Time Nick Message 00:47 hmmmm back 00:48 hmmmm whoop, i see the problem 00:48 hmmmm man i feel dumb 00:48 VanessaE see? it had time to "develop" 00:48 VanessaE sometimes you just have to take a break. 00:48 hmmmm nah didn't take any time 00:48 VanessaE (also, wb) 00:48 hmmmm i just looked at it right now 00:49 VanessaE well at least you solved it. 00:53 hmmmm i should work in a nuclear power plant :( 00:55 VanessaE that bad huh? 00:55 VanessaE ok, I'll bite. 01:11 hmmmm you see, i forgot to divide the coordinates by the spread value. 01:11 hmmmm i had forgotten that i originally intended to leave it up to the caller to do that 01:11 VanessaE oops. :-) 01:11 hmmmm so the coordinates that the perlin noise functions would be getting were crazy 01:12 VanessaE not that I'm all that well-versed in the mapgen code, but it sounds like the same sort of braino I've made many times in other languages. 01:12 hmmmm should work right now 01:14 VanessaE sounds like you're about ready to have this merged to minetest master? 01:14 hmmmm noo 01:14 VanessaE oh ok heh 01:14 hmmmm it needs polish, but it is working 01:15 hmmmm i have a lot of hacky things that need to be smoothed out like hardcoded values added for the sake of debugging 01:15 felix20 how fast is it? 01:15 VanessaE I'm curious - has the base algorithm changed? I mean, will the same seed make the same map with your code as it would with 0.4.4? 01:15 VanessaE (biomes aside) 01:16 hmmmm i can't give a fair assessment since i'm running it in debug with only -O1 01:18 hmmmm at this point i'd have to say that the major bottleneck in mapgen performance is the lighting update and allocating/deallocating about a million v3fs per chunk 01:18 hmmmm er v3s16s 01:19 hmmmm that's going to take a bit of thinking to find an elegant solution for 01:19 hmmmm before optimizing that i'd like to add the threading code in for the emerge queue 01:23 VanessaE here's a question: are you running the mapgen all in a single thread now? Would there be any benefit to, say, launching one thread per mapblock to be generated (up to some reasonable max)? 01:23 hmmmm that's the plan 01:23 VanessaE oh ok 01:24 VanessaE guess I missed the memo :-) 01:30 hmmmm although i question its utility unless the distance at which new blocks are generated is increased 01:30 VanessaE well, in my worlds, I have the program set to generate 10 blocks away to go along with my increased view distance. 01:31 hmmmm right now even when you're really busy, in -O1, you'll only ever get like, i dunno, 6 chunks in the queue 01:31 hmmmm which get ripped up instantaneously 01:31 VanessaE hm 01:31 hmmmm so when this goes to master, the default sendblock distance is going to need to be increased in order for multithreading to have any utility 01:32 VanessaE what about changing it to spawn threads dynamically? 01:32 hmmmm lol :/ 01:32 hmmmm that's pretty wasteful 01:32 VanessaE based on the output of the auto fps code 01:33 hmmmm this is server side so it has nothing to do with FPS 01:33 VanessaE yeah Iknow 01:33 hmmmm do you mean the auto client-tick-sync? 01:33 VanessaE but that code calculates view distance (or is supposed to) based on a desired FPS, can't you use that to *also* determine how many threads should be launched to run the mapgen? 01:34 hmmmm not really 01:34 VanessaE I mean, what's the point of generating more land than you can see (or too much faster than you can load) 01:34 hmmmm i determine how many mapgen threads to create based on how many cores you have 01:34 VanessaE hrm. well it was just a thought 01:34 hmmmm that's all the threads you'd want 01:34 hmmmm and if they aren't used, well, then they just sleep 01:34 hmmmm if they're needed, then they'll be used 01:35 hmmmm hmm now's a good time to fix up the RUN_IN_PLACE relative filenames for OSX/FreeBSD 01:36 VanessaE well this sounds fine then, but there is one caveat: 01:36 hmmmm i don't have access to mac OS X 01:36 hmmmm do you know anybody who does? 01:36 VanessaE what about hyperthreading systems? do those report a virtual number of cores? 