files 20:37 RealBadAngel and some users love to delete those files ;) 20:37 eythen hehe 20:37 RealBadAngel filling up bug reports then 20:37 eythen we dont touch mods file system architecture 20:37 eythen we store only user based information and other stuff to mysql 20:38 dsimon eythen, so i take it that you have a map server or something like that? 20:38 RealBadAngel so take care also of mods.db 20:38 dsimon php can actually read from sqlite as well 20:38 RealBadAngel with lua ways to read/write to it 20:38 eythen sfan5: is there any command for bt to export the bt full command to a piped file? 20:38 dsimon your CMS hopefully has support for PDO classes, that would make it easy 20:39 sfan5 eythen: no, just copy it from the console 20:39 eythen that informations we store isnt relevant to the map, except the location of each player 20:39 eythen urgh **** thats much 20:40 RealBadAngel https://github.com/minetest-technic/unified_inventory/issues/11 20:41 RealBadAngel ^^ thats why i need database for mods 20:41 eythen sfan5: http://pastebin.com/aMADSNSc 20:42 sfan5 hm 20:42 sfan5 why is block == NULL? 20:42 sfan5 this is not supposed to happen 20:42 sfan5 eythen: thanks 20:42 eythen dont know 20:42 eythen maybe we have a corrupt working db? ;) 20:43 RealBadAngel http://wiki.mudlet.org/w/Manual:Lua_Functions#Database_Functions 20:43 sfan5 I don't think so 20:43 RealBadAngel this is example of such thing in another project 20:44 eythen RealBadAngel: but your not going to build nw extra dbs with seperate files each on each do you? 20:45 RealBadAngel one database for all the mods should be enough 20:45 eythen hm ... how do you import than thos multiple mods? 20:46 sapier one database doesn't have only benefits ;-) 20:46 RealBadAngel i mean one single database accesible from lua per world 20:46 eythen but you dont need several @ sapier 20:46 RealBadAngel so any mod can use it to store its data 20:47 eythen RealBadAngel: i was thinking about to store all the mods settings into a db ... :D 20:47 RealBadAngel its absolutely necesary 20:47 eythen and then load them from the db with different options you made on any backend your choice 20:47 sapier on the other hand everything is lost on a single error ;-) 20:47 eythen nope 20:48 RealBadAngel i dont mean settings 20:48 eythen master/slave dbs if you paranoid :D 20:48 RealBadAngel i mean data created by mods 20:48 RealBadAngel by now we have to use own systems and store data in files 20:48 eythen you extract the data out of minetest ... currently lets say ... the bell mod places a bell and the current information is passed to a text file 20:49 RealBadAngel what? 20:49 eythen the server needs to load the textfile at the beginning of the upstart 20:49 RealBadAngel what are you talking about? 20:49 sfan5 eythen: please attach it to the github issue, I won't have time to work on it today 20:49 eythen i am trying to foloow you 20:50 RealBadAngel so read now: i need ability to acess database from lua and store there data : lua tables or blobs 20:51 RealBadAngel and i dont mean some settings and recording actions 20:51 eythen sfan5: i attached it 20:51 sfan5 thanks 20:51 eythen your welcome 20:52 eythen RealBadAngel: we also access databases from lua and store informations into it 20:52 eythen out of a mod 20:52 RealBadAngel where can i see it? 20:52 eythen because minetest does not ship this function 20:53 RealBadAngel thats why i do request such thing 20:53 RealBadAngel generic way provided by core 20:54 RealBadAngel so lua mod wont need some foreigner libs, own solutions or whatever 20:54 LemonLake I hope you guys aren't planning to access an online database within a mod 20:54 RealBadAngel shit 20:54 eythen i cant post code within here ... 20:55 RealBadAngel no i dont plan such things :P 20:55 RealBadAngel i need data storage for mods 20:55 LemonLake gooood 20:55 RealBadAngel period 20:55 LemonLake yeah, I've not been paying attention and tl;dr but I saw some snippets 20:56 RealBadAngel best if it was transparent for lua 20:56 RealBadAngel i mean it shall be accesible like a table 20:57 eythen http://pastebin.com/CB8xSDdr 20:57 eythen thats the class to connecto to the db 20:57 RealBadAngel so one could get an element from it like mods_db.unified_inventory.players.waypoint.pos 20:57 RealBadAngel to get the pos element 20:58 RealBadAngel function mysql_connect() 20:58 RealBadAngel return assert(mysql:connect(db_name, db_user, db_pass, db_host)) 20:58 RealBadAngel end 20:58 RealBadAngel this is transparent? 