Time Nick Message 16:30 BlockMen any comments on #1114? 16:30 ShadowBot BlockMen: Error: Delemiter not found. 16:32 VanessaE #1114 16:32 ShadowBot VanessaE: Error: Delemiter not found. 16:32 VanessaE meh. 16:34 VanessaE some people have been asking for that mode, seems nice 16:34 VanessaE might be a good idea to make the head able to pitch/yaw properly first though 16:34 VanessaE it'll look strange if you don't, I think 16:42 ShadowNinja Hmmm, has GitHub changes their issue format?... 16:42 ShadowNinja +Ttitle 16:42 CiaranG Why would the head pitch and yaw? 16:43 VanessaE CiaranG: watch any Minecraft video where third-person mode is in use and you'll see what I mean by that 16:43 ShadowNinja Oh, ,,(title ) 16:43 ShadowBot ShadowNinja: Page is too big. 16:44 BlockMen VanessaE, i think its a nice feature in the future, but it is not necessary for 3rd person view 16:44 VanessaE the head should react to the movements of the camera, just as players see each other turn side-to-side when they move around, so should their heads turn, because it looks better 16:44 VanessaE BlockMen: oh, no it isn't strictly necessary, just thinking of aesthetics. 16:47 CiaranG Hmm, see what you mean. Looks kind of weird to me though. 16:47 CiaranG In any case, I'm pretty sure it's not a reason to block a perfectly good feature that could have been merged 3 months ago. 16:48 VanessaE nope, no reason to block. just saying. 16:49 ShadowNinja !issue 1114 16:49 ShadowBot https://github.com/minetest/minetest/issues/1114 16:51 ShadowNinja Seems good. I agree with the head pitch+yaw thing. 16:52 BlockMen I guess we all do :P But agree and implementing are two different pair of shoe 16:53 Calinou head only is not simple to do 16:53 Calinou pitching whole model by 25% is easier 16:53 Calinou but, we need to at least have entity pitch for that 16:54 BlockMen if we would want such a feature we should drop the multiple model support for player IMO 16:55 Calinou why? 16:58 ShadowNinja We can have support for other models, we just might require that you define the head region and the like. 16:59 BlockMen because you need many information for the head entity to then, or can a part of an entity moved? 16:59 BlockMen then it would not be necessary ofc 16:59 ShadowNinja That should be done already actually, and the engine should move the parts. 16:59 Calinou multipart models are obsolete and scary for artists 17:00 VanessaE there was a mod some time back that could move the head separately. 17:00 VanessaE I think kaeza made it 17:00 ShadowNinja Eg, you define a arm, name it "larm" and define a path for it to follow when it moved to dig. 17:00 VanessaE it used a fuckload of network traffic to do it, but it DID work 17:00 VanessaE it didn't do anything special, it manually moved the head in the model 17:02 kaeza https://www.youtube.com/watch?v=edzDAH6-WQU 17:02 VanessaE speak of the devil 17:02 kaeza :) 17:03 kaeza but as noted by Taoki, there are some issues with irrlicht (not being able to animate *and* manually place bones at the same time) 17:04 kaeza so I just manually placed every bone in the animation, resulting in this "fuckload of network traffic" due to set bone messages :) 17:08 BlockMen it should be localy handled like the animations for local player model in 3rd person view. but then just the own player would be able to move the head (without great lags) =\ 17:25 sapier #1143 another request for comment ... I'm gonna merge it in a couple of hours 17:25 ShadowBot https://github.com/minetest/minetest/issues/1143 -- Protocol version <= 22 compatibility mode + clean new client init by sapier 17:37 proller + 1dfa415...9224f2a desertcobble -> sfan5/desertcobble (forced update) fuck push-f ! 17:37 sfan5 lel 19:17 sapier I'm quite sure someone will complain within the next 24 hours but I'm gonna merge #1143 in 10 minutes anyway 19:17 ShadowBot https://github.com/minetest/minetest/issues/1143 -- Protocol version <= 22 compatibility mode + clean new client init by sapier 19:18 proller who test it ? 19:19 sapier vanessaE is running this for about 2 months now on all of her servers 19:20 sapier guess there aren't many pull requests tested this intense 19:20 VanessaE the only caveat is that players using non-official tablet clients often can't see players that were already online when they joined a server. 19:20 VanessaE because those unofficial clients are busted. 19:21 VanessaE all official clients, old or current, seem to work fine after the first or second try. 19:21 VanessaE (all current ones work fine on the first try, every time) 19:21 sapier that's why I made the android port ... of course I have to update it to this changes to to gain full media fetch performance 19:21 VanessaE right 19:22 VanessaE someone buy sapier the appropriate hardware so he can make an iOS port also :) 19:23 sapier no way I won't do any work for apple devices ;-) 19:24 sapier and I have to do some work for mobf once the android things are merged ;-) 19:24 sapier lots of good ideas are waiting ;) 19:25 VanessaE well okay, if you wanna field all 2 support requests ;) 19:25 VanessaE ('cuz I'm forwarding all unofficial clients here :D ) 19:26 proller "seem to work fine after the first or second try" - good quality 19:26 sapier no problem I'm targeting full master merge till end of month as of that time improvements can be done step by step 19:27 sapier proller it's a race condition not fixable server side 19:27 proller when you make downloadind speed more 300kbps ? 19:28 sapier old version did run in it every now and then, old <-> inofficial combination causes it once , new <-> new never 19:28 proller old connection with 3 lines tune transfer 1-1.5mbps.. 19:29 sapier once someone is willing to support protocol testing 19:29 VanessaE current connection with current code I can tune the transfer to ONE HUNDRED MEGABITS. 