01:36 VanessaE no, not off the top of my head, at least not anyone who uses minetest. 01:37 VanessaE (my sis in law has a fairly decent mac system though) 01:38 hmmmm it'd return the number of threads 01:38 VanessaE ok 01:38 VanessaE so a non-issue then 01:39 hmmmm hmmmm 01:39 hmmmm maybe it's different for Linux 01:39 hmmmm i don't have anything linux so i've never tried it.... 01:40 VanessaE I run it here, though I don't think my particular processor has hyperthreading. 01:40 hmmmm yeah it doesn't 01:40 hmmmm someone who has Linux and hyperthreading, have them run this: http://pastebin.com/xdUqwb9Y 01:59 hmmmm hrm 01:59 hmmmm is there an ARRAYLEN() type macro defined somewhere in minetest already? 01:59 VanessaE actually I think there is 02:00 VanessaE one sec 02:00 VanessaE well strike that - in the Lua there is anyway/ 02:00 VanessaE doesn't do you any good though I guess 02:00 hmmmm nah 02:00 hmmmm well i don't really need it 02:23 hmmmm mmmm.. 02:24 hmmmm celeron.... you can't #include headers inside of code blocks such as you do in initializePaths() 02:24 hmmmm i guess it works for windows.h and unistd.h, but it doesn't for most other headers 02:42 hmmmm https://github.com/kwolekr/minetest/commit/b3db3362e3832d8fb9b694338a4f20d01d387373 03:59 SitDog_ i can volunteer for minetest on the mac. i have a low end and high end mac mini. a single core mac and the best mac mini you could get last fall. 04:02 SitDog_ i need to get a linux box setup next. my main pc is a beast. i also have a 6core amd phenom 2, a dual core athlon, and a dual processor pentium 4. 04:03 SitDog_ probably going to put linux on the pentium box 04:41 hmmmm that's awesome 04:41 hmmmm if it's not asking too much, could you test that patch i just pasted? 04:42 hmmmm patch porting.cpp, any version really (that file hasn't changed in forever), compile with -DRUN_IN_PLACE, see if it works 04:42 hmmmm can you tell me what kind of processor is in that mac mini? 04:43 hmmmm both of them actually 04:51 SitDog_ sorry for the delay. the mac mini is a core solo 1.5ghz. 04:52 SitDog_ the other is a intel core i5 at 2.5ghz. 04:53 SitDog_ i dont' have the build setup on the machines right now. so it will take me a bit to get it going. 04:55 hmmmm is that i5 a mobile processor? 04:59 SitDog_ it is a desktop processor i believe. let me check. 05:00 SitDog_ midrange and mobile computers 05:02 hmmmm argh 05:02 hmmmm well 05:02 hmmmm probably easier to just tell me the model number of it 05:08 SitDog_ MC816XX/A for the new mini and i have the absolute first mac mini intel (core solo) MA205*/A 05:09 SitDog_ feb 2006 and july 2011 models. 05:10 hmmmm ah, i5-2520M 05:10 hmmmm you have hyperthreading 05:10 hmmmm can you run this and tell me the number it spits out? http://pastebin.com/xdUqwb9Y 05:15 SitDog_ one sec 05:18 SitDog_ 4 is the answer :D 05:18 hmmmm okay, thanks 05:18 hmmmm it's probably the same for Linux too 05:19 hmmmm so, to sysconf, it'll return the number of threads, not the number of cores.... good to know 05:19 SitDog_ i have mad updates. wow 05:19 hmmmm i am sorry they're so upset 05:21 SitDog_ i have many of them... :D 05:47 SitDog_ is there a way for you to tell the mac to update everything and beep when you are done? :D 05:57 hmmmm shrug 08:22 celeron55 hmmm: test.c:2:33: error: ‘_SC_NPROCESSORS_CONF’ undeclared (first use in this function) 08:23 celeron55 oops... 08:23 celeron55 forgot the headers 8) 08:23 celeron55 it says 4 for a dualcore i3 (with hyperthreading) 08:24 celeron55 on linux 09:11 celeron55 hmmmm: related to your porting.cpp stuff: https://github.com/celeron55/minetest/issues/203 09:13 celeron55 actually... your link modifies the run-in-place behavior and that is about the global behavior 11:51 celeron55 https://github.com/celeron55/minetest/commit/8bb36f465000f0fdf5e9202d9f76b170c330c49f 11:52 celeron55 the thing works like magic once again! 11:56 celeron55 i wonder how many games implement derivative estimation based render distance tuners 11:59 celeron55 it actually calculates how much it would need to modify the render distance in order to get a certain frametime change 12:02 celeron55 when it's trying to tweak the frametime, it gets the estimated frametime/range derivative from measuring how much a previous change in range affected the frametime 8) 12:03 celeron55 i made it when it turned out anything simpler would not work in a wide enough range of situations 12:06 celeron55 for example, if you are wathing down from a high tower, nothing else will work at all 12:06 celeron55 they'll just oscillate 12:06 celeron55 watching* 12:09 celeron55 (or alternatively are so slow that they're equal to having nothing) 13:41 celeron55 i am testing the MSVC build at the moment 13:45 celeron55 it appears to build at least 14:00 celeron55 is there anything that exists, is not upstream and should still go in 0.4.4, in someone's opinion? 14:28 RealBadAngel hi 14:28 thexyz ttf fonts 14:30 RealBadAngel celeron55, are you there? 14:31 celeron55 thexyz: where is that currently available? 14:32 RealBadAngel celeron55, i have a question bout function that get craft recipes 14:32 thexyz celeron55: it's not up-to-date, but http://minetest.ru/ttf/ttf-25cf3757b29a0397e201a4719788ead3dc76ec94.patch 14:32 RealBadAngel currently is returnin only one recipe even when multiple are aviable 14:34 darkrose that's exactly what I wrote it to do 14:34 RealBadAngel darkrose: you talkin bout get_craft_recipe? 14:35 celeron55 haha, the perfect answer 14:35 darkrose yes 14:35 RealBadAngel so its wrong, it shall return alll possible ones 14:36 RealBadAngel i tried to catch them all with overriding register function, but its seems to be not working 14:37 RealBadAngel some mods get executed before i do override 14:37 celeron55 somebody could implement minetest.get_craft_results() and minetest.get_craft_recipes() 14:38 RealBadAngel get recipes is done, but when get 1st one it returns it 14:38 RealBadAngel doesnt lookin for another ones 14:38 celeron55 ehm 14:39 RealBadAngel darkrose, since you made the original one, could you go further? 14:39 celeron55 no, there is no plural forms of neither of those functions 14:40 RealBadAngel so you mean plural to get multiple and one we already have to get first found? 14:44 RealBadAngel i have another path ready i would like to merge. its about edit boxes. by now enter closes the formspec, which is very annoyin. i propose when editin first enter to lose fucus, return to formspec, then if nothin is focused next enter hit will close the formspec 14:44 RealBadAngel *patch 14:47 darkrose eww, use shift-enter if you don't want to submit, write a patch for that, it's fairly common practice and won't result in people complaining that 'enter key doesnt work' 14:48 RealBadAngel im using search box in my inventory mod 14:48 RealBadAngel all the folks complainin when hit enter it closes 14:48 celeron55 why not add some kind of an enter_function[submit/close/none] 14:49 RealBadAngel i can 14:49 RealBadAngel but enter (not +shift) is most natural way to submit a query 14:50 celeron55 hmm, more specifically, set_enter_function 14:50 darkrose field[x,y;w,h;name;label;default;onenterpress] 14:50 RealBadAngel its made already another way 14:50 RealBadAngel if enter then shutdown formspec 14:51 celeron55 darkrose: no 14:51 RealBadAngel imho enter shall cause send, lose focus 14:51 RealBadAngel esc leave 14:52 celeron55 ehm 14:52 celeron55 there are many single-field formspecs in where enter should accept and close 14:52 celeron55 it should be the default 14:52 darkrose celeron55: why? make it an optional field, default behaviour is what it is now? 14:52 celeron55 darkrose: because that makes unnecessary ambiguous situations 14:53 darkrose ? 14:53 celeron55 for example, what if you set it to many fields? 14:53 celeron55 better have a separate command for setting what to do on enter press 14:53 celeron55 it is way more logical 14:53 celeron55 it's form-global anyway 14:55 darkrose that was kind of the point... one field could be submit on enter, but not the rest, or vice-versa, which I could see uses for... like with multiple fields where enter gives focus to the next field, then submits when you're on the last one 14:56 celeron55 hmm 14:58 celeron55 so then we would set the only behavior of pressing enter with no fields selected to nothing, and pressing enter when a field is selected to submit the form and close, and allow specifying what to do for a field by using the format you mentioned 15:01 celeron55 thexyz: https://gist.github.com/4225065 15:02 celeron55 the patch can be applied without problems, but it produces that kind of stuff 15:03 celeron55 maybe you should update it to be included to 0.4.5 8) 15:07 celeron55 hmm, then there was the problem of my MSVC being too old to produce stuff without SSE2 15:07 celeron55 ...maybe 15:08 RealBadAngel celeron55, i see your point but its leftover from editin text for sign 15:08 RealBadAngel we are building now more complex formspecs 15:09 thexyz celeron55: i've just applied it to master branch, built without any problems 15:09 RealBadAngel please rethink one enter lose focus, 2nd exit 15:09 RealBadAngel this way is easy 15:09 RealBadAngel i have it done already and it works 15:10 darkrose RealBadAngel: it breaks the default mod, this does not qualify as 'it works' 15:10 celeron55 RealBadAngel: ...can't you understand a thing called compatibility? 15:11 celeron55 things can be added, as long as they won't break existing mods with their default behavior 15:11 RealBadAngel i do, thats why im asking for ideas 15:11 RealBadAngel instead of tryin to push just my own] 15:11 celeron55 i proposed a scheme that would work for you and for old stuff 15:12 celeron55 (or, well, wrote down more specifically what darkrose suggested) 15:12 thexyz celeron55: what irrlicht version do you use? 15:12 RealBadAngel so definitions on_enter? 15:13 celeron55 thexyz: it appears arch updated it to 1.8, so that one now 15:13 celeron55 it should definitely work on 1.7 and 1.8 15:13 thexyz ok, i'll port it to 1.8 15:14 RealBadAngel celeron55, are you familiar with L-systems? 15:14 celeron55 i can see one problem with irrlicht 1.8 already; it makes the text fields of the UI white like the font is and seeing anything is a bit hard 15:14 celeron55 gotta fix that sometime... 15:15 celeron55 RealBadAngel: i have no idea what that even means 15:15 RealBadAngel celeron55, jeija has the patch for it ready 15:15 celeron55 RealBadAngel: where is it 15:15 RealBadAngel i got it but its about week old 15:16 RealBadAngel i think he have to rebase it 15:16 celeron55 i have no problems rebasing stuff 15:16 celeron55 just give me the changeset/commit/whatever 15:16 RealBadAngel i will find it and send 15:17 RealBadAngel L-Systems are fractal ways to imitate growing of plants 15:18 RealBadAngel heve you seen another open source voxel game called Terasology? 15:18 celeron55 googled: http://minetest.net/forum/viewtopic.php?id=3920 15:18 RealBadAngel its using L-systems to generate epic trees 15:19 RealBadAngel http://blog.movingblocks.net/wp-content/uploads/screen1_0-522658_960x332.jpg 15:20 RealBadAngel and the code that makes them: 15:20 RealBadAngel http://pastebin.com/TwzcDina 15:21 RealBadAngel celeron55, yeah, thats it. simply font and background had almost very same color 15:23 celeron55 hmm, i wonder how i can check irrlicht's version to not try to do those for 1.7 15:24 celeron55 ha 15:24 celeron55 IrrCompileConfig.h has version defines 15:27 celeron55 i guess this is quite right: https://gist.github.com/4225305 15:27 hmmmm i think i've seen terasology before, i know the guy who made it 15:27 hmmmm it's cheating because the nodes are much smaller than those in minecraft 15:27 hmmmm so you can obviously makre more complex and better looking trees and what not 15:27 celeron55 well, actually, that'll not work with irrlicht 2.x... 15:29 hmmmm at our node scale, unless the tree is absolutely enormous, using an L-system to generate them would be useless since all the detail would be removed 15:29 celeron55 ha, the bulletproof version: https://gist.github.com/4225332 15:30 celeron55 at this rate, irrlicht 2.0 is probably coming in 10 years 15:30 celeron55 (actually, probably never) 15:36 doserj Can someone try out https://github.com/celeron55/minetest/pull/335 ans see if that works for them? 15:37 doserj for example, set (preload_item_visuals = 0 and page through the creative inventory 15:38 PilzAdam testing 15:41 PilzAdam the blackscreen is gone but the HUD and chat arent displayed and i get an FPS drop 15:43 bulletrulz help 15:43 bulletrulz i got muted on the main channel for some reason 15:43 Calinou why would that happen? 15:43 bulletrulz idk 15:44 bulletrulz i was fucking around with the bot yesterday but VanessaE was cool with she told me to stop after a little and i did 15:46 thexyz that's not a valid topic for -dev 15:46 doserj PilzAdam: but they are redrawn after the rendering is finished? at least that is happening for me 15:46 PilzAdam yep 15:47 bulletrulz well sorry but had no where else to go 15:48 PilzAdam doserj, its better than the current blackscreen but its not the perfect solution 15:49 doserj yeah. 15:49 hmmmm wait, what blackscreen 15:49 doserj I am looking what i can do 15:58 celeron55 ...wtf is irrlicht doing if that gets rid of the black screen 15:58 doserj hmmmm: every time the game renders to a target texture, for example when a item image is created, the screen turn black for the time it needs to render that. This is especially noticeable when you don't preload the item visuals. otherwise it is just done all at startup. 15:59 doserj celeron55: i have no idea... 15:59 hmmmm that's probably why i never noticed it.. 16:00 RealBadAngel confirm 16:00 celeron55 i added the on-demand rendering of them just a week or two back 16:00 RealBadAngel that option f**ck up also all creative mods 16:01 RealBadAngel even default one 16:01 celeron55 before that it wasn't even an option to make them afterwards 16:02 Calinou "Raise default maximum render distance to 240" < what if you look at a wall near you (which increases FPS and brings view distance to 240) then turn around 180 degrees, on a slow computer? 16:02 Calinou (if the blocks behind you are already loaded) 16:05 celeron55 that question doesn't make any sense in the context of raising from 160 to 240 16:05 celeron55 if it was from 40 to 240, then it would 16:06 celeron55 but the answer is that it works relatively well in practice, because stuff loads up before it raises that high 16:07 celeron55 also, there is a limit to the amount of more blocks it can suddenly draw 16:08 celeron55 so rather than lagging like hell, it will render stuff from a bit odd places for a few frames and then settle to good values 16:08 celeron55 or at least should 16:09 celeron55 princesses can set their own maximums in the config as they see fit 16:17 celeron55 i'll merge this render-to-texture patch 16:18 celeron55 Calinou: can you link to your changelog 16:19 PilzAdam http://dl.dropbox.com/u/82342922/0.4.x-changelog.txt 16:21 celeron55 i think i need to add some news section on minetest.net as c55.me/blog is down too much currently... 16:22 thexyz celeron55: i've fixed that patch 16:23 celeron55 where is it 16:25 doserj it seems you can fix it by changing the order in which you draw things 16:25 thexyz celeron55: https://github.com/minetest/minetest/tree/ttf 16:26 celeron55 how is it to compile on windows? 16:28 thexyz painful, as usual 16:28 celeron55 i'm not even building gettext because it sucks 16:28 celeron55 never have 16:28 thexyz just point cmake to freetype2 include dirs & libs 16:28 Calinou "celeron55 authored in an hour" < github... 16:28 celeron55 does it build without freetype? 16:29 thexyz no 16:29 celeron55 that is bad 16:29 celeron55 we can't afford having building for windows being any harder than now; it sucks away all windows contributors 16:30 thexyz not having unicode support in 2012 is worse 16:30 thexyz are there any? 16:30 celeron55 hmmmm is 16:30 celeron55 i'm not against having unicode support in official builds, but one should be able to build a version without freetype and without unicode support 16:31 thexyz hmmmm: are you ok with freetype2 requirement? 16:32 doserj ttf-support would allow adding TextSceneNodes properly (with proper selection of fonts). Signs with in-game text done properly. in-game waypoints, etc... 16:33 celeron55 it's not about whether hmmmm is okay with it now; it is all about if he had started even trying to do anything if faced with freetype 16:34 celeron55 is adding such support hard? 16:34 celeron55 or are you just being extra lazy 16:34 hmmmm huh? freetype? 16:35 hmmmm i have nothing to do with that 16:35 hmmmm you might be thinking of someone else 16:35 celeron55 hmmmm: we are maybe going to put unicode support into minetest by using the freetype2 library 16:35 hmmmm ohh, thexyz's freetype patch... i never got around to doing anything like that 16:35 hmmmm hmm 16:35 hmmmm sounds like a lot of work 16:35 celeron55 and it currently is a pain in the ass on windows 16:36 hmmmm yes i remember that whole thing 16:36 celeron55 it already works, but it does not have support for building without the freetype library, which i don't like and thexyz is too lazy to add support for doing that 16:36 celeron55 i'm worried that we wouldn't be getting any windows contributors anymore after doing that 16:36 thexyz i didn't say that 16:36 hmmmm i thought there was a slight problem with drawing on the formspec 16:37 hmmmm or something along those lines 16:37 celeron55 ...well, is there? 16:37 thexyz sure 16:37 celeron55 okay, this isn't going into 0.4.4 anwyay 16:37 celeron55 maybe we can get this sorted out for 0.4.5 16:38 hmmmm i think xyz said something about the unicode characters showing up as blocks or 16:38 thexyz it only works well with chat; windows fucks up text filled into formspec fields 16:38 hmmmm oh 16:38 hmmmm meaning it doesn't convert from multibyte to wide char 16:38 hmmmm is that a freetype2 windows bug? 16:39 hmmmm or does it encompass more platforms 16:39 thexyz i guess it's some windows-specific bug which affects string conversion 16:39 celeron55 i assume there are no objections for releasing the current github master of minetest and minetest_game as 0.4.4 16:39 hmmmm when you say "fucks up text", what do you mean precisely? can you show a screenshot? 16:39 hmmmm i bet i can tell what the actual problem is right away 16:40 thexyz it sends some shit to server instead of actual string 16:41 hmmmm does this happen with all text or just text with unicode characters? 16:41 thexyz only unicode (non-latin/numbers/etc) ones 16:41 hmmmm does it have like an A with an omlaut on the top along with some other weird characters? 16:42 doserj celeron55: I added a fix for the hotbar disappearing issue. 16:45 hmmmm off-topic, but what would the best place to define biomes be? i currently have that stuff in builtin/misc.lua but that's obviously wrong 16:45 celeron55 anyone do a quick test too? it seems to work for me https://github.com/celeron55/minetest/pull/335 16:45 celeron55 specially drawn texts do still flicker 16:45 celeron55 hmmmm: paste how that looks like currently 16:45 doserj1 celeron55: Seems I did get disconnected. 16:45 doserj1 celeron55: I added a fix for the hotbar disappearing issue. 16:46 celeron55 i saw that and requested somebody else to test it too 16:46 thexyz hmmmm: ah, i don't remember; will build windows client now and tell you results 16:46 celeron55 i don't understand why it works though, but i guess it's better than nothing once again... 16:47 PilzAdam tested 16:47 PilzAdam works 16:48 hmmmm oh frig you can't be seriouis 16:48 hmmmm i lost last night's work 16:48 PilzAdam with preload_visuals you dont get a blackscreen but the media load screen 16:48 hmmmm unless i saved it in a different directory 16:49 thexyz used previous version, hmmmm, when there's an unicode character it displays nothing (using sign, for example) 16:56 RealBadAngel darkrose: can i talk to you PM? 17:01 hmmmm oh 17:01 hmmmm i put it misc_register 17:01 hmmmm http://pastebin.com/vRAFchZ1 17:02 hmmmm yeah that made more sense 17:02 darkrose RealBadAngel: sure 17:02 hmmmm i don't really get a lot of the Lua stuff, is there any documentation on how it works? i've really just been mimicing the existing code 17:03 hmmmm like what's the 'index' parameter about in the callbacks for lua? 17:03 celeron55 http://www.lua.org/manual/5.1/manual.html#3 17:03 hmmmm ah okay, thanks 17:04 celeron55 for more insight, this book is the... THE book: http://www.lua.org/pil/ 17:04 hmmmm also scriptapi.cpp is so unwieldy... what if it was broken up into envref functions, nodemetaref, etc. 17:07 Calinou patch, this makes shader-less lights look better: http://dl.dropbox.com/u/82342922/mapblock_mesh.patch -- way better than adding lighting to top, better remove lighting to sides/bottom... 17:07 Calinou (I used -ruN format, is that good?) 17:07 Calinou was too lazy to make pull request for a small change like this :p 17:07 hmmmm people do pull requests for one-line changes 17:08 celeron55 sounds like it will make things unplayably dark on bad displays 17:08 Calinou minecraft does similar stuff, btw... it only darkens sides and bottom 17:08 celeron55 yes, and it makes it unplayably dark on bad displays 17:09 doserj minecraft has a brightness setting, though 17:09 hmmmm yeah, don't make things too dark... eduke32 does that by default and it makes me want to kill something 17:09 hmmmm or we could just add a NIGHT VISION POTION huehehueh 17:09 Calinou it looks much better... not my fault if people use very high contrast for their screens 17:09 celeron55 many displays aren't configurable at all, like laptop screens 17:10 celeron55 and have very screwed gamma and no fidelity at all at dark colors 17:10 Calinou then the driver's control panel lets you fix it :P 17:11 celeron55 of course i have an IPS screen both on my laptop and desktop (bought one yesterday...), but that isn't the case for most people 8) 17:11 Calinou also, light of top side is unaffected, so it would not really harm gameplay on bad displays, maybe help it because you can see the edges 17:13 Calinou http://dl.dropbox.com/u/82342922/screenshot_1898498005.png < how it looks 17:13 Calinou if you enable shaders, it looks similar, but water is darker, and top surface is brighter 17:13 Calinou no longer hard-to-see edges :p 17:13 celeron55 too large image for my sucky cellphone connection 17:14 hmmmm maybe it's just me but i don't see the difference 17:14 doserj Calinou: how does it look in caves? 17:14 hmmmm also did anybody ever say that the new default textures are way oversaturated 17:14 celeron55 funny; now i have a 24" 1920x1200 display, but now i don't have ADSL :-D i quite prefer the ADSL 17:15 Calinou http://dl.dropbox.com/u/82342922/screen.jpg 17:15 celeron55 hmmmm: most aren't, but for example bricks were replaced already 17:15 Calinou celeron55: what specs does it have, where did you buy it btw? 17:15 celeron55 Calinou: dell u2412m 17:16 celeron55 pretty much the only sane thing to buy 17:16 hmmmm >not getting a catleap 17:16 Calinou that's a screen, not a computer :P i was talking about computer 17:16 celeron55 computer? 17:16 Calinou desktop 17:16 celeron55 the laptop? 17:16 celeron55 i bought it a year ago 17:17 Calinou "desktop (bought one yesterday...)" 17:17 celeron55 screen of desktop 17:18 celeron55 how do you expect i would have said that sentence to make it mean a desktop display 17:18 Calinou ah ok 17:18 celeron55 silly french! 17:18 hmmmm maybe he thought you were one of those people who have a dual monitor laptop setup 17:20 celeron55 i would if i had a laptop fast enough 17:34 celeron55 oh god this release-making stuff... 17:34 celeron55 horrible, horrible work 17:34 Calinou since when releasing stuff is not horrible? :) 17:36 celeron55 1) update version in source, 2) update branches and tags of minetest and minetest_game on github, 3) update source downloads on minetest.net, 4) build windows version 5) tweak windows package a bit, 6) upload windows package to github, 7) update it to minetest.net, 8) write a release notice somewhere with a bit of changelog and stuff 17:37 celeron55 and worry all the time that something goes wrong and you have to start from the beginning 17:37 celeron55 this wouldn't be so headache-prone if it could go wrong 17:39 celeron55 ah, and i forgot stuff 17:39 celeron55 9) update stuff on launchpad to get ubuntu builds for the new version 17:40 celeron55 at least i don't have to compile C++ to x86 assembly by hand... 17:48 celeron55 currently at 6 17:54 celeron55 http://minetest.net/download.php 18:04 celeron55 added this for now: http://minetest.net/news.php 18:16 celeron55 aaand launchpad packaging updated, should be ready in some tens of minutes 18:16 celeron55 then begins the wait for angry users 18:25 darkrose all the files in the tarball from github are timestamped 'in the future' (so complains tar when I extracted) 18:26 darkrose not that much can be done about that I'm guessing: github's mess 18:26 celeron55 umm 18:26 celeron55 well, dunno 18:27 celeron55 my zip packet might be messed up time-wise at least 18:27 celeron55 probably a few hours in the future 8) 18:27 doserj your commits were also timed in the future on github 18:27 darkrose 1 hour I think, a little under 18:27 celeron55 my clock doesn't seem to be quite right 18:29 celeron55 well, at least it's hours, not years... 18:29 darkrose eh, just boast about it being "ahead of it's time" and everyone will think it's awesome 18:29 celeron55 8) 18:30 celeron55 more like "double time-zoned" 18:30 celeron55 it's like being in a timezone, but two times better 18:31 darkrose "tomorrow's minetest today" 18:33 celeron55 it would be useful if somebody'd care to annotate the changelog with "who has done what" 18:33 celeron55 i'm too tired to even try 18:39 PilzAdam I will do this 18:55 PilzAdam celeron55, https://gist.github.com/4227081 18:57 thexyz PilzAdam: "Improve map generation speed a lot" is not mine, it's hmmmm's 18:57 thexyz i also prefer to be called xyz 18:57 celeron55 that's quite hastily done indeed 8) and instead of marking me at the end, you could just put my name there like everyone else's 18:57 hmmmm i don't really care about credits that much 18:57 hmmmm i mean, if someone cares, they'll see it in the repository 18:59 celeron55 "Tweaked some textures: apple, nyan cat, bricks, papyrus, steel sword (Calinou)" <- the nyan cat was made by vanessa, and i don't even know who's brick texture ended up in the game after the debate... 19:00 PilzAdam Calinou's brick textures are used in the default mod 19:00 celeron55 i guess i'll edit that to my liking 19:00 PilzAdam updated 19:01 PilzAdam but feel free to edit it 19:01 celeron55 it's also missing the "* Fixed the automagic render distance tuner" that i added on the news page 19:01 PilzAdam also Im not sure about the server tick and player pos send thing 19:02 PilzAdam I took Calinou's dropbox version as staring point 19:02 thexyz > * Fixed empty bucket being named "emtpy bucket" (khonkhortisan and PilzAdam) 19:02 thexyz lol, that one required two people to fix 19:02 celeron55 updated the page 19:03 celeron55 that changelog is terribly wrong at many parts but it's better than nothing 8) 19:03 PilzAdam there is inconsistent use of ", " and "and" 19:04 celeron55 some things are just, like, facts, but completely inverted ones 19:05 celeron55 well, good enough now 19:07 darkrose "changelog - an irc production loosely based on the changes to minetest" 19:11 celeron55 hmm... i'm actually quite proud of this release 19:12 celeron55 it's just so good considering everything 19:13 PilzAdam there are many good reasons to be 19:14 celeron55 altough a large bunch of known bugs are spinning in my head 8) 19:15 PilzAdam but nothing that breaks everything