20:58 eythen what you mean 20:58 RealBadAngel you expect mods to setup database on a server? 20:59 RealBadAngel fuck, game is already using db 20:59 eythen no ... thats the way we use it .. the mod is not for public btw 20:59 RealBadAngel add new one and make lua ways to access it just 20:59 RealBadAngel is it that hard? 20:59 eythen that would be a benefit 21:00 eythen somtheing into a core to gain access to other dbs on the fly 21:01 eythen i guess it wouldnt be that hard .. but i think you cant add as much you like to have in minetest .. someone probably says ... enough ;-) 21:02 eythen dsimon: as you have seen ... we use mysql ... we have made a own rollback mod which uses mysql .... 21:03 eythen whats guess how fast those thing work? 21:03 RealBadAngel i dont care what some blind man will say about that feature 21:04 eythen who is the blind man? didnt see it so far :D 21:04 RealBadAngel generic datastorage for mods is needed 21:04 RealBadAngel if no1 is willing to code that i will do that 21:04 eythen you should connect with addi RealBadAngel ... he has the same intention you do 21:05 eythen he is our housekeeping modder :D 21:05 sapier you should write a api doc for review prior implementing everything 21:05 RealBadAngel i do have two mods that are generating lotsa data files in world folder 21:06 RealBadAngel im sick of problems with such solution 21:07 sapier just try to write an improvement usefull for everyone and not only suitable for your special usecase ;-) 21:07 eythen its all about use cases ... :D 21:08 * VanessaE pokes RealBadAngel regarding meta_set_nodedef and last night's discussion 21:08 RealBadAngel sapier, i mean generic way for any mod out there 21:08 eythen but i a gree with RealBadAngel ... some sort of integrated db api should be there ... 21:09 RealBadAngel so such db should be like a table for mods 21:09 RealBadAngel ie minetest.mods_db = {} 21:09 RealBadAngel and now any mod can write/read it 21:10 eythen ummm i really suggest RealBadAngel to open this as an topic on minetest forum ... cuz we probably would attend our ideas to yours any way 21:10 eythen so we could design the ideas on that plattform befoe handinmg it out to the devs? 21:11 RealBadAngel eythen topics or whatever wont help 21:11 RealBadAngel it was discussed already ages ago 21:12 RealBadAngel somebody just have to code it for christ sake 21:12 eythen maybe sfan5 will be able to help ... 21:12 eythen i have no knowledge of c++ coding at all 21:13 eythen if i would i would pachted mt anyway ;) 21:13 RealBadAngel i dont know nothing bout databases yet, but im pissed enough to stop working on gfx and code that shit 21:13 RealBadAngel i keep asking for that since months already 21:14 RealBadAngel each fucking project around has such thing but for us it is not obvious "feature" 21:14 eythen just do me a favor ... ill keep that topic on mind and forward this chat log to addi ... he will take his time to review it and bring some ideas as well to the topic and we place it to the dev team. there is no other way i guess 21:15 RealBadAngel its not a feature, its the thing missing for modders 21:15 eythen i know 21:15 eythen we are on the same boat 21:15 RealBadAngel so modders have to fiddle with lua IO and filesystem 21:16 eythen we suffer of such "builtin core feature" as well 21:16 RealBadAngel its sick 21:16 RealBadAngel each mod that needs to store something is creating own files 21:16 eythen we have almost 17mb for our mods on file .... 21:17 eythen correcntion 30 MB 21:18 eythen no wonder if people got bored before connecting to the server ;) 21:19 RealBadAngel technic_game itself is way bigger 21:19 RealBadAngel but we are not talkin bout mods size 21:25 eythen ok i am off for now ... thanks for the diskussion ... lets see what we can do together ... 21:53 CraigyDavi Suggestion: make texture packs and mods use the same information file. Texture packs use info.txt and mods use description.txt. Would make sense to maek them the same.