19:29 VanessaE now shaddup proller 19:29 proller wat? 19:29 VanessaE yeah. 19:29 VanessaE I can push about 12-15 Mbps to any one user over that pipe 19:30 VanessaE that's the average speed over http and about...what was it sapier? 9? over udp 19:30 proller im about udp 19:30 proller http have nothing with connection 19:30 VanessaE but because most of my players already have all the media and I don't constantly add stuff, they don't need to transfer most of it anymore 19:30 * proller faceplam 19:31 sapier sorry to much time but I remember you thougt someone downloading http and have been surprised it beeing udp instead 19:31 VanessaE but if you would read the second part, I said 9. 19:31 VanessaE as in 9 Mbps over UDP 19:31 proller why i have 300-400k when joining to LOCAL game ?? 19:31 VanessaE proller: no idea. I blame your hardware 19:32 proller i have 50+mbps with enet on this hardware 19:32 VanessaE when I play a local game, it's going to/from a good SSD 19:32 VanessaE so bandwidth is never an issue 19:33 sapier proller I said multiple times I need help to benchmark enet but none of you enet supporting guys ever did do a single piece to gain any progress at this task 19:34 sapier proller those 300 kbit happen for you on current master too correct? 19:36 VanessaE and I wonder if proller bothered to try out those settings at all 19:36 VanessaE e.g. max_simultaneous_block_sends_server_total 19:38 VanessaE better than two orders of magnitude speed boost from your code alone, then another order of magnitude by changing just that one setting to a value appropriate to my pipe 19:40 sapier proller are you testing right now or did you just leave? 19:41 sfan5 https://github.com/sfan5/minetest/tree/redis-db redis database backend 19:41 sapier benchmarks? 19:41 sfan5 none yet 19:41 sapier or some other benefit over our current variants? 19:41 sfan5 you can have the db on a remote server 19:42 sfan5 performance also depends on the redis config 19:42 sapier well that's true for all databases ... why redis and not something more standard like mysql? 19:42 sapier never heared about "redis" 19:42 sfan5 redis is a key-value db 19:43 sfan5 heavily optimized for that 19:43 sfan5 s/that/such stuff/ 19:43 sfan5 Redis is an open source, BSD licensed, advanced key-value store. http://redis.io/ 19:43 celeron55 does it have better support than leveldb? 19:43 sfan5 what is "better support"? 19:44 celeron55 from, say, python and other linux packages 19:44 celeron55 and windows 19:44 sapier as our database interface is quite modular I guess it won't be a issue 19:44 sfan5 IIRC redis has more support than LevelDB 19:44 celeron55 with leveldb you can't just make a python script that does something with minetest's data 19:45 sfan5 redis-cli would be enough to delete some mapblocks 19:45 sfan5 (from an sysadmin perspective) 19:45 sfan5 windows & python should work with redis 19:45 celeron55 (or can you?) 19:45 celeron55 well whatever 19:45 celeron55 redis-cli sounds good 19:45 sapier celeron55 do you have a problem with version reading in debug builds? from mods supposed to be debug mods only 19:45 sfan5 http://redis.io/clients 19:46 celeron55 does leveldb have a similar one? 19:46 sfan5 you can use it from practically everywhere 19:46 sfan5 I don't think so 19:47 celeron55 sapier: i don't care (i don't have time or motivation for caring at the moment) 19:49 celeron55 it looks like leveldb bindings *are* widely available for python 19:49 sapier ok in this case I'm gonna merge it this was and fight against anyone trying to enable it in release builds too 19:49 celeron55 but i can't quickly find a proper list of for what they actually are available 19:50 sapier "redis crashes without /dev/urandom" ... does this issue mean it's linux only? 19:51 sfan5 if you don't have /dev/urandom you have a problem anyway 19:51 celeron55 https://github.com/mythz/redis-windows 19:51 celeron55 looks like a windows port is maintained by... microsoft 8) 19:52 sfan5 don't forget that redis is not directly usable as a normal backend as you need to start the redis server manually and set redis_address (and optionally redis_port) in world.mt 19:53 celeron55 actually this https://github.com/MSOpenTech/redis 19:53 celeron55 yes, it's not actually usable as an embedded thing like leveldb and sqlite 19:54 sfan5 but it may still be very useful for (bigger) servers 19:54 sapier sfan5 we've got 2014 ;-) 19:56 sfan5 should I create a pull or are we still waiting for libmtmap (which doesn't seem to be happening soon)? 19:58 sapier I suggest creating a pull and doing some benchmarks .. in best case create a pull for those benchmarks too 19:58 sapier ok I'm merging #1143 now 19:58 ShadowBot https://github.com/minetest/minetest/issues/1143 -- Protocol version <= 22 compatibility mode + clean new client init by sapier 20:00 sfan5 https://github.com/minetest/minetest/pull/1210 20:00 sfan5 ~title 20:00 ShadowBot sfan5: Add redis database backend by sfan5 · Pull Request #1210 · minetest/minetest · GitHub 20:01 sfan5 sapier: does the server now receive the info from porting::get_sysinfo() ? 20:02 sapier no 20:02 sapier unchanged to the pull request standing there for 2 monthe ... which I asked for final comments for about 2 days now 20:02 sfan5 add it please 20:02 sapier nope 20:02 sfan5 :-( 20:03 sapier I removed the version infor for release clients too ... they're not meant to be used for anything else then debugging 20:04 sapier but you're free to add that information and create a pull request of course 20:04 sapier except the version things, I'll not support adding it to release versions 20:04 sfan5 I do not intend to do that 22:40 PilzAdam lighting seems to have no effect on water 22:40 PilzAdam might be another bug of RBAs shaders 22:43 PilzAdam reverting 0dc1aec50940140e28f434c524296e284e73d623 fixes it 22:48 VanessaE or someone could just find the glitch and fix